ICRMCE

Lecture Notes in Civil Engineering
Stefanus Adi Kristiawan

Buntara S. Gan
Mohamed Shahin
Akanshu Sharma Editors
Proceedings of the
5th International
Conference on
Rehabilitation and
Maintenance in Civil
Engineering
ICRMCE 2021, July 8–9, Surakarta,
Indonesia
Volume 225
Series Editors
Marco di Prisco, Politecnico di Milano, Milano, Italy
Sheng-Hong Chen, School of Water Resources and Hydropower Engineering,
Wuhan University, Wuhan, China
Ioannis Vayas, Institute of Steel Structures, National Technical University of
Athens, Athens, Greece
Sanjay Kumar Shukla, School of Engineering, Edith Cowan University, Joondalup,
WA, Australia
Anuj Sharma, Iowa State University, Ames, IA, USA
Nagesh Kumar, Department of Civil Engineering, Indian Institute of Science
Bangalore, Bengaluru, Karnataka, India
Chien Ming Wang, School of Civil Engineering, The University of Queensland,
Brisbane, QLD, Australia
Lecture Notes in Civil Engineering (LNCE) publishes the latest developments in
Civil Engineering - quickly, informally and in top quality. Though original research
reported in proceedings and post-proceedings represents the core of LNCE, edited
volumes of exceptionally high quality and interest may also be considered for
publication. Volumes published in LNCE embrace all aspects and subfields of, as
well as new challenges in, Civil Engineering. Topics in the series include:
• Construction and Structural Mechanics
• Building Materials
• Concrete, Steel and Timber Structures
• Geotechnical Engineering
• Earthquake Engineering
• Coastal Engineering
• Ocean and Offshore Engineering; Ships and Floating Structures
• Hydraulics, Hydrology and Water Resources Engineering
• Environmental Engineering and Sustainability
• Structural Health and Monitoring
• Surveying and Geographical Information Systems
• Indoor Environments
• Transportation and Traffic
• Risk Analysis
• Safety and Security
To submit a proposal or request further information, please contact the appropriate
Springer Editor:
- Pierpaolo Riva at [email protected] (Europe and Americas);
- Swati Meherishi at [email protected] (Asia - except China, and
Australia, New Zealand);
- Wayne Hu at [email protected] (China).
All books in the series now indexed by Scopus and EI Compendex database!
More information about this series at https://link.springer.com/bookseries/15087

Stefanus Adi Kristiawan
•
Buntara S. Gan Mohamed Shahin

• •
Akanshu Sharma
Editors
Proceedings of the 5th

International Conference
on Rehabilitation
and Maintenance in Civil
Engineering
ICRMCE 2021, July 8–9,
Surakarta, Indonesia
123
Editors
Stefanus Adi Kristiawan Buntara S. Gan
Department of Civil Engineering College of Engineering
Universitas Sebelas Maret Nihon University
Surakarta, Indonesia Tokyo, Japan
Mohamed Shahin Akanshu Sharma

Department of Civil Engineering Institute of Construction Materials
Curtin University University of Stuttgart
Perth, WA, Australia Stuttgart, Germany
ISSN 2366-2557 ISSN 2366-2565 (electronic)

ISBN 978-981-16-9347-2 ISBN 978-981-16-9348-9 (eBook)
https://doi.org/10.1007/978-981-16-9348-9
© The Editor(s) (if applicable) and The Author(s), under exclusive license
to Springer Nature Singapore Pte Ltd. 2023
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether
the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and
transmission or information storage and retrieval, electronic adaptation, computer software, or by similar
or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, expressed or implied, with respect to the material contained
herein or for any errors or omissions that may have been made. The publisher remains neutral with regard
to jurisdictional claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd.
The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721,
Singapore
Foreword
The International Conference on Rehabilitation and Maintenance in Civil

Engineering (ICRMCE) is a triennial conference that aims to provide a forum for
researchers, academicians (professors, lecturers, and students), government agen-
cies, consultants, and contractors to exchange experiences, technological
advancement, and innovations in the world of civil engineering, specifically in the
fields of rehabilitation and maintenance. The previous four ICRMCE conferences
took place successfully in 2009, 2012, 2015, and 2018. Hundreds of researchers
worldwide attended these events to present their scientific papers in various areas of
civil engineering such as material engineering, structural engineering, geotechnical
engineering, transportation engineering, and construction management.
This year’s conference was organized by Sebelas Maret University in collabo-
ration with Mataram University. The conference was initially scheduled offline in
Mataram, Indonesia. However, due to the escalating coronavirus (COVID-19)
outbreak and the need for social distancing, we decided to hold the conference
online. Some reputable universities and institutions are participating in the current
ICRMCE as partners. Among them are Nihon University, University of Stuttgart,
National Taiwan University, TU Delft, Hiroshima University, Diponegoro
University, Muhammadiyah University of Yogyakarta, Jenderal Soedirman
University, University of Jember, UPN Veteran East Java, the National Center for
Research on Earthquake Engineering (NCREE) Taiwan, Himpunan Ahli
Konstruksi Indonesia (HAKI), and Himpunan Ahli Teknik Tanah Indonesia
(HATTI).
The ICRMCE 2021 was successfully held on July 8–9. Presenters who joined
this conference came from Japan, Singapore, Malaysia, China, Vietnam, Taiwan,
England, the Netherlands, Kuwait, and Indonesia. Furthermore, several outstanding
keynote speakers gave a presentation of the state-of-the-art findings in the field of
civil engineering. Our esteemed speakers are Prof. Shyh-Jiann Hwang (National
Taiwan University), Prof. Buntara Sthenly Gan (Nihon University), Dr. Edgar
Bohner (VTT Technical Research Centre of Finland), and Prof. Mohamed Shahin
(Curtin University).
v
vi Foreword
In the process of organizing this conference, we received invaluable motivation,

advice, and support from several individuals and institutions. I intend to express my
gratitude and appreciation to all of them. First, my most profound appreciation goes
to all organizing committee members who worked day and night preparing this
conference. Special thanks to the conference and media partners for their generous
support. We also express our gratitude to Prof. S.A. Kristiawan (Sebelas Maret
University), Dr. Ing. Akanshu Sharma (University of Stuttgart), Prof. Mohamed
Shahin (Curtin University), and Prof. Buntara Sthenly Gan (Nihon University) for
their willingness to serve as the editors of the 5th ICRMCE proceedings.
Halwan Alfisa Saifullah

The 5th ICRMCE Chairman
Preface
Civil engineering infrastructures are the backbone for the continuous development
of civilization. Managing these infrastructures is essential in keeping the quality of
services they provide to the community. A decline in the performance of key
infrastructure will have an impact on the quality of these services, which in turn can
cause social and economic problems. A variety of factors affects the performance of
infrastructure. In each case, the declining performance of infrastructure requires an
appropriate and adaptive response to offer effective solutions. Protection, mainte-
nance, repair, and retrofitting are part of the various solutions that can be imple-
mented. All of these solutions are assisted by technological developments related to
repair materials, methodologies, systems, management, and operational efficiency,
as well as economic and social considerations.
Infrastructure performance is also inevitably affected by exposure to hazards
originating from natural and environmental conditions such as earthquakes, land-
slides, and floods, among others. Therefore, hazard mitigation is also an interesting
topic of discussion. In addition, risk reduction and safety are among the most
important issues of infrastructure management. Finally, various perspectives on
sustainability in civil engineering are also covered in this conference.
This book is a collection of papers presented at the 5th International Conference
on Rehabilitation and Maintenance in Civil Engineering (ICRMCE) 2021 that deals
with the issues stated above. The papers are grouped into sequential themes rep-
resenting the structure of this book:
• Part I: Factors affecting performance of buildings and infrastructures
• Part II: Assessment, protection, maintenance, repair, and retrofitting of buildings
and infrastructures
• Part III: Maintenance management of buildings and infrastructures
• Part IV: Hazard mitigation
• Part V: Risk reduction and safety management
• Part VI: Sustainability aspects in transportation engineering
• Part VII: Sustainability aspects in construction projects
vii
viii Preface
• Part VIII: Sustainability aspects in water resources management

• Part IX: Construction materials for sustainable infrastructures
Postgraduate students, researchers, and practitioners who would like to update
their knowledge on the topics above will find this book very useful.
Surakarta, Indonesia Stefanus Adi Kristiawan

Chief Editor
Contents
Factors Affecting Performance of Buildings and Infrastructures

A Review on Application of Machine Learning in Building
Performance Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
R. W. Triadji, M. A. Berawi, and M. Sari
The Effect of P-Delta and P-Delta Plus Large Displacements Modelling
on Lateral and Axial Displacement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
Jonie Tanijaya and Robby S. Kwandou
Effect of Regulatory Change in Earthquake Load Analysis on
Structures with Irregular Shapes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Hendramawat Aski Safarizki and Dini Ayu Saputri
Seismic Performance of Tall and Slender Minaret Structure
with Hexagonal RC Wall Section by Means Fragility
Curve Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Erik Wahyu Pradana, Senot Sangadji, and Angga Destya Navara Noor
Role of Diagonal Bars in Reinforced Concrete Deep Beams Tested
Under Static Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Erwin Lim and Rahmat Ramli
Analysis of Reinforced Concrete Capacity for Irregular
Cross-Sections Using Numerical Methods . . . . . . . . . . . . . . . . . . . . . . . . 57
Nuroji
Assessment of Retaining Wall Design in Harris Skyline Tower’s
Basement, Surabaya . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Siti Nurlita Fitri and Ahmad Soimun
Additional Vertical Movement of the Single Pile Foundation with
Combined Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Sumiyati Gunawan, Niken Silmi Surjandari, Bambang Setiawan,
and Yusep Muslih Purwana
ix
x Contents
The Influence of the Number and Height Steps of Terraced Model on

Slope Stability Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
Niken Silmi Surjandari, Siti Nurlita Fitri, and Fenty Madani
Validation of TRMM Rainfall Data on Slope Stability in
Karanganyar, Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
David Raja Simare Mare, Rr Rintis Hadiani, and Raden Harya Dananjaya
Rubberized Asphalt Pilot Road Trial in Kuwait . . . . . . . . . . . . . . . . . . 117
H. Al-Baghli, Z. S. Awadh, and S. E. Zoorob
Application of the Updated PSHA on the Stability Analysis of the
Meninting Diversion-Spillway Tunnel in Lombok Island—Indonesia . . . 129
Didi S. Agustawijaya, Ria R. Marlaningtyas, Suryawan Murtiadi,
Mudji Wahyudi, Muhajirah, Hartana, and Ausa R. Agustawijaya
Risk and Stability Evaluation of Klego Dam, Boyolali, Indonesia . . . . . 141
Suharyanto, Kresno Wikan Sadono, Rizqi Iqbal Maulana, P. Arie Bagus,
and Dyah Ari Wulandari
Impact of Climate Change on Dam Safety . . . . . . . . . . . . . . . . . . . . . . . 157
Heri Sulistiyono, Ery Setiawan, and Humairo Saidah
Assessment, Protection, Maintenance, Repair, and Retrofitting

of Buildings and Infrastructures
Investigating Materials for Refurbishment Strategies of Heritage
Buildings: A Case Study of Soesman Kantoor, Semarang . . . . . . . . . . . . 169
Ferry Hermawan, Didi Wibowo Tjokro Winoto, Ismiyati,
Bambang Purwanggono, and Robby Soetanto
Self-monitoring and Localization of Crack of Concrete Beam with
Fibers and Carbon Black Subjected to Bending . . . . . . . . . . . . . . . . . . . 179
Genjin Liu and Yining Ding
Seismic Performance Analysis of Multi-story Buildings with Addition
of Bracing Based on SNI 1726: 2019 (Case Study: Airlangga
University Parking Building) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
Krisnamurti, Willy Kriswardhana, and Achmad Wahyu Ramadiyan
Various Strut—Macro Modelings for Infilled Frame Analysis . . . . . . . . 199
Isyana Ratna Hapsari, Marwahyudi, Edy Purwanto, Senot Sangadji,
and Stefanus Adi Kristiawan
Strengthening of Non-engineered Building Beam-Column Joint to
Increase Seismic Performance with Variation of Steel Plate Width . . . . 215
Edy Purwanto, P. Amarta Adri, S. A. Kristiawan, Senot Sangadji,
and S. Halwan Alfisa
Contents xi
A Proposed Method of FRP Anchorage for FRP Confined

Over-Reinforced Concrete Beam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 225
Nuroji, Ay Lie Han, Sri Tudjono, Lena Tri Lestari,
and Tiara Murtisari
Experimental Investigation on the Shear Behavior of Patched RC
Beams Without Web Reinforcements: Efficacy of Patching Position
with Respect to the Shear Span . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
Adji P. Abrian, Stefanus A. Kristiawan, Halwan A. Saifullah,
P. Muhammad Rafi, R. Muhammad Hafizh, Andreas M. Simanjuntak,
and D. Abel Bismo
Secondary AE Analysis of Pre-corroded Concrete Beam . . . . . . . . . . . . 243
Ahmad Zaki and Zainah Ibrahim
Cable Force Prediction Technique Using Subspace and Effective
Vibration Length Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257
Muhammad Ibnu Syamsi, Hao-Lin Wang, and Chung-Yue Wang
Implementation of Microbially Induce Calcite Precipitation (MICP)
by Bacillus Subtilis and Adding Sand in Repairing Shear Strength
Parameters of Peat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
Firman Syarif and Dian M. Setiawan
Sensitivity Analysis on the Effect of Reinforcement Materials Addition
for Soil Stabilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
Ida Agustin Nomleni, Raden Harya Dananjaya,
Numerical Simulation of Slope Stability for Soil Embankment
Reinforced with Inclined Bamboo Piles . . . . . . . . . . . . . . . . . . . . . . . . . 287
Ngudiyono and Tri Sulistyowati
Analysis of Community Satisfaction Level on the Road Rehabilitation
and Reconstruction Project (Learn from Palu Disasters Area) . . . . . . . . 297
Andri Irfan Rifai, Eko Prasetyo, and Rhismono
Seepage and Piping Control of Earth Fill Dam . . . . . . . . . . . . . . . . . . . 311
Muhammad Zainal Arifin, Yusep Muslih Purwana,
and Raden Harya Dananjaya
Maintenance Management of Buildings and Infrastructures

State-of-the-Art of Artificial Intelligence Methods in Structural
Health Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
I. G. E. A. Putra
xii Contents
Development of Preventive Maintenance Guidelines

for Simple-Classification Government Buildings Based
on Work Breakdown Structure Within the DKI Jakarta
Provincial Government . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
Dyah Ayu Pangastuti and Yusuf Latief
Development of Preventive Maintenance Guidelines for Electrical
Components on Government Building Based on Work Breakdown
Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
Azhara Yudha Pradipta, Yusuf Latief, and Rossy Armyn Machfudiyanto
Evaluation of the Implementation of Fire Safety Management Based
on Work Breakdown Structure Affecting the Insurance Premium
Costs of High-Rise Lecture Buildings . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
E. P. Mahardika, F. Muslim, Y. Latief, and P. S. Nugroho
A Critical Review of Bridge Management System in Indonesia . . . . . . . 381
Surya Dewi Puspitasari, Sabrina Harahap, and Pinta Astuti
Crack and Corrosion Inspections for Coastal and Marine Concrete
Infrastructure: A Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391
Sabrina Harahap, Surya Dewi Puspitasari, and Ahmad Aki Muhaimin
e-Peralatan System as an Equipment Management for Disaster
Mitigation on Indonesia National Roads . . . . . . . . . . . . . . . . . . . . . . . . 401
Adityo Budi Utomo, P. Gitaning, and S. Tosan Kunto
The Implementation of Functional Road Assessment on Pramuka
Road Section in Klaten District . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 411
Amirotul Musthofiah H. Mahmudah, Syahrul Anggara Wuryatmaja,
and Ary Setyawan
The Evaluation of Irrigation Maintenance in Pacal Irrigation Area
at Bojonegoro Regency, East Java . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
Mahdika Putra Nanda, Rintis Hadiani, and Antonius Suryono
Hazard Mitigation
Landslide Analysis Subject to Geological Uncertainty Using Monte
Carlo Simulation (A Study Case in Taiwan) . . . . . . . . . . . . . . . . . . . . . 437
Joni Fitra, Wen-Chao Huang, and Yusep Muslih Purwana
The Performance of Horizontal Drain as a Landslide Mitigation
Strategy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
Putu Tantri K. Sari
The Analysis of Impact and Mitigation of Landslides Using Analytical
Hierarchy Process (AHP) Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 457
A. Andriani, B. M. Adji, and S. Ramadhani
Contents xiii
The Stability of a Slope on Soft Soil Using the Hardening

Soil Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467
Yerry Kahaditu Firmansyah and Maharani Putri Dewanty
Software Performance of Risk-Targeted Maximum Considered
Earthquake (MCER) Calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479
Windu Partono
Seismic Microzonation of Yogyakarta Province Based on 2019
Risk-Targeted Maximum Considered Earthquake . . . . . . . . . . . . . . . . . 489
Windu Partono, Ramli Nazir, Frida Kistiani, and Undayani Cita Sari
Liquefaction Potential of Volcanic Deposits During Lombok
Earthquake in 2018 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
Muhajirah
Fault Structure Interpretation on the Western Part of East Java Using
Second Vertical Derivative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 511
Wien Lestari, Amien Widodo, Dwa Desa Warnana, Firman Syaifuddin,
Rusba Saputra Rivensky, and Bagoes Idcha Mawardi
Lesson Learned from Weathering Clay Shale Residual Interface
Shear Strength Testing Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 523
Fathiyah Hakim Sagitaningrum, Samira Albati Kamaruddin, Ramli Nazir,
Budi Susilo Soepandji, and Idrus M. Alatas
Shear Strain Evaluation on Analysis of Additional Clay Liner Layer
Modeling in Ngipik Landfill, Gresik . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533
Siti Nurlita Fitri
Tsunami Hazard in Cilacap City Due to the Megathrust of West-
Central Java Segment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 543
Wahyu Widiyanto and Sanidhya Nika Purnomo
Assessment of the Conditioning Factor for Flash Flood Susceptibility
Potential Based on Bivariate Statistical Approach in the Wonoboyo
Watershed in East Java, Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . 553
Entin Hidayah, Gusfan Halik, and Wiwik Yunarni Widiarti
Flood Management Strategies in Indonesia: A Lesson Learned from
Pepe River, Central Java . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575
Rian Mantasa Salve Prastica and Amalia Wijayanti
Comparison of Suitable Drought Indices for Over West Nusa
Tenggara . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 587
Humairo Saidah, Heri Sulistiyono, and I Dewa Gede Jaya Negara
Small Debris Flow Simulation Using MORPHO2DH . . . . . . . . . . . . . . . 601
Puji Harsanto, Dandy Darvin Septiandy, Berli Paripurna Kamiel,
and Nursetiawan
xiv Contents
Simulation of Debris Flow Using “SIMLAR” in the Watershed

of Gendol River, Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 609
Hendy Dwi Cahyo, Jazaul Ikhsan, and Ani Hairani
Wind-Generated Wave Simulation on Payangan Beach Utilizing
DELFT3D . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 619
Enggar Setia Baresi, Retno Utami Agung Wiyono,
and Wiwik Yunarni Widiarti
Risk Reduction and Safety Management

Incorporating Cultural Attributes into Disaster Risk Reduction-Based
Development Plans in Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 631
Yusron Saadi
Identification Risk Potential Hazard of Railway Project Based on the
Work Breakdown Structure to Improve Safety Performance . . . . . . . . . 641
D. V. Aryanto, L. S. Riantini, R. A. Machfudiyanto, and Y. Latief
Identification Factors of Safety Climate, Awareness, and Behaviors to
Improve Safety Performance in Telecommunication Tower
Construction at PT X . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 651
B. F. Athaya, L. S. Riantini, and R. A. Machfudiyanto
Cost Structure of Construction Safety on High Residential Buildings
in Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 659
Ratih Fitriani, Yusuf Latief, and Rossy A. Machfudiyanto
Adaptive Traffic Signal Control Using Fuzzy Logic Under Mixed
Traffic Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 669
Budi Yulianto
Application of Fuzzy Inference System Mamdani
at Pelican Crossing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 681
Salsabila Naura Putri
Determining the Maximum Speed Limit in Residential Area . . . . . . . . . 693
Gito Sugiyanto, Fadli Wirawan, Eva Wahyu Indriyati, Yanto,
and Mina Yumei Santi
Identification of Traffic Accident Hazardous Location and Cost of
Accidents in Developing Country (Case Study of Tabanan Regency,
Bali-Indonesia) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 705
Putu Alit Suthanaya and Made Oka Sugiana
Characteristics of Foreign Motorcyclists in Tourism Areas in Bali . . . . 717
Cokorda Putra Wirasutama, Putu Alit Suthanaya,
Dewa Made Priyantha Wedagama, and Anak Agung Gde Agung Yana
Contents xv
Analysis of the Behaviour Model of Foreign Motorcyclists in Tourism

Areas in Bali . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 729
Knowledge and Practice of Helmet Usage Among Senior High School
Students in Klaten Regency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 743
Dewi Handayani, Putri Dewi Prasetianingrum,
and H. M. Amirotul Musthofiah
Sustainability Aspects in Transportation Engineering

Investigation of CO2 Emissions on Two Local Streets by Means
of IPCC and Direct Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 753
Florentina Pungky Pramesti, Annisa Tri Utami, Iva Yenis Septiariva,
and Fajar Sri Handayani
Reducing the Release of Greenhouse Gases in the Rigid Pavement
Material Transportation Process Unit . . . . . . . . . . . . . . . . . . . . . . . . . . 763
Fajar Sri Handayani, Florentina Pungky Pramesty, and Ary Setyawan
Cost Effectiveness Analysis of Greenhouse Gas Emissions Reduction
in the Flexible Pavement Material Transportation Process Unit . . . . . . 771
Fajar Sri Handayani, Florentina Pungky Pramesti, and Ary Setyawan
A Probabilistic Model of Container Port Demand in Java Concerning
the Port Hinterland Connectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 781
Lydia Novitriana Nur Hidayati, Gerard de Jong, Anthony Whiteing,
and Munajat Tri Nugroho
Evaluation of Logistics System Performance-Based on Indonesian
Government Policy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 791
Muhammad Rizky Prakoso, Mohammed Ali Berawi, and Gunawan
Analysis of the Influence of Region Development Factors, Individual
and Activity, Internal Operator and External Operator on the
Demand for the Jakarta Bandung High-Speed Rail . . . . . . . . . . . . . . . . 799
Samijan, Mohammed Ali Berawi, and Andyka Kusuma
Investigating the Factors Influencing the Demand of School Bus . . . . . . 811
Willy Kriswardhana, Syamsul Arifin, and Ainal Akbar
Sustainability Aspects in Construction Projects

Key Performance Indicator for Analytical Hierarchy Process Used for
Determining the Effect of Reverse Supply Chain Toward Green
Building Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 823
Hermawan, Jati Utomo Dwi Hatmoko, and Jovana Neilkelvin
xvi Contents
Development of Blockchain and Machine Learning System in the

Process of Construction Planning Method of the Smart Building to
Save Cost and Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 833
Christiantono Tedjo, Mohammed Ali Berawi, and Mustika Sari
The Development of Blockchain Based Knowledge Management
System Model at EPC Projects to Improve Project
Time Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 843
D. Y. Priyambodo, M. A. Berawi, and M. Sari
Risk Allocation Implementation Analysis of Public-Private
Partnership for Infrastructure Project (Case Study of the
Solo-Yogyakarta-NYIA Kulon Progo Highway Project) . . . . . . . . . . . . . 853
Widi Hartono, Aloysia Putri Hastari Purnomo, and Sunarmasto
Risk Assessment of Construction Project Scheduling . . . . . . . . . . . . . . . 863
Zetta Rasullia Kamandang
Investigating Construction Project Delay Using Fault Tree Analysis
Based on Its Dominant Risk on Private Project . . . . . . . . . . . . . . . . . . . 873
Jojok Widodo Soetjipto, Amalia Martha Sukmana, and Syamsul Arifin
Implementing a Relational Database in Processing Construction
Project Documents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 891
Mik Wanul Khosiin and Ardian Umam
Evaluation of the Public Procurement Principles Implementation
in Surabaya Construction Projects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 901
Patrisius Valdoni Sandi and Mohammad Arif Rohman
Relationship Model Between Conceptual Cost Estimation Process of
Flyover Development in the Provincial Government of DKI Jakarta
with the Accuracy Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 913
Putika Yussi, Yusuf Latief, and Rossy Armyn Machfudiyanto
The Sustainability Aspect of the Consulting Firm in Terms of Its
Competitiveness in Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 933
Nofita Harwin, Mairizal, Zelmi Sriyolja, Abd Rahman Mohd Sam,
and Mohd Zaimi Abd Majid
Sustainability Aspects in Water Resources Management

Sustainability Analysis of Minimization of Spills from a Reservoir . . . . 945
Syamsul Hidayat, Ery Setiawan, Ida Bagus Giri Putra, M. Bagus Budianto,
and Salehudin
Infiltration Wells as an Alternative Eco Drainage System a Case Study
in Mangkubumen Surakarta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 953
Retnayu Molya, R. R. Rintis Hadiani, and Adi Yusuf Muttaqien
Contents xvii
Analysis of Leb Irrigation Patterns of Pipe System in Sorgum Plants

in Sand Dry Lands Akar Akar Village . . . . . . . . . . . . . . . . . . . . . . . . . . 965
I Dewa Gede Jaya Negara, Sasmito Soekarno, Suwardji, Humairo Saidah,
and Atha Adi
Hourly Rainfall Simulation Using Daily Data . . . . . . . . . . . . . . . . . . . . . 975
Suroso, Fatimatus Sholihah Marush, Purwanto Bekti Santoso,
Irfan Sudono, Edvin Aldrian, and Nelly Florida Riama
Water Quality Mapping on the Coast of Bangkalan Madura Based on
the Acidity Value from Aqua MODIS Satellite Imagery . . . . . . . . . . . . 989
Hendrata Wibisana, Zetta Rasullia Kamandang, and Kartini
Multi-attribute Analysis of Raw Water Treatment from Deep Wells at
PDAM Tirta Mahottama, Klungkung Regency, Bali . . . . . . . . . . . . . . . 999
Ni Kadek Dian Utami Kartini, Nurulbaiti Listyendah Zahra,
Ariyanti Sarwono, Intan Rahmalia, Almira Davina Nastiti,
Iva Yenis Septiariva, and I. Wayan Koko Suryawan
Determination of Produced Wastewater Treatment Systems for
Reclaim Water in the Oil and Gas Industry . . . . . . . . . . . . . . . . . . . . . . 1009
Novena Lany Pangestu, Nurulbaiti Listyendah Zahra, Ariyanti Sarwono,
Intan Rahmalia, Iva Yenis Septiariva, and I. Wayan Koko Suryawan
Preference of Sludge Treatment Plan in IPA II Pejompongan Water
Treatment Plant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1019
Nailatul Fadhilah, Nurulbaiti Listyendah Zahra, Fatimah Dinan Qonitan,
Imroatus Sholikhah, Intan Rahmalia, Iva Yenis Septiariva,
and I. Wayan Koko Suryawan
Design of Typical Rainwater Harvesting Storage Tanks Based on
Housing Type (Case Study in Indonesia) . . . . . . . . . . . . . . . . . . . . . . . . 1029
Lina Indawati, Setyo Budi Kurniawan, Siti Rozaimah Sheikh Abdullah,
Risk Analysis of Shared Marine Space in the View of Traditional
Fishermen Perceptions in the National Tourism Strategic Area of
Lombok, Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1043
Ida Ayu Oka Suwati Sideman, R. M. Nyoman Budiartha,
Ida Bagus Putu Adnyana, and Ngakan Ketut Acwin Dwijendra
Construction Materials for Sustainable Infrastructures

Residual Stress Evaluation on Cold-Formed Steel C-Section by X-Ray
Diffraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1057
T. Widya Swastika, Heru Purnomo, M. R. Muslih, R. Apriansyah,
Henki W. Ashadi, and Mulia Orientilize
xviii Contents
Mechanical Properties of Fine-Grained Concrete Using Fine-Red

Sand and Fly Ash for Road Construction: A Case Study
in Vietnam . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1067
Nguyen Thanh Sang, Thai Minh Quan, May Huu Nguyen, and Lanh Si Ho
A Systematic Review of Concrete Material for Noise Reduction of
Transportation Sectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1077
Ecky Ferry Ferdyan, Dewi Handayani, Sholihin As’ad, and Ubaidillah
Mapping Literature of Reclaimed Asphalt Pavement Using
Bibliometric Analysis by VOSviewer . . . . . . . . . . . . . . . . . . . . . . . . . . . 1085
Mochammad Qomaruddin, Han Ay Lie, Widayat, Bagus Hario Setiadji,
and Mochamad Agung Wibowo
Literature Study: Alternative Materials for Hot Rolled Sheet-Wearing
Course (HRS-WC) Pavement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1095
Elsa Eka Putri, Purnawan, Bayu Martanto Adji, and Bobby Herman
Marshall Characteristics of Asphalt Mixture with Water Hyacinth
Ash as Filler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1109
Dony Rohmad Dony, Bagus Hario Setiadji, and Bambang Riyanto
The Setting Time of Portland Composite Cement Mixed with Calcium
Stearate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1119
A. Maryoto, P. Hardini, and R. Setijadi
Evaluation of Silt Content in Natural Sand Used as Building
Materials: A Statistical Analysis Approach . . . . . . . . . . . . . . . . . . . . . . 1133
Jauhar Fajrin, Mulyadi, Hariyadi, and Agung Prabowo
Compressive and Flexural Strength Behavior of Banana Tree Fiber
Hybrid Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1143
Fadillawaty Saleh, Fanny Monika, Hakas Prayuda,
Bella Lutfiani Al Zakina, Martyana Dwi Cahyati, Adira Aldi,
and Feri Adri Wibowo
Investigation of Polypropylene Fiber Reinforced Concrete After
Elevated Temperature Using Color Quantification and Alkalinity
Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1153
Ni Nyoman Kencanawati, Suryawan Murtiadi, and Zul Aida Nur
The Effect of 12.5% Metakaolin and Variations of Silica Fume on Split
Tensile Strength and Modulus of Rupture of High Strength
Self-Compacting Concrete (HSSCC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1165
Endah Safitri, Wibowo, Halwan Alfisa Saifullah,
and Farhan Gilang Septian
Contents xix
The Strength and Modulus of Elasticity of High Strength

Self-compacting Concrete (HSSCC) with 12.5% Metakaolin and
Variations of Silica Fume . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1173
Endah Safitri, Wibowo, Halwan Alfisa Saifullah,
and Fernanda Sarwatatwadhika Perdana
Factors Affecting Performance
of Buildings and Infrastructures
A Review on Application of Machine
Learning in Building Performance
Prediction
R. W. Triadji, M. A. Berawi, and M. Sari
Abstract Designers usually use Building Performance Simulation (BPS) to support

decision making in facing design requirements and expected building performance.
However, the fact is that BPS still experiences several limitations, such as BPS
requires high computation time in assessing various design options. Machine learning
is considered capable of solving the problem that the existing BPS has. Research on
this problem has been conducted to provide solutions and prove the reliability of
machine learning in predicting building performance. Therefore, this paper aims
to discuss the research and overview of how machine learning has been used in
predicting building performance. The results show that, performance prediction using
machine learning has been developed on energy and environmental performance.
Also, machine learning can significantly reduce the prediction time without reducing
its accuracy.
Keywords Machine learning · Building performance · Energy performance ·

Environmental performance
1 Introduction
The importance of digital technology and automation systems in playing a role in

industry 4.0 makes every device now equipped with machine learning [1]. As the
core of artificial intelligence [2], machine learning is a process where machines can
learn various things by themselves based on their experiences [3]. Machines learn
the relationship between input and output data received as information and turn it
R. W. Triadji (B) · M. A. Berawi · M. Sari

Civil Engineering Department, Faculty of Engineering, Universitas Indonesia, Depok, Indonesia
e-mail: [email protected]
M. A. Berawi
M. Sari
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 3
S. A. Kristiawan et al. (eds.), Proceedings of the 5th International Conference on
Rehabilitation and Maintenance in Civil Engineering, Lecture Notes in Civil Engineering 225,
https://doi.org/10.1007/978-981-16-9348-9_1
4 R. W. Triadji et al.
into an experience to offer solutions that can adapt according to given situations
[4, 5]. Machine learning has been developed in the construction industry to solve
complicated and difficult problems [6]. It is shown by the potential of machine
learning in improving building performance [3].
Building performance is considered since the design stage; this step is taken as
an effort for the building to perform its functions optimally when it is operating
[7]. Designers usually use Building Performance Simulation (BPS) to support deci-
sion making in facing design requirements and expected building performance [8].
However, BPS is still experiencing several deficiencies in its application; first, BPS
requiring a high computation time to assess various design options [8, 9]. Moreover,
some factors have not been considered in implementing the simulation, thus making
the predictions offered by BPS less accurate [7]. Machine learning is considered to
have the potential and capability to answer these deficiencies [8]. Its ability to predict
various designs in a second and incorporate factors not covered by existing BPS will
improve the prediction results and replaced the existing Building performance Simu-
lation [7, 9]. Below, we made the conceptual diagram to give a glance of insight (see
Fig. 1). Building performance no longer predicts by the existing BPS; it replaced
by machine learning and its algorithm to predict those predictions such as energy
performance and environmental performance.
Therefore, this paper aims to examine studies related to machine learning in
building performance and an overview of how machine learning has been used to
predict building performance. The review expected can provide more extensive refer-
ence and insight in involving machine learning in the construction industry to enhance
building performance prediction.
Fig. 1 Conceptual diagram of machine learning and existing BPS

A Review on Application of Machine Learning … 5
2 Methods
In this study, we collected literature through Scopus as the central database for
academic publications. The keywords used are “machine learning”, “building perfor-
mance”, “energy performance”, “energy consumption”, “environment” “indoor envi-
ronment”, “thermal comfort”, “visual comfort”, the selected documents are limited
to journal articles and conference papers that have been published in the last ten years
(2011–2021). The search structure example for Scopus is as follows:
TITLE-ABS-KEY ("machine learning” AND "building performance") AND (LIMIT-
TO (DOCTYPE, "ar”) OR LIMIT TO (DOCTYPE, "cp”)) AND (EXCLUDE (PUBYEAR,
2010))
The authors conducted the first screening by reviewing abstracts and matching
predetermined keywords so irrelevant documents could be separated from the search
results. 37 documents are deemed relevant to the topics we discussed. Then, the
second screening will be carried out by reading the entire document. 17 papers were
not selected to be discussed, after the second screening. Thus, 20 papers will be
discussed and grouped based on two primary performances that become the main
concern in a building: energy performance and environmental performance [10, 11].
The algorithm of machine learning in each paper will also be discussed in this review.
3 Result and Discussion
3.1 Energy Performance
Several building energy performance tools such as TrnSys, EnergyPlus, ESP-r, DOE-
2 and Modellica have been widely used, but these tools still require high computation
time for various designs with multiple inputs; this also makes simulation tools difficult
in delivering precise result [9, 12, 13]. In this section, the discussion will be grouped
into two main parts, which are thermal and electricity; due to the current study that we
discussed in this section focuses on thermal performance and follows by electricity
performance.
Thermal Energy
In previous studies, machine learning has been developed to predict the thermal
(cooling and heating energy) performance of buildings using an Artificial Neural
Network (ANN) [9, 12, 14–16]. Comesaña et al. [12] developed a machine learning
model based on outdoor and indoor temperatures and radiation levels. Meanwhile,
Geyer and Singaravel’s [14] model based on building components such as walls,
windows and floors. Geyer and Singaravel [14] stated that the computation time to
make predictions using machine learning is reduced drastically. In line with this,
Ascione et al. [16] ensure a computation time savings of 98% can be achieved using
machine learning, this has been proven in their research. They develop a model for
predicting energy performance and the energy retrofitting scenario for each building
element. Similarly, Seyedzadeh et al. [17] also made a model to predict energy perfor-
mance and support decision-making about energy retrofit scenarios, the algorithm
used is gradient boosted regression trees.
Singh et al. [9] make energy predictions by creating four schematic methods for
making predictions. The first three using EnergyPlus, and the last method is using
machine learning. The result is that machine learning can achieve the objectives of
reducing computation time than EnergyPlus. Westermann et al. [18] using the Convo-
lutional Neural Network (CNN) algorithm to develop a machine learning model to
displace a building energy simulation tool. It can predict thermal energy based on
input from various design and climate variations. The model is representative enough
to be used in various climates of building locations and does not affect its accuracy.
On the other hand, Ciulla and D’Amico [19] developed a model without the inten-
tion of replacing the existing simulation model. The model was developed with the
Multiple Linear Regression algorithm, built as simple as possible, so even non-expert
users can use it.
Chakraborty and Elzarka [20] developed their model using three different algo-
rithms (XGBoost, ANN and Degree-day-based OLS regression); the result is
XGBoost the most accurate algorithm in predicting thermal performance. In contrast
to other studies, Attanasio et al. [21] developed a model only to predict heating energy
by comparing four machine learning algorithms. Meanwhile, Yu et al. [13] combine
two algorithms, genetic algorithm (GA) and back-propagation (BP), to optimize
energy performance predictions and thermal comfort based on the building design
and design envelope in residential building. Similarly, Robinson et al. [22] devel-
oped a model for commercial building using CBECS (Commercial Building Energy
Consumption Survey) data released by the US Energy Information Administration
(EIA). It is intended that the proposed model can be used in any city in the US.
Electricity Energy
Zeng et al. [23] developed machine learning to predict electrical energy consumption
in a building using the Gaussian Process algorithm. The model developed is based
on three types of buildings: offices, shopping centres, and hotels with various design
configurations. Later, the model can be used in several kinds of buildings. Moreover,
Pangaribuan et al. [24] also build a model to predict electrical energy using a support
vector regression algorithm for residential homes.
3.2 Environmental Performance
Environmental performance is also one of the criteria that need to be considered in

the design to balance each building’s performance and achieve optimal sustainability
goals [25, 26]. In this case, the environmental performance consists of an indoor or
outdoor environment. The indoor performance focuses on the comfort of occupants
in the building [13]. Meanwhile, outdoor performance pays more attention to the
impact caused by buildings on the outdoor environment, such as CO2 gas emissions
[11]. Mazuroski et al. [27] developed a recurrent neural network-based model to
predict indoor temperature based on density, specific heat and thermal conductivity
of each building material that will be used. The same idea was done by Kamel et al.
[28]. The difference between these two studies is the data used as input to predict
indoor temperature, i.e. humidity, radiations, airspeed, door open/close status, and
motion from sensors installed in the room. In their research, the model developed by
Yu et al. [13] with GA-BP algorithm is not only to predict energy performance but
also to predict thermal comfort.
On the other hand, Symonds et al. [29] used an artificial neural network to predict
indoor environmental quality. The model developed can make predictions in over-
heating metric, PM2.5 ratio, relative humidity and heating energy use with a reason-
able degree of accuracy. In his study, Chen [30] said that difference results between the
building performance analysis predictions and reality could not be avoided. Machine
learning was developed to improve the accuracy of green BIM by using an artificial
neural network algorithm and daylight luminance level as an outcome of the predic-
tion. The model can also predict climate analysis, thermal comfort, energy calcula-
tion, and other BPA dimensions. Chatzikonstantinou and Sariyildiz [31] focused their
study on office space’s visual comfort; a model developed can predict Daylight Glare
Probability and Daylight Autonomy. Algorithms feed-forward networks (FFNs) were
chosen as the algorithm with the highest level of accuracy compared to support vector
machines (SVMs) and random forests (RFs).
While some of the studies focused on indoor performance, Feng et al. [11] devel-
oped a model to discover the environmental impact from the building design. They
were using GWP/unit area as an indicator of environmental performance. The input
of the model is the early design parameter, and fuzzy c means algorithm approach
was used to perform the model. The model expected can help designers in testing
several design scenarios and the resulting performance.
4 Conclusion
Machine learning has been used to predict buildings performance in terms of energy
and the environment. From our review, energy performance is a performance that is
widely used in developing machine learning models. Also, artificial neural networks
are found as an algorithm that is widely used in model development. The potential
of machine learning in predicting performance in a second and accurately can help
designers make decisions of a wide variety of designs. It reveals that machine learning
can be a promising and powerful tool, mainly when used at the design stage. However,
while machine learning has promising potential, it still has its limitations. First, to
make good predictive, machine learning requires big data to train the algorithm [3].
Moreover, some studies apply machine learning to only one building type, and this
makes the model unable to be used for other building types. It is expected that with
these reviews, the construction industry can take advantage of machine learning
capabilities in predicting building performance.
Acknowledgements The author is grateful to Universitas Indonesia for giving support in this
research.
References
1. Muhuri PK, Shukla AK, Abraham A (2019) Industry 4.0: a bibliometric analysis and detailed
overview. Eng Appl Artif Intell 78:218–235. https://doi.org/10.1016/j.engappai.2018.11.007
2. Jordan MI, Mitchell TM (2015) Machine learning: trends, perspectives, and prospects. Science
(80-)349:255–260. https://doi.org/10.1126/science.aaa8415
3. Hong T, Wang Z, Luo X, Zhang W (2020) State-of-the-art on research and applications of
machine learning in the building life cycle. Energy Build 212.https://doi.org/10.1016/j.enb
uild.2020.109831
4. Dimiduk DM, Holm EA, Niezgoda SR (2018) Perspectives on the impact of machine learning,
deep learning, and artificial intelligence on materials, processes, and structures engineering.
Integr Mater Manuf Innov 7:157–172. https://doi.org/10.1007/s40192-018-0117-8
5. Yücel M, Nigdeli SM, Bekdaş G (2021) Artificial intelligence and machine learning with
reflection for structural engineering: a review. In: Studies in systems, decision and control.
Springer Science and Business Media Deutschland GmbH, Department of Civil Engineering,
Istanbul University-Cerrahpaşa, Avcılar, Istanbul, 34320, Turkey, pp 23–72
6. Darko A, Chan APC, Adabre MA et al (2020) Artificial intelligence in the AEC industry:
scientometric analysis and visualization of research activities. Autom Constr 112:103081.
https://doi.org/10.1016/j.autcon.2020.103081
7. Chokwitthaya C, Zhu Y, Mukhopadhyay S, Collier E (2020) Augmenting building perfor-
mance predictions during design using generative adversarial networks and immersive virtual
environments. Autom Constr 119:103350. https://doi.org/10.1016/j.autcon.2020.103350
8. Singaravel S, Suykens J, Geyer P (2018) Deep-learning neural-network architectures and
methods: using component-based models in building-design energy prediction. Adv Eng
Inform 38:81–90. https://doi.org/10.1016/j.aei.2018.06.004
9. Singh MM, Singaravel S, Klein R, Geyer P (2020) Quick energy prediction and comparison
of options at the early design stage. Adv Eng Inform 46:101185. https://doi.org/10.1016/j.aei.
2020.101185
10. Kim J, Hong T, Hastak M, Jeong K (2020) Intelligent planning unit for the artificial intelligent
based built environment focusing on human-building interaction. J Asian Archit Build Eng
00:1–18. https://doi.org/10.1080/13467581.2020.1803076
11. Feng K, Lu W, Wang Y (2019) Assessing environmental performance in early building design
stage: an integrated parametric design and machine learning method. Sustain Cities Soc
50:101596. https://doi.org/10.1016/j.scs.2019.101596
12. Comesaña MM, Febrero-Garrido L, Troncoso-Pastoriza F, Martínez-Torres J (2020) Prediction
of building’s thermal performance using LSTM and MLP neural networks. Appl Sci 10:1–16.
https://doi.org/10.3390/app10217439
13. Yu W, Li B, Jia H et al (2015) Application of multi-objective genetic algorithm to optimize
energy efficiency and thermal comfort in building design. Energ Build 88:135–143. https://doi.
org/10.1016/j.enbuild.2014.11.063
14. Geyer P, Singaravel S (2018) Component-based machine learning for performance prediction in
building design. Appl Energ 228:1439–1453. https://doi.org/10.1016/j.apenergy.2018.07.011
15. Hwang JK, Yun GY, Lee S et al (2020) Using deep learning approaches with variable selection
process to predict the energy performance of a heating and cooling system. Renew Energ
149:1227–1245. https://doi.org/10.1016/j.renene.2019.10.113
16. Ascione F, Bianco N, De Stasio C et al (2017) Artificial neural networks to predict energy
performance and retrofit scenarios for any member of a building category: a novel approach.
Energy 118:999–1017. https://doi.org/10.1016/j.energy.2016.10.126
17. Seyedzadeh S, Pour Rahimian F, Oliver S et al (2020) Machine learning modelling for predicting
non-domestic buildings energy performance: a model to support deep energy retrofit decision-
making. Appl Energ 279:115908. https://doi.org/10.1016/j.apenergy.2020.115908
18. Westermann P, Welzel M, Evins R (2020) Using a deep temporal convolutional network as a
building energy surrogate model that spans multiple climate zones. Appl Energy 278:115563.
https://doi.org/10.1016/j.apenergy.2020.115563
19. Ciulla G, D’Amico A (2019) Building energy performance forecasting: a multiple linear
regression approach. Appl Energ 253:113500. https://doi.org/10.1016/j.apenergy.2019.113500
20. Chakraborty D, Elzarka H (2019) Advanced machine learning techniques for building perfor-
mance simulation: a comparative analysis. J Build Perform Simul 12:193–207. https://doi.org/
10.1080/19401493.2018.1498538
21. Attanasio A, Piscitelli MS, Chiusano S et al (2019) Towards an automated, fast and interpretable
estimation model of heating energy demand: a data-driven approach exploiting building energy
certificates. Energies 12.https://doi.org/10.3390/en12071273
22. Robinson C, Dilkina B, Hubbs J et al (2017) Machine learning approaches for estimating
commercial building energy consumption. Appl Energ 208:889–904. https://doi.org/10.1016/
j.apenergy.2017.09.060
23. Zeng A, Ho H, Yu Y (2020) Prediction of building electricity usage using Gaussian Process
Regression. J Build Eng 28:101054. https://doi.org/10.1016/j.jobe.2019.101054
24. Pangaribuan AB, Octa A, Pradnyana IWW, Afrizal S (2019) Predictive analytic for estimating
electric consumption of smart grid platform in residential single-family building using support
vector regression approach. J Phys Conf Ser 1196.https://doi.org/10.1088/1742-6596/1196/1/
012022
25. Geng Y, Ji W, Wang Z et al (2019) A review of operating performance in green buildings:
energy use, indoor environmental quality and occupant satisfaction. Energy Build 183:500–514.
https://doi.org/10.1016/j.enbuild.2018.11.017
26. Jalaei F, Jalaei F, Mohammadi S (2020) An integrated BIM-LEED application to automate
sustainable design assessment framework at the conceptual stage of building projects. Sustain
Cities Soc 53.https://doi.org/10.1016/j.scs.2019.101979
27. Mazuroski W, Berger J, Oliveira RCLF, Mendes N (2018) An artificial intelligence-based
method to efficiently bring CFD to building simulation. J Build Perform Simul 11:588–603.
https://doi.org/10.1080/19401493.2017.1414880
28. Kamel E, Javan-Khoshkholgh A, Abumahfouz N et al (2020) A case study of using
multi-functional sensors to predict the indoor air temperature in classrooms. In: ASHRAE
Transactions, pp 3–11
29. Symonds P, Taylor J, Chalabi Z et al (2016) Development of an England-wide indoor over-
heating and air pollution model using artificial neural networks. J Build Perform Simul
9:606–619. https://doi.org/10.1080/19401493.2016.1166265
30. Chen SY (2019) Enhancing validity of green building information modeling with artificial-
neural-network-supervised learning—taking construction of adaptive building envelope based
on daylight simulation as an example. Sens Mater 31:1831–1845. https://doi.org/10.18494/
SAM.2019.2147
31. Chatzikonstantinou I, Sariyildiz S (2016) Approximation of simulation-derived visual comfort
indicators in office spaces: a comparative study in machine learning. Archit Sci Rev 59:307–
322. https://doi.org/10.1080/00038628.2015.1072705
The Effect of P-Delta and P-Delta Plus
Large Displacements Modelling
on Lateral and Axial Displacement
Jonie Tanijaya and Robby S. Kwandou
Abstract The structural analysis programs have become increasingly sophisticated

along with the advancement of science. However, engineers need to be more careful
and understand the use of the options in the program. One program that can incorpo-
rate non-linear effects is SAP2000. In the SAP2000 program, non-linear modeling
options are classified into two options, namely non-linear and non-linear plus large
displacements. The SAP2000 program is a finite element-based program. The finite
element method is a method that solves problems by dividing a large element into
several small element segments or commonly referred to as meshing. The smaller the
meshing segment used, the more accurate the output will be. However, its significance
remains to be studied further. Therefore, this study was conducted to determine the
significance of differences in modeling options (linear, non-linear, non-linear plus
large displacements), as well as the effect of segment division in the analysis. The
analysis results show that the linear model cannot capture the effect of the lateral
displacement that changes due to the incremental of the compression force. The
compression force has significant effect to moment value especially for the higher
compression force. Therefore, the P-Delta effect should be analyzed carefully, espe-
cially for element with high compression force since the linear model could not
capture this effect. The meshing by divide segment does not provide significance
difference for both P-Delta and P-Delta plus large displacement model in this case.
Keywords P-Delta · P-Delta plus large displacements · Lateral displacement
J. Tanijaya (B)
Civil Engineering Department, Universitas Kristen Indonesia Paulus Makassar, Makassar,
Indonesia
R. S. Kwandou
Postgraduate Alumni of Hasanuddin University, Makassar, Indonesia
12 J. Tanijaya and R. S. Kwandou
1 Introduction
The structural analysis programs have become increasingly sophisticated along with
the advancement of science. However, engineers need to be more careful and under-
stand the use of the options in the program. The assumptions of modelling need to
be considered carefully. The effect of lateral load on horizontal displacement should
be checked in analysis. The elastic-linear analysis condition can produce an under-
estimated result, especially in the case of structural elements that experience large
deformation, such as thin glass panel structures and mono-poles that are exposed to
the wind load. A structural analysis that can take into account the large deforma-
tions effect will certainly provide a more accurate result. When using linear elastic
analysis, the effect of changes in geometry due to large deflection cannot be taken
into account in the analysis [1]. One program that can incorporate non-linear effects
is SAP2000. In the SAP2000 program, non-linear modeling options are classified
into two options, namely non-linear and non-linear plus large displacements. The
SAP2000 program is a finite element-based program. The finite element method is a
method that solves problems by dividing a large element into several small element
segments or commonly referred to as meshing. The smaller the meshing segment
used, the more accurate the output will be. Different output values would be provided
when including the meshing effect. However, its significance remains to be studied
further. Therefore, this study was conducted to determine the significance of differ-
ences in modeling options (linear, non-linear, non-linear plus large displacements),
as well as the effect of segment division in the analysis.
2 Literature Review
The P-Delta effect occurs due to gravity load (P) that produces the increasing of hori-
zontal displacement () [2]. The eccentricity of the gravitational load (P) against
the column axis causes an increase in horizontal displacement when the structure
is subjected to lateral loads. Thus, the P-Delta effect will cause an increase in the
moment value as well as the structural drift [3]. Indonesia is surrounded by active
tectonic faults so that the lateral load due to earthquakes needs to be considered
[4]. The lateral force due to earthquake loads which is influenced by the P-Delta
effect becomes critical, especially when the P-Delta effect needs to be consid-
ered in the analysis. Based on the research results of Isitono and Ramadhan, the
increasing moment value is about 10% on the building structure model [2]. The
average percentage increase in moment value that occurs due to P-Delta is between
19 and 27% based on the results of Suhana and Pello’s research on a 15-story building
[5].
P-Delta effect as one type of geometric nonlinearity, typically produce relatively
small additional displacement due to large external force. If deformations become
The Effect of P-Delta and P-Delta Plus Large … 13
Fig. 1 P-Delta on column
sufficiently large as to break from linear compatibility relationships, then Large-

Displacement analyses may become necessary [6]. Typically, there are two P-Delta
effect, namely P- or P-big delta and P-δ or P-small delta (see Fig. 1). P-δ is asso-
ciated with local deformation between end nodes while P- is associated with the
member ends displacements. The difference between P-Delta and P-Delta plus large
displacements is that the P-Delta option only capture P- effect while P-Delta plus
large displacements captures both P- and P-δ.
The geometric stiffness matrix is used in SAP2000 for nonlinear analysis [7].
The use of the geometric stiffness matrix is a general approach to include secondary
effects in the static and dynamic analysis of all types of structural systems. Geometric
stiffness in this case is also refer to geometrical nonlinearity. Secondary order effect
is not presented in linear analysis results [8]. Hence, the nonlinearity option should
be activated to account this effect in analysis. In civil structural engineering, this
geometric stiffness is also commonly referred to P-Delta Analysis. The additional
moment resulting from the P-Delta effect is called the secondary moment or second-
order moment (see Fig. 2).
The elastic force deformation relationship for a prismatic bending element without
shearing deformations is given by the following equation:
FE = k E v
⎡ ⎤ ⎡ ⎤⎡ ⎤
Fi 12 6L −12 6L vi
⎢ Mi ⎥ ⎢ ⎥⎢
⎢ ⎥ E I ⎢ 6L 4L 2
−6L −2L 2 ⎥⎢ i ⎥
⎥
⎣ Fj ⎦ = L3 ⎢ ⎣ −12 −6L 12 −6L
⎥⎣
⎦ vj ⎦
(1)
Mj −6L −2L −6L 4L 2
2 j
The force–displacement relationship of bending element is given by the following

equation:
Fig. 2 Secondary moment
FG = k G v
⎡ ⎤ ⎡ ⎤⎡ ⎤
Fi 36 3L −36 3L vi
⎢ Mi ⎥ ⎢ 3L 4L 2 ⎥⎢
⎢ ⎥ T ⎢ −3L −L 2 ⎥⎢ i ⎥
⎥
⎣ F j ⎦ = 30L ⎢
⎣ −36 −3L 36 −3L
⎥⎣
⎦ vj ⎦
(2)
Mj 3L −L 2 −3L 4L 2 j
Therefore, the total forces acting on the bending element will be the summary of
linear and nonlinear force:
FT = FE + FG = [kE + kG ]v = kT v (3)
P-Delta effect is correlated with compression axial force. For concrete structures
as stated in [9, 10], the compression force should be taken less than 0.75 critical
buckling load (0.75Pcr). The theoretical value of Pcr is given by the formula:
π2E I
Pcr = (4)
(k L)2
π2E I
For k = 2, Pcr = (5)
4L 2
Equation (5) is used for cantilevered element where the effective length factor k
= 2 [11]. In calculating the critical buckling load, effective stiffness (EI)eff is used
for reasonable approximation due to cracking and nonlinearity effect of concrete.
Effective stiffness is presented in [10], which is given by the formula:
Ec I
(E I )e f f = (6)
1 + βns
The inertia I in Eq. (5) for column element is expressed as:

Ast Mu Pu
0.35Ig ≤ I = 0.80 + 25 1− − 0.5 Ig ≤ 0.875Ig (7)
Ag Pu h Po
where, Ec = concrete modulus (MPa), Ig = moment of inertia of gross concrete

section (mm4 ), L = span length (mm), βns = maximum factored axial sustained
(dead) load/maximum factored axial total load, Ast = total area of non-prestressed
longitudinal reinforcement (mm2 ), Ag = gross area of concrete section (mm2 ), Mu
= factored moment (N mm), Pu = factored axial load (N), h = depth of member
(mm), Po = nominal axial strength at zero eccentricity (N).
3 Methodology
The analysis model is a single cantilevered column with a height of 5 m with a free-
fixed condition (see Fig. 3). A simple model is used in this case in order to speed
up the analysis process and also to review different options that would take a large
amount of time process if a more complex model is used. The model is analyzed using
SAP2000 program. There are three analysis options used, which is linear, P-Delta and
P-Delta plus large displacements. The compression force (P) and horizontal force (F)
are applied to the column for all models. In this case, the horizontal force F = 1 ton is
applied as the same final horizontal load while the final compression force is varied
for each model. SNI 2847:2019 [9] article 18.6 states that the beam is an element with
a compression force that does not exceed 0.10Agf’c while the it exceeds 0.10Agf’c
for column element. For beam element, the compression force can be neglected due
to its small value (less than 0.10Agf’c). Therefore, the minimum compression force
Fig. 3 Analysis model

Table 1 Parameter model

Item Units Value
Concrete compression strength f’c MPa 25.00
Concrete modulus Ec MPa 23,500.00
Column dimension mm 400 × 400
Concrete gross area Ag mm2 160,000
Column reinforcement Ast mm2 3402.34 (12D19)
Ratio Ast /Ag % 2.13
Column length L mm 5000.00
Gross stiffness = (EI)g N.mm2 5.0133E+13
Effective stiffness calculated N.mm2 4.2178E+13
(EI)eff = 0.84(EI)g
Hand calculation for critical axial buckling load Pcr using (EI)eff N 4,162,804.39
ton 424.49
in this case is taken as 0.10Agf’c = 40 ton to match the column definition in SNI
2847:2019.
The parameter used in model can be seen in Table 1. The critical buckling load Pcr
needs to be calculated by trial– error method to match the limit Pu = 0.75Pcr since the
effective stiffness value is varied due to the change in compression force and moment
value. The model is then analyzed using buckling analysis in SAP2000 to validate
the critical axial buckling load Pcr between hand calculation with model analysis. It
can be seen that the critical buckling load produce almost the same value with the
difference is smaller than 1% (see Table 2). Hence, it has no significance difference
between hand calculation with model analysis. For Pcr = 422.50 ton, 0.75 Pcr is
316.87 ton. Therefore, the P applied in the model should be taken less than 316.87
ton. The compression forces and horizontal force applied in model can be seen in
Table 3. The effective stiffness for P1, P2, P3 and P4 is calculated using Eq. (6). P2,
P3 and P4 has the same effective stiffness since it is limited by 0.875 Ig as given in
Eq. (7).
Table 2 Buckling analysis result

Stiffness Buckling factor Pinitial (ton) Pcr model = Pcr using hand Pcr model /Pcr
calculate output buckling factor calculation (ton)
✕ P (ton)
Using (EI)eff 422.50 1.00 422.50 424.49 0.9953
Table 3 Force and stiffness applied in model

Notation Compression Horizontal Correlation Correlation Effective
force P (ton) force F (ton) with Ag f’c with Pcr stiffness (EI)eff
P1 40.00 1.00 0.10Ag f’c 0.09Pcr 0.852(EI)g
P2 120.00 1.00 0.30Ag f’c 0.28Pcr 0.875(EI)g
P3 200.00 1.00 0.50Ag f’c 0.47Pcr 0.875(EI)g
P4 280.00 1.00 0.70Ag f’c 0.66Pcr 0.875(EI)g
Nonlinear result (P-Delta or P-Delta plus large displacement) refers to as NL while

the linear result refers to as LIN (see Figs. 4 and 5). The lateral displacement value
for linear case has the same value between P2, P3 and P4 due to its similar effective
stiffness value. It indicates that the lateral displacement in linear analysis is not
affected by different compression forces. Figures 4 and 5 also show that the curve
line is formed as the increasing of compression force due to the effect of geometric
nonlinearity while the straight line is formed for linear case. It can also be seen that
the same moment value is obtained for all models in linear analysis (see Table 4). This
indicates that the moment value is only provided by the horizontal force acting on
the top column in linear case, while the compression force does not give contribution
to it.
The analysis results show that the linear model cannot capture the effect of the
lateral displacement that changes due to the incremental of the compression force.
P-Delta vs Linear Analysis

30
25
Lateral Displacement (mm)
20
15
10
0
0 0.2 0.4 0.6 0.8 1 1.2
Horizontal Force (ton)
NL-P1 NL-P2 NL-P3 NL-P4

LIN-P1 LIN-P2 LIN-P3 LIN-P4
Fig. 4 Lateral displacement for P-Delta model

P-Delta plus Large Displacement vs Linear Analysis

30
25
20
15
10
0
0 0.2 0.4 0.6 0.8 1 1.2

Fig. 5 Lateral displacement for P-Delta plus large displacement model
The increasing of lateral displacement due to larger compression force can only be
seen when activating geometric nonlinearity option for both compression and lateral
force. Table 4 also shows that the second-order moment increases as the increasing
of compression force. It can also be seen that the moment for nonlinear case is about
1.1–2.4 times the moment value for linear case. This concludes that the compression
force has significant effect to moment value especially for the higher compression
force. In this case, the moment value of P-Delta plus large displacement is larger
than the moment value of P-Delta. However, the difference is smaller than 3%. The
larger moment of P-Delta plus large displacement comparing to P-Delta indicates
that the large displacement effect takes account to the increasing of moment value.
Table 5 shows that there is no significant difference between lateral displace-
ment of P-Delta plus large displacement with P-Delta model (smaller than 0.10%).
However, the difference of lateral displacement is quite significance when comparing
between nonlinear with linear model. The difference ratio of lateral displacement
when comparing P-Delta plus large displacement to linear case is about 1.10–2.71.
Therefore, the P-Delta effect should be analyzed carefully, especially for element
with high compression force since the linear model could not capture this effect.
Table 6 shows that axial displacement between linear and nonlinear model (P-Delta
and P-Delta plus large displacement) has no significant difference (smaller than
3%). This indicates that P-Delta effect only affect the lateral displacement but has
no significant effect for the axial displacement case.
Figures 6 and 7 shows the axial displacement value for the same incremental
horizontal force with variation in compression force. It can be seen that the graph
for nonlinear model for both P-Delta and P-Delta plus large displacement has almost
coincides with the gradient line for linear case. However, it is quite different for the
Table 4 Second-order moment for P-Delta dan P-Delta plus large displacement model
Notation Mfirst-order (t m) P-Delta P-Delta plus large displacement
Mnonlinear (t m) Msecond-order (t m) Mnonlinear /Mfirst order Mnonlinear (t m) Msecond-order (t m) Mnonlinear /Mfirst order
The Effect of P-Delta and P-Delta Plus Large …
P1 5.000 5.433 0.433 1.0866 5.433 0.433 1.0866

P2 5.000 6.544 1.544 1.3088 6.550 1.550 1.3100
P3 5.000 8.376 3.376 1.6752 8.408 3.408 1.6816
P4 5.000 11.859 6.859 2.3718 12.115 7.115 2.4230
19
Table 5 Lateral displacement difference between P-Delta and P-Delta plus large displacement
Notation Lateral displacement (mm) Difference ratio
Linear P-Delta P-Delta plus large P-Delta plus large P-Delta plus large
displacement disp/P-Delta disp/linear
P1 9.604 10.604 10.598 0.9994 1.1035
P2 9.352 12.836 12.825 0.9992 1.3714
P3 9.352 17.039 17.023 0.9991 1.8202
P4 9.352 25.329 25.354 1.0010 2.7110
Table 6 Axial displacement difference between P-Delta and P-Delta plus large displacement
Notation Axial displacement (mm) Difference ratio
Linear P-Delta P-Delta plus large P-Delta plus large P-Delta plus large
displacement disp/P-Delta disp/linear
P1 0.534 0.534 0.545 1.0209 1.0209
P2 1.577 1.579 1.594 1.0095 1.0107
P3 2.621 2.635 2.653 1.0071 1.0125
P4 3.664 3.664 3.732 1.0185 1.0185

4
3.5
3
Axial Displacement (mm)
2.5
1.5
0.5
0
0 0.2 0.4 0.6 0.8 1 1.2

Fig. 6 Axial displacement for P-Delta model
lateral displacement graph (see again for Figs. 4 and 5). This confirms that P-Delta
effect has significant effect to lateral displacement but not significant for the axial

4
3.5
3
Axial Displacement (mm)
2.5
1.5
0.5
0
0 0.2 0.4 0.6 0.8 1 1.2

Fig. 7 Axial displacement for P-Delta plus large displacement model
displacement when comparing between nonlinear and linear model as we can see for
each load step has no significant difference.
As stated before, the smaller meshing used in finite element program could
produce different output, but it needs to be studied further in this case. To observe the
meshing effect, the single line model is then divided into 5 segments with a length
per segment of 1 m for the both P-Delta and P-Delta plus large displacement model.
Figures 8 and 9 shows that the meshing by divide segment does not provide signifi-
cance difference for both P-Delta and P-Delta plus large displacement model in this
case. The marking point (data for meshing by divide frame) almost coincides with
the line (data for no meshing frame) for each load step. The difference ratio is smaller
than 1% for top lateral displacement (see Table 7). Hence, it can be concluded that
the meshing option has no significant effect comparing to no mesh option in this
case.
Figures 10 and 11 produce a different curve due to the difference compression
force versus horizontal force applied in each load step. As stated before, the same
final horizontal force 1 ton is applied while the final compression force is different for
P1, P2, P3 and P4. The graph lines show that the larger the value of the compression
force, the more visible the curve line due to the nonlinear effect. When the small
compression force is applied (for P1 case), the line formed tends to be almost straight
(linear). Then, the curve line takes forms as the increasing of compression force. It
can also be seen that for linear model, the lines formed tend to coincide with each
other even when applied with different compression force. This concludes that the
lateral displacement is not affected by the compression force in linear model analysis.
P-Delta Analysis
(Mesh vs No Mesh)
30
25
20
15
10
0
0 0.2 0.4 0.6 0.8 1 1.2
P1 (No Mesh) P2 (No Mesh) P3 (No Mesh) P4 (No Mesh)

P1 (Mesh Frame 1 m) P2 (Mesh Frame 1 m) P3 (Mesh Frame 1 m) P4 (Mesh Frame 1 m)
Fig. 8 Meshing effect for P-Delta model
P-Delta plus Large Displacement Analysis

(Mesh vs No Mesh)
30
25
20
15
10
0
0 0.2 0.4 0.6 0.8 1 1.2
P1 (No Mesh) P2 (No Mesh) P3 (No Mesh) P4 (No Mesh)

P1 (Mesh Frame 1 m) P2 (Mesh Frame 1 m) P3 (Mesh Frame 1 m) P4 (Mesh Frame 1 m)
Fig. 9 Meshing effect for P-Delta plus large displacement model

Table 7 Meshing effect for P-Delta and P-Delta plus large displacement
Lateral displacement for P-Delta model Lateral displacement for P-Delta plus
(mm) large displacement (mm)
Notation No mesh Mesh by No No mesh Mesh by No
(mm) divide mesh/mesh (mm) divide mesh/mesh
segment segment
(mm) (mm)
P1 10.604 10.595 1.0009 10.598 10.595 1.0003
P2 12.836 12.820 1.0012 12.825 12.827 0.9999
P3 17.039 16.991 1.0028 17.023 17.046 0.9986
P4 25.329 25.072 1.0103 25.354 25.525 0.9933

300
250
Compression Force (ton)
200
150
100
50
0
0 5 10 15 20 25 30

Fig. 10 P-Delta model (compression force versus lateral displacement)
The meshing model by divide frame for 1 m is used to observe the P-Delta plus
large displacement effect for local bending moment at each joint (see Tables 8 and 9).
It can be seen that the bending moment at each local joint increase as the increasing
of compression force. Hence, the P-Delta plus large displacement effect tends to have
more effect for the larger compression force. However, the difference is not significant
in this case. The largest difference is about 2.65% for the largest compression force
applied (P4), while it has the same value for the smallest compression force applied
(P1). It can also be seen that each local joint has different moment ratio, this indicates
that the P-Delta plus large displacement effect is taking account for local deformation.

300
250
Compression Force (ton)
200
150
100
50
0
0 5 10 15 20 25 30

Fig. 11 P-Delta plus large displacement model (compression force versus lateral displacement)
Table 8 Local moment at joint for P1 and P2

P1 P2
Distance Linear M P-Delta M P-Delta Ratio a/b P-Delta M P-Delta Ratio a/b
from (t m) (t m) (a) plus large (t m) (a) plus large
support disp M disp M
(t m) (b) (t m) (b)
0m 5.000 5.431 5.431 1.0000 6.541 6.548 1.0011
(base)
1m 4.000 4.406 4.406 1.0000 5.457 5.463 1.0011
2m 3.000 3.340 3.340 1.0000 4.225 4.230 1.0012
3m 2.000 2.243 2.243 1.0000 2.880 2.884 1.0014
4m 1.000 1.126 1.126 1.0000 1.459 1.461 1.0014
5 m (top) 0.000 0.000 0.000 1.0000 0.000 0.000 1.0000
5 Conclusion
1. The lateral displacement in linear analysis is not affected by different compres-

sion forces. It can also be seen that the same moment value is obtained for
all models in linear analysis. This indicates that the moment value is only
provided by the horizontal force acting on the top column in linear case, while
the compression force does not give contribution to it.
Table 9 Local moment at joint for P3 and P4

P3 P4
Distance Linear M P-Delta M P-Delta Ratio a/b P-Delta M P-Delta Ratio a/b
from (t m) (t m) (a) plus large (t m) (a) plus large
support disp M disp M
(t m) (b) (t m) (b)
0m 5.000 8.380 8.419 1.0047 11.903 12.166 1.0221
(base)
1m 4.000 7.199 7.236 1.0051 10.543 10.793 1.0237
2m 3.000 5.697 5.729 1.0056 8.531 8.744 1.0250
3m 2.000 3.943 3.966 1.0058 5.995 6.149 1.0257
4m 1.000 2.016 2.028 1.0060 3.091 3.173 1.0265
5 m (top) 0.000 0.000 0.000 1.0000 0.000 0.000 1.0000
2. The analysis results show that the linear model cannot capture the effect of
the lateral displacement that changes due to the incremental of the compres-
sion force. The increasing of lateral displacement due to larger compression
force can only be seen when activating geometric nonlinearity option for both
compression and lateral force.
3. The moment for nonlinear case is about 1.1–2.4 times the moment value for
linear case. This concludes that the compression force has significant effect to
moment value especially for the higher compression force.
4. The moment value of P-Delta plus large displacement is larger than the moment
value of P-Delta. However, the difference is smaller than 3%. The larger moment
of P-Delta plus large displacement comparing to P-Delta indicates that the large
displacement effect takes account to the increasing of moment value.
5. There is no significant difference between lateral displacement of P-Delta
plus large displacement with P-Delta model (smaller than 0.10%). However,
the difference of lateral displacement is quite significance when comparing
between nonlinear with linear model. The difference ratio of lateral displace-
ment when comparing P-Delta plus large displacement to linear case is about
1.10–2.71. Therefore, the P-Delta effect should be analyzed carefully, espe-
cially for element with high compression force since the linear model could not
capture this effect.
6. The axial displacement between linear and nonlinear model (P-Delta and P-
Delta plus large displacement) has no significant difference (smaller than 3%).
This indicates that P-Delta effect only affect the lateral displacement but has no
significant effect for the axial displacement case.
7. The meshing by divide segment does not provide significance difference for both
P-Delta and P-Delta plus large displacement model in this case. The difference
ratio is smaller than 1%.
8. The bending moment at each local joint increase as the increasing of compres-
sion force. Hence, the P-Delta plus large displacement effect tends to have more
effect for the larger compression force. However, the difference is not significant
in this case. The largest difference is about 2.65% for the largest compression
force applied (P4), while it has the same value for the smallest compression
force applied (P1). It can also be seen that each local joint has different moment
ratio, this indicates that the P-Delta plus large displacement effect is taking
account for local deformation.
References
1. Dewobroto W, Chendrawan W (2011) Effect of modeling and analysis on behavior of glass

structures against wind pressure. J Nat Seminar-1 BMPTTSSI—KoNTekS 5 USU, Medan
2. Istiono H, Ramadhan AY (2020) Analysis of the P-Delta effect due to the height difference
in earthquake resistant buildings (Case Study: Non-Highrise Building). J REKAYASA SIPIL
14(3). ISSN 1978-5658
3. Avinash T, Pandian G (2017) Investigation of the effects of P-Delta on tubular tall buildings. J
IJCIET 8:487–495
4. Adityawarman G (2014) Planning of evacuation buildings in earthquake and tsunami prone
areas. J Kajian Teknologi 10(2)
5. Suhana N, Pello O (2015) P-Delta effect analysis on fifteen stories building due to the increasing
of helipad load. J Rekayasa Infrastruktur 1(2):44–105. ISSN 2460-335X
6. CSI technical knowledge base. https://wiki.csiamerica.com/display/kb/P-Delta+effect. Last
accessed January 2021
7. Wilson EL (2002) Three-dimensional static and dynamic analysis of structures, 3rd edn.
Computers and Structures Inc., California
8. SCIA support. https://www.scia.net/en/support/faq/calculation-non-linear-dynamics/second-
order-analysis-questions-part-2. Last accessed May 2021
9. National Standardization Agency (2019) SNI 2847:2019, Structural Concrete Requirements
for Buildings with Commentary. BSN, Jakarta (2019).
10. Computers & Structures Inc (2019) Concrete frame design manual ACI 318-14 SAP2000.
Computers and Structures, Inc., California
11. Computers & Structures Inc (2017) CSI analysis reference manual for SAP2000. ETABS,
SAFE and CSIBridge, California
Effect of Regulatory Change
in Earthquake Load Analysis
on Structures with Irregular Shapes
Hendramawat Aski Safarizki and Dini Ayu Saputri
Abstract The effect of the earthquake must be reviewed in the design of the building
structure. The development of ASCE 7-10 and ASCE 7-16 regulations was followed
by the development of earthquake-resistant building regulations in Indonesia, namely
SNI 1726-2019. The developments of Indonesia SNI 1726-2019 was accompanied by
the publication of the Indonesian Spectra Design 2021, which replaced the Indonesian
Spectra Design 2011. The development of this regulation has raised concerns that
the existing buildings designed by referring the previous regulation would behave
differently when they were analyzed with the latest SNI 1726-2019 regulations. In
this study, a response spectrum analysis was carried out based on the Indonesian
Spectra Design 2021 for buildings that were designed according to the Indonesian
Spectra Design 2011. The building under study has an L shape with a difference in
length on the side of the X direction and the Y direction. The shape of a building falls
into a category of high irregularity. The results of this analysis indicate that there is
a change in the level of safety in the structure. It is necessary to conduct a study on
the planned building before the implementation of the SNI 1726-2019 regulations
which are accompanied by Indonesian Spectra Design 2021.
Keywords Earthquake · Regulation · Response spectra
1 Introduction
Indonesia is an archipelago whose geographic position is on the most active earth-

quake route in the world because it is surrounded by the Pacific Ring of Fire (Ring
of Fire) which is above three collisions of continental plates, namely the Indo-
Australian, Eurasian and Pacific plates. This geographical condition makes Indonesia
a region prone to natural disasters such as volcanic eruptions, earthquakes, and
tsunamis.
H. A. Safarizki (B) · D. A. Saputri

Universitas Veteran Bangun Nusantara, Sukoharjo, Indonesia
28 H. A. Safarizki and D. A. Saputri
The basic shape of the structure in a building generally has to contribute to resisting
the lateral forces caused by earthquakes. The magnitude of the earthquake load is
greatly influenced by the condition of the building structure such as the building
plan and the mass of the building itself [1]. Recent research states that the higher
the horizontal irregularity, the higher the shear force and max drift [2–4]. Based on
research on a building with an L plan, there will be a critical area located at the back
end of the building [5].
The effect of the earthquake must be reviewed in the planning of the building struc-
ture. The structure as a whole must remain standing even though it is on the verge of
collapse in the event of an earthquake. In general, structural analysis of earthquakes
is divided into 2 types, namely equivalent static analysis, namely in the form of hori-
zontal force (Px, Py) applied to the floor of the structure, and dynamic analysis (time
history and response spectrum) in the form of waves based on previous earthquake
data. Which is applied to the base structure, and analyzed under non-linear condi-
tions. Spectrum response analysis has been widely applied in earthquake-resistant
structure analysis [6–12]. The dynamic analysis aims to determine the distribution
of the level shear forces due to ground motion by earthquakes and can be done by
analyzing the response spectrum [13].
The development of ASCE 7-10 and ASCE 7-16 regulations was followed by
the development of earthquake-resistant building regulations in Indonesia, namely
SNI 1726-2019. On developments in Indonesia SNI 1726-2019 was followed by the
publication of the Indonesian Spectra Design in 2021 which replaced the Indone-
sian Spectra Design in 2011. The development of this regulation was followed by
concerns that buildings designed before this regulation were implemented would
behave differently when they were tested with the latest SNI 1726-2019 regulations.
In this study, a response spectrum analysis was carried out based on the Indonesia
Spectra Design 2021 against the planned building using the Indonesian Spectra
Design 2011. It is hoped that the results of this study can illustrate changes in the
behavior of earthquake-resistant building structures with the implementation of SNI
1726-2019 followed by the 2021 Spectra Design.
2 Methods
2.1 Structure Plan
The structure reviewed in this research is an L-shaped 12-story building in Surakarta

on moderate soil conditions which is planned as a high-rise building. The structure
consists of 12 floors with a total height of floors 1 to 12 is ±38.10 m.
The loading of the building can be determined by adjusting the use of the building,
that is, the function of the room in each different building is determined based
on the determination of the load following SNI 03-1727-2013 Minimum Load for
Effect of Regulatory Change in Earthquake Load Analysis … 29
Fig. 1 Typical floor plan 1–12
Designing Buildings and Other Structures and SNI 03-2847-2013 Concrete Require-
ments Structural for Buildings. Effect of an earthquake and earthquake-resistant
planning for building structures following the Guidelines for Earthquake Resistance
Planning based on SNI 03-1726-2012 with Indonesian Spectra 2011 and SNI 03-
1726-2019 with Indonesian Spectra Design 2021. The plan of the building being
reviewed appears in Fig. 1.
The building under study has an L shape with a difference in length on the side
of the X direction and the Y direction. The shape of a building like this has a high
irregularity. Irregularity considerably affects the seismic and the seismic response of
the structures [3].
2.2 Structural Modeling
Structural modeling using SAP 2000 software is used to facilitate analysis of the
building structure under study. In the model created, material specifications and other
data are also applied to calculate the load on the building. The planned building struc-
ture is an L-shaped 12-story apartment building with a building height of ±38.10 m
(Fig. 1).
Fig. 2. 3D modeling of
building structures
2.3 Load on Structure
The planning of loading is guided by SNI 03-1727-2013 concerning the Planning of

Loads for Houses and Buildings. The load calculation includes dead load, live load,
and wall load. Based on SNI 03-1726-2012, the building risk category is included
in group II with the earthquake priority factor (Ie) of 1.0. Based on the acceleration
response parameter in the short period (S) of 0.602 with the class II risk category
being the seismic design category D (KDS D). The seismic design category is based
on the acceleration response parameter in the 1 s (S) period of 0.570 with the class
II risk category being the seismic design category D (KDS D). So that the seismic
design category based on SDS, SD1, and risk category II is included in KDS D.
Based on SNI 03-1726-2019, the building risk category is included in group II
with the earthquake priority factor (Ie) of 1.0. Based on the acceleration response
parameter in the short period (S) of 0.672 with the class II risk category being the
seismic design category D (KDS D). The seismic design category based on the
acceleration response parameter in the 1 s period (S) is 0.633 with the class II risk
category being the seismic design category D (KDS D). So that the seismic design
category based on SDS, SD1, and risk category II is included in KDS D. Table 1
describes the differences in spectra design from the Indonesian Spectra Design in
2011 and 2021. If Table 1 is then made in the form of a response spectrum curve, it
will look like in Fig. 3. In this study did not consider the torque magnification factor.
Site Coefficients and Risk-Targeted Maximum Considered Earthquake (MCER)
Spectral Response Acceleration Parameters for short periods (SMS) and at 1 s (SM1),
adjusted for Site Class effects, shall be determined by Eqs. 11.4.1 and 11.4.2 [14],
respectively.
SMS = FaSS (1)

Table 1 The difference in

Variable Value
value between Indonesian
Spectra Design 2011 and Indonesian Spectra Indonesian Spectra
Indonesian Spectra Design Design 2011 Design 2021
2021 PGA (g) 0.3560 0.3503
SS (g) 0.7360 0.8009
S1 (g) 0.3090 0.3883
FA 1.2270 1.2593
FV 2.7630 2.4468
SMS (g) 0.9040 1.0086
SM1 (g) 0.8540 0.9501
Fig. 3 Indonesian Spectra Design 2011 and 2021 response spectrum curve
SM1 = FvS1 (2)
where:
SS = the mapped MCER spectral response acceleration parameter at short periods
as determined by Section 11.4.1 ASCE 7-10, and
S1 = the mapped MCER spectral response acceleration parameter at a period of
1 s as determined by Section 11.4.1 ASCE 7-10
3 Results
3.1 Based on Regulations SNI 03-1726-2012
The deviation control between design floors () shall be calculated as the difference
in deflection at the center of mass at the top and bottom levels under review. Based on
SNI 03-1726-2012, the allowable deviation between floors for risk category II (other
structures) for this structure is 0.020 hsx. This type of structure is not a structure that
uses shear walls.
a = 0.020h x
= 0.020 × 3100
= 62 mm → 0.062 m
From Table 9 on SNI 03-1726-2012, the Cd factor in the earthquake resistance

system, the designed structure is a special moment-bearing reinforced concrete frame
system, the value of Cd = 5.5. Calculating the deviation that occurs, take floor 12
joint number 6233 and floor 11 joint number 5936 viewed in the X-direction. Points
6233 and 5936 are the same point on different floors. Number joint 6233 δ2 =
0.030.153 m, number joint 5936 δ1 = 0.029.492 m. The complete calculation results
for all floors are in Table 2. The control of the maximum deviation between floors
for the X direction based on calculating using SNI 03-1726-2012 is 0.00847 m <
0.062 m (the allowable deviation limit between floors) and the drift ratio on the top
floor (0.12%) which is calculated based on the deviation shows that the drift ratio
that occurs is still smaller than the drift. The maximum required in SNI 1726-2019
is 2% so that it still meets the requirements.
Table 2 The deviation between the floors of the design level () in the X direction based on
calculating using SNI 03-1726-2012
Floor Height (m) No. joint Deflection/δ (m) Deviation/ (m) Drift ratio (%)
12 3.1 6233 0.030.153 0.003.635 0.12
11 3.1 5936 0.029.492 0.005.797 0.19
10 3.1 5639 0.028.438 0.008.201 0.26
09 3.1 5342 0.026.947 0.010.516 0.34
08 3.1 5045 0.025.035 0.012.678 0.41
07 3.1 4748 0.022.730 0.014.647 0.47
06 3.1 4451 0.020.067 0.016.423 0.53
05 3.1 4154 0.017.081 0.017.969 0.58
04 3.1 3857 0.013.814 0.019.239 0.62
03 3.1 3560 0.010.316 0.020.026 0.65
02 3.1 3263 0.006.675 0.036.713 1.18
01 4.0 33 0 0 0
Table 3 The deviation between the floors of the design level () in the X direction based on
calculation using SNI-03-1726-2019
Floor Height (m) No. joint Deflection/δ (m) Deviation/ (m) Drift ratio (%)
12 3.1 6233 0.062.562 0.0.084.700 0.27
11 3.1 5936 0.061.022 0.0.126.775 0.41
10 3.1 5639 0.058717 0.0.175.835 0.57
09 3.1 5342 0.055520 0.0.224.180 0.72
08 3.1 5045 0.051444 0.0.269.500 0.87
07 3.1 4748 0.046544 0.0.311.080 1.00
06 3.1 4451 0.040888 0.0.348.260 1.12
05 3.1 4154 0.034556 0.0.380.325 1.23
04 3.1 3857 0.027641 0.0.404.690 1.31
03 3.1 3560 0.020283 0.0.414.700 1.34
02 3.1 3263 0.012743 0.0.700.865 2.26
01 4.0 33 0 0 0
3.2 Based on Regulations SNI 03-1726-2019
From Table 12 (page 50) on SNI 03-1726-2019, the Cd factor for earthquake resis-
tance systems, the designed structure is a special moment-bearing reinforced concrete
frame system, the value of Cd = 5.5. Calculating the deviation that occurs as shown
in Table 3, take floor 12 joint number 6233 and floor 11 joint number 5936 viewed
in the X-direction. Points 6233 and 5936 are the same point on different floors. The
control of the maximum deviation between floors for the X direction is 0.00847 m <
0.062 m (the allowable deviation limit between floors) and the drift ratio on the 2nd
floor (2.26%) which is calculated based on the deviation shows that the drift ratio
that occurs is greater than the maximum drift which is required in SNI 1726-2019
which is 2% so it does not meet the requirements.
4 Discussion and Conclusions
If Tables 2 and 3 are made in the form of a graph of the relationship between the drift
ratio analysis results with SNI 1726-2012 and SNI 1726-2019, it will appear that
there is an increase in the deviation of each floor (Fig. 4). This increase in deviation
causes the drift ratio of each floor to increase above 90% if the analysis results with
SNI 1726-2019 are compared with the planning using SNI 1726-2012.
The results of the calculation of the critical area as reviewed by SNI 1726-2012
are located at the back end of the building with maximum control of the deviation
between floors for the X direction of 0.036713 m < 0.062 m (allowable deviation
between floors) and has a drift ratio of 0.12% < 2% (maximum drift required in SNI
1726-2012) indicates that the building structure has met the required limits.
Fig. 4 Comparison of drift ratio values
The results of the calculation of the critical area as reviewed by SNI 1726-2019
are located on the 2nd floor with a maximum control of the deviation between floors
for the X direction of 0.00847 m < 0.062 m (allowable deviation between floors)
and has a drift ratio of 2.26% > 2% (maximum drift required in SNI 1726-2019).
This indicates that the building structure does not meet the limits set out in SNI
1726-2019.
The results of this analysis indicate that there is a change in the level of safety
in the structure. It is necessary to conduct a study on the planned building before
the implementation of the SNI 1726-2019 regulations which are accompanied by
Indonesian Spectra Design 2021.
References
1. Batu ML, Dapas SO, Wallah SE (2016) Efisiensi Penggunaan Dinding Geser untuk Mereduksi
Efek Torsi pada Bangunan yang Tidak Beraturan. Sipil Statik
2. Prijasambada, Hafifah V (2018) Analisa Gaya Diafragma, Kord Dan Kolektor Pada Bangunan
Gedung Sesuai Dengan Sni 1726:2012. IKRAITH-TEKNOLOGI. 2:41–49
3. Naveen SE, Abraham NM, Kumari ASD (2019) Analysis of irregular structures under
earthquake loads. Procedia Struct Integr 14:806–819. https://doi.org/10.1016/j.prostr.2019.
07.059
4. Ahmed MMM, Abdel Raheem SE, Ahmed MM, Abdel Shafy AGA (2016) Irregularity effects
on the seismic performance of L-shaped multi-story buildings. JES J Eng Sci 44:513–536.
https://doi.org/10.21608/jesaun.2016.111440
5. Saputri DA, Safarizki HA (2021) Perencanaan Struktur Gedung Apartemen 12 Lantai
Berbentuk L di Surakarta. In: Simposium Nasional Teknologi Infrastruktur Abad ke-21,
Yogyakarta, pp 201–206
6. De Domenico D, Falsone G, Ricciardi G (2018) Improved response-spectrum analysis of base-

isolated buildings: a substructure-based response spectrum method. Eng Struct 162:198–212.
https://doi.org/10.1016/j.engstruct.2018.02.037
7. Cacciola P, D’Amico L, Zentner I (2014) New insights in the analysis of the structural response
to response-spectrum-compatible accelerograms. Eng Struct 78:3–16. https://doi.org/10.1016/
j.engstruct.2014.07.015
8. Mostafijur Rahman M, Jadhav SM, Shahrooz BM (2018) Seismic performance of reinforce
concrete buildings designed according to codes in Bangladesh, India and U.S. Eng Struct
160:111–120. https://doi.org/10.1016/j.engstruct.2018.01.010
9. Cacciola P, Zentner I (2012) Generation of response-spectrum-compatible artificial earthquake
accelerograms with random joint timefrequency distributions. Probabilistic Eng Mech 28:52–
58. https://doi.org/10.1016/j.probengmech.2011.08.004
10. Milani G, Valente M (2015) Failure analysis of seven masonry churches severely damaged
during the 2012 Emilia-Romagna (Italy) earthquake: non-linear dynamic analyses versus
conventional static approaches. Eng Fail Anal 54:13–56. https://doi.org/10.1016/j.engfailanal.
2015.03.016
11. Islam ABMS, Jameel M, Ahmad SI, Salman FA, Jummat MZ (2011) Engendering earthquake
response spectra for Dhaka region usable in dynamic analysis of structures. Sci Res Essays
6:3519–3530. https://doi.org/10.5897/sre11.218
12. Hassaballa AE (2013) Seismic analysis of a reinforced concrete building by response spectrum
method. IOSR J Eng 03:01–09. https://doi.org/10.9790/3021-03930109
13. Mulyo A, Utami D, Dinamik A (2012) Metode Respon Spectrum Ditinjau Pada Drift Dan
Displacement Menggunakan Software ETABS (Performance Analysis of Building Struc-
tural with Spectrum Response Method Reviewed on Management and Control Using ETABS
Software). J Infras 4:65–71
14. ASCE (2016) Minimum Design Loads for Buildings and Other Structures (ASCE 7-10)
Seismic Performance of Tall and Slender
Minaret Structure with Hexagonal RC
Wall Section by Means Fragility Curve
Development
Erik Wahyu Pradana , Senot Sangadji ,

and Angga Destya Navara Noor
Abstract Currently, there are 278,629 mosques registered in the information system
of the Indonesian Ministry of Religion. This number is believed to keep growing.
One of them is a mosque that was being built in Surakarta, namely Taman Sriwedari
Mosque. This building is equipped with 114 m height minaret and is claimed to
be the tallest in Indonesia. Considering that Indonesia is located in an earthquake-
prone region, it is important to assess the seismic performance of this structure.
This research aims to evaluate the seismic performance of the minaret using the
fragility function. Fragility function is a function that correlates the damage level
with an intensity measure. The structural model of the minaret was developed using
Seismostruct. Furthermore, the structural response is obtained from the structural
model analyzed using static nonlinear pushover. Based on this analysis result, a
structural capacity curve is obtained which is then used in constructing the fragility
curve. The fragility curve shows that the damage probability of minaret structure
on spectral acceleration, Sa = 0.098 g based on the material strain limits criteria for
damage states DS1, DS2, DS3, and DS4 is 57%, 29%, 19%, and 12% respectively. On
the other hand, based on the maximum base shear criteria for the damage states DS1,
DS2, and DS3 is 50%, 20%, and 12% respectively. Perceiving the fragility curve for
the structure under review, building owners can predict the probability of structural
damage due to other earthquake intensity scenarios and carry out evaluations to
determine retrofitting strategies rationally.
Keywords Seismic performance · Minaret structure · Fragility curve
E. W. Pradana (B) · S. Sangadji · A. D. N. Noor

SMARTQuake Research Group, Civil Engineering Department, Faculty of Engineering,
Universitas Sebelas Maret, Surakarta, Indonesia
S. Sangadji
38 E. W. Pradana et al.
1 Introduction
An earthquake with a magnitude of 9.1 that occurred in Northern Sumatra, Indonesia

in 2004 was the third-largest earthquake in the world, recorded since 1900 [1]. In the
list, 7th and 11th positions are earthquakes that also occurred in Indonesia, namely the
Northern Sumatra earthquake with a magnitude of 8.6 and the Banda Sea earthquake
with a magnitude of 8.5. A number of major earthquakes that happened in Indonesia
followed by large tsunamis, such as the 2004 Aceh; 2005 Nias; 2006 Yogyakarta,
West Java, and West Sumatra; and 2007 Bengkulu and West Sumatra earthquake,
cause loss of life and structural damage [2]. These facts indicate that Indonesia is a
country with a high risk of seismicity.
Most earthquakes in Indonesia are caused by tectonic activity. The relative
movement of tectonic plates causes the collision of the Indo-Australian Plate, the
Eurasian Plate, and the Pacific Plate. This tectonic activity led to the formation of a
volcano series along with the islands of Sumatra, Java-Bali-Nusa Tenggara, northern
Sulawesi-Maluku, to Papua which became known as the ring of fire. In the forearc
where the two tectonic plates meet, many active faults often cause earthquakes.
Currently, there are 278,629 mosques registered in the information system of the
Ministry of Religion in Indonesia [3]. This number is believed to keep growing. One
of them is a mosque that was being built in Surakarta, namely the Taman Sriwedari
Mosque. This building is equipped with five minarets. One of the minarets is 114 m
in height and is claimed to be the tallest in Indonesia. Considering that Indonesia is
located in an earthquake-prone region, it is important to assess the seismic perfor-
mance of the minaret. This research aims to evaluate the seismic performance of the
minaret by developing its structural fragility curve. The fragility function eluci-
dates the probability of exceedance for multiple damage states as a function of
ground-motion intensity [4].
The structural model of the minaret was developed using Seismostruct [5].
Furthermore, the structural response is obtained from the structural model analyzed
using a nonlinear static procedure (pushover analysis). Based on this analysis, a struc-
tural capacity curve is obtained which is then used to construct the fragility curve.
The fragility curve obtained shows the probability of structural damage (exceed the
predetermined damage states) related to the intensity of ground motions during its
service life. The fragility curve shows that the damage probability of minaret struc-
ture on spectral acceleration, Sa = 0.098 g based on the material strain limits criteria
for damage states DS1, DS2, DS3, and DS4 is 57%, 29%, 19%, and 12% respectively.
On the other hand, based on the maximum base shear criteria for the damage states
DS1, DS2, and DS3 is 50%, 20%, and 12% respectively. Perceiving the fragility curve
for the structure under review, building owners can predict the probability of struc-
tural damage due to other earthquake intensity scenarios and carry out evaluations
to determine retrofitting strategies rationally.
Seismic Performance of Tall and Slender Minaret Structure … 39
2 Research Method
In this research, the structural model of the minaret was developed using Seismostruct
(see Fig. 1a). Seismostruct is a finite element-based structural analysis software
that has the capability for predicting the large displacement behavior of 3D frames
under static or dynamic loadings by considering both geometric nonlinearities and
material inelasticity [6]. The load–displacement relationship (capacity curve) of the
modeled structure is obtained from static nonlinear pushover analysis. Furthermore,
the capacity curve obtained is used to determine the fragility curve of the structure
under study.
The minaret structure has been built in 2019 using the slip-form construction
method. The minaret structure consists of two layers (inner and outer) of reinforced
concrete shear walls with a hexagon geometry. The height of the minaret is 114 m.
The material properties of the concrete are as follows: compressive strength f c

is 30 MPa and the modulus of elasticity is taken as 4700 f c . While the material
properties of the reinforcing steel are as follows: yield strength f y is 390 MPa and
the modulus of elasticity is 200,000 MPa. Detailed dimensions and reinforcement of
the minaret structure are presented in Fig. 1b–g.
The live load design of this structure is 4.79 kN/m2 in accordance with SNI
1727:2013 [7]. Meanwhile, the earthquake load is determined based on SNI
1726:2012 [8] with the Peak Ground Acceleration value (Sa) = 0.602 g. This structure
is located in Surakarta with SD (medium soil) site class, the risk-targeted maximum
considered earthquake (MCER) spectral response acceleration at short period (Ss )
= 0.754 g, and at a 1 s period (S1 ) = 0.315 g. Based on the structural model, the
fundamental period of the structure is 3.782 s. Furthermore, this value corresponds
to the spectral acceleration value, (Sa) = 0.098 g, in the Response Spectrum Design
for the Surakarta region based on SNI 1726:2012. The obtained spectral accelera-
tion value is then used to evaluate the seismic performance of the minaret using the
fragility function.
3.1 Capacity Curve and Capacity Spectrum
A capacity curve is a curve that describes the relationship between base-shear and
displacement at the top of the observed structure. In this research, the capacity curve
of the minaret structure is obtained through static nonlinear pushover analysis (see
Fig. 2). Furthermore, this curve is converted into a capacity spectrum in ADRS
(Acceleration Displacement Response Spectrum) format which is used to determine
the median spectral displacement (See Fig. 3). The capacity spectrum describes the
(a)
(b) SW1-A-400 (-4.50 to +45.60) (c) SW2-B-300 (+46.80 to +74.40)

A1 = D19-100 Vertical (Crossties) A1 = D19-100 Vertical (Crossties)
A2 = 2D19-100 Vertical (Hoops) A2 = 2D19-100 Vertical (Hoops)
B1 = D25-300 Horizontal B1 = D22-300 Horizontal
B2 = D19-300 Vertical B2 = D19-300 Vertical
C = 12D25 (L=1100 mm) C = 12D22 (L=1100 mm)
T = 400 mm T = 300 mm
Fig. 1 a Structural model in seismostruct, b reinforcement details of SW1-A-400, c reinforcement

details of SW2-B-300, d reinforcement details of SW2-C-300, e reinforcement details of SW3-A-
250, f reinforcement details of SW3-B-250, and g reinforcement details of SW3-C-250
(d) SW2-C-300 (+75.60 to +104.40) (e) SW3-A-250 (-4.50 to +45.60)

C = 12D19 (L=1100 mm) C = 8D25 (L=900 mm)
T = 300 mm T = 250 mm
(f) SW3-B-250 (+46.80 to +74.40) (g) SW3-C-250 (+75.60 to +104.40)

C = 8D22 (L=900 mm) C = 8D19 (L=900 mm)
T = 250 mm T = 250 mm
Fig. 1 (continued)
8000
7000
Base Shear (kN) 6000
5000
4000
3000
2000
1000
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
Displacement (m)
Fig. 2 Capacity curve of minaret structure
0.30
0.25
Spectral Acceleration, Sa (g)
0.20
0.15
0.10
0.05
0.00
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Spectral Displacement, Sd (m)
Fig. 3 Capacity spectrum of minaret structure
relationship between spectral displacement (Sd) and spectral acceleration (Sa) that
occurs in a structure when resisting a certain ground acceleration.
3.2 Damage States and Uncertainty Parameters
In this research, the damage states are determined based on material strain limits and
maximum base shear criteria. The material strain limits are determined by the strain
limits of the concrete and reinforcing steel materials. In general, these damage states
are classified into DS1, DS2, DS3, and DS4. DS1, DS2, and DS3 were obtained
Table 1 Damage states based on material strain limits and maximum base shear criteria
Based on material strain limits Based on maximum base shear
DS1 DS2 DS3 DS4 DS1 DS2 DS3
Sa , ds Sa Sa Sa Sa Sa Sa Sa
0.0840 0.1538 0.2013 0.2610 0.0974 0.1925 0.2586
from the points in the capacity spectrum where the first yielding occurred in the
reinforcement, cracks in the unconfined concrete section, and cracks in the confined
concrete core. Meanwhile, DS4 is obtained from the point on the capacity spectrum
where the first chord rotation occurs on the structure that is automatically calculated
by the Seismostruct [6].
Damage states based on maximum base shear are determined based on research
conducted by Silva et al. [9]. These damage states are classified into DS1, DS2,
and DS3. DS1 is obtained from the condition when the base shear reaches 75% of
the maximum base shear. DS2 is obtained from the condition when the base shear
reaches its maximum condition. Furthermore, DS3 is obtained from the condition
when the base shear drops 20% from its maximum condition. Table 1 provides details
of the damage states based on the material strain limits and maximum base shear
criteria.
The standard deviation of uncertainty parameters needs to be taken into account
in constructing the fragility curve. This standard deviation includes the spectrum
demand uncertainty (βd ), structural capacity spectrum uncertainty (βc ), and damage
states value uncertainty (βM(ds) ). Furthermore, the total uncertainty based on Hazus-
MH MR5 can be calculated according to Eq. (1) [10]:
2
βds = [(C O N V [βc , βd ])]2 + β M(ds) (1)
where:
βc = standard deviation for structural capacity spectrum uncertainty

βd = standard deviation for spectrum demand uncertainty
β M(ds) = standard deviation for damage states value uncertainty.
The standard deviation for structural capacity spectrum uncertainty can be

calculated using Eq. (2):

s2
βc = ln +1 (2)
m2
where:
m = average spectral structure acceleration capacity

s = standard deviation of spectral structure acceleration capacity.
Table 2 Uncertainty response parameter

Damage States Sa , ds (g) Parameter
βC βD β M(ds) β ds
Based on material strain DS1 0.084 0.551 0.45 0.4 0.816
level DS2 0.154 0.551 0.45 0.4 0.816
DS3 0.201 0.551 0.45 0.4 0.816
DS4 0.261 0.551 0.45 0.4 0.816
Based on max. base shear DS1 0.097 0.551 0.45 0.4 0.816
DS2 0.193 0.551 0.45 0.4 0.816
DS3 0.259 0.551 0.45 0.4 0.816
Furthermore, the calculation results of uncertainty parameters for each damage

state of the minaret structure are presented in Table 2.
3.3 Fragility Curve
The fragility function elucidates the probability of exceedance for multiple damage
states as a function of ground-motion intensity [4]. In this research, the fragility curve
was developed using the fragility function according to Hazus-MH MR5 [10]. This
function is presented in Eq. (3):

1 Sa
P[ds|Sa ] = ln (3)
βds Sa , ds
The fragility curve of the minaret structure under study is presented in Fig. 4. The
fragility curve depicts the probability of structural damage (beyond the predeter-
mined damage states) related to the intensity of ground shaking over its service life
for the specified damage states. In this research, the damage states are determined
based on the material strain limits and maximum base shear criteria. Based on the
material strain limits criteria, damage states are classified into DS1, DS2, DS3, and
DS4. Meanwhile, based on the maximum base shear criteria, the damage states are
classified into DS1, DS2, and DS3.
The fragility curve may be understood as a building fingerprint, which helps
to assess the vulnerability of a structure during its service life. For example, in
this research, the minaret structure with a natural period of 3.782 s has a spectral
acceleration, Sa = 0.098 g according to the spectrum response design [8]. Based on
the fragility curve, the damage probability of the minaret structure for DS1, DS2,
DS3, and DS4 is 57%, 29%, 19%, and 12% respectively, based on the material
strain limits criteria. Meanwhile, based on the maximum base shear criteria, the
damage probability of the structure for DS1, DS2, and DS3 is 50%, 20%, and 12%
respectively.
1
0.9
0.8
0.7
0.6
Probability
0.5
0.4
0.3
0.2
0.1
0
0 0.2 0.4 0.6 0.8 1
Spectral Acceleration, Sa (g)
DS1 DS2 DS3 DS4 (Based on Material Strain Limits)
DS1 DS2 DS3 (Based on Maximum Base Shear)
Fig. 4 Fragility curve of minaret structure
By comparing the two damage states criteria used, it can be concluded that the
damage for the material strain limits criteria provides a more sensitive prediction
with four damage levels compared to the maximum base shear criteria which only
provides three damage levels. In general, the maximum base shear criteria presents
a lower probability of damage prediction than the material strain limits criteria. The
damage probability difference of the minaret structure for spectral acceleration, Sa
= 0.098 g for the damage level DS1, DS2, DS3 of both damage states is 7%, 9%,
and 7% respectively. Perceiving the fragility curve for the structure under review,
building owners can predict the probability of structural damage due to other earth-
quake intensity scenarios and carry out evaluations to determine retrofitting strategies
rationally.
4 Conclusion
In this research, a fragility curve has been developed for the minaret structure under
study. The fragility curve is developed using two damage states criteria, based on
material strain limits and maximum base shear. The natural period for the minaret
structure is 3.782 s. By plotting the natural period of the structure in the spectrum
response design, the acceleration spectrum is obtained, Sa = 0.098 g. Based on the
developed fragility curve, it is concluded that the damage probability of the minaret
structure on spectral acceleration, Sa = 0.098 g based on the material strain limits
criteria for damage states DS1, DS2, DS3, and DS4 is 57%, 29%, 19%, and 12%
respectively. On the other hand, based on the maximum base shear criteria for the
damage states DS1, DS2, and DS3 is 50%, 20%, and 12% respectively. Perceiving
the fragility curve for the structure under review, building owners can predict the
probability of structural damage due to other earthquake intensity scenarios and
carry out evaluations to determine retrofitting strategies rationally.
Acknowledgements This research was financially supported by Universitas Sebelas Maret,

Indonesia through Hibah Penelitian Unggulan, 2020.
References
1. Villaverde R (2009) Fundamental concepts of earthquake engineering, 1st edn. CRC Press,
United States of America
2. Pribadi K, Kusumasturi D, Rildova (2008) Learning from recent indonesian earthquakes:
an overview to improve structural performance. In: 14th world conference on earthquake
engineering, Beijing, China
3. Mosque Information System (The Indonesian Ministry of Religion) Homepage. https://simas.
kemenag.go.id/. Last accessed 2021/6/16
4. Elnasai AS, Sarno LDi (2015) Fundamentals of Earthquake Engineering (From Source to
Fragility). 2nd ed. Wiley, United Kingdom
5. Seismosoft Homepage. https://seismosoft.com/product/seismostruct/. Last accessed 2021/3/18
6. Seismosoft (2020) Seismostruct User Manual. Seismosoft Ltd., Italy
7. Badan Standardisasi Nasional (BSN) (2013) 1727:2013 Beban Minimum untuk Perancangan
Bangunan Gedung dan Struktur Lain. BSN, Indonesia
8. Badan Standardisasi Nasional (BSN) (2012) SNI 1726:2012 Tata Cara Perencanaan Ketahanan
Gempa untuk Struktur Bangunan Gedung dan Non Gedung. BSN, Indonesia
9. Silva V, Varum H, Crowley R, Sousa R, Pinho R (2012) Evaluation of analytical methodologies
to derive vulnerability functions. In: 15th world conference on earthquake engineering, Lisboa,
Portugal
10. Federal Emergency Management Agency (FEMA) (2010) HAZUS-MH MR5 Technical and
User’s Manual. FEMA, United States of America
Role of Diagonal Bars in Reinforced
Concrete Deep Beams Tested Under
Static Load
Erwin Lim and Rahmat Ramli
Abstract Four deep beam specimens were tested to investigate the role of diagonal
bars. This study shows that the presence of diagonal bars increases the shear strength
and ductility of deep beam specimens. In addition, test results also show that spec-
imens loaded through stub, instead of bearing plate, generally gives higher shear
strength due to steeper strut inclination angle. The strength of a deep beam can be
predicted using ACI 318 strut-and-tie model with conservatism.
Keywords Deep beam · Diagonal bar · Strut-and-tie
1 Introduction
The use of diagonal bars has been widely adopted in the seismic design of a coupling
beam [1]. The seismic behavior of a coupling beam is greatly enhanced due to
the presence of diagonal bar, especially for that with span-to-depth ratio less than
2.0 where shear behavior is dominant [2–6]. The main reason for the successful
application of diagonal bars being that they increase the beam’s shear strength and
therefore allowing the full development of its ductile flexural behavior.
The structural element susceptible to shear failure is not merely limited to that
under seismic load. Some elements under static gravity load might also experience
similar issue. However, the use of diagonal bars in those elements is not common and
limited to some tests of corbel [7]. Therefore, in this paper, the authors examine the
use of diagonal bars in deep beam specimens under static loading. As many as four
deep beam specimens were tested in the laboratory and the test results are discussed
and evaluated using the strut-and-tie model.
E. Lim (B) · R. Ramli

Institut Teknologi Bandung, Ganesha 10, Bandung, Indonesia
E. Lim
Lab. Rekayasa Struktur, Pusat Rekayasa Industri ITB, Ganesha 10, Bandung, Indonesia
48 E. Lim and R. Ramli
2 Experimental Study
2.1 Design of Test Specimens
As many as four deep beam specimens were designed and tested under static loading
test. Two of them, labeled with “C” are specimens tested through column stub repre-
senting real column in the building, while the other two specimens, labeled with “B”
are specimens tested through bearing plate. In each group (both “C” and “B”), one
specimen was designed with the presence of diagonal bars, while the other one is
none.
Specimens without diagonal bars (B-15–0 and C-15–0) were deliberately designed
to fail in shear so that the beneficial effect of a diagonal bar can be more distinct.
The main flexural reinforcement was 3D19 and 3D16 (flexural reinforcement ratio
ρ equals 1.73%). Meanwhile, the compressive longitudinal reinforcement ratio, ρ’
equals 0.47% and the vertical (ρ v ) and horizontal (ρ h ) reinforcement ratios were kept
to be 0.37% and 0.12%, respectively as shown in Fig. 1a and b.
Fig. 1 Specimen’s geometry and reinforcement layout

Role of Diagonal Bars in Reinforced Concrete … 49
Table 1 Parameters of deep beam specimens

No Spec. ID f ’ c (MPa) a/d d (mm) ρ (%) ρ ’ (%) ρ v (%) ρ h (%) ρ d (%)
1 C-15–0 21.19 1.56 336 1.73 0.47 0.37 0.12 –
2 B-15–0 20.21 1.56 336 1.73 0.47 0.37 0.12 –
3 C-15–2 19.63 1.56 336 1.73 0.47 0.37 0.12 0.48
4 B-15–2 24.84 1.56 336 1.73 0.47 0.37 0.12 0.48
The hypothesis of this study is that the presence of diagonal bars could alter
the failure mode of a shear dominated deep beam becoming a flexure-dominated
beam. So, for specimens with diagonal bars (C-15–2 and B-15–2), the flexural rein-
forcement ratio was maintained similar to those without diagonal bars. The only
difference was simply the presence of diagonal bars (2D16) as shown in Fig. 1c and
d. All specimen parameters are summarized in Table 1.
2.2 Test Setup and Results
Before testing of a deep beam specimen, tensile test of reinforcement bars were
carried out. For each bar diameter, as many as two coupon tests were tested and the
average yield strength is reported in Table 2.
Several strain gages were attached to the flexural, vertical, horizontal, and diagonal
bars of each specimen as indicated in Fig. 2a as monitoring device. After that, the
vertical load was applied slowly using displacement controlled actuator as shown in
Fig. 2b. Several LVDTs were also mounted as presented in Fig. 2b.
The load–deflection curve of each specimen is given in Fig. 3a. The plotted load
is the force measurement from actuator with respect to displacement at mid span.
Specimen C-15–0 showed elastic behavior when the load was still below 590 kN
with displacement of 6.19 mm. The first crack on this specimen occurred at the
load of 210.28 kN and formed flexure crack pattern with yielding of longitudinal
reinforcement as well. When the load reached the maximum load, a large crack
occured and formed a diagonal crack and then the load started to drop suddenly.
The maximum load achieved was 748.2 kN with corresponding displacement of
Table 2 Yield strength of

No. Diameter (mm) f y (MPa)
rebars
1 6 338.28
2 8 387.07
3 10 564.15
4 13 412.04
5 16 370.81
6 19 386.36
Fig. 2 Instrumentation layout and test setup
Fig. 3 Load displacement curve testing of deep beams
9.41 mm. The dominant crack pattern that occurs was shear crack. The failure mode
of this specimen was shear with diagonal splitting after yielding of longitudinal
reinforcement as shown in Fig. 3b. Due to the presence of 2D16 diagonal bars,
specimen C-15–2 attained the maximum load at 827.4 kN and the corresponding
displacement was 13.91 mm. The failure mode of this specimen was flexural failure
with slight indication of concrete crushing at the vicinity of column stub as indicated
in Fig. 3c.
Specimens loaded through bearing plate in general show similar behavior, but with
lower maximum attained load than specimens loaded through stub. The maximum
load of specimen B-15–0 and B-15–2 was 687.9 kN and 796.8 kN, respectively.
Again, specimens with diagonal bars possessed higher strength than that without
diagonal bars. The failure mode of B-15–0 is immediate shear with diagonal splitting
after flexural yielding of longitudinal reinforcement (Fig. 3d), while that of B-15–2
was flexural failure (Fig. 3e). Summary of test results is given in columns (b)–(d) of
Table 3.
Comparing specimens with and without diagonal bars, Fig. 3a clearly shows that
the presence of diagonal bar increased the maximum strength and ductility of a deep
Table 3 Strength results
Spec. ID Test values Strut-and-tie model
V test (MPa) test (mm) FMtest θ (deg) V ns (kN) V nn (kN) V nt (kN) V STM (kN) V test :2/V STM FMpredicted
(a) (b) (c) (d) (e) (f) (g) (h) (i) (j) (k)
C-15–0 748.2 9.41 FS* with diag. 31.3 186.9 249.2 341.4 186.9 2.00 Upper strut crushing
splitting
Role of Diagonal Bars in Reinforced Concrete …
B-15–0 687.9 8.18 FS* with diag. 27.3 222.2 296.3 289.4 222.2 1.55 Upper strut crushing
splitting
C-15–2 827.4 13.91 Flexure 30.9 259.8 323.9 335.2 259.8 1.59 Upper strut crushing
B-15–2 796.8 10.99 Flexure 28.3 317.1 400.2 302.4 302.4 1.32 Tie yielding
Average 1.62
COV 0.18
*FS indicates: immediate shear after flexural yielding of longitudinal reinforcement
51
Fig. 4 Illustration of shear element for specimens with column stub and bearing plate
beam specimen, despite some early yielding of main longitudinal reinforcement in all
four specimens. The failure mode also changed from flexural shear failure to ductile
flexural failure. The major crack of specimen with column stub (θ stub ) was observed
to occur mainly at the vicinity of the stub. Meanwhile, major inclination crack of
specimens with bearing plate (θ bearing ) was observed to occur under the point load
as seen in Fig. 3b–e and illustrated in Fig. 4. Therefore, the crack inclination angle
of specimens loaded through column stub in general was steeper than those loaded
through bearing plate. This could be the reason why the maximum load of specimens
with column stub was higher than those with bearing plate.
3 Analytical Study
In this study, the strength of a deep beam is evaluated analytically using the strut-
and-tie provision in ACI 318-14 [1]. The strength of a deep beam can be taken as the
minimum among the strength of concrete strut (V ns ), nodal zone (V nn ), or yielding
of tie (V nt ) as shown in Eq. (1):
VST M = min{Vns ; Vnn ; Vnt } (1)
Before determining the strength of respective components, it is important to firstly

determine the size of shear element through the macro model [8]. The size of shear
element also represents the inclination angle of diagonal strut, θ as shown in Eq. (2):

−1 v
θ = tan (2)
h
where v is the lever arm determined from the compressive force to the tensile force,
and h is the horizontal boundary of a shear element. For deep beam loaded under
column stub, the h should be bounded by the geometry discontinuity of the concrete
stub, while for beams loaded under bearing plate, the h is bounded by the force
discontinuity as illustrated in the following Fig. 4. Similar determination of shear
element size was also proposed by Lu and Chu [9]. The calculated strut angle for all
four specimens can be seen in column (e) of Table 3.
Strength of a Strut
The strength of a diagonal strut, F ns can be determined from the contribution of
concrete strut and diagonal bar (if presence) as shown in the Eq. (3):
Fns = 0.85 f c βs Acs + Ad f yd (3)
where β s is the ACI 318 strut efficiency factor, Acs is area of concrete strut, Ad is
area of diagonal bar, and f yd is yield strength of a diagonal bar.
Due to the difference of discontinuity as illustrated in Fig. 4, the area of concrete
strut (Acs ) at upper part for specimens with column stub and bearing plate can be
calculated as shown in Eqs. (4a) and (4b):
Column stub : Acs = (ws cos θ )b (4a)
Bearing plate : Acs = (ws cos θ + a p /2 sin θ )b (4b)
Since the lower part of the specimen uses bearing plate for all specimens, therefore
the area of concrete strut is determined using Eq. (5):
Acs = (wt cos θ + lb sin θ )b (5)
The width of strut at upper part, ws and width of tie at lower part, wt can be
determined from horizontal force equilbirum as suggested by Tjhin and Kuchma
[10]. Meanwhile, ap and b equals to 150 mm is the depth of bearing plate and
column stub, respectively. These parameters are illustrated in Fig. 5.
Finally, the shear strength of a deep beam due to crushing of a diagonal strut is
presented in Eq. (6):
Vns = Fns sin θ (6)
Strength of a Nodal Zone

The strength of a nodal zone, F nn can be determined from the contribution of concrete
strut and diagonal bar (if presence) as shown in the Eq. (7):
Fnn = 0.85 f c βn Acn + Ad f yd (7)

a) Strut area for beam with stub b) strut area for beam with bearing plate
Fig. 5 Illustration of strut area
where β n is the ACI 318 nodal efficiency factor, Acn is area of nodal zone similar
to area of concrete strut.
Therefore, the strength of a deep beam due to crushing of nodal zone is shown in
Eq. (8):
Vnn = Fnn sin θ (8)
Normally, since the lower node governs because it is considered as a CCT node.
Strength of a Yielding Tie
The strength of a yielding tie can be determined using Eq. (9):
Fnt = Ast f y (9)
where Ast is the area of longitudinal flexural reinforcement, f y is the yield strength
of main longitudinal beam reinforcement.
The shear strength of a deep beam due to yielding of longitudinal bar can be
written as:
Vnt = Fnt tan θ (10)
Finally, Table 3 summarizes the calculated strengths for all four specimens along
with the strength ratio and predicted failure mode in columns (j) and (k), respec-
tively. The average strength ratio is 1.62 with COV of 0.18, which gives conserva-
tive strength estimation. This conservatism is expected because the ACI 318 STM
might not properly capture the overall behavior which leaves possibility of further
refinement of this model.
4 Conclusion
Based on the results of this study, it can be concluded that:

1. The presence of diagonal bars could alter the failure mode of deep beam speci-
mens which were originally designed to fail in shear. The diagonal bars increase
the overal shear strength and therefore allowing development of beam’s flexural
behavior. Correspondingly, the ductility of the deep beam is also enhanced.
2. Specimens loaded through stub generally showed higher capacity because the
inclination angle of the strut is steeper than specimens loaded through bearing
plate
3. ACI 318 strut-and-tie model gives conservative strength estimation of the tested
four specimens.
Acknowledgements This research was carried out using the grant from Program Penelitian,
Pengabdian kepada Masyarakat, dan Inovasi (P3MI) ITB of which the authors are grateful.
References
1. American Concrete Institute (2016) Building code requirements for structural concrete (ACI
318–14) and commentary (ACI 318M-14). Farmington Hills, pp 623
2. Paulay B, Binney JR (1974) Diagonally reinforced coupling beams of shear walls. ACI J
579–598
3. Naish D, Fry A, Klemencic R, Wallace J (2013) Reinforced concrete coupling beams—Part I:
testing. ACI Struct J 110:1057–1066
4. Lim E, Hwang SJ, Wang TW, Chang YH (2016) An Investigation on the seismic behavior of
deep reinforced concrete coupling beams. ACI Struct J 113:217–226
5. Park WS, Kang THK, Kim S, Yun HD (2020) Seismic performance of moderately short concrete
coupling beams with various reinforcements. ACI Struct J 117:141–154
6. Ameen S, Lequesne RD, Lepage A (2020) Diagonally reinforced concrete coupling beams
with grade 120 (830) high-strength steel bars. ACI Struct J 117:199–210
7. Foster SJ, Powell RE, Selim HS (1996) Performance of high-strength concrete corbels. ACI
Struct J 93:555–563
8. Lim E, Hwang SJ (2016) Modeling of the strut-and-tie parameters of deep beams for shear
strength prediction. Eng Struct 108:104–113
9. Lu WY, Chu CH (2019) Tests of high-strength concrete deep beams. Mag Concr Res 71:184–
194
10. Tjhin TN, Kuchma DA (2002) Example 1b: alternative design for non-slender beam (deep
beam). ACI Special Publication (SP-208). Ed. Karl-Heinz Reineck, pp 242
Analysis of Reinforced Concrete
Capacity for Irregular Cross-Sections
Using Numerical Methods
Nuroji
Abstract Although reinforced concrete member usually forms the rectangular

shape, for architectural reasons or an optimization purpose the section may be formed
on the nonrectangular or irregular shape. Analysis of the sectional capacity in the
case of non-square or irregular sections is very complex and takes time. This paper
offers a cross-sectional analysis with a numerical approach that is implemented in a
computer program. Several cross-sectional forms from the results of previous studies
were adopted for validation purposes. The selection of the cross-section considers the
representation of the cross-section shape and the configuration of the reinforcement.
In addition to the cross-sectional shape, the method of analysis is also a consider-
ation for selection. From the results of the analysis using numerical methods and
comparison of the analysis of previous researchers, it shows a fairly good level of
accuracy with an average deviation of 2.35%. The largest deviation was in Section 7
and 8 with a deviation of 5.03 and 8.62%, respectively. This deviation is more due
to the analysis method. Analysis using finite element method for sections, 9 and 10
show slightly higher than the numerical method due to neglecting tensile strength of
concrete and strain hardening of steel on the numerical method. On the other side,
the analysis with the cross-section conversion approach shows lower results. This
software may be used to solve the nonrectangular or irregular sections.
Keywords Irregular section · Numerical · Flexural
1 Introduction
In contrast to steel materials that have the same strength between compressive and
tensile conditions. Generally, the tensile strength of concrete does not exceed 10%
of the compressive strength [1, 2]. The low tensile strength of the concrete makes
this material often combined with reinforcing steel to withstand the tension so that it
forms a reinforced concrete composite structure. In general, the structural elements
Nuroji (B)
Diponegoro University, Semarang, Indonesia
58 Nuroji
of reinforced concrete are in the form of a square section. However, due to architec-
tural rea-sons or structural optimization considerations, reinforced concrete sections
can be non-square, such as circle, octagonal, hexagonal, trapezoidal, and triangular.
Moreover, a square section that is subjected to the biaxial moment can also behave
as a non-square section due to the rotation of the section in receiving the biaxial
resultant moment.
In a bending element, the internal forces acting in a section will take balance, so
that the compressive forces will be the same as the tensile forces. The compressive
force in the cross-section is generated from compressed concrete and compressive
reinforcement, while the tensile force is generated from tensile reinforcement. The
compressive force of the concrete is the resultant stress of the concrete which forms a
stress block. By assuming the plane section remains plane before and after bending,
the strain distribution from the neutral axis to the extreme compression fiber can be
considered a linear function of the neutral axis distance. Therefore, the stress distribu-
tion will also be identical to the shape of the stress–strain relationship curve. A stress
block in a reinforced concrete section is a volume formed by the compression area
of the concrete and its stress. In a square section where the width of the compression
area is constant, the concrete compressive force of the stress block can be simpli-
fied to be an equivalent stress block in a square shape [3]. However, for non-square
sections, the simplification of square stress blocks to calculate the compressive force
of concrete may be wrong.
The calculation of non-square sections will be more appropriate if using the
numerical approach than the stress block simplification. This paper discusses the
moment capacity analysis on irregular sections of reinforced concrete beam by using
a numerical approach implemented in a computer program to overcome design
problems, especially in the case of non-square beam sections or biaxial bending
cases.
1.1 Material Model
1.1.1 Concrete Compressive Stress–Strain
The compressive strength of concrete is generally determined from the compres-

sive test of a concrete cylinder with 150 mm diameter and 300 mm length in the
longitudinal direction at the age of 28 days. Numerous approaches are used to deter-
mine the shape of the concrete compressive stress–strain relationship curve before
maximum stress in a second-order parabolic [4]. Hognestad proposed the compres-
sive stress–strain curve of concrete as a parabolic function up to the maximum stress
and subsequently decreasing linearly until it reaches the ultimate strain as shown in
Fig. 1a [5]. The concrete compressive Stress–strain curve is described in two regions
as shown in Eqs. 1 and 2.
Analysis of Reinforced Concrete Capacity … 59
Fig. 1 Stress–strain relationship. a Concrete in compression. b Steel
2
2εc εc
fc = f c − , 0 ≥ εc ≥ εc0 (1)
ε0 ε0

0.15εc + 0.85εco − εcu
fc = f c , εc0 ≥ εc ≥ εcu (2)
εco − εcu
2 f c
Where, ε0 = Ec
and εcu = 0.0038.
1.1.2 Steel
In general, the behavior of the steel is determined from the stress–strain relationship
of the uniaxial tensile test results in the laboratory. The stress–strain relationship
curves of steel for the compressive conditions are considered to be the same and
identical to the tensile [4]. Chen determined the stress–strain relationship of steel by
dividing three parts, the linear elastic region starting from the point of origin until the
steel reaches the yield strain εy , horizontal plateau starting from yield-strain εy up to
8–15 times its elastic range εy , and strain hardening where the steel stress increases
and reaches a maximum at the ultimate strain then gradually decreases until the failure
strain [6]. In this study, the behavior of the stress–strain relationship of reinforcing
steel is modeled as a bilinear function with neglecting the strain hardening effect as
shown in Fig. 1b.
1.1.3 Section Model and Reinforcement Configuration
The irregular sections are modeled by multilinear from nodal points coordinate forms
closed polygon. The coordinates of each polygon boundary point are denoted (X(i) ,
Y(i) ), where i = 1, 2, 3, … n, and n is the number of polygon coordinates. The
configuration and location of the reinforcement in the section are also determined by
coordinates (Xr(i) , Yr(i) ), where i = 1, 2, 3, … nr, and nr is the number of reinforce-
ments in the section. From the cross-sectional coordinates, it can be determined Ymax
which is the outermost compressive point that undergoes the maximum compressive
60 Nuroji
Fig. 2 a Section. b Strain diagram. c Concrete stress block. d Forces of reinforcements
strain. Meanwhile, from the coordinates of the location of the reinforcement can be
determined the outermost tensile reinforcement. Irregular cross-sectional shape and
reinforcement configuration as shown in Fig. 2a.
2 Cross Section Analysis
In the analysis of flexural beam with a rectangular section, the compressive force
of concrete, and the tensile force of the reinforcement can be obtained from the
equilibrium of forces, where the compressive force C c is defined based on a simplified
equivalent stress block of concrete. While the irregular cross-section, determination
of compressive force C c becomes very complex because it has to be solved through
the integration of the stress–strain relationship function on the compression concrete
area. Therefore, it will be more accurate in case irregular cross-section analysis
undertakes numerically. The analysis of reinforced concrete sections in this study is
based on several basic assumptions on the following flexural sections [7].
1. The first presumption is plane sections remain plane before and after bending
as Bernoulli’s principle. This assumption implies that the strain in the section is
proportional to the distance to the neutral axis. The result of this assumption is
that the distribution of compressive stress in the compressive region is identical
to the stress-strain relationship curve.
2. The bond between concrete and steel reinforcement is assumed as full-bonded,
and there is no slip between both concrete and steel. It means that the strain of
reinforcement is the same as concrete strain at the same level.
3. The tensile stress of the concrete is not more than10% of its compressive
strength, thus the force of the tensile concrete below the neutral axis is also
small. The lever arm of the concrete tension force to the neutral axis is small.
So that, the contribution of concrete tension force to the bending capacity of
the cross-section is considered very small. For this reason, the tensile stress of
concrete can be neglected.
4. Concrete is considered only to be able to withstand compressive stress until
the ultimate strain. ACI sec. 10.2.3 defined the ultimate compressive strain of
the concrete to 0.003 for design [8]. However, in this study, the ultimate strain
of compression concrete was determined according to Hognestad’s model, i.e.,
0.0038.
Resultant of internal forces in a section can be derived by Eq. 3.

nr
Pt = Cc + T(i) (3)
i=1
Where:
C c : Concrete compression force.

nr
T(i) : Sum of the reinforcement forces, where nr is number of reinforcement.
i=1
The resultant of internal forces P t in Eq. 3 are obtained by determining the value of
the outer tensile reinforcement as shown in Fig. 3. The process of determining P t is
carried out with the following procedure.
1. Define the strain of the outer tensile reinforcement εs .
2. Assuming the strain distribution is linear, the depth of the neutral axis c can
be defined. (see Fig. 2b).
3. The compressed concrete area is divided into small slices extending in the X
direction in a number of ns slices. The thickness of each slice is Δ = c/ns.
4. Calculate the coordinates of the intersection points between the edge of slice
and the polygon to calculate the width of the slice. b(i) , i = 0, 1, 2, …, ns, and
ns is number of slices.
Fig. 3 Bisection iteration

62 Nuroji
5. Average slice width can be derived using Eq. 4.
b(i−1) + b(i)
b1(i) = (4)
2
6. The center of gravity of each slice can be assumed exist at midpoint of slice
thickness, thus the arm of these points to neutral axis is approximated by Eq. 5.
Y(i) = (i − 0.5) (5)
7. The concrete compressive strain on each slice edge is proportional to the

distance of the edge to neutral axis, and the strain can be defined by Eq. 6.
i
εc(i) = × 0.0038 (6)
n
Where: i = 0, 1, 2, …, ns, ns is number of slices

8. Furthermore, the compressive stress of the concrete on each slice edge can be
determined based on the stress-strain relationship according to Eqs. 1 and 2.
9. The average stress of slice can be calculated with Eq. 7.
f c(i−1) + f c(i)
f c1(i) = (7)
2
10. Compression force for each slice can be defined by Eq. 8.

Cc1(i) = b1(i) × f c1(i) (8)
11. The resultant of concrete compressive force C c is the sum of all the compressive
forces of the on each slice as shown in Eq. 9.

n
Cc = Cc1(i) (9)
i=1
12. The space of the reinforcement to the neutral axis is determined based on
Eq. 10. The negative sign indicates the location of the reinforcement is below
the neutral axis.
Yr 1(i) = Yr (i) + c − Ymax (10)
13. The reinforcement strain proportional with respect to the distance of the rein-
forcement to the neutral axis, the reinforcing strain can be calculated by Eq. 11.
In this case the negative sign indicates the tensile strain.
0.0038
εs(i) = Yr 1(i) (11)
c
14. Base on Eq. 11, then reinforcement stresses in each reinforcing bar is obtained
using expression of Eq. 12.
f s(i) = εs(i) × E s , where E s = 2 × 105 Mpa (12)
if εs(i) > ε y , then f s(i) = f y .

if εs(i) < −ε y , then f s(i) = − f y .
15. The forces on each reinforcement are the product of the reinforcement area
and the stress as shown in Eq. 13.
T(i) = As(i) × f s(i) (13)
16. From the compression force C c and the reinforcement forces T (i) can be
calculated the resultant internal forces of section using Eq. 3.
2.1 Internal Forces Equilibrium
To achieve a balance of internal forces of the cross section can be obtained iteratively
by using the bisection method. The concrete strain at top fiber is set 0.0038, and the
outer reinforcing steel strain is entered twice, i.e., εsa = 0 and ε sb = 0.04 as the initial
strain. Iteration process of bisection method as shown in Fig. 3.
1. The first step is to enter εsa = 0. By using the calculation procedure to find
the resultant of internal forces for εs = ε sa will be found P ta that produce the
compressive force, (see Fig. 4c).
Fig. 4 a Section. b Strain iteration. c Internal forces at εs = εsa . d Internal forces at εs = εsb
64 Nuroji
2. By using the same procedure enter εsb = 0.04 as ε s to get P tb that produce the
force in tension, (see Fig. 4d). Where ε sa and ε sb are the initial strain given to
outer reinforcement which has not produce the balance condition (P t = 0). sb
3. The Next Step Using εsc as Mid-Point Between ε sa and ε sb , εsc = εsa +ε 2
.
Then Calculate the Resultant of Internal Forces for εs = ε sc , i.e., P tc .
If Ptc × Pta < 0, then εsb = εsc .
If Ptc × Ptb < 0, then εsa = εsc .

a)

(εc −εb )
4. If
(εc ε−ε
c

≤ T olerance or
εc
≤ T olerance, then go to step 5. If No,

back to step 3.
5. The ε s =ε sc , in this iteration the equilibrium occurs.
The balanced condition occurs when the strain of outer tension reinforcement εs
produces internal forces resultant Pt which very small or less than the tolerance value
defined earlier. Then calculate the concrete compressive moment from the number
of slice forces multiplied by its lever arm to the neutral axis as shown in Eq. 14, and
the moment of reinforcement which is the result of an addition of moment of each
reinforcement to the neutral axis as shown in Eq. 15.

ns
Mc = Cc1(i) × Y1(i) (14)
i=1

nr
Mr = T(i) × Yr 1(i) (15)
i=1
The flexural capacity can be determined by using Eq. 16
M = Mc + Mr (16)
Numerical solution to determinate the reinforced concrete capacity presented in

Fig. 5.
3 Program Validation
Before the program is used, a validation process is required to ensure whether the
program is correct or still requires improvement. In this study, the validation process
was undertaken by comparing the results of program running and analysis from
previous studies. The sections used for validation represent the geometric shape of the
square and irregular sections. Reinforcement ratio and reinforcement configuration
are also considered in section selection. The selected sections in this study as shown
in Fig. 6 (Table 1).
Fig. 5 Flow chart of irregular reinforced concrete beam section analysis
Fig. 6 Selected cross section of beam

66 Nuroji
Fig. 6 (continued)
Table 1 The results of analysis

No Section f c (MPa) f y (MPa) φ μ (kN.m) Deviation (%) Analysis
Analysis Reference method
1 Figure 6a 20.7 276 603.31 595.00 1.40 Approximation

analysis [4]
2 Figure 6b 30 400 1141.14 1139.00 0.19 Flexural
equation [9]
3 Figure 6c 30 420 108.56 109.00 0.41 Flexural
equation [9]
4 Figure 6d 30 420 202.22 203.00 0.38 Flexural
equation [9]
5 Figure 6e 30 420 131.99 132.00 0.00 Flexural
equation [9]
6 Figure 6f 30 420 181.45 180.00 0.81 Flexural
equation [9]
7 Figure 6g 30 400 93.38 88.90 5.03 Equivalent
square shape
[9]
8 Figure 6h 30 400 152.72 140.60 8.62 Equivalent
square shape
[9]
9 Figure 6i 41.8 400 11.88 12.10 1.84 Finite element
analysis [10]
10 Figure 6j 33.23 309.5 122.36 128.50 4.78 Finite element
analysis [10]
Based on the numerical analysis results and comparing with the previous studies, it is
apparent that the cross-sectional analysis using the numerical method implemented
in a computer program is very accurate with an average deviation of 2.35%.
The largest deviation in this study is the comparison between cross-sections 7
and 8 analysis in this study and reference with deviation 5.03 and 8.62%, respec-
tively. These deviations are more due to the analysis method distinction, where
hexagonal and circular sections are simplified with square sections. Cross-sections
analyzed with FEM i.e., 9 and 10 indicate greater capacity than this study. This is
because the numerical analysis of cross-section ignores the concrete tensile stress
as stated in one of the analysis assumptions. The cross-sections 7 and 8 analyzed
with the section area approximation, where the hexagonal and circular sections are
approached by converting the square section to simplify the analysis. The other
cross-sections analyzed with flexural section analysis show a very small deviation
i.e., 0.36%.
A similar numerical simulation was also developed by Dundar and Sahin that
assume the irregular cross-sections as a closed polygon [11]. Even though the simu-
lation model can analyze arbitrary reinforced concrete sections subjected to axial load
68 Nuroji
and biaxial bending. But, the model still uses an equivalent stress block to compute
concrete compression force that may be inappropriate for the non-square section.
On the other hand, Oscar Fitrah Nur undertook a numerical analysis of rectangular
reinforced concrete beam with divide compression area into small slices to generate
the stress block which appropriates with actual stress–strain [12].
The simulation in this study considers the cross-section as a closed polygon and the
stresses in compression concrete as a function of strain which linear to the distance of
the neutral axis. So, this study indicates that the numerical simulations implemented
in a computer program show progress in research.
5 Conclusion
The numerical method by using a multi-slice approach with bisection iteration to

calculate the moment capacity of the reinforced concrete section shows a good result
with an average deviation of 2.35% compared with other methods. Implementation
of this method as a computer program is very precise to analyze reinforced concrete
beams for the irregular section including circular, hexagonal, T shapes, etc.
References
1. O’Neil EF, Neeley BD, Cargile JD (1999) Tensile properties of very-high-strength concrete
for penetration-resistant structures. Shock Vib 6:237–245, ISSN 1070-9622
2. Arιoglu N, Girgin ZC Arιoglu E (2006) Evaluation of ratio between splitting tensile strength
and compressive strength for concretes up to 120 MPa and its application in strength criterion.
ACI Material Journal/January-February 2006
3. Whitney CS (1937) Design of reinforced concrete members under flexure or combined flexure
and direct compression. J ACI 33(3):483–498
4. Park R, Paulay T (1975) Reinforced concrete structures. Jhon Wiley & Sons, Canada
5. Hognestad E (1951) A study of combined bending and axial load in reinforced concrete.
University of Illinois Engineering Experiment Station Bulletin Series No. 399 University of
Illinois, Urbana
6. Chen WF (1982) Plasticity in reinforced concrete. McGraw-Hill, New York
7. MacGregor JG (1997) Reinfoced concrete mechanics and design. Inc. USA, Prentice Hall
International
8. ACI 318–05: building code requirements for structural concrete (ACI 318–05) and commentary
(ACI 318R-05), 2005
9. Al-Ansari MS, Afzal MS (2019) Simplified irregular beam analysis and design. Civ Eng J
5(7):1577–1589
10. Nuroji, Besar MS, Imran I (2010) Rotated discrete crack model for reinforced concrete
structures. In: 35th conference on our world in concrete & structures. 25–27 August 2010,
Singapore
11. Dundar C, Sahin B (1993) Arbitrarily shaped reinforced concrete members subject to biaxial
bending and axial load. Comput Struct 49(4):643–662
12. Nur OF (2009) Analisa Pengaruh Penambahan Tulangan Tekan Terhadap Daktilitas Kurvatur
Balok Beton Bertulang. Jurnal Rekayasa Sipil 5(1)
Assessment of Retaining Wall Design
in Harris Skyline Tower’s Basement,
Surabaya
Siti Nurlita Fitri and Ahmad Soimun
Abstract The project of Harris skyline tower in Surabaya was designed as a hotel and
office tower concept. The total area of this project is 38.821 m2 with 29 floors and 2nd
levels of basement. The substructure design is necessary to 7.5 m depth of basement
construction. Furthermore, Surabaya has a high level of groundwater table which
affects the risk of the project with the depth excavation. This condition encourages
the retaining wall construction design to prevent the structure from collapse and
failure in overall stability. The cost for basement projects included in the cost of
the superstructure, thus it became essential to reduce this cost by investigating pile
varieties and settling on the most cost-effective material type. This study aims to
know the suitable design of the retaining wall in Harris Skyline tower. The material
structure was considered in optimum result (safety factory and deformation) and best
cost value. Substructure calculation of this research was conducted in 3 alternatives;
sheet pile, steel pipe pile, and bored pile with the high level of groundwater table
design. The deformation analysis was carried out in Plaxis simulation modeling. The
result of this paper is a 21.5 m length calculation of a retaining wall with a 14 m
embedded structure. The final displacement is 3,7 cm as SF = 2.66. The appropriate
alternative design is a bored pile with an analysis of material price Rp 11.369.860.000.
Keywords Retaining wall · Sheetpile · Bored pile · Plaxis · Displacement
1 Indtroduction
The development of Surabaya as the capital city of east java rises rapidly with the
establishment of high-rise buildings in all of the area. Harris Skyline Tower is a
prestigious project from PT. Bumi Sejahtera Tower combines the concept of hotel
S. N. Fitri (B)
Civil Engineering, Universitas Sebelas Maret, Surakarta, Indonesia
A. Soimun
Polytechnic of Land Transportation, Bali, Indonesia
70 S. N. Fitri and A. Soimun
and office tower. The Harris Skyline Tower project location is on Jl. Maj. Gen.
Sungkono No. 178, Dukuh Pakis, Surabaya City, East Java.
The construction of office buildings, apartments, and public areas require a large
area of parking to accommodate visitor vehicles. This is a problem for developers
to make parking zones; therefore a substructure construction (basement) is the best
alternative to provide a parking area. Basement construction has its own require-
ments criteria and stages of construction implementation. The design of retaining
walls for the material of basements under the construction method used has a consid-
erable affected. The structure method for the basement influences the implementa-
tion schedule for the project because the first process of building is substructure
construction.
Harris skyline tower was designed with 2 levels of basement and 29 floors. The
basement construction is 7.5 m depth of levels. Moreover, Surabaya has a high level
of the groundwater table. Hence, the project with deep excavation has a high-risk
condition. Design of the basement construction is necessary to prevent the failure
and sliding of the soil behind the substructure [1, 2].
There are many alternatives basement construction designs; sheet pile, bored
pile, steel pipe pile, and many more. Many studies in Indonesia success to design
the basement construction with the selected material [3–6]. The other reinforcement
to prevent the sliding slope is geotextile and terracing [7, 8]. However, the last
improvement does not suitable for this case. The decision of retaining walls to support
deep excavations for basement construction is affected by a combination of variables,
including soil types, allowable movement, construction methods, and material and
equipment availability [9].
This research aims to determine the best alternatives of retaining wall design to
Harris Skyline towers basement based on the material prices. According to Lateef and
Burhan [10], The material price of pile foundations to basement project is included
in the cost of the superstructure, and it became necessary to decrease this cost by
studying pile types and then deciding on the best pile type in terms of cost. The output
of this study is expected as guidance to a similar construction basement project.
2 Methodology
This study has simulated the performance design of 3 alternatives basement materials.
The embedded pile has been carried out in manual calculation. Next, the Safety factor
value in order to installation the basement pile is designed in Plaxis calculation.
Moreover, The maximum deformation value has been obtained as Plaxis output.
Furthermore, to conclude the best alternative structure material was look in the lowest
material cost.
Assessment of Retaining Wall Design in Harris Skyline Tower’s … 71
Table 1. N-SPT test for

Depth (m) Soil Average SPT
Harris skyline tower
0–6 Brown clay 3
6–10 Brown silt 6
10–25 Silt sand 40
25–35 Loamy sand 50
Table 2. Soil data calculation for Harris skyline tower

Depth Soil Average Cu Gs γb Gsat e n K E
(m) SPT (t/m2 ) (t/m2 )
0–6 Brown 3 1.5 2.62 1.66 1.704 1.29 0.57 3 2× 400
clay 10-10
6–10 Brown 6 2.6 2.69 1.641 1.67 1.49 0.62 6 2× 900
loam 10-10
10–25 Clayed 40 – 2.68 1.987 2.00 0.78 0.44 38 2× 2000
sand 10-9
25–35 Loamy 50 – 2.68 2.057 2.25 0.72 0.3 41 2× 4000
sand 10-9
2.1 Soil Data
The SPT investigation is described in Table 1. According to Table 1, The SPT test was
conducted in Ahmad soimun and Arif [11]. The dominant soil in clay with different
consistency (0–10 m) and Sand until 35 m depth. After the result of SPT test analysis,
the next step is to calculate the other parameters (Specific gravity, unit weight, and
shear stress based on Bowles [12]. The detail of soil data is shown in Table 2.
2.2 Lateral Earth Pressure
The lateral earth pressure is analyzed according to Rankine and Coulomb Theory
[13]. The calculation considered two conditions of the groundwater table; normal
and maximum level. The model was shown in Fig. 1.
Figure 1 describes the model of lateral earth pressure calculation. The first picture
(a) illustrates the maximum level, this condition realizes the maximum range of the
embedded pile requirement, while the second (b) considers the lowest result. The
soil data were analyzed in 3 types. Soil 1 was model in 6 m depth, Soil 2 in 4 m, and
15 m for the third soil. The external load 1 t/m2 represents a load of heavy equipment
and the material. Do is the calculation result of sheet pile embedded depth.
Fig. 1. Model of lateral 1 ton/m2

earth pressure calculation. a
Normal groundwater table. b 1
Maximum ground water
table Soil I
6.00
7.50
2
excavation base 3
Soil II 4.00
2.50
4
5
DO ? Soil III
6 15.00
(a)
1 ton/m2
1
Soil I 4.10
7.50 2
3
4
1.90
5
Elevation base
Soil II 4.00
2.50
6
7
DO ? Soil III
8
15.00
(b)
3.1 Lateral Earth Pressure Analysis
The analysis of lateral earth pressure was conducted by manual calculation. The
process result is shown in Fig. 2 in a normal level of groundwater table and Fig. 3
in maximum level. The figure depicts the diagram of active and passive force from
Tables 3, 4, 5 and 6. The active force is located on the right side of the graph (behind
the sheet pile), and the passive force is in the front.
After the calculation of active and passive force, the next step is to determine the
moment from force and distance based on the diagram. The simulated was carried
out to how long the embedded depth of substructure requirement (do). Based on the
calculation in Tables 3 and 4, the analysis describes as follows:

MPassive = 28.84 + 27.37do + 4.795do2 + 0.863do3

Mfinal = Mactive − Mpassive
Fig. 2. Lateral earth 1 ton/m2

pressure diagram in
maximum level of ground 1
3.07
hc
water table
6.00 Brown clay
7.50 Ea 1
2
excavation base 0.82 3
0
Ea 2 4.00 Brown silt
2.50 Ep 2
Ep 14
5
Silty Sand
Ep 3 Ea 3 Ewa 1
DO ? DO ?
Ewp 1
Ep 4 6
Ea 4
Fig. 3. Lateral earth 1 ton/m2

pressure diagram in normal
level of ground water Table 1.20 hc 1
Ea 1
4.10 Brown clay
7.50 2
3 Ea 2 Ea 3 1.90
excavation base 4
5
0
Ea 4 4.00 Brown silt
Ep 1
2.50 Ep 2 6
Ea 5
DO ? Ep 3 Ea 6 DO ? Silty sand
Ewp 1 Ewa 1
Ea 7
Ep 4 8
Table 3. Active force calculation in maximum level of ground water table

σv ’ σ’ H (t/m2 ) H (m) Active force (ton) Distance (m) Moment (tm)
(t/m2 )
1.00 −1.95 3.07
5.22 1.86 2.93 2.721229595 4.98+do 13.54+2.72do
5.22 −0.45 0.82
7.92 1.74 3.18 2.769944132 1.06+do 2.94+2.77do
7.92 1.88 do 1.88xDo 0.5 do 0.94 do2
7.92+ 1.901+0.24do do 0.001Do+0.06Do2 0.33do 0.00033 do2 +0.0198
(1x do) do3
0.5 x 10+do 50+10Do+0.5Do2 3.33+0.33do 166.5+49.80do+4.97
(10+do)2 x1x1 do2 +0165 do2
Table 4. Passive force calculation in maximum level of ground water table

σv ’ (t/m2 ) σ’ H (t/m2 ) H (m) Passive Force Distance (m) Moment (tm)
(ton)
00 5.78 2.5 14.43796269 1.25 + do 18.05 + 14.44do
1.69 7.86 2.5 9.820811 0.833 + do 8.18 + 9.82do
1.69 7.09 Do 7.09do 0.5do 3.545 do2
1.69 + do 7.09 + 4.2do Do 2.1do2 0.33do 0.693 do3
0.5 x (2.5 + do)2 3.13 + 2.5do + 0.833 + 0.33do 2.61 + 3.11do +
×1×1 do2 1.25
do2 + 0.17do3
= 154.14 + 27.92do + 1.12do2 − 0.678do3

or = −0.678do3 + 1.12do2 + 27.92do + 154.14
Hence,
do = 9.012 m
and total the depth of embedded substructure is 9.012 + 2.5 = 11.512 m
for safety reason, the result must times with Safety factor
Thus,
11.512 m × 1.2 = 13.81 m = 14 m
The final result of substructure is 14 m + 7.5 m(open cut) = 21.5 m
Next, the maximum momen is calculated as follow :

Mfinal = − 0.678do3 + 1.12do2 + 27.92do + 154.14
= − 0.678(4.29)3 + 1.12(4.29)2 + 27.92(4.29) + 154.14
= 240.905 Ton.m = 2409.05 Kn.m
Table 5. Active force calculation in normal level of ground water table
σv ’ (t/m2 ) σ’ H (t/m2 ) H (m) Active force (ton) Distance (m) Moment (tm)
7.81 4.18 2.9 6.07 6.87+do 41.65+6.07do
7.81 4.18 1.9 7.948511606 4.95+do 39.35+7.95do
9.14 5.39 1.9 1.144326245 4.63+do 5.302+1.14do
9.14 2.73 4 10.9261824 2+do 21.85+10.93do
11.84 4.92 4 4.377928192 1.33+do 5.84+4.38do
11.84 2.82 do 2.82do 0.5do 1.41do2
11.84+do4.97+0.42do do 1.24do+0.105do2 0.33do 0.409do2 +0.035do3
Assessment of Retaining Wall Design in Harris Skyline Tower’s …
5.9+do 5.9+do 5.9+do 17.41+5.9do+0.5Do2 0.33do 34.3+17.4do+2.37do2 +0.17do3

75
76
Table 6. Passive force calculation in normal level of ground water table

σv ’ (t/m2 ) σ’ H (t/m2 ) H (m) Passive force (ton) Distance (m) Moment (tm)
0.00 5.78 2.5 14.43796269 1.25+do 18.05+14.44do
1.69 7.86 2.5 9.820811367 0.83+do 8.18+9.82do
1.69 7.09 do 7.09do 0.5do 3.545do2
1.69+do 7.09+4.2do do 2.1do2 0.33do 0.693do3
2.5+do 2.5+do 2.5+do 3.13+2.5do+o.5do2 0.833+0.33do 2.61+3.11do+1.25do2 +0.17do3
S. N. Fitri and A. Soimun
The same procedure was applied in the calculation in the normal level of the
groundwater table. The result analysis of the embedded depth and the maximum
moment is shown as follows:

MPassive =28.84 + 27.37do + 4.795do2 + 0.863do3

Mfinal = Mactive − Mpassive
= 119.45 + 20.47do − 0.606do2 − 0.658do3
Hence do = 7.185
Consired the SF do = 9.685
Final do = 7.5 + 12 = 19.5 m.
The maximum moment :

Msum = − 0.658Do3 − 0.606Do2 + 20.47do + 119.45
= − 0.658(2.92)3 − 0.606(2.92)2 + 20.47(2.92) + 119.45
= 157.647 Ton.m = 1576.47 Kn.m
3.2 Retaining Wall’s Structure Calculation
The basement substructure analysis was conducted in 3 types of material. The first
is steel pipe pile with diameter = 0.9 m, thick = 0.015 m, the area of section = 417
cm2 , inertia moment = 409000 cm4 , and section modulus = 9080 cm3 . The profile
determines by the calculation as follows :
σ = 310000kN
Mdesign 2409.05
Zo = × 16 = × 106 = 7771.12cm3
σa 310000
Hence, the applied profile with w = 9080 cm3 > 7771.12 cm3
The second calculation is bored pile with fy = 400 Mpa, fc = 30 Mpa, diameter
= 1 m, and result of bored pile analysis :

Ag fç
ρs = 0, 45 −1
Ach f yt

785398.16 30
= 0, 45 −1 = 0.00979
664761.01 250
as π (Dc − db ) 132.73π (920 − 13)
s= = = 58.05 mm (1)
(π/4)Dc2 ρs (π/4)9202 x0.00979
Thus the reinforcement is D13-60mm

The last is sheet pile Z-BOX Piles type CAZ Z-Box
σt = 265000 knE = 200000 Mpa
B = 1400 mm I = 475810 cm4
H = 922 mm S = 408 cm3
W = 10285 cm3
Z o = Mdesign
σa
× 106 2409.05
265000
× 106 = 9090.745 cm3
Control Rowe reduction : w = 13140 cm3 > 9090.745 cm3
The result of the basement substructure design will be the basic calculation to
analysis the material price.
3.3 Plaxis Modeling Analysis
The Finite element method analysis was conducted to know the deformation of the
basement model. The result is shown in Figs 4, 5 and 6.
Based on Fig. 4, the total displacement of steel pipe pile design is 4.1 cm. Figure 5
describes the output of total displacement of the bored pile, the value is 5.05cm, and
sheet pile analysis is 4.12 cm of displacement in Fig. 6. Meanwhile, the result of
Safety factor value is 1.66, 1,56 and 1.6 respectively. Based on Bowles [12], all of
the SF result value indicated a safety condition.
The highest deformation is design of bored pile, and the others have a similar
value. while, in SF assessment, the lowest is also bored pile structure. This result
appropriate with [16], which the SF is corresponding with the deformation analysis.
Fig. 4. Total displacement steel pipe pile

Fig. 5. Total displacement bored pile
Fig. 6. Total displacement sheet pile
The analysis use Mohr-Coulomb model, according to Mahmood et al. [17] has shown
in Table 1. The structure is represented as a linear elastic material, and the model is
simulated using the Mohr-Coulomb model. It has been discovered that embedding
the pile reduces the maximum displacement amplitude.
Table 7. Material price

Type Price
result
Steel pipe pile Rp 31.234.158.000,-
Bored pile D = 1.0 m Rp 11.369.860.000,-
Sheet pile Z-box Rp 17.747.111.000,-
3.4 Material Price Analysis
The material volume of the basement structure design was analyzed with the material
requirement. The calculation is considered to cover all the area of the basement site.
The result of the final material price has shown in Table 7.
Table 7 depicts a material price with 3 type categories. The bored pile material
price is the lowest material cost, and the steel pipe pile is the highest. The steel pipe
pile price is expensive in the market construction
4 Conclusions
The result shows different parameters in 3 alternative retaining wall structures. The
Steel Pipe pile has the lowest deformation and highest Safety factor calculation.
However, the material cost is very expensive. The bored pile is the lowest material
cost with Safety Factor value 1.56 and 5.05 cm deformation model. In the case of
the Harris skyline basement project, the considerable structure that applicable to
construct is a retaining wall with a bored pile structure.
References
1. Ng WM, Khor EL, Tiong LK, Lee J (1998) Simulation modeling and management of large
basement construction project. J Comput Civ Eng 12(2):101–110. https://doi.org/10.1061/(asc
e)0887-3801(1998)12:2(101)
2. Wen YW, Liu SY, Hu ML, Zhang GZ (2013) Deformation control techniques for existing build-
ings during construction process of basement. Yantu Gongcheng Xuebao/Chinese J Geotech
Eng 35(10)
3. Ray N, Setia L, Wibowo B, Cahyono MSD (2018) Perencanaan Dinding Penahan Tanah Pada
Basement Grand Dharmahusada Lagoon Surabaya, pp 1–6
4. Widyarini EL (2016) Perencanaan Basement 5 Lantai Dan Pondasi Gedung Spazio Tower Di
Surabaya Barat. Tugas Akhir Jur. Tek. Sipil ITS
5. Kurniawan DA (2017) Perencanaan Dinding Penahan Tanah Pada Basement Midtown Point
And Ibis Styles Hotel Jakarta. Tugas Akhir Jur. Tek. Sipil Its
6. Livando R, Kawanda A (2020) Perancangan Dinding Penahan Tanah Pada Basement Dengan
Dukungan Strut-Beam. JMTS J. Mitra Tek. Sipil 3(3). https://doi.org/10.24912/jmts.v3i3.8748
7. Surjandari NS et al (2021) Slope stability analysis in various Terraces model (case study :
Sendangmulyo, Tirtomulyo District, Wonogiri Regency). J Phys Conf Ser 1858:2021. https://
doi.org/10.1088/1742-6596/1858/1/012005
8. Nurlita Fitri S (2021) Perkuatan Oprit Jembatan Kali Jubang Jalan Tol Pejangan-Brebes Timur
menggunakan Geotekstil. BENTANG J Teor dan Terap Bid Rekayasa Sipil 9(1):37–46
9. Tan Y, Chow C (2008) Design of retaining wall and support systems for deep basement
construction–A Malaysian experience. Deep Excav Retaining Walls
10. Lateef HH, Burhan AM (2019) Time-cost-quality trade-off model for optimal pile type selection
using discrete particle swarm optimization algorithm. Civ Eng J 5(11):2461–2471. https://doi.
org/10.28991/cej-2019-03091424
11. Ahmad Soimun YL, Arif M (2016) Perencanaan dinding penahan tanah untuk pembuatan
basemnet Harris Skyline Tower Surabaya. Tugas akhir Inst. Teknol. Sepuluh Nop
12. Bowles JE (1996) Foundation analysis and design 5th edn. McGraw-Hill handbooks
13. Das BM (2006) Principles-of-geotechnical-engineering. Thomson, pp 1–10
14. Kanoun F, Haffoudhi S, Bouassida M (2017) Design and Follow up of 20 m Depth Nailed
Wall. In: ICSMGE 2017—19th international conference on soil mechanics and geotechnical
engineering, vol 2017
15. Jasmine Nisha J, Muttharam M (2017) Deep excavation supported by diaphragm wall: a case
study. Indian Geotech J 47(3):373–383. https://doi.org/10.1007/s40098-017-0230-1
16. Srivastava A, Kumar P, Sivakumar GL (2012) Stability analysis of 18 M deep excavation using
micro piles. In: Proceeding Indian geotechnology conference
17. Mahmood M, Al-Wakel S, Abdulwahhab I (2018) Three-dimensional analysis for piled raft
machine foundation embedded in sand. In: MATEC Web of conferences, vol 162. https://doi.
org/10.1051/matecconf/201816201023
Additional Vertical Movement
of the Single Pile Foundation
with Combined Loads
Sumiyati Gunawan, Niken Silmi Surjandari, Bambang Setiawan,

Abstract Pile-foundation support the upper structure to the subgrade, it generally

supports vertical loads, occasionally when the horizontal loads are found to be more
dominant, it becomes an important consideration in the design. Pile foundations with
combined loads are generally analyzed separately, however on the site, the two loads
work simultaneously. In-Indonesia, particularly, loading tests do not performed once.
Therefore, it does not account for the additional vertical displacement and lateral
deflection. This research will investigate the analysis and experimental laboratory,
effect of vertical displacement on single pile foundation, as the effect of combined
loads, and get a relationship that can be used as a reference for analyzing additional
vertical displacement due to combined loads on single pile foundations. The process
was in three stages with the first being the the preliminary analysis through 2D Finite
Element Method (2D-FEM) Plaxis2D8.6 while the second was an experimental test
in the laboratory, and third, being the analysis through 3DFEM Plaxis3D1.1 for
validation with soil data and loading test on several projects in the field. The results
obtained show that after adding vertical load (Pu) about 5 × lateral load (Hu), that on
(Pu/Hu > 5), the vertical displacement stays, (as the result of the loading test) until
it collapses, when (Pu/Hu ≤ 5), the lateral load will cause a significant increase in
vertical movement due to the combination of the load.
Keywords Vertical-horizontal load · Single pile · Vertical displacement
S. Gunawan (B) · N. S. Surjandari · B. Setiawan · Y. M. Purwana

Department of Civil Engineering, Sebelas Maret University, Jawa Tengah, Indonesia
N. S. Surjandari
B. Setiawan
Y. M. Purwana
84 S. Gunawan et al.
1 Introduction
The vertical loads of a structure are usually carried by pile-foundation but horizontal
loads are dominant in some other structures. This means it is very important to
calculate the horizontal loads in pile foundations [1]. Combined loads are seldomly
analyzed in simultaneity. This involves calculating the axial load first to determine
the axial carrying capacity and vertical displacement followed by determining the
lateral load to evaluate the lateral bearing capacity and deflection. Meanwhile, the
two loads work simultaneously on the site [2]. In Indonesia, these loading tests are
not usually conducted simultaneously and this means the additional lateral deflection
due to the combined loads is not considered in construction as required by the ASTM
D3966-07 [3].
Previous studies reported the reduction of lateral deflection in pile foundation
under combined loads due to axial loads [4] while lateral loading was discovered
not to be causing any vertical movement but has the ability to increase the move-
ment in combined loads [5]. A study also showed lateral bearing was reducing as
the embedded part of the pile decreased while lateral deflections reduced with the
increase in vertical load on the pile head [6]. A three-dimensional finite element
analysis was also conducted to determine the influence of combined axial and lateral
loads on homogeneous clay and sandy soils. The results showed a significant increase
in the effect of axial load on the lateral bearing capacity in sandy soil and a slight
decrease in clay soil but a substantial influence of axial load was recorded for sandy
soils, even for piles with 30D in length, and a less significance impact was found
with clay soils for piles above 15D in length [7].
Another test on poorly graded sand with variations in the pile and loading also
showed an increase in the lateral bearing capacity as the vertical load was increased
[8]. A numerical study analysis of pile-soil interactions which were subjected to
axial and lateral loads simultaneously using LPILE, a finite element (FE) model with
Abaqus/Cae and SAP 2000, showed an increase in axial load caused a reduction in the
induced bending moment and lateral displacement and this subsequently increased its
capacity to withstand lateral forces [9]. It was also discovered in another research that
the influence of axial loads on the lateral bearing capacity pile increased significantly
in sandy soils but less significant in loamy soils [10].
Vertical loads were also reported to have less effect on lateral resistance in sandy
soil but the progression was observed to be increasing as the soil density increased
[11, 12]. Moreover, the influence of axial load on lateral bearing capacity significantly
increased in sandy soils and slightly in clay soils while the square-shaped pile was
found to have the ability to withstand 1.3 times more load than the round pile [13].
A numerical analysis also showed the effects of combined loads are beneficial but
the interactions were very complex and, depending on the load conditions, there is
a possibility of a contrary effect on system rigidity and max load [14]. A limited
experimental research discovered the application of a static axial load has a minimal
effect on the lateral behavior of micropiles fixed in rigid clay soil [15]. Furthermore,
Additional Vertical Movement of the Single Pile Foundation … 85
some studies also provided evidence that lateral loads were decreased in combined
loads even though deflection was reduced by axial loads due to their presence.
Although previous studies reported that the axial loads reduce deflection due to
lateral loads, lateral loads also increase vertical movement in the load combination,
this has been proven by the preliminary 2D Finite Element Method test (2D-FEM).
This research will investigate the analysis and experimental laboratory, effect of
vertical displacement on single pile foundation, as the effect of combined loads, and
get a relationship that can be used as a reference for analyzing additional vertical
displacement due to combined loads on pile foundations.
The process was divided into three stages, the first being the preliminary analysis
conducted using the 2D Finite Element Method (2D-FEM) Plaxis2D8.6 approach,
the second was the Laboratory Model Experimentation test. The model used was a
(1,5 × 1,5 × 1,2) m3 box filled with silty-sand soil, pile foundation steels at 0,5 and
0,6 m lengths and 0,015 and 0,002 m diameters, load variations. The third with being
the analysis conducted using the 3D Finite Element Method (3D-FEM) Plaxis3D1.1
approach based on the results of loading tests in several projects in the field.
2 Metodology
2.1 Stage I, Analysis with 2D Finite Element Method

(2D-FEM)
Keywords: The preliminary analysis with 2D Finite Element Method (2D-FEM),

modeling of a single pile foundation using the axisymetry menu on 2D Finite Element
Method (2DFEM) comprised of variation on loading direction (axial, lateral, combi-
nation of axial and lateral), variation on length and diameter of the pile, variations
on soil type (soft, medium, dense, homogeneous and layered), variations on ground-
water table condition, (submerged and not submerged). Variation of soil parameter
and loading direction in this research which was conducted using 2D-FEM presented
in Tables 1 and 2.
2.2 Stage II, Laboratory Model Experimental Test
In this study, The model a test box with a size of (1,5 × 1,5 × 1,2) m3, it filled with
silt sand soil, as a description of conditions in the field (samples taken in Karang,
Kalitirto, Berbah, Sleman, Special Region of Yogyakarta), the single pile foundation
using steel with length of 0,5 and 0,6 m, diameter of 0,015 and 0,02 m. Given loads,
axial, horizontal, and combined loads. The laboratory model can be seen in Fig. 1.
The variation of the length and diameter of the pile, and combined load is described
in Table 3,
Table 1 Variation of soil parameter on 2D-FEM

Name T3 T5 T6 Pile
Layer 1 Layer 2 Layer 1 Layer 1 Layer 2 Layer 3 Layer 4
Material MC MC MC MC MC MC MC LE
model
Condition D D D D D D D Non Porous
γk kN/m3 16 17 16,5 11,86 12,22 12,86 13,57 24
γsat 20 21 20 17,50 18,03 17,67 17,83
kN/m3
E kN/m2 1,2.105 1,2.105 8.104 1,6.106 1,6.106 1,6.106 1,6.106 25,33
Poisson 0,3 0,3 0,3 0,3 0,3 0,3 0,3 0,3
Ratio, μ
c kN/m2 1 1 1 2,05 2,30 0,05 0,02
ϕ(o ) 30 33 31 30 30 35 40
Interface, 1 0,7 1 0,4 0,35 0,7 0,7
Rinter
Note D = Drained, NP = Nonporous, MC = Mohr Coulomb, LE = Linear Elastic
Table 2 Variation of load on 2D-FEM

Type of soil Pile dimensions Load Kn, until it collapses
D (m) L (m) Lateral Axial Combined
(5–600) KN (5–12.000) kN (kN)
√ √ √
T5 (soft soil) 1 layer 0,65 17
√ √ √
30
√ √ √
T3 (medium) 0,65 22
2 layer √ √ √
26
√ √ √
30
√ √ √
T6 (dense) 0,6 17
4 layer
2.3 Stage III, Analysis with 3D Finite Element Method

(3D-FEM)
The analysis using the 3D Finite Element Method (3D-FEM) comprised of modeling
a single pile foundation soil data from several projects in the field, namely: Citarum
Bridge, South Jakarta Cikampek II; Dompak Bridge, Tanjung Pinang, Riau; PLTU
Batang, Central Java, The variables used including axial, lateral, and combined loads,
of when it collapses, are presented in Tables 4 and 5.
Fig. 1 Laboratory test model
Table 3 Variation of load on laboratory test model

Type of soil Pile dimensions Load, until it collapses
Dm Lm Lateral gr Axial gr combined loads gr
√ √ √
Silty sand 0,015 0,5
√ √ √
Silty sand 0,02 0,6
Table 4 Variation of soil parameter on 3D-FEM

Project Depth Type Gs w γb γk γsat Soil parameters
(m) of soil (%) (kN/m3 ) (kN/m3 ) (kN/m3 ) ϕ (o ) C
(kN/m2 )
PLTU 1 to 8 Silty 2,6 40,0 16,6 11,86 17,50 25,0 2,05
Batang sand
8 to 12 Silty 2,6 35,0 16,5 12,22 18,03 25,0 2,30
sand
12 to Sand 2,7 40,0 18,0 12,86 17,67 30,0 0,05
20
20 to Sand 2,7 40,0 19,0 13,57 17,83 35,0 0,02
30 gravel
Dompak 1 to 6 Silty 2,5 20,0 15,5 12,92 19,62 15,0 2,20
Bridge sand
6 to 12 Sandy 2,5 20,0 15,8 13,17 19,62 20,0 2,10
silt
12 to Sand 2,7 20,0 16,6 13,83 20,39 25,0 0,05
22
(continued)
Table 4 (continued)
Project Depth Type Gs w γb γk γsat Soil parameters
(m) of soil (%) (kN/m3 ) (kN/m3 ) (kN/m3 ) ϕ (o ) C
(kN/m2 )
22 to Sand 2,7 20,0 19,5 16,25 20,64 35,0 0,02
34 gravel
Citaru 0 to 1 Top 2,5 20,0 15,5 12,92 19,62 15,0 2,20
Bridge soil
1 to 4 silty 2,5 20,0 15,8 13,17 19,62 17,0 3,00
clay
4 to 14 Sand 2,7 20,0 19,7 16,42 20,64 35,0 0,02
14 to silty 2,6 20,0 17,6 14,67 20,14 35,0 0,02
60 sand
Table 5 Variation of the load on 3D-FEM

Project Pile dimensions Load (kN) (until collapsed)
D (m) L (m) Lateral Axial Combined
√ √ √
Citarum bridge 1,2 20,0
√ √ √
Dompak bridge 1,0 18,0
√ √ √
PLTU batang 0,6 17,0
3.1 Result
The results of the preliminary analysis with 2D-FEM, with axial loading, lateral
loading, combined loads (axial and lateral), variation of length and diameter of the
pile, soft soil type, medium and dense, homogeneous soil and layered, submerged
and not submerged in water, illustrates movement at the head of the pile due to axial
and lateral loading are presented in Figs. 2 and 3.
Experimental tests of laboratory scale models, with silty sand as a description of
conditions in the field, while the single pile foundation uses steel with lengths of 0,5
and 0,6 m, diameters of 0,015 and 0,02 m, movement at the head of the pile with
variations in axial loads and lateral loads in Fig. 4 and vertical movement because
of the combination of loads in Fig. 5.
Analysis with 3D-FEM, with axial loading, lateral loading, and combined loads
(axial and lateral), based on soil data from several projects in the field, movement at
the head of the single pile due to axial and lateral loading as in Figs. 6 and 7.
Fig. 2 Movement at the head of the pile (2D-FEM). a Displacement vertical without lateral load.
b Deflection horizontal without axial load
Fig. 3 Vertical displacement With Combined Loads (2D-FEM)

Fig. 4 Movement at the head of the pile (laboratory). a Displacement vertical without lateral load.
Fig. 5 Vertical displacement with combined loads silty sand (laboratory)

Fig. 6 Movement at the head of the pile (3D-FEM). a Displacement vertical without lateral load.
3.2 Discussion
Previous studies reported the la teral deflections reduced with the increase in vertical
load on the pile head [6], another test on poorly graded sand with variations in the
pile and loading also showed an increase in the lateral bearing capacity as the vertical
load was increased [8]. It was also discovered in another research that the influence
of axial loads on the lateral bearing capacity pile increased significantly in sandy
soils but less significant in loamy soils [10]. Although previous studies reported
that the axial loads reduce deflection due to lateral loads, lateral loads also increase
vertical movement in the load combination, this has been proven by the preliminary
2D Finite Element Method test above. Moreover, something new was found with
the preliminary FEM 2D analysis and this was that at certain comparisons of Pu
(axial load) and Hu (lateral load), whereas the addition of the vertical load averages
5.0 × the lateral load, at (Pu/Hu > 5.0) the vertical decrease tends to remain ((as
the result of the loading test) until it collapses, when (Pu/Hu ≤ 5), the lateral load
will cause a significant increase in vertical movement due to the combination of the
load. Following up on the preliminary 2D-FEM analysis test above, the research was
continued with a experimental test in the laboratory and validation by modeling the
3DFEM using based on loading test data on several projects in the field. It shows
things that confirm the same thing at certain comparisons of Pu (axial load) and Hu
(lateral load), whereas the addition of the vertical load averages 5.0 × the lateral load,
at (Pu/Hu > 5.0) the vertical decrease tends to remain ((as the result of the loading
Fig. 7 Vertical displacement

with combined loads
(3D-FEM)
test) until it collapses, when (Pu/Hu ≤ 5). This, therefore, needs to be considered in
calculating and planning the pile foundations’ bearing capacity.
Full-scale models also need to be verified through further studies to obtain
satisfactory results.
4 Conclusion
This research focuses on investigate the analysis and experimental laboratory, effect
of vertical displacement on single pile foundation, as the effect of combined loads,
and get a relationship that can be used as a reference for analyzing additional vertical
displacement on single pile foundations. Following the preliminary 2D-FEM analysis
test, the analysis was continued laboratory experiment and further validated by 3D-
FEM model using loading test data of several projects in the field. The test results
for experiment in the laboratory and vali-dation by modelling the 3D-FEM not much
different from 2D-FEM, that when the vertical movement was increased significantly
at (Pu/Hu > 5.0), the lateral load has no effect on vertical movement. Hence the
behaviour, needs to be considered in designing the pile foundations bearing capacity.

The results obtained showed that after adding vertical load (Pu) of about 5 × lateral
load (Hu), that is (Pu/Hu > 5), the vertical displacement remain constant (like the
loading test) until it collapsed. Also when (Pu/Hu ≤ 5), the lateral load will increase
the vertical movement significantly due to the load combination, with axial load is
5 times greater than the lateral load, there is no increase in vertical displacement
observed. However, at an axial load less than 5 times the lateral load, there is an
increase in the vertical displacement that needs to be taken into account.
Acknowledgements The authors appreciate the Soil Mechanics Laboratory, Faculty of Engi-
neering, Universitas Sebelas Maret Surakarta, and Soil Mechanics Laboratory, Faculty of Engi-
neering, Universitas Atma Jaya, Yogyakarta, for supporting the research as well as the Universitas
Atma Jaya Yogyakarta, Universitas Sebelas Maret Surakarta, and The Slamet Riyadi Foundation
for all funding allowances for education and research.
References
1. Brown DA, Morrison C, Reese LC (1988) Lateral load behavior of pile group in sand. J Geotech
Eng 114(11):1261–1276
2. Mandolini GR, Viggiani C (2005) Pile foundations: experimental investigations, analysis and
design. In Proceedings of the international conference on soil mechanics and geotechnical
engineering, vol 16, no 1, p 177
3. Standard A, “D3966 (2007) Standard test methods for deep foundations under lateral load.
ASTM International, West Conshohocken, pa, 2007. https://doi.org/10.1520/d3966-07
4. Ibrahim SF, Al-Soud MS, Al-Asadi FI (2018) Performance of a single pile under combined
axial and lateral loads in layered sandy soil. J Eng Sustain Dev 22(1):121–136
5. Anagnostopoulos C, Georgiadis M (1993) Interaction of axial and lateral pile responses. J
Geotech Eng 119(4):793–798
6. Zhu M-X, Zhang Y, Gong W-M, Wang L, Dai G-L (2017) Generalized solutions for axially and
laterally loaded piles in multilayered soil deposits with transfer matrix method. Int J Geomech
17(4):04016104
7. Karthigeyan S, Ramakrishna V, Rajagopal K (2007) Numerical investigation of the effect of
vertical load on the lateral response of piles. J Geotech Geoenviron Eng 133(5):512–521
8. Nugroho SA et al Pengaruh beban vertikal terhadap daya dukung lateral pondasi tiang. Ph.D.
dissertation, Riau University
9. Khodair Y, Abdel-Mohit A (2014) Numerical analysis of pile-soil interaction under axial and
lateral loads. Int J Concr Struct Mater 8(3):239–249
10. Rajagopal K, Karthigeyan S (2008) Influence of combined vertical and lateral loading on the
lateral response of piles. In: International association for computer methods and advances in
geomechanics (IACMAG), pp 3272–3282
11. Hazzar L, Hussien MN, Karray M (2016) Investigation of the influence of vertical loads on the
lateral response of pile foundations in sands and clays 2
12. Hazzard (2017) Influence of vertical loads on lateral response of pile foundations in sands and
clays. J Rock Mech Geotech Eng 9(2):291–304
13. Abbas JM, Chik Z, Taha MR (2018) Modelling and assessment of a single pile subjected to
lateral load. Studia Geotechnica et Mechanica 40(1):65–78
14. Achmus M, Thieken K (2010) On the behavior of piles in non-cohesive soil under combined
horizontal and vertical loading. Acta Geotech 5(3):199–210
15. Kershaw KA, Luna R (2014) Full-scale field testing of micropiles in stiff clay subjected to
combined axial and lateral loads. J Geotech Geoenviron Eng 140(1):255–261
The Influence of the Number and Height
Steps of Terraced Model on Slope
Stability Analysis
Niken Silmi Surjandari, Siti Nurlita Fitri, and Fenty Madani
Abstract Wonogiri is an area with a high volume of private, public, and logistics
vehicles because it is located in the border connecting East Java, Central Java, and the
Special Region of Yogyakarta. This area also has high susceptibility to landslide due
to its topography. These landslides occur due to the weight of the slope and excess
external traffic loads. Therefore, this research aims to determine the influence of the
number and height steps of the terraced model on slope stability analysis in Wonogiri.
Data were collected from 3 types of traffic loads, namely Own Weight (BS) +
pavement + truck’s load on two paths, BS + pavement + truck’s load on one path, and
BS + pavement. The road class was designed with an assumed average daily traffic
(ADT) of 3000 pcu/day (local road) with a width of 3 m (Direktorat Jenderal Bina
Marga in Tata Cara Perencanaan Geometrik Jalan. Antar Kota, No.038/T/BM/1997.
Badan Penerbit Pekerjaan Umum, Jakarta, 1997 [1]). Furthermore, the slope stability
analysis was determined using the limit equilibrium method. The result showed that
the addition of steps increases the value of the safety factor (SF) significantly. The
greater the steps of the terracing model, the higher the SF value of slope stability.
The height variation with the same number of steps proves that it affects the safety
factor value. Therefore, the more step height designs, the higher the steps cloud the
ability to resist the traffic load.
Keywords Terracing · Steps · Safety factor · Slope
N. S. Surjandari (B) · S. N. Fitri · F. Madani

S. N. Fitri
96 N. S. Surjandari et al.
1 Introduction
Landslide is one of the natural disasters that often occur in Indonesia due to internal
as properties of a slope and external factors found outside, such as seismicity, rain-
fall, vegetation, rock/soil morphology, soil moisture level, seepage, and geological
activity [2].
Therefore, this research determined the effect of landslides in various terracing
models in Sendangmulyo Village, Tirtomoyo District, Wonogiri Regency and a
continuation of the research carried out by Lugut Tri Pramudo et al. [3]. Numerous
methods of slope reinforcement have been carried out in Indonesia, such as soil
nailing, gabions, vegetation, geotextiles, grouting, and terracing. This study was
developed using traffic loads with the most critical safety factor in previous studies
used as a reference to improve the new model. Angles 30° and 60° were used to
determine the slope before and after the rain. Finally, the model was simulated by
two-dimensional slope modeling analysis.
2 Literature Review
Listyawan et al. analyzed the slope stability of the Gondang Dam Spillway in
Karanganyar Regency using the Fellenius Method. The slope was divided into 25
slices with a landslide angle ϕ of 42°. The results of the analysis obtained a safety
factor (SF) of 0.742, which indicated that the slope is unsafe hence reinforcement
is needed. After terracing was made for slope improvement, the safety factor (SF)
value was raised to 1.609 and was safe from landslides (SF ≥ 1.5) [4].
Haryadi and Razali analyzed slope stability using terracing with Fellenius Method
analysis in Geostudio Slope Program. The type 1 variation had 3 steps with the effect
of pore water pressure, at a slope of 19° to obtain SF = 1.608 > 1.5 (safe slope).
Type 2 with a slope angle of 29° obtained SF = 1.601 > 1.5 (safe slope), while type
3 with a slope angle of 29° obtained SF = 1.539 > 1.5 (safe slope). Slope terracing
with 3 steps and without the effect of pore water pressure on types 1, 2, and 3 with
slope angles of 19°, 29°and 29° obtained SF values of 2.833 > 1.5, 2.499 > 1.5, and
2.697 > 1.5 [5].
Terracing is soil and water management systems mechanically built to decrease
the slope length by digging and filling the dirt. Terraces are useful to increase water
penetration into the soil, decrease the amount of surface runoff, and reduce the
possibility of water erosion on the soil [6].
The Influence of the Number and Height Steps of Terraced Model … 97
According to [7], the highway is classified based on its function, class, and status.
The classification of roads based on their function is divided into the following:
1. Arterial roads, such as public roads that serve major transportation with the
characteristics of long-distance trips, high average speed, and the number of
efficiently limited entrances.
2. Collector roads, such as public roads that serve collectors or dividers with the
characteristics of moderate travel, moderate average speed, and a limited number
of roads.
3. Local Roads, such as public roads that serve local transportation with the charac-
teristics of short trips, low average speed, and an unlimited number of entrances.
This road network is usually used for regional activities and to connect different
groups.
The type of pavement model used is asphalt concrete, with a homogeneous mixture
of aggregate and asphalt as a binder at a certain temperature, and is used to accept
high traffic loads. Construction of Hot Mix Asphalt Pavements [8] stated that asphalt
concrete is a mixture consisting of hard asphalt as a binder and coarse, fine, and
filler aggregates by mixing and compaction under hot conditions at certain temper-
atures. Table 1 is the Guidelines for Planning for Flexible Pavement Thickness,
which consists of a minimum thickness of the asphalt concrete surface layer and the
aggregate base layer.
The materials used for asphalt concrete pavement need to comply with established
regulations, such as a List of the weight of contents (γ). Load weight for dead
load (kN/m3 ) based on [1] for asphalt concrete is 22.0 kN/m3 and for concrete is
22.0–25.0 kN/m3 .
Table 1 Minimum thickness of asphalt concrete surface and subbase course

Traffic (ESAL) ACWC Surface course AC-Base Base course
inci cm inci cm inci cm inci cm
< 50,000* 1.0* 2.5 2 5 2 5 4 10
50,001–150,000 2.0 5.0 – – – – 4 10
150,001–500,000 2.5 6.25 – – – – 4 10
500,001–2,000,000 3.0 7.5 – – – – 6 15
1,000,001–7,000,000 3.5 8.75 – – – – 6 15
> 7,000,000 4.0 10.0 – – – – 6 15
* or surface maintenance condition
Fig. 1 Pavement model structure
3 Method
3.1 Data
The soil parameter data used [9, 10] was conducted in Sendangmulyo Village, Tirto-
moyo District, Wonogiri Regency with its analysis consisting of γb, γsaturated, ϕ,
and c. The road class model has an ADT assumption of 3000 pcu/day with a width
of 3 m [1]. The road structure model is shown in Fig. 1.
3.2 Pavement Model
The pavement in this study is asphalt concrete (γbeton = 24 kN/m3 ), while the lower
foundation is mashed concrete with a bottom thickness foundation of 10 cm.
3.3 Vehicle
The design is assumed when two vehicles are parallel with the double-axle MST (2
axles) of 8 tons used to ensure the load for each wheel is 40 kN [1].
3.4 Traffic Load Model
This study simulated 3 types of traffic loads that refer to the calculations above.
A. Type 1
The type 1 load used the road load plus the load of 2 trucks on both roads as
shown in Fig. 2.
Load type 1(2 truck + pavement) = qpavement + 4(p)

= 2.95 + 4(172.7513)
= 693.9552 kN/m2
B. Type 2
Load type 2 is the same model as the previous type, with an additional truck-load
in one path (Fig. 3).
Load Type 2(1 truck + pavement) = qpavement + 2(p)

= 2.95 + 2(172.7513)
= 348.4526 kN/m2
C. Type 3
Type 3 load uses road loads only as shown in Fig. 4.
Type 3(Pavement) = qpavement

= 2.95 kN/m2
Fig. 2 Type 1 traffic model

The results of the original 30° slope angle modeling output and the Type 1 traffic
load (BS + pavement + truckload on 2 roads) before rainfall conditions are shown
in Fig. 5.
Figure 6 shows that at 30° the slope terraces with the addition of Type 1 traffic
loads before the conditions.
4.1 SF Relationship with Existing Slopes, Load Type,

and Season
The slope stability analysis was carried out in 2 types of modeling conditions. The
first was before and after rainfall, while the second was before and after terracing at
30 and 60 slope degrees. The types of loading due to traffic loads are as follows: own
weight, type 1 load (BS + pavement load + truckload on two roads), type 2 (BS +
pavement load + truck on one section, and type 3 (BS + pavement load) as shown
in Fig. 7.
Based on the analysis in Fig. 7, there is an increase in traffic loads during the rainy
season with a decrease in SF when the slope becomes steeper. The analysis carried
out in rainy conditions shows that all safe numbers are below the safe value required
by Bowles (SF = 1.0). Therefore, the vulnerability of landslides in the study area
increases in accordance with the amount of rainfall.
Fig. 5 Analysis of existing slope of 30° + traffic load Type 1 before rain
Fig. 6 Analysis of 30° slope with Type 1 traffic load terracing before rain
Fig. 7 Relationship between SF and slope under several loads and season conditions
4.2 SF Relationship Between SF and Number of Step

with Variation Terraces
In this research, the terracing used 10 variations with 2, 3, 4, and 5 numbers of steps
and varying trap heights. The relationship between the safety factor (SF) value and
the number of steps in many variations are shown in Figs. 8 and 9.
Furthermore, the addition of steps affects the value of the safety factor (SF), as
indicated by the trend line which always increases in each graph. The greater the
number of steps in the slope, the increase in the soil masses on the slopes, therefore
the vulnerability of landslides is also low. In rainy conditions, the number of steps
affects the speed of rainwater flow through the slopes. The more steps in the terracing
model, the slower the speed of rainwater flow, therefore it is not easily eroded by
rainwater. Terracing is proven to increase the slope safety factor with traffic loads as
evidenced by the increase in the Safety Factor with each increase in the number of
steps.
Fig. 8 The relationship between SF and the number of steps in several load conditions (before
rainfall)
Fig. 9 The relationship between SF and the number of steps in several load conditions (after
rainfall)
Fig. 10 The relationship between SF and the step height of the first terracing layer in the 3 steps,
before the rain for slope angles of 30° and 60° (before rainfall)
before the rain for slope angles of 30° and 60° (after rainfall)
before the rain for slope angles of 30° and 60° (before rainfall)
4.3 Relationship Between SF and Steps Height
Figures 10, 11, 12, 13, 14 and 15 compare the heights of the first layer of terraces in
slope angles 30 and 60 under several load types and seasons.
before the rain for slope angles 30° and 60° (before rainfall)
Figures 10, 11, 12, 13, 14 and 15, shows that the same variation of the step height
significantly affects the safety factor (SF) value and the number of steps. The height
of the first layer of the step is higher than the next trap with the ability to withstand
vertical loads in this case traffic loads.
5 Conclusion
The addition of steps significantly affects the value of the safety factor of slopes with
a significant number unable to reached the Bowles’s Safety Number (SF = 1.0). The
greater the steps, the higher the slope safety factor value. Furthermore, the variation
in height in the same number of steps proves that the height affects the safety factor
value with the ability to withstand vertical loads in traffic loads.
Acknowledgements The authors are grateful to LPPM UNS for funding this research through
Research Group grants.
References
1. Direktorat Jenderal Bina Marga (1997) Tata Cara Perencanaan Geometrik Jalan. Antar Kota,
No.038/T/BM/1997. Badan Penerbit Pekerjaan Umum, Jakarta
2. dan Kesumadhama A (1991) Konstruksi Jalan di daerah Pegunungan tropis, Makalah Ikatan
Ahli Geologi Indonesia, PIT ke-20, Desember 1991, pp 471–481
3. Lugut Tri Pramudo H, Djarwanti N, Surjandari NS (2016) Analisis stabilitas lereng dengan
terasering di desa sendangmulyo, Tirtomoyo, Wonogiri, pp 470–475
4. Listyawan AB, Wiqoyah Q, Rahmawan RA (2019) 155–272, pp 155–162
5. Haryadi D, Mawardi M, Razali MR (2019) Analisis Lereng Terasering Dalam Upaya Penang-
gulangan Longsor Metode Fellenius Dengan Program Geostudio Slope. Inersia, Jurnal Teknik
Sipil 10(2):53–60
6. Sukartaatmadja S (2004) Perencanaan dan Pelaksanaan TeknisBangunan Pencegah Erosi.
Institut Pertanian Bogor (IPB), Bogor
7. Indonesian Law no. 22 of 2009 about traffic and road
8. Construction of Hot Mix Asphalt Pavements (1970) Manual ... The Asphalt Institute: thickness
design—full depth asphalt pavement
9. Hawin (2014) Pengaruh Hujan 2 Harian Terhadap Stabilitas Lereng Di Das Tirtomoyo Wono-
giri (Studi Kasus: Desa Sendang Mulyo, Tirtomoyo, Wonogiri). Universitas Sebelas Maret,
Surakarta
10. Surjandari NS et al (2021) Slope stability analysis in various Terraces model (case study:
Sendangmulyo, Tirtomulyo District, Wonogiri Regency). J Phys Conf Ser 1858:2021. https://
doi.org/10.1088/1742-6596/1858/1/012005
Validation of TRMM Rainfall Data
on Slope Stability in Karanganyar,
Indonesia
David Raja Simare Mare , Rr Rintis Hadiani ,

Abstract The use of TRMM rain data is one source of solving the lack of rain
data sources in Indonesia. TRMM satellite data has not been widely used, so the
validation of the data needs attention. TRMM GSMap_NRT is one type of satellite
data that can be used as a source of rain data. However, research on the validation of
TRMM GSMap_NRT rain data has not been widely discussed. This study presents the
validation of the GSMap_NRT TRMM satellite data against the rain gauge’s regional
rain data in the Karanganyar area, Central Java. The data validation was also carried
out on soil stability at the site, considering that the research location is landslide-
prone. This validation compares the safety factor (SF) value from calculating the two
types of rain data (TRMM GSMap_NRT and rain gauge). Rain indicators for regional
rain are Jatipuro, Tawangmangu, and Karangpandan. Rainfall processing used the
Thiessen Polygon method. Infiltration processing used the SCS-CN method. The data
validation used the r correlation equation approach and the calibration coefficient was
obtained. The correlation coefficient and error values are 0.76 and 251.63 mm/month
(20.97 mm/day). The TRMM rainfall data were then analyzed to find the value of
slope stability by analyzing the annual maximum two-day rainfall. The results of this
study are a map of the SF for 7 years (2014–2020). The conclusion of this study is
the SF map for seven years does not show a significant difference. However, further
research in this regard is highly recommended.
Keywords TRMM validation · TRMM GSMap_NRT · Rainfall · Landslide
D. R. S. Mare (B) · R. R. Hadiani · R. H. Dananjaya

Department of Civil Engineering, Sebelas Maret University, Ir. Sutami 36A Street, Surakarta
57126, Indonesia
108 D. R. S. Mare et al.
1 Introduction
1.1 Preliminary
Disaster Information Data of Indonesia The National Disaster Management Agency

(BNPB) of the Republic of Indonesia (2018) noted that 151 landslides occurred in
Central Java Province. One of the mountainous areas that often occurs in landslides
is the area around Mount Lawu, precisely in Karanganyar district, an area located in
the arch of volcanic mountains [1].
Landslides can be caused by rainfall. Water falling down the slope will become a
burden to the soil, affecting its stability. The need for rainfall data is essential in this
case. The limited availability of rain gauges is a significant problem in Indonesia.
Therefore, other data sources are needed to be processed in particular case studies; one
of the data sources is the Tropical Rainfall Measuring Mission (TRMM) data. TRMM
is a joint project between NASA, America, and the Japan Aerospace Exploration
Agency. The TRMM theory was developed to find a rain plan in a watershed with a
rain register station [2].
However, satellite data is considered less representative of the actual conditions
in the field. Therefore, validation of rain data needs to be done. The point of view
carried out in this study is to validate the TRMM data to be used for safety factor
analysis.
2 Research Significant
2.1 TRMM and Rainfall Gauge Data
Research conducted by Syaifullah [3] shows that GSMap_NRT has a better corre-
lation value to data on rainfall at breeding stations compared to data from 3B42RT.
Therefore, this study will use TRMM JAXA GSMap_NRT rain data [3]. The rain
gauges used came from three stations around the study area: Karangpandan, Tawang-
mangu, Jatipuro. Rainfall data was taken from 2014 to 2020 in the wet months
(November–April).
2.2 Homogeneity
Test Data homogeneity test was carried out to ensure that the data used were classi-
fied as homogeneous. According to Buishand (1982), homogeneity is based on the
number of partially adjusted or mean deviations [4].
Validation of TRMM Rainfall Data on Slope Stability … 109
Table 1 Critical value of Q_RAPS and R_RAPS

√ √
n Q/ n R/ n
CI = 90% CI = 95% CI = 99% CI = 90% CI = 95% CI = 99%
10 1.05 1.14 1.29 1.21 1.28 1.38
20 1.10 1.22 1.42 1.34 1.43 1.60
30 1.12 1.24 1.46 1.40 1.50 1.70
40 1.13 1.24 1.50 1.42 1.53 1.74
50 1.14 1.27 1.52 1.44 1.55 1.78
100 1.17 1.29 1.55 1.50 1.62 1.86
∞ 1.22 1.36 1.63 1.62 1.75 2.00
Source Harto [5]
Rescaled Adjusted Partial Sums (RAPS) is a method that determines a relatively

homogeneous method by comparing the calculated value with the value of criticism
by the table of values criticism. QRAPS and RRAPS data are classified as homo-
geneous if the calculated value is smaller than the critical value according to the
Confidence Interval (CI). The critical value of QRAPS and RRAPS is provided in
Table 1.
2.3 The Coefficient of Rain Regional Thiessen
The rainfall intensity is the depth of rainwater per unit time [6]. This study uses the
Thiessen Polygon method to calculate rainfall intensity. The Thiessen method is also
known as the weighted mean method. The area of influence is formed by drawing the
axes perpendicular to the two nearest measuring posts. It is assumed that the variation
of rainfall between stations is linear and that any post is considered to represent the
nearest area [6].
2.4 Calculation of the Depth of Saturated Soil
The use of the SCS-CN method is considered possible to analyze the infiltration
because this method also considers soil conditions in determining the Curve Number
(CN). The basic equation for calculating the depth of adequate rain is formulated in
the following equation [7]:
(P − 0.2S)2
Pe = (1)
P + 0.8S
The maximum potential retention value has the following equation:

25,400
S= − 254 (2)
CN
CN or Curve Number is a function of watershed characteristics such as soil type,
land use, cover crops, and soil management methods [8].
The CN value varies from 0 to 100. The higher the CN value, the more imperme-
able the land surface. An increase in the CN value indicates a decrease in the land’s
ability to store precipitation water. The impact of this is a reduced infiltration volume,
thereby increasing the volume of surface runoff or increasing peak discharge [9].
Research conducted by Amalia [10] concluded that the CN value would affect
the flood discharge. In his research in the Progo watershed, the CN value change in
2006 was 83.5–86.2 in 2020, making the measured hydrograph unit value 16.3%.
The most dominant CN value change is caused by land-use assuming the soil type
will be relatively constant [10].
2.5 The Validity of TRMM Rainfall Data
The analysis was conducted by comparing the annual maximum two-days. The anal-
ysis was performed by calculating the correlation coefficient value and root mean
square error (rmse). Research conducted by Raharjo [2] revealed a close relationship
between the use of TRMM data and local rain stations [2]. This research found that
this value has a pattern that follows the actual rainfall even though the value tends
to be below the estimate. The correlation coefficient (R) value is expressed by a
coefficient that states a linear relationship between two variables. The equation is as
follows:

n xy − x y
R= 2
2 (3)
n x − 2 x n y − 2 y
If the value (R) is approaching 1, the relationship between the two and the closer
to the actual value. The error that occurs is defined as the difference between the two
variables, the estimated value, and the observed value. Root mean square error (rmse)
shows the degree of bias in the estimation model’s estimation. The rmse equation
can be seen as follows:

n 2
i=1 (yi − y )
r mse = (4)
n
If the rmse value between the two is getting smaller, the difference between the
two is getting smaller, and then the data is getting more accurate.
2.6 Slope Stability with the Method Infinite Slope
Alternatively, soil stability can be shown by the value of the safety factor (SF).
In Hardiyatmo Soil Mechanics (2007), Terzaghi divides slope landslides’ causes
consisting of internal and external influences. What is meant by external influence
here is any factor that causes an increase in shear forces without a change in the soil’s
shear strength. The safety factor (SF) is defined as the value of the ratio between the
holding force and the driving force [11].
τ
SF = (5)
τd
Generally, the SF value is taken to be greater than or equal to 1.2–1.5. Table 2 is
the value of the safety factor according to Bowles [12].
This study uses the infinite slope method. This method assumes the variable value
of soil b has the same size in each part and the forces acting on each side. Figure 1
is shown to illustrate the model [13].
Based on the model above, the equation can be modified to be:
c γ tanϕ
SF = + (6)
γ sat Z w Sinα Cosα γ sat tanα
Table 2 Classification of the

Value of the Intensity or occurrence of landslides
value of the safety factor
SF < 1.07 Landslides are common/frequent (unstable
slopes)
1.07 < SF < 1.25 Landslides have occurred (critical slopes)
SF > 1.25 Landslides rarely occur (slopes are relatively
stable)
Fig. 1 Infinite slope method

model
Slope Field
3 Research Methods
The intended research location is in the Karanganyar area. The taken soil data consists
of four points locations spread over three districts: Matesih, Karangpandan, and
Tawangmangu. The central point of the location, which is the reference for TRMM
rain data collection, is at coordinates (− 7.657557, 111.081424). Soil sample data
was taken at four points around the center point, and the samples were processed at
the Soil Mechanics Laboratory, Sebelas Maret University, Surakarta.
Secondary data goes through a data similarity process first to determine its homo-
geneity. Rainfall data TRMM and rainfall gauge experienced homogeneity test with
the RAPS method with a CI value of 95%. Regional rainfall analysis uses the Thiessen
Polygon method with the Quantum GIS application’s help to obtain regional coef-
ficients. Rainfall data of TRMM is not analyzed for regional rainfall. Infiltration
and saturated soil depth values for safety factor analysis using the SCS-CN method
consider land use in the study area. The validation test is validating rainfall data and
use that valid data to analyze the safety factor.
4 Analysis and Discussion
4.1 Soil Data Analysis
Results show that the soil is categorized as sandy soil. Based on the undisturbed
sample analysis, the soil has an average cohesion value and an average shear angle
of 0.073 kN/m2 and 35.84°.
4.2 Rainfall Data
Analysis Homogeneity analysis shows that the data for rainfall gauge and TRMM
rain are homogeneous. Thiessen Polygon method is used to get the area coefficient.
The polygon coefficient has a value of 0.65; 0.25; 0.10 in Karangpandan Station,
Tawangmangu Station, and Jatipuro Station.
Jatipuro Station is considered to have a too-small coefficient so that the coeffi-
cient of Jatipuro Station is combined with the nearest station, namely Tawangmangu
Station. The latest coefficient data obtained are 0.65 and 0.35 at Karangpandan Station
and Tawangmangu Station. Two consecutive daily rainfall analyses were carried out
with the latest coefficient, and two daily rainfall data were obtained. The maximum
annual two-day rainfall values are obtained as shown in Table 3.
Table 3 Annual maximum

Year Rainfall gauge (mm) TRMM rainfall (mm)
two-day rainfall value
2014 68.85 93.64
2015 73.90 81.30
2016 90.35 72.80
2017 101.67 159.63
2018 96.46 52.10
2019 132.65 125.64
2020 128.65 52.30
4.3 Rainfall Station Data
The next analysis is to process the maximum consecutive two-day TRMM rain data
based on the annual maximum two-day rainfall data. The recapitulation of station
rainfall data and TRMM two-day maximum annual rainfall show in Table 3.
4.4 Rainfall Validation and Calibration
Validation analysis uses maximum monthly data from 2014 to 2020. The daily rainfall
data had a value that tends to vary between satellite and actual data. Analysis of
the two-days rainfall data shows the correlation coefficient 0.22 and error value
46.47 mm/day. The Analysis of total monthly rainfall data shows better value than
two-days rainfall data. The correlation coefficient and error values are 0.76 and
251.63 mm/month (20.97 mm/day).
By eliminating data that is far apart from other data, data is obtained with a
tendency to approach the linear line formed. The equation you get is y = 1.0811x
+ 31.205. With the equation above, the calibration is carried out on each annual
maximum rainfall data so that the maximum annual two-day rainfall data is obtained
as follows (Fig. 2; Table 4).
4.5 Safety Factor Analysis Based on Calibrated TRMM

Rainfall Data
Slope stability analysis using a newly developed web-based application called

UNSlide. The required parameter values are the cohesion value, internal shear angle,
slope, saturated soil depth, and the value of the safety factor. Saturated soil depth
depends on the curve number of the field. The curve number value if the field is
74.23.
200 1400 y = 0.5771x + 150.68

y = 0.2833x + 19.614
R² = 0.5797
R² = 0.0497 1200
rmse = 251.63
150 rmse = 46.47 mm/day 1000 mm/month
TRMM
800
100
TRMM
600
50 400
200
0 0
0 50 100 150 0 1000 2000
Actual Actual
(a) (b)
Fig. 2 a Maximum two-days rainfall data. b Monthly rainfall data
Table 4 Annual maximum

Year Before calibration (mm) After calibration (mm)
two-day rainfall value after
calibration 2014 93.64 132.44
2015 81.30 119.10
2016 72.80 109.91
2017 159.63 203.78
2018 52.10 87.53
2019 125.64 167.03
2020 52.30 87.75
The results show that the difference in the value of rain that occurs each year does
not give much difference to slope stability. However, the greater the rain that occurs
each year and occurs in a long period, the greater the potential for landslides to occur.
Figures 3 and 4 show the 2017 and 2018 SF map, which are the map with the peak
and lowest rainfall values.
Fig. 3 Safety factor map in 2017
Fig. 4 Safety factor map in 2018

5 Conclusion
TRMM rain data can be used to overcome the problem of limited sources of rain
data. If the study area does not have reliable rain data, TRMM data can be used.
Validation based on the total monthly rainfall data has a high enough correlation
coefficient and a minimal error than the validation based on two-days rainfall data.
The author suggests validating with more comprehensive rainfall data and various
methods for its utilization. It is expected that more comprehensive rainfall data will
provide more reliable accuracy and validation. The use of various methods to validate
the TRMM rain data is expected to provide a new approach to lead to the most reliable
results in the future.
References
1. Karnawati D (2011) Landslide hazard and community-based risk reduction effort in

Karanganyar and the surrounding area, Central Java, Indonesia. J Mt Sci 8(2):149–153
2. Raharjo SR, Suprapto M, Ikhsan C (2018) Rain design analysis using TRMM (tropical rainfall
measuring mission) method (case study: Cisadane watershed). In: AIP Conference proceedings,
vol 2014. American Institute of Physics Inc.
3. Syaifullah MD (2014) Validation of TRMM data with actual rainfall data in three watersheds
in Indonesia. J Meteorol Geophys 15(2):109–118
4. Buishand TA (1982) Some method for testing the homogeneity of rainfall records. J Hydrol
58(1–2):11–27
5. Harto S (1993) Hydrological analysis. Gramedia Pustaka Utama, Jakarta
6. Suripin (2004) Sustainable urban drainage systems. Indonesia
7. Chow (1988) Applied hydrology. McGraw Hill
8. Triatmodjo B (2010) Applied hydrology. Indonesia
9. Dasanto dan Risyanto BD (2006) Evaluation of landuse change impact on run-off volume case
study: Ciliwung Hulu watershed, West Java. J Agromet Indonesia (2):1–13
10. Amalia M (2011) Analysis of increasing curve number value on flood debit in Progo watershed.
Infoteknik 12(2)
11. Hardiyatmo HC (2007) Soil mechanics 2, 4th edn. Indonesia
12. Bowles JE (1989) Physical properties of soil geology (soil mechanics). Erlangga, Jakarta
13. Dananjaya RH, Djarwanti N, Gunawan BJ (2021) Development of web-based mapping program
for spatially distributed slope stability analysis. In: IOP Conference series: materials science
and engineering, vol 1144. IOP Publishing
Rubberized Asphalt Pilot Road Trial
in Kuwait
H. Al-Baghli, Z. S. Awadh, and S. E. Zoorob
Abstract Results of an investigation carried out at the Kuwait Institute for Scien-
tific Research (KISR) are presented showing the advantages of incorporating Pellet-
Pave (a rubber-bitumen additive in pelletized form) as partial substitute to conven-
tional bitumen in rubberized hot mix asphalt (rubberized-HMA). The Pelletpave was
composed of 18% tire derived crumb rubber blended with 60/70pen grade bitumen
and processed by USA technology provider Phoenix Industries LLC into easy to
handle granular/pellet form. Laboratory mix designs were followed up with full
scale asphalt plant trials and the construction of a pilot trial (mill and replace an
old wearing course) to test out the technology. The trial comprised 3 different test
sections to compare the performance of rubberized-HMA with polymer modified
Marshall asphalt and polymer modified SuperPave asphalt mix formulations. The
test sections were laid on a low-medium trafficked road and one year following the
trial, the performance of the rubberized-HMA appears to be on par with Polymer
Modified HMAs.
Keywords Hot mix asphalt · Rubberized asphalt · PelletPave · Pelletized
1 Background and Justification
Conventional rubberized hot mix asphalt (rubberized-HMA) is a process of blending

waste tire derived crumb rubber (minimum 15% by mass of bitumen) into hot bitumen
H. Al-Baghli (B) · Z. S. Awadh · S. E. Zoorob

Kuwait Institute for Scientific Research, P.O. Box 24885, 13109 Safat, Kuwait
Z. S. Awadh
S. E. Zoorob
118 H. Al-Baghli et al.
(wet process) to produce a highly resilient rubber modified binder (rubberized-

bitumen) for use in road paving. Rubberized-HMAs have an excellent record in rela-
tion to rutting, fatigue, reflective and thermal cracking and environmental durability
[1].
In 2019, the State of Kuwait introduced new and improved specifications
(QCS-2014) for asphalt materials that encompass performance specifications, and
encourage the adoption of performance graded polymer modified binders (PMBs)
and SuperPave mix design methodology [2].
In this investigation it was decided to formulate rubberized-HMA using locally
available 60/70pen grade Kuwaiti bitumen and crumb rubber sourced from discarded
tyres. The objective was to utilize local materials and introduce a viable competitor
to high performance polymer modified HMAs.
Crumb rubber was hence blended with 60/70pen grade bitumen to produce a
rubberized-bitumen blend. The optimum blend, based on a target PG grade of 76 °C,
maximum elastic recovery and minimum non-recoverable creep performance (using
Dynamic Shear Rheometer testing) was obtained at 18% rubber content by mass
of bitumen. This blend (along with a small quantity of hydrated lime stiffener) was
subsequently converted into a pelletized form (PelletPave) via a patented process by
USA technology provider Phoenix Industries LLC [3].
One major advantage of PelletPave is that it can be introduced directly into the
mixer at the asphalt plant during the HMA production process as a partial substitute
to the bitumen content, furthermore, unlike polymer and wet mix rubber modified
bitumens, PelletPave is designed to be stored and handled as granular matter [4, 5].
2 PelletPave Rubberized Asphalt Mix Design Methodology
In this investigation, several Gabbro aggregate trial blends were formulated by KISR
team and investigated at a local asphalt producer’s laboratory which has an ISO/IEC
Standard 17025:2017 (General requirements for the competence of testing and cali-
bration laboratories) valid certificate. The source of all the Gabbro aggregates (coarse
aggregate and crushed sand) used in all the trials was the United Arab Emirates
(Ahgar Al-Fujairah). An example of key properties of the 20 mm Gabbro coarse
aggregates are shown in Table 1. Properties of all coarse and fine aggregates satisfy
local QCS-2014 highway materials specifications.
The design gradations were all in compliance with Type III wearing course HMA
and with nominal maximum aggregate size of 19 mm. For each of the trial asphalt mix
blends, the following key properties were analyzed: Aggregate Effective (Gse ) and
Bulk (Gsb ) specific gravities, Volume of absorbed binder (Vba ), and Voids in mineral
aggregate (VMA). The blend that satisfied all asphalt mix volumetric requirements
had the following particle size distribution (% passing): 25 mm (100%), 19 mm
(99.7%), 12.5 mm (86%), 9.5 mm (77.5%), 4.75 mm (53.1%), 2.36 mm (35%),
1.18 mm (23.7%), 0.6 mm (15.4%), 0.3 mm (11.3%), 0.15 mm (8%), 0.075 mm
Rubberized Asphalt Pilot Road Trial in Kuwait 119
Table 1 Physical properties 20 mm crushed Gabbro test results

Test method Results
Relative density; (oven dry), (SSD) 2.913, 2.926
Apparent relative density 2.952
Water absorption (%) 0.5
% of Flat and elongated particles (ASTM D4791-10 method B) 0
Aggregate crushing value (%) (BS 812-1990) 12
Los Angeles abrasion (%) (ASTM C535-16) 11
Soundness for coarse aggregate (%) (ASTM C88-18) 1
Particles with the specified no. of fracture faces (ASTM D5821-13) 2 or more100%
Flakiness index (%), (BS 818-105:1) and (BS 818-105:2) 8, 22
(6.5%). The filler was composed entirely of Gabbro powder from asphalt plant dust
collectors (Baghouse fines).
In the following stage, PelletPave plus 60/70pen grade bitumen was introduced
to the optimum aggregate blend to produce rubberized-HMAs. The objective at this
stage was to determine the optimum combination of PelletPave and 60/70pen bitumen
to satisfy key volumetric and compaction properties. A gyratory compactor was used
to compact the HMAs which facilitated monitoring the reduction in volume of each
asphalt specimen as the number of compaction gyrations were increased [6].
The laboratory compaction effort selected for this part of the investigation was
suitable for HMAs expected to carry “Medium to Heavy Traffic (i.e., 3 to < 30 million
standard axles (ESALs)”. The Initial, Design and Maximum gyratory compaction
effort was pre-set at Nini = 8, Ndes = 100 and Nmax = 160 gyrations respectively. The
mixing temperature was specified at 180 °C whilst the target compaction temperature
range was 157–158 °C.
In compliance with the SuperPave based local HMA specifications (QCS-2014),
the required gyratory compacted mix density (%Gmm ) values and their equivalent
(% voids) were as follows:
• Nini ≤ 89% (i.e., 11% voids at initial compaction), Ndes = 96% (i.e., 4% voids for
design), and Nmax ≤ 98% (i.e., 2% voids at refusal density).
• At Ndes ; Voids in Mineral Aggregate (VMA) ≥ 13%, and Voids Filled with
Bitumen (VFB) = 65–75%.
Figure 1 shows gyratory mix densification curves of 3 rubberized-HMAs, all
containing 3% PelletPave by mass of total mix. Each curve represents the average
results obtained from 2 gyratory compacted specimens. The 60/70pen bitumen
content was gradually increased from 1.75 to 2.5% in order to assess impact on
mix volumetrics.
Close observation during laboratory mixing showed that for all 3 mixes, the Pellet-
Pave did not disperse adequately and uniformly throughout the asphalt mix, even after
Fig. 1 Gyratory compaction profiles of 3% PelletPave content mixes at various 60/70pen contents.
Each densification line shows the average result from 2 gyratory specimens
several minutes of continuous high temperature mixing. This was a cause for concern,
as the probability of having a non-uniform mix on the road was unacceptable.
More importantly, none of the 3% PelletPave blends shown in Fig. 1 had accept-
able compaction profiles (voids too high) and they all failed to satisfy the QCS-2014
volumetric requirements at Ndes and Nmax . It was therefore decided to produce alterna-
tive blends at a reduced PelletPave content. Figure 2 shows the gyratory compaction
profiles of 2.5% PelletPave mixes at varying 60/70pen content ranging from 3.25 to
4.85% by mass of total mix.
Based on the data shown in Fig. 2, the rubberized asphalt mix with 2.5% PelletPave
plus 4.55% 60/70pen bitumen was found to satisfy voids requirements at Ndes and
Nmax . This mix was reproduced and was shown to result in volumetric properties that
were highly repeatable.
Figure 3 presents a comparison of the average density values at Nini , Ndes and
Nmax for all rubberized asphalt mix trials as a percentage of the maximum theoretical
density values (%Gmm ) for each formulation. As a reminder, the QCS-2014 require-
ment are as follows; Nini ≤ 89%, Ndes = 96%, Nmax ≤ 98%, and mix (2.5% PelletPave
+ 4.55% 60/70pen) was found to possess optimum volumetric properties.
Fig. 2 Gyratory compaction profiles of 2.5% PelletPave content mixes at various 60/70pen
contents. Each densification line shows the average result from 2 gyratory specimens
Fig. 3 %Gmm versus rubberized asphalt mix composition

3 Discussion on the Performance of Rubberized Asphalt

Mix
It must be noted that the total binder content (i.e., 2.5% PelletPave + 4.55% 60/70pen
bitumen) for the optimized rubberized asphalt mix was equal to 7.05%. To ensure
adequate volumetric properties (i.e., acceptable voids, VMA, VFB), typical non-
rubber modified continuously graded mixes have optimum binder contents in the
range of 5–6%, thus the rubberized-HMA designed in this investigation can be clas-
sified as a binder rich mix. Binder rich mixes are more durable to environmental
exposure (i.e., better resistance to water stripping and oxidation) due to the thicker
binder layer coating the aggregates, on the other hand, a thicker binder layer has
the potential to reduce the mix resistance to externally applied shear stresses (i.e.,
reduced rutting resistance).
In order to assess the mix performance to combined water resistance and rutting,
Immersion wheel tracking tests at 60 °C (20,000 cycles) in compliance with
(AASHTO T324) were carried out on the optimized rubberized-HMA described
above. Figure 4 shows the result on the (2.5% PelletPave + 4.55% 60/70pen) as
shown by the black line. The pass/fail criteria in QCS-2014 is a maximum of 12 mm
rut depth at 20,000 cycles and the results (see black line) indicate that unfortunately
the mix is susceptible to high temperature rutting (though visual inspection of the
tracked specimens did not reveal any binder stripping which was a very positive
indicator).
Fig. 4 Immersion wheel tracking test results on optimized 2.5% PelletPave mix plus 4.55%
60/70pen bitumen (black line) in comparison to 2 runs of 2.5% PelletPave plus 4.55% PMB mix
(yellow and blue lines)
Thus far it was possible to conclude that the optimized rubberized asphalt mix
(2.5% PelletPave + 4.55% 60/70pen) would not be ideal for medium to heavy volume
trafficked roads (> 3 million standard axles). On the other hand, for low intensity
trafficked roads which the mix was primarily designed for, the mix is expected to
perform to a very high standard with no stripping or fretting or raveling which are
primary causes for concern with local asphalt mixes utilizing Gabbro aggregates.
An additional variation on the optimized mix (2.5% PelletPave + 4.55% 60/70pen)
was investigated by replacing the 60/70pen bitumen component by the same mass
of polymer modified bitumen (PMB). The objective was to assess the capability of
PelletPave mixes to be designed for heavily trafficked roads. The polymer used was
Elvaloy and the performance grade of the PMB was PG 76H-10. The mix designation
thus became (2.5% PelletPave + 4.55% PMB).
Figure 4 shows wheel tracking results of the (2.5% PelletPave + 4.55% PMB) mix
(colored lines). Compared to the (2.5% PelletPave + 4.55% 60/70pen) (black line),
a significant improvement in rutting performance was observed. The PMB modi-
fied rubberized-HRA readily satisfied QCS-2014 SuperPave rutting and stripping
requirements.
The resistance of the optimized rubberized-HMA was also tested using Tensile
Strength Ratio (TSR) test (AASHTO T283). Six nearly identical gyratory specimens
were produced, 3 were tested in a dry state whilst the other 3 were conditioned in water
at 60 °C for 24 h and all specimens tested for tensile strength at 25 °C. For SuperPave
mix designs, the minimum TSR must be 80% which was readily satisfied by the (2.5%
PelletPave content + 4.55% 60/70pen bitumen) rubberized-HMA formulation.
4 Full Scale Mixing Trials at the Asphalt Plant
Preparation work for a full-scale asphalt plant production run were initiated with
careful batching of PelletPave (20 kg batches) into polyethylene quick melt bags to
enable direct manual feeding of the PelletPave into the mixer during production. The
plant operators were faced with the problem of PelletPave granules adhering into
clumps (during material stacking and storage under high ambient temperatures). A
team of workers spent an entire week manually sieving out the clumps and retaining
the correct sized granules in order to weigh the material into the required 20 kg
batches.
Eventually, a mixing trial was carried out at the contractor’s Asphalt Plant. In the
first step, 2 batches of dry graded aggregates (no added bitumen) were produced to
ensure that the asphalt plant drying drum were heating up the aggregates to the correct
target temperature range (approx. 190 °C). The aggregate gradation was identical to
the optimum designed in the laboratory designed as described earlier.
In the second step, 2.5% PelletPave was manually introduced into the mixer in
20 kg pre-batched bags and mixed with the graded hot aggregates in the temperature
range between 180 and 190 °C. Seven trial batches were produced with different
mixing times, namely 40, 50, 60 and 70 s. Each completed (PelletPave + hot aggre-
gate) batch was dumped on the ground for inspection. The temperature of each
trial batch was measured and close visual inspection carried on each batch out
to identify the presence of PelletPave semi-molten dark clumps. Clumping is an
indicator that the particular combination of mixing time and/or temperature was
insufficient to evenly and uniformly disperse the PelletPave throughout the mix.
This process was repeated until everyone were satisfied with the outcome of the
“semi-dry” PelletPave-aggregate mix.
The final optimum Plant mixing temperature was achieved when batches were
maintained at 187–190 °C. The hot graded aggregate and 2.5% PelletPave were
initially dry mixed for 75 s, this stage was followed by introducing 4.55% 60/70pen
bitumen followed by 30 s wet mixing, resulting in a laying temperature of around
177 °C. This protocol ensured a well dispersed PelletPave fully coated mix.
A trial strip (25 m × 3.5 m) of rubberized-HMA was next laid and compacted
adjacent to the asphalt plant. No problems were encountered laying the material
and no issues with binder adherence to the steel compactors (which were lightly
lubricated with vegetable oil). Visual inspection of cores taken from the trial show
that the rubberized asphalt layer to be very well compacted with no visible mix
segregation or large voids.
5 Pilot Road Trial
In the following phase, a full-scale pilot road trial was conducted in order to assess the
performance of PelletPave rubberized-HMA during full scale production, transporta-
tion, paving and compaction operations. The secondary objective was to compare the
long-term performance of the PelletPave mix with other high-performance polymer
modified HMAs. An old internal road carrying medium–low volume traffic was
selected for this trial. The aged wearing course on a stretch 450 m in length was
milled to a depth of 5 cm, the milled sections was then divided into 3 sections and
these were subsequently re-laid with 3 new HMA wearing course formulations as
follows:
Section 1 (100 m in length): Rubberized-HMA (2.5% PelletPave + 4.55%
60/70pen),
Section 2 (140 m in length): SuperPave HMA (PG 76H-10 PMB binder),
Section 3 (210 m in length): Conventional Marshall HMA (PG 77H-10 PMB
binder).
Details of the final gradations and binder contents of the Rubberized-HMA,
Marshall and SuperPave mixes as laid during the KISR Pilot Road trial is shown
in Fig. 5.
The 60/70pen bitumen used in the rubberized-HMA mixes had the following
properties: Density at 15 °C (1.0508 kg/L), Cleveland open cup flash point (318 °C),
Loss on heating (0.05% by mass), Ductility at 25 °C (> 150 cm), Penetration at
Fig. 5 Gradations of Marshall, SuperPave and rubberized asphalt mixes laid during the KISR pilot
trial. The dashed lines are the Kuwait QCS-2014 upper and lower limit specifications
25 °C (65 dmm), Softening point (53 °C), Solubility in trichloroethylene (99.85%

by mass).
Rheologically the PMB produced was equivalent to a PG grade 76H-10 and
primary test results obtained from the unaged PMB, which were all within QCS-
2014 limits are as follows: Density at 15 °C (1.037 kg/L), Cleveland open cup flash
point (317 °C), Viscosity at 135 °C (2.17 Pa s), G*/sinδ @ 76 °C and 10 rad/s
(1.98 kPa), Separation test (ASTM D7173-20) difference in G* @ 76 °C between
top and bottom of oven conditioned binder specimen (2.8%).
For the SuperPave and Modified Marshall mixes, the same PMB binder grade was
utilized for both mix types. The polymer type was Elvaloy 5160 (Dupont Elvaloy
RET) at a concentration of 1.7%. At the asphalt plant, blending of the polymer with
bitumen was carried out using a Marini PMB High Shear Mill unit using 2 passes
through the mill at 5000 rpm followed by 10 min in an agitation tank. From the
agitation tank each batch of PMB was next fed into the storage tanks and kept under
additional circulation for a minimum of 35 min. The entire blending/agitation/storage
process was maintained at 185 °C. For each working day, the PMB was kept at high
temperature in storage tanks (with constant circulation) for up to 6 h.
Tables 2 and 3 show the key volumetric and mechanical properties of the Super-
Pave and Modified Marshall wearing course mixes, both designed at optimum binder
contents using the same Gabbro aggregates and PMB. It is interesting to note that
the Superpave mix designed at 4% voids is 0.2% richer in binder content compared
to the Marshall mix at 6.8% voids, despite both mixes having very similar aggregate
gradations (see Fig. 5). Compared to the SuperPave mix, the VMA result was higher
Table 2 SuperPave mix

Test Design value QCS-2014
design summary, Type III
wearing course Ninitial , Ndesign , Nmax 8, 100, 160 Design ESALs 10 to <
30 msa
Design binder content 4.2 –
(%)
SG of combined 2.847 –
aggregate (Gsb )
Effective SG of 2.861 –
combined aggregate
(Gse )
Effective binder 4.0 Report
content (Pbe )
Max. SG of mixture 2.664 –
(Gmm )
% Gmm @ Nini 87.7 ≤ 89.0
% Gmm @ Ndes 96.3 96.0
VMA @ Ndes (%) 13.6 ≥ 13.0
VFB @ Ndes (%) 73 65–75
Test Design value QCS-2014
% Gmm @ Nmax 96.8 ≤ 98.0
Dust proportion based 1.1 0.6–1.2
on effective b.c. (%)
Indirect tensile strength 828 Report
dry set (IDT) (kPa)
Tensile strength ratio 87 ≥ 80
(TSR) (%)
Immersion wheel 2.91 12.5
tracking rut depth (mm)
by 1.8% for the Marshall mix, whilst the VFB was lower by 17%. Hence, the Super-
Pave mix design skeleton was slightly denser, whilst the interstitial spaces within the
aggregate skeleton were significantly more saturated with binder, approaching the
upper limit of specifications.
6 Results of Core and Loose Mix Testing
For each mix type, hot loose mix asphalt samples were collected from behind the
paver for specific gravity determinations and following the compaction process,
cores were extracted from the 3 trial sections and the bulk density was measured and
air voids calculated. The range of measured compacted voids for the various mix
types were as follows: PMB Marshall Mix (4.4–5.9% voids), PMB SuperPave Mix
Table 3 Marshall PMB Type III asphalt mix design, key mix properties
Item o.b.c. QCS-2014 limits
Design bitumen content (% by mass of total mix) 4.0 3.4–4.4
Effective binder content (Pbe ) (%) 3.5 –
Air voids (Va ) (%) 6.8 5.0–8.0
Voids in mineral aggregate (VMA) (%) 15.4 14.0 min
Voids filled with bitumen (VFB) (%) 56 50–75
Marshall stability (kN) 19.1 11.5 min
Marshall flow (mm) 2.71 2–4
Air voids after 400 BPF @ o.b.c. (%) 4.5 4.0 min
Retained stability (%) 87 75 min
Tensile Strength ratio (TSR) at 25 °C, at o.b.c. 76 Reported
(Filler/Binder) ratio 1.12 0.75–1.35
Dust proportion (filler/effective binder content) 1.26 –
Theoretical maximum specific gravity (Gmm ) 2.693 –
Compacted mixture bulk specific gravity (Gmb ) 2.511 –
Immersion wheel tracking rut depth (mm) 5.48 12.5
(6.1–7.3% voids), PelletPave Rubberized Asphalt (7.3–7.6% voids). All results were
within acceptable specification values for field cores.
Loose hot mix samples obtained from the various trial sections were laboratory
compacted. For the PMB Marshall mix compacted at (75 blows Marshall hammer),
the stability obtained was 21.7 kN. For the PelletPave rubberized asphalt mix (75
blows), the stability was 13.5 kN. The rubberized mix was also compacted to refusal
(400 blows Marshall hammer) to assess the effect of densification on the mix mechan-
ical performance, and the stability measured was 22.2 kN. The stability of the rubber-
ized asphalt mix increased significantly with increased compaction effort and hence
the mix is expected to react positively to extra heavy compaction during laying
operations.
The 3 trial sections have been visually monitored for over 1 year and partly as a
result of the low-medium traffic levels, no visible signs of distress were observed.
Falling weight deflectometer (FWD) testing was also carried out after 1 year of traf-
ficking and no significant difference in deflection values were observed between the 3
sections. Thus far, the performance of the 3 trial sections have been indistinguishable.
7 Conclusions
An investigation was carried out to evaluate the advantages of incorporating Pellet-

Pave (a pelletized rubber-bitumen additive) as partial substitute to conventional
bitumen in the formulation of rubberized hot mix asphalt (rubberized-HMA). A
formulation composed of 2.5% PelletPave plus 4.55% of 60/70pen bitumen satisfied

all volumetric requirements according to Superpave based specifications. The binder
rich mix was shown to have excellent resistance to water stripping as evidenced by
both immersion wheel tracking and Tensile Strength Ratio tests. On the other hand,
rutting resistance was compromised making the mix more suitable for up to medium
traffic levels.
Test results on loose mix samples and cores obtained from a road trial showed
that volumetric properties were all within specifications. Marshall stability of the
rubberized-HMA was sensitive to compaction level and hence the mix is expected
to react positively to heavy compaction during laying operations.
Following one year of visual monitoring, no visible signs of distress were observed
as a result of the low-medium traffic levels on this pilot road and falling weight
deflectometer testing revealed comparable deflection values between the rubberized-
HMA section and two other polymer modified HMA sections laid as part of the same
trial.
References
1. Rubber Asphalt Foundation (RAF). http://www.ra-foundation.org/about-raf/. Accessed 29 Mar

2021
2. Qatar Construction Specifications (QCS-2014). Section 06: roadworks, Part 05; Asphalt works.
Fully adopted by Ministry of Public Works, State of Kuwait
3. Amirkhanian S, Sockwell K (2012) Development of polymerized asphalt rubber pelleted binder
for HMA mixtures. In: Asphalt rubber conference 2012, Munich. http://www.ra-foundation.org/
library/
4. Zoorob SE, Al-Bahar SK, Al-Otaibi SF (2018) Design & optimization of a rubber-bitumen blend
in preparation for a rubberized-asphalt road trial in the State of Kuwait. In: 4th International
conference on rehabilitation & maintenance of civil engineering, Solo, Indonesia, 11–12 July
2018. MATEC Web of conferences, vol 195
5. Zoorob S, Al-Bahar S (2018) Preliminary rheological characterization of tyre derived crumb
rubber blended with Kuwaiti bitumen. In: 4th International conference on civil, offshore &
environmental engineering (ICCOEE-2018), 13–15 Aug 2018, Malaysia. MATEC Web of
conference, vol 203
6. Al-Baghli H (2020) Characterization and design of a crumb rubber modified asphalt mix
formulation. World Acad Sci Eng Technol Int J Civ Environ Eng 14(2)
Application of the Updated PSHA
on the Stability Analysis of the Meninting
Diversion-Spillway Tunnel in Lombok
Island—Indonesia
Didi S. Agustawijaya, Ria R. Marlaningtyas, Suryawan Murtiadi,

Mudji Wahyudi, Muhajirah, Hartana, and Ausa R. Agustawijaya
Abstract The Meninting diversion-spillway tunnel is a part of the Meninting dam

project located in West Lombok District. The construction commenced in 2017; then,
a series of severe Lombok earthquakes halted the construction of the tunnel for a while
in 2018, where some landslides occurred around the site. This possibly caused rock
masses around the tunnel to be sheared off during the earthquake events. Updated
probability seismic hazard analysis shows increased seismic parameters, including
peak ground, short spectrum and long spectrum accelerations, which then resulted
in increased stresses around the tunnel. Thus, the stability of the tunnel decreased
in terms of its factors of safety. Although, the overall factor of safety was about 2.5,
the tunnel was still stable in current conditions; the tunnel certainly requires more
stability improvements to face possible similar severe earthquakes in future.
Keywords Lombok earthquake · Meninting tunnel · Updated PSHA · Stress

around tunnel · Factor of safety · Stability improvement
D. S. Agustawijaya (B) · R. R. Marlaningtyas · S. Murtiadi · M. Wahyudi · Muhajirah · Hartana

Post-Graduate Study Program, Department of Civil Engineering, University of Mataram,
Mataram, Indonesia
S. Murtiadi
Muhajirah
Hartana
D. S. Agustawijaya
Geo-Engineering Research Group, Department of Civil Engineering, University of Mataram,
Mataram, Indonesia
A. R. Agustawijaya
Department of Geomatics Engineering, ITS, Surabaya, Indonesia
130 D. S. Agustawijaya et al.
1 Introduction
The Mininting dam project is located in Dasan Gria, West Lombok District, which
is only at 50 km distance from the epicenter of the 5th August 2018 earthquake.
During a series of shallow earthquakes from July to September 2018, the seismic
events have ruined thousands of houses and buildings in Lombok Island [1–3]. These
current events occurred at depths of less than 30 km, which were triggered by the
Flores Fault located at the north part of Lombok Island [4]. The intensity of the
earthquake event was from MMI III to IX, and was MMI VII at the project site [5]
(Fig. 1), which caused some collapsed area around the site [2].
Agustawijaya et al. [6] have updated probabilistic seismic hazard analysis (PSHA)
for Nusa Tenggara Region based on earthquake data of 1922–2018, which resulted in
higher peak ground acceleration, spectral accelerations SS and S1, compared to those
of the SNI 1726:2019 [7]; whereas, the Meninting tunnel was possibly designed based
on the older standards for calculating its design earthquake forces [8]. Therefore, this
paper recalculated the seismic design of the Meninting diversion-spillway tunnel
according to the current PSHA.
Meninting Dam
Fig. 1 The intensity of the earthquake event in 5th August 2018 around Lombok Island, the location
of Meninting dam included into MMI VII [5]
Application of the Updated PSHA on the Stability Analysis of … 131
2 Tunnel Design
The Meninting dam located at Dasan Gria, where the three rock formations: Kali-
palung, Kalibabak, and Lekopiko covered the area [9]. These sedimentary rock mate-
rials contained volcanic breccias, lava and poor compacted volcanic rock materials of
agglomerates and colluviums. Most rock materials were from erupted rock materials
of the Rinjani volcano [10].
The Meninting dam was designed to have two tunnels: intake, and diversion-
spillway. The diversion-spillway tunnel was designed to be of two types: Type 1
diversion tunnel, and Type 2 spillway tunnel. The type 1 was connected to the type
2 with a connecting shaft to become a diversion-spillway tunnel.
As be seen Fig. 2, the particular shoe-shaped type 2 was analyzed. The tunnel had
a dimension of 9.40 × 9.40 × 252.5 m. The elevation of the tunnel base would be at
+ 147.8 m above the MSL, while the top surface of the tunnel would be at + 211.9 m
above the MSL. The tunnel was planned to have fully support systems, including
wiremesh with a 0.15 m thickness of shotcrete, steel H-beams installed with a spacing
A
1.20
θ Earthquake force kP
C
7.00
r
1.20
Concrete K225
E D
σθ
1.20 7.00 1.20
σ rθ
9.40 σ rθ
σr
Fig. 2 Typical type 2 of diversion tunnel with stresses working around tunnel: Crown A, Crown B,
Wall C, Base D and Base E are parts of tunnel being analyzed. Earthquake forces are horizontally
working perpendicular to the tunnel axis, where a is radius and r is a point distance of estimations
[8]
of 1 m, a concrete lining of 0.6 m thickness, consolidation and curtain grouts. Then,

far field stresses, P on vertical and kP on horizontal directions, were estimated on
each part of the tunnel: A, B, C, D, E; and additional stresses of earthquake were
estimated working on perpendicular to tunnel axis.
3 Method
The construction of the Meninting tunnel followed the New Austrian Tunneling
Method (NATM) [11]. The support systems of the tunnels utilized the geomechanics
classification by Bieniawski [12], and suggested geological data by the ISRM [13].
For the current stability analyses in this paper, the estimations of rock mass
strength followed the non-linear Hoek–Brown (HB) [14, 15]:
a
σ3
σ1 = σ3 + σci mb +s (1)
σci
According to Eq. (1), the strength of rock masses will depend on parameters
mb , s and a, which can be obtained from the rock mass rating (RMR) proposed by
Bieniawski [12].
For a seismic area, additional uncertainties will be from earthquake forces. When
earthquake forces work on rock, the rock is under dynamic shear conditions, which
may cause a stress drop up to 10% from the static conditions [16].
Then, the estimation of stresses around the tunnel followed the Kirsch solutions
[14]. Within these estimations of Eq. (2), the far field stresses are P in the vertical
direction, and kP in the horizontal direction. Then, stresses at a point (r, θ) around
the tunnel will be calculated as radial (σr ), tangential (σθ ) and shear stresses (σrθ ),
where θ is the angle between the vertical and r line from the axis point, as in Fig. 2:

a2 3a4 4a2
σr = 0.5P (1 + k) 1 − 2 + (1 − k) 1 + 4 − 2 cos 2θ
r r r
2 4
a 3a
σr = 0.5P (1 + k) 1 + 2 − (1 − k) 1 + 4 cos 2θ
r r
4 2
3a 2a
σrθ = 0.5P −(1 − k) 1 − 4 + 2 sin 2θ (2)
r r
A specific earthquake design standard for tunnels has not been established yet
by the BSNI; but, two national standards area available for embankment dam (Pd
T-14-2004-A) [17], and for building and non-building [7]. These standards might be
adapted for the estimations of earthquake forces for the Meninting diversion-spillway
tunnel. Equation (3) is adapted from the Pd T-14-2004-A to estimate earthquake
forces from peak ground acceleration on the surface and weight of rocks around the
tunnel. Then, the equation uses the earthquake zone coefficient Z, which is depending
on the location on the map of Indonesia earthquake zone [17].
ad
F = a1 × ×W (3)
g
F = horizontal earthquake force

a1 = correction factor for construction type
ad = peak ground acceleration on the surface
. = Z × ac × V
Z = earthquake zone coefficient
ac = peak ground acceleration
V = correction factor for rock/soil type
g = gravity
W = weight.
However, the SNI 1726:2019 [7] has already applied the seismic parameters of
peak ground and spectral accelerations to calculate shear forces. The consideration
in utilizing the standards [7] to the tunnel is that a shoe-shaped underground structure
(Fig. 2) is constructed into rocks; then, seismic forces are transmitted to the structure
through surrounding rocks.
Subsequently, Agustawijaya et al. [6] updated seismic parameters for Lombok
and surrounding islands based on the earthquake data of up to 2018. The updated
PSHA for the event period of 1922–2018 determined parameters for peak ground
acceleration of bedrock (PGA), spectral acceleration at T = 0.2 s (Ss ), and spectral
acceleration at T = 1.0 s (S1 ) for the applied exceedance probability of 2% in 50 years
[6]. Results shows that the PGA values of bedrock were 0.6–0.7 g; the Ss values
were in the range of 1.0–1.2 g, and the S1 values were 0.25–0.4 g for Lombok and
surrounding islands. These values were therefore applied for the current stability
analysis of the Meninting tunnel.
4 Results
4.1 Rock Mass Strength
The Meninting diversion-spillway tunnel was mainly excavated into volcanic brec-
cias; although on the top of the tunnel, rocks contained lava, agglomerates and collu-
viums. The RMR should be different for each rock type; and of course, other rock
material parameters should also be different, such as cohesion, friction, and unit
weight. In general, these rocks had a RMR of 40, which was included into poor rock
mass quality, thus the stand up time for the tunnel was one week without any support
[14].
Table 1 Results of the rock mass strength of the Mininting tunnel type 2
Parameter Crown A Crown B Wall C Base D Base E
Rock V Breccia V Breccia V Breccia V Breccia V Breccia
Unit weight, γ (MN/m3 ) 0.021 0.021 0.021 0.021 0.021
Depth of axis, H (m) 26.6 26.6 26.6 26.6 26.6
σci (MPa) 5.9 5.9 5.9 5.9 5.9
σ3 (MPa) 0.06 0.05 0.06 0.07 0.07
mi 19 19 19 19 19
RMR 40 40 40 40 40
σ1 (1) (MPa) 0.36 0.32 0.36 0.41 0.39
Using this RMR value, mi of 19 and confining pressures of 0.06 MPa, the rock
mass strength of Eq. (1) was within the range of 0.32–0.41 MPa depending on the
rock position on the tunnel (Table 1). Such low σ1 values were very much influenced
by low confining pressures of 0.06 MPa. Also, a RMR of 40 did not really increase
the rock mass strength, as it should be [18].
4.2 Stresses Around the Tunnel
Each part of the tunnel had a different stress concentration working on each part.
On the boundary when the radius a equals the distance of the point estimated r, the
radial stress σr and the shear stress σrθ were zero, so the only tangential stress σθ had
a non-zero value. When r = a + 1 m, stress values on each part of the tunnel can be
seen in Table 2.
The tangential stresses worked significantly on the Crown B, Wall C and Base D;
shear stress concentrations should be on the Crown B and Base D; while, the Crown
A should be tensioned, and some heave might occur on the Base E.
Table 2 Stresses working on each part of the tunnel for r = a + 1 m

Stress (MPa) Crown A Crown B Wall C Base D Base E
σr 0.01 0.08 0.18 0.07 0.01
σθ −0.09 0.22 0.63 0.29 −0.09
σrθ 0.00 −0.20 0.00 0.20 0.00
Table 3 Stresses after earthquake working on each part of the tunnel for r = a + 1 m
Stress (MPa) Crown A Crown B Wall C Base D Base E
Estimations from Eq. (3)
σr 0.013 0.10 0.20 0.08 0.01
σθ −0.09 0.29 0.69 0.31 −0.09
σrθ 0.00 −0.26 0.00 0.21 0.00
Estimations from Eq. (3) with updated PGA [6]
σr 0.013 0.12 0.25 0.09 0.01
σθ −0.09 0.34 0.87 0.37 −0.09
σrθ 0.00 −0.31 0.00 0.25 0.00
Estimations from [6, 7]
σr 0.013 0.14 0.29 0.11 0.01
σθ −0.09 0.39 1.01 0.42 −0.09
σrθ 0.00 −0.35 0.00 0.29 0.00
4.3 Earthquake Stresses
When earthquake forces working on the tunnel were considered to be horizontally

perpendicular to the tunnel axis, the influence of the earthquake forces (3) on the
tunnel depended on the position on the tunnel. The earthquake stresses on the peak
Crown A and Base E were minimum; otherwise, on the Wall C were maximum of
0.04 MPa. Thus, these earthquake forces resulted in the increase of stresses working
around the tunnel (Table 3).
However, when the tunnel has been fully constructed as an underground structure,
the earthquake force estimate procedures of the SNI 1726:2019 [7] may still be
relevant to investigate the influence of earthquake forces on the tunnel. Then, updated
PSHA [6] provided higher peak ground, short and long acceleration values than that
of [7], which were also applied to estimate earthquake stresses working around the
tunnel. Results show higher estimations than that of Eq. (3), (Table 3).
From all estimations, the earthquake stresses certainly increased radial and tangen-
tial stresses working on the Crown B, Wall C and Base D. Particularly, the Wall C
was under higher radial and tangential stresses, such tangential stresses were up to
1.01 MPa.
4.4 Factor of Safety
The stability of the tunnel was described in terms of a factor of safety (FoS) of each
part of the tunnel. The FoS of the tunnel after excavation and under the influence
of earthquakes was calculated in terms of a ratio between rock mass strength and
forces working around the tunnel subject to Kirsch formulations [14], for each part
Table 4 Factor of safety on

Factor of safety Crown B Wall C Base D
each part of the tunnel
Estimations from [17]
σ1 /σr 3.04 1.78 5.10
σ1 /σθ 1.10 0.52 1.30
Estimations from [6, 17]
σ1 /σr 2.57 1.42 4.31
σ1 /σθ 0.93 0.41 1.09
of the tunnel: Crown B, Wall C and Base D, where stresses under the influence of
earthquakes were concentrated.
When earthquake forces were considered to work on the horizontal line perpen-
dicular to the tunnel axis, the most deformed part of the tunnel was the Wall C. The
wall C had only an FoS of 0.52 against the σθ stress. When using the updated data of
PGA [6], the FoS value for the Wall C was higher for about 27%, from that of [17],
which was 0.41 (Table 4).
According to Hoek and Brown [14], considering the RMR of 40, rocks around
the tunnel with such an FoS might have a stand up time of up to 1 week prior to
the installation of tunnel supports. However, this estimated stand up time would
be too long; and it would shorten, probably down to be 25 min when earthquake
forces were added. Thus, supports should be immediately installed as primary and
secondary supports [11, 19].
After the completion of the construction, the tunnel had an increased FoS to
become over 2.0; although, the Wall C had still a lower value than other parts of the
tunnel (Table 5). The FoS value for the Wall C estimated from the updated PGA was
slightly lower than that from the older PGA value [7], which reduced for about 7%.
Overall FoS values may still be relevant to the tunnel, since all tunnel supports were
fully installed, and the underground structure had fairly stability under earthquakes.
The overall FoS of the structure was about 2.5.
Table 5 Factor of safety for

Factor of safety Crown B Wall C Base D
fully constructed Meninting
tunnel Estimations from [6, 7]
σ1 /σr 3.65 1.93 6.01
σ1 /σθ 1.32 0.56 1.52
Estimations from [7]
σ1 /σr 3.84 2.06 6.32
σ1 /σθ 1.39 0.60 1.60
5 Discussion
The stability of the Meninting tunnel was influenced by earthquake forces. The FoS of
the tunnel reduced significantly, particularly on the tangential directions of 0° on the
horizontal line, 45° up and down the horizontal line, when earthquakes forces worked
on horizontal direction. The HB strength criterion seems to have less sensitive to the
influence of earthquakes forces, rather depends on rock properties. One suggested
stability improvement of many suggestions [16] would be the application of grouting
and rock bolting [20]. This could increase the shear strength; also reduce deformation
of rock masses [21, 22].
The Meninting tunnel construction might be sufficient to stand earthquake
stresses, but since it was excavated into a hill of various volcanic rock materials,
the stability of rock masses would be a problem [23, 24]. On the top of the tunnel,
loose bouldery agglomerates and colluviums had low strength, particularly with a
high coefficient of permeability; they could be easily to lose their shear strength under
earthquake forces. As the events in 2018, the surrounding area of the Meninting dam
collapsed, where many landslides occurred during the events [2]. Estimations show
that the upper tunnel cross section could be more vulnerable than the lower part; but,
it might due to the PGA < 0.3 g [25]. Therefore, the stability of the tunnel would not
only depend on the support systems installed to the tunnel; but, it also could depend
on the stability of residual rock mass strength around the tunnel. The residual shear
strength of the rocks could drop 41% after the earthquake events in 2018 [3], this
should be problems for the tunnel stability in future [26].
6 Conclusions
The stability of Meninting diversion-spillway tunnel has been recalculated using the
updated data of peak ground acceleration, short and long spectral accelerations of
Lombok Island. The recalculations resulted in the reductions of the factor of safety of
the tunnel for all calculation procedures of earthquake forces for embankment dam
and building and non-building designs. The FoS reduction could be down to 35%;
although, overall stability of the tunnel should be fine regarding the overall FoS of
2.5. For future stability, the tunnel certainly requires some improvements since the
current stability should be depending on the residual rock strength against possible
similar severe earthquakes.
Acknowledgements Authors acknowledge BWS Nusa Tenggara I (NTB) for data supports and
access to the Meninting dam site.
References
1. Pramono S (2018) Perencanaan Rekonstruksi Wilayah Lombok Berbasis Mitigasi Gempabumi.

Badan Meteorologi, Klimatologi dan Geofisika, Jakarta (2018) (in Indonesian). https://cdn.
bmkg.go.id/web/Artikel_20181002115022_u1dpuf_-Perencanaan-Rekontruksi-Wilayah-
Lombok-Berbasis-Mitigasi-Gempabumi.pdf#viewer.action=download. Last accessed 22 Oct
2018
2. Irsyam M, Hanifa NR, Djarwadi D (2018) Kajian Rangkaian Gempa Lombok, Provinsi Nusa
Tenggara Barat. Pusat Studi Gempa, Pusat Litbang Perumahan dan Pemukiman, Kementerian
Pekerjaan Umum dan Perumahan Rakyat, Jakarta (2018) (in Indonesian)
3. Agustawijaya DS, Sulistyowati T, Layli BA, Agustawijaya AR (2019) The ground deformation
of the Luk Barat Bridge after Lombok earthquakes 2018. IOP Conf Ser Earth Environ Sci
389(012051). https://doi.org/10.1088/1755-1315/389/1/012051
4. Pranantyo IR, Cummins PR (2019) Multi-data-type source estimation for the 1992 Flores
earthquake and tsunami. Pure Appl Geophys. https://doi.org/10.1007/s00024-018-2078-4[las
taccessed23/1/2020]
5. Agustawijaya DS, Taruna RM, Agustawijaya AR (2022) Earthquake forces on the stability of
the Meninting diversion tunnel design based on current seismic conditions of Lombok Island.
Indonesian J Geosci 9(1):105–117
6. Agustawijaya DS, Taruna RM, Agustawijaya AR (2020) An update to seismic hazard levels
and PSHA for Lombok and surrounding islands after earthquakes in 2018. Bull N Z Soc Earthq
Eng 53(4):215–226
7. Badan Standardisasi Nasional (BSNI) (2019) SNI 1726:2019: Tata Cara Perencanaan Keta-
hanan Gempa Untuk Struktur Bangunan Gedung dan Non Gedung (in Indonesian)
8. Balai Wilayah Sungai (BWS) (2017) Sertifikasi Disain Bendungan Meninting di Kabupaten
Lombok Barat. Laporan Utama, Satuan Kerja Balai Wilayah Sungai Nusa Tenggara I, Mataram
(in Indonesian)
9. Andi Mangga S, Atmawinata S, Hermanto B, Setyogroho B (1994) Geological map of the
Lombok sheet, West Nusa Tenggara, scale 1:250.000. Geological Research and Development
Centre, Bandung
10. Lavigne F, Degeai JP, Komorowski JC, Guillet S, Robert V, Lahitte P, Wassmer P (2013) Source
of the great AD 1257 mystery eruption unveiled, Samalas volcano, Rinjani volcanic complex,
Indonesia. Proc Natl Acad Sci 110(42):16742–16747
11. Harjomuljadi S (2010) Terowongan dengan NATM. Mediatama Saptakarya, Jakarta, 90 pp (in
Indonesian)
12. Bieniawski ZT (1989) Engineering rock mass classifications. Wiley, New York, p 272
13. International Society for Rock Mechanics (ISRM) (1981) Rock characterization, testing and
monitoring. In: Brown ET (ed) ISRM suggested methods. Pergamon Press, Oxford
14. Hoek E, Brown ET (1994) Underground excavations in rock. Chapman & Hall, London
15. Hoek E, Carranza-Torres C, Corkum B (2002) Hoek-Brown failure criterion. In: Proceedings
of NARMS-TAC conference, 2002 edn, Toronto, pp 267–273
16. Brady BHG, Brown ET (1993) Rock mechanics for underground mining, 2nd edn. Chapman
and Hall, London
17. Department of Public Work (2004) Analisis stabilitas bendungan tipe urugan akibat beban
gempa, Pedoman Konstruksi dan Bangunan Pd T-14-2004-A (in Indonesian)
18. Priest SD (1993) Discontinuity analysis for rock engineering. Chapman & Hall, London
19. De Farias MM, Moraes AH Jr, De Assis AP (2004) Displacement control in tunnels excavated
by the NATM: 3-D numerical simulations. Tunn Undergr Space Technol 19:283–293
20. Aldiamar F, Kartikasari S, Desyanti (2015) Metode perencanaan penggalian dan sistem
perkuatan terowongan jalan pada media campuran tanah-batuan. Pedoman Bahan Konstruksi
Bangunan Dan Rekayasa Sipil, Kementerian Pekerjaan Umum, Jakarta (in Indonesian)
21. Zhang B, Li S, Xia K, Yang X, Zhang D, Wang S, Zhu J (2016) Reinforcement of rock mass
with cross-flaws using rock bolt. Tunn Undergr Space Technol 51:346–353
22. Agustawijaya DS, Sulistiyono H, Wahyudi M, Yasa IW, Ashari KF, Agustawijaya AR
(2020) Grouting performance of the Pandan Duri dam shacked by Lombok earthquake 2018.
In: MATEC Web of conference, vol 331(02004). https://doi.org/10.1051/matecconf/202033
102004
23. Agustawijaya DS (2019) Practical applications of strength criteria in civil engineering design
for shallow tunnel and underground structures in weak rocks. Int J Technol 5(2):93–105
24. Agustawijaya DS (2018) Influence of rock properties in estimating rock strength for shallow
underground structures in weak rocks. Indonesian J Geosci 5(2):93–105
25. Wen K, Shimada H, Sasaoka T, Zhang Z (2018) Numerical study of plastic response of urban
underground rock tunnel subjected to earthquake. Int J Geo-Eng 8(28). https://doi.org/10.1186/
s40703-017-0066-7. Last accessed 2/11/2018
26. Stiros SC, Kontogianni VA (2009) Coulomb stress changes: from earthquake to underground
excavation failures. Int J Rock Mech Min Sci 46:182–187
Risk and Stability Evaluation of Klego
Dam, Boyolali, Indonesia
Suharyanto, Kresno Wikan Sadono, Rizqi Iqbal Maulana, P. Arie Bagus,

and Dyah Ari Wulandari
Abstract Klego Dam was constructed in the 1990s and has been operated since
1993. Its main function is to supply Irrigation water of 13.53 km2 . It is located right
at the boundary between Bengawan Solo Watershed and Jratunseluna Watershed.
Previously, the dam was managed by BBWS Bengawan Solo. In the 2010s, the
dam’s management is relinquished to Pemali Juana River Basin Authority. This
management’s change contributes to some discontinuity in the dam’s operation and
maintenance, which contributes further to the loss of some important historical data
and information of the dam. This unfavorable situation for proper operation and
maintenance of the dam is further aggravated by the publication of Indonesian’s
new ground speed (earthquake) map for seismic design. It is, therefore, urgent to
evaluate the stability and safety of Klego Dam with the new earthquake map even
under limited available data and information. The paper presents the evaluation on
the status of the dam’s safety considering the seepage and slope stability. It uses
Geostudio software to evaluate stability of the dam under various conditions. It also
performed seepage analysis uses limited Piezometers’ records. The analysis shows
that the safety of Klego Dam is barely sufficient. However the dam needs some
rehabilitation on its embankments, dam’s crest, and its instrumentations as well as the
proper implementation of Maintenance and Operation procedures. The availability
on the SOP and its implementation will reduce the risk of dam’s failure.
Keywords Klego Dam · Stability analysis · Seepage analysis
Suharyanto (B) · K. W. Sadono · R. I. Maulana · P. A. Bagus · D. A. Wulandari

Civil Engineering Department, Engineering Faculty, Diponegoro University, Semarang, Indonesia
K. W. Sadono
D. A. Wulandari
142 Suharyanto et al.
1 Introduction
Klego Dam was built in the 1990s and operated since 1993. Its main function is to
supply irrigation water of 13.53 km2 . Its location is right at the boundary between
Bengawan Solo Watershed and Jratunseluna Watershed. Previously, the dam was
managed by BBWS Bengawan Solo, while its water distribution is operated by
Public Works Office in Boyolali Regency. In the 2010s, the dam’s management is
relinquished to BBWS Pemali Juana, however the Public Works office in Boyolali
still responsible for water allocation and distribution. This management change
contributes to some discontinuity in its operation and maintenance, which contributes
further to the loss of some important historical data and information of the dam.
This unfavorable situation for proper operation and maintenance of the dam is
further aggravated with the publication of new ground speed graph for seismic design
(Indonesian’s earthquake map, 2010). It is, therefore, urgent to evaluate the stability
and safety of Klego Dam using the new earthquake map even under limited data and
information.
2 Literature Review
The inherent risk of a dam’s failure, has stipulated that a dam be evaluated its state of
safety regardless of its limited data/information. Under this limited data/information,
the safety state cannot be analyzed using more rigorous statistical methods, rather a
more simplified one. It uses a method of total risk factor, TRF [1]. The method is
simple and does not require adequacy of historical data or complete probability of
hazards. It uses the risk factors accounting for risk contributing by the dam (Capacity,
Height, and Age), the Downstream hazard (population and property at risk), and
the predicted Damage factor (the dam type and the site-specific seismic hazard),
respectively [1]. In Indonesia, it adopts similar principle and modified into: Guideline
for stability analysis of earth and rock fill dams under earthquake condition [2] which
is referred to PD-T-14-2004 A.
TRF = CRF + HRF + ERF + DRF
where:
TRF: Total risk factor,
CRF: Influence of Dam’s Capacity to risk factor,
HRF: Influence of Dam’s Height to risk factor,
ERF: Influence of Potential Evacuated Persons to risk factor,
DRF: Influence of Downstream damages to risk factor.
Risk and Stability Evaluation of Klego Dam, Boyolali, Indonesia 143
2.1 Risk Classification
In the analysis of total risk, it considers four risk factors in determining risk level of a
dam in Indonesia. They are factors related to dam’s capacity, dam’s height, potential
evacuated persons, and downstream damages. Table 1 defines each factor scores [2].
2.2 Risk Level
Based on the TRF, the dam’s risk level can be classified as low risk (class I), moderate
risk (class II), high risk (class III), and extreme risk (class IV) and defined as follow
(Table 2).
This risk level classification is used to determine the scale of design earthquake
such as shown in the Table 3. The earthquake design uses MDE (Maximum Design
Earthquake) condition.
Table 1 Risk factor and its

Risk factor Classification (in the brackets is weight)
classification
Extreme High Moderate Low
Dam Capacity >100 100–1.25 1.00–0.125 <0.125
(106 m3 ), (6) (4) (2) (0)
CRF
Dam Height (m), >45 45–30 30–15 <15
HRF (6) (4) (2) (0)
Evacuation Need >1000 1000–100 100–1 0
(person), (12) (8) (4) (0)
ERF
D/S damage VH H M Nul
DRF (12) (10) (4) (0)
Table 2 Risk level for a dam

TRF Risk level
safety [2]
0–6 I (low)
7–18 II (moderate)
19–30 III (high)
31–36 IV (extreme)
Table 3 Earthquake design

Risk class No damage allowed Allowable damage
level [2]
condition
T (years) Analysis T Analysis
method (years) method
IV (Ext.) 100–200 Earthquake 10.000 Earthquake
Coeff. (MDE) Coef. Or
III (High) 50–100 5.000 Dynamic
(MDE)
II (Mod) 3.000
(MDE)
I (Low) 1.0 (DE)
Table 4 Design ground

Return period, T (years) ac (gal)
acceleration [2]
10 90
20 120
50 160
100 190
200 220
500 250
1000 280
5000 330
10,000 330
2.3 Design Ground Acceleration (ac)
The ground acceleration (ac) is determined based on the return period of the earthquake
used in the analysis. The ground acceleration is presented as follow (Table 4):
2.4 Loads of Earthquake
The design load of earthquake is usually represented as horizontal forces acting on

the structures, which is called coefficient approach.
Coefficient Approach
The horizontal force due to an earthquake’s dynamic is represented using the
following expressions:
ad = Z ∗ ac ∗ v
Kh = ad/g
where:
Z = Earthquake Zone Coefficient.
ac = Basic ground acceleration (gal).
ad = corrected ground acceleration (gal).
v = correction factor due to type of soil/rock at site.
Kh = horizontal earthquake coefficient (gal).
G = gravitational acceleration (cm/s2 ) = 981 cm/s2 .
In this approach, it assumes that the structures’ components, on the base or either at
higher position, will subject to equal earthquake forces horizontally. This assumption
is not quite realistic. Another approach is to apply segmental coefficient.
Modified Coefficient for Earthquake
These modified coefficients are endorsed in [2] as well as in Seismic design guideline
for earth and rock fill dams [3]. The design earthquake coefficient Kh = a/g. The
design earthquake coefficient at each height can be calculated as follow:
Ko = α2 × Kh
where:
Ko: design earthquake coefficient at the basis of the structure (a dam).
α2: correction based on the type of structures (for earth-fill dam α2 = 0.5).
Kh: basic earthquake coefficient (depend on the return period of the analysis).
The design earthquake coefficient at each depth of the dam is calculated for Y =
0.25H; 0.50H; 0.75H; (where Y is the depth measured from the crest of the dam and
H is the dam’s height) using formula as follow.
For 0 < Y/H ≤ 0.4 :

K = Ko × (2.5 − 1.85 × (Y/H))
For 0.4 < Y/H ≤ 1.0 :
K = Ko × (2.0 − 0.60 × (Y/H))
3 Stability Analysis
Stability analysis of dams has long been conducted using mathematical model of
finite element method subject to earthquake dynamic forces [4]. Since then, the
consideration of earthquake forces has also been practiced [5, 6]. The conditions
of analysis are: at (1) the end of construction, (2) the steady seepage, and (3) rapid
drawdown. Early example is by [7] who analyzed the during-construction and early
impounding stage.
In this study, it needs to collect secondary and primary data. The secondary data
available is very limited and mostly on water distribution. The data on cross section
of the dam, its zonation, and their properties are not available. Therefore, it needs to
perform field measurement, dam boring and samplings.
The dam boring and sampling was conducted at six locations in the dam, i.e., 3
at upstream slope and 3 at downstream slope such as shown in Fig. 1 and Table 5.
The results of laboratory analysis on undisturbed samples taken at several depths are
Fig. 1 Location of Bor machine along the Dam
Table 5 Location for soil

Bor No. Location Coordinat
boring
X Y
BM1 D/S slope, north 467,477 9,186,402
BM3 D/S slope, mid 467,385 9,186,105
BM5 D/S slope, south 467,276 9,185,813
BM2 U/S slope, north 467,464 9,186,404
BM4 U/S slope, mid 467,364 9,186,108
BM6 U/S slope, south 467,262 9,185,818
shown in Tables 6 and 7. From Table 8, the embankment material of Klego Dam
is hard clay, sandy clay, and clay with fine sand. Therefore, the embankment of the
dam is uniform (non zonation) (Fig. 2).
From topographic survey and detail measurement combined with the result of
embankment boring, a typical cross section of Klego Dam is shown in Fig. 3 which
is the highest cross section. It is shown that there is origin hard rock formation at the
downstream side which was kept and thus adding more stability.
During the survey, at Klego Dam, there are 17 standpipe piezometers installed,
two of which are not functioning. The piezometer readings are performed once a
Table 6 Embankment material at downstream slope

No. Parameter Unit BM.1 BM.3 BM.5
1 Wet density gr/cm3 1.71–1.73 1.62–1.63 1.60–1.72
2 Dry density gr/cm3 1.43–1.48 1.36–1.39 1.35–1.44
3 Wc % 17.30–19.57 17.58–19.28 19.13–19.28
4 SG 2.67–2.69 2.68–2.69 2.68–2.69
6 E 0.82–0.87 0.93–0.97 0.86–0.99
7 Finer # 200 % 87.34–94.44 89.23–92.28 89.83–90.53
8 Cc 0.80–0.85 0.28–0.32 0.22–0.37
9 Cv ×10–2 cm2 /s Stiff 0.11–0.63 0.11–0.24
10 Mv ×10–2 cm2 /kg Stiff 6.31–9.17 7.92–9.06
11 Cohesion c kg/cm2 Stiff 0.09–0.26 0.14–0.147
12 Phi ° >35 19.42–19.28 19.69–20.03
13 Permeability ×10–6 cm/s 8.49–3.75 8.44–3.69 7.21–6.20
Table 7 Embankment material at upstream slope

No. Parameter Unit BM.2 BM.4 BM.6
1 Wet Density gr/cm3 1.66–1.74 1.66–1.74 1.61–1.72
2 Dry Density gr/cm3 1.40–1.48 1.39–1.47 1.35–1.44
3 Wc % 17.30–18.57 18.14–18.85 18.99–19.86
4 SG 2.67–2.69 2.68–2.69 2.68–2.68
6 E 0.80–0.91 0.82–0.92 0,86–0,99
7 Finer # 200 % 89.34–89.64 87.44–93.04 89.63–91.52
8 Cc 0.21–0.33 0.22–0.315 0.23–0.38
9 Cv ×10–2 cm2 /s 1.14–1.30 0.013–0.099 0.016–0.08
10 Mv ×10–2 cm2 /kg 9.06–9.08 8.92–9.99 1.08–9.06
11 Cohesion c kg/cm2 0.09–0.31 0.05–0.15 0.305–0.311
12 Phi ° 19.30–>35 18.57–>35 19.44–19.69
13 Permeability ×10–6 cm/s 6.48–3.66 8.77–7.23 6.94–6.33
Table 8 Embankment material at Klego Dam

Bor location Depth of Boring (m) Depth from surface (m) Description
BM1 10.00 m 0.00–2.20 Clay with fine sand, brown,
X: 467,477 (Water table at −4.0) hard
Y: 9,186,402 2.20–4.60 Clay hard sand, coral,
brown and hard
4.60–6.50 Sandy clay, coral, greyish
brown, hard to very hard
6.50–10.00 Sandy clay, coral, grey,
very hard
BM2 10.00 m 0.00–4.20 Clay with fine sand,
X: 467,464 (Water table at −4.0) reddish brown, rather hard
Y: 9,186,404 4.20–5.40 Clay with fine sand,
reddish, brown hard
5.40–6.00 Clay with fine sand, fine,
reddish brown, hard
6.00–7.20 Sandy clay, grey, hard
7.20–8.20 Hard clay, sand, brown,
very hard
8.20–10.00 Sandy clay, coral, grey,
very hard
Fig. 2 Boring at Klego Dam
Fig. 3 Typical cross section adopted of the Klego Dam

Fig. 4 Observed phreatic line based on piezometric data
Table 9 Values for factor of

Normal Storage 2.33 × 106 m3
risk (RTD Klego 2016)
Height of the Dam 10 m
People requires evacuation 9802 persons
Potential damages or losses 113 Billion Rupiah (Very High)
month. Observation on the piezometer no 1, 2, and 3 at the highest recorded water

level (+239.28) is shown in Fig. 4. It can be seen that the water pressure falls at the
existing hard rock formation downstream. It indicates that the adopted typical cross
section of Klego Dam is represented.
3.1 Risk Evaluation
Based on Table 1 and data for Klego Dam, the risk level for the dam is:
CRF = 4 (capacity 2.33 × 106 m3 )
HRF = 0 (Height is 10 m)
ERF = 12 (potential evacuee 9000 persons)
DRF = 12 (potential loss Rp. 113 billion)
TRF = CRF + HRF + ERF + DRF = 28
Based on the total risk factor (TRF) = 28, the risk level of Klego Dam is considered
High (19–30) (Table 9).
3.2 Earthquake Load
Based on the High risk level (III), the earthquake analysis is performed with: (1)
design ground acceleration of 100 year return period using OBE for no damages
analysis (ac = 190 gal) and (2) design ground acceleration of 5.000 year return
period using MDE for allowed damages with no failures (ac = 330 gal).
OBE Design Ground Acceleration

ad = Z * ac * v = 0.7 * 190 * 1 = 133 gal
Kh = ad /g = 133/981 = 0,136 g
K0 = α2 × Kh = 0.5 × 0.136 = 0.068 g
y/H = 0.25 : K = K0 × (2.5–1.85 y/H) = 0.13855 g
y/H = 0.50 : K = K0 × (2.0–0.60 y/H) = 0.1156 g
y/H = 0.75 : K = K0 × (2.0–0.60 y/H) = 0.1054 g
y/H = 1.00 : K = K0 × (2.0–0.60 y/H) = 0.0952 g
MDE Design Ground Acceleration
ad = Z * ac * v = 0.7 * 330 * 1 = 231 gal
Kh = ad /g = 133/981 = 0.235 g
K0 = α2 × Kh = 0.5 × 0.235 = 0.1175 g
y/H = 0.25 : K = K0 × (2.5–1.85 y/H) = 0.2395 g
y/H = 0.50 : K = K0 × (2.0–0.60 y/H) = 0.1998 g
y/H = 0.75 : K = K0 × (2.0–0.60 y/H) = 0.1821 g
y/H = 1.00 : K = K0 × (2.0–0.60 y/H) = 0.1645 g
3.3 Stability Analysis Using Geostudio Software
The stability analysis uses Geostudio software. The input for the model is material of
embankment and idealized as in Fig. 5. The embankment is homogenous provided
with toe drain along its downstream end. The embankment lies over a stiff clay
bed (layer 1). The properties of the material in the model is shown in Table 10.
The analysis is performed for normal condition, OBE earthquake loads, and MDE
earthquake loads.
Factor of Safety during normal condition
See Table 11.
Fig. 5 Embankment material at Klego Dam

Table 10 Properties of embankment material

Zona γd (kN/m3 ) γsat (kN/m3 ) C (kPa) (o ) E (kPa)
Embankment 13.95 18.72 30 25 15
Toe drain 17 20 0 45 30
Stiff clay 14.315 18.96 40 27 30
Fondation 16 18 40 27 30
Table 11 Factor of safety

No Condition FS
under normal condition
Geostudio 2007
1 Steady state
A U/S Face 2.287
B D/S Face 3.229
2 Rapid draw down
A U/S Face 2.858
B D/S Face 1.949
Factor of Safety During Earthquake (OBE)

See Table 12.
Factor of Safety Under Earthquake (MDE)
See Table 13.
Downstream face
See Figs. 6, 7, 8 and 9.

No Y/H; K FS
during earthquake (OBE)
1 0.25; 0.139 g
a U/S face 3.274
b D/S face 3.329
2 0.5; 0.116 g
a U/S face 2.659
b D/S face 2.516
3 0.75; 0.105 g
a U/S face 2.255
b D/S face 2.202
4 1.0; 0.095 g
a U/S face 2.212
b D/S face 2.232

No Y/H; K SF Geo Slope 2012
during earthquake (MDE)
1 0.25; 0.240 g
A U/S face 2.399
B D/S face 2.647
2 0.5; 0.20 g
A U/S face 2.012
B D/S face 2.046
No Y/H; K SF Geo Slope 2012
3 0.75; 0.182 g
A U/S face 1.781
B D/S face 1.895
4 1.0; 0.165 g
A U/S face 1.777
B D/S face 1.899
Fig. 6 FS at y/H = 0.25; K = 0.240 g (SF = 2.647)
Fig. 7 FS at y/H = 0.5; K = 0.20 g (SF = 2.046)

Fig. 8 FS at y/H = 0.75; K = 0.182 g (SF = 1.895)
Fig. 9 FS at y/H = 1; K = 0.165 g (SF = 1.899)
Upstream face
See Figs. 10, 11, 12 and 13.
Seepage Rate
The result of the model gives the seepage rate is 171.87 × 10–9 m3 /sec/m (=14.84
L/day/m), which is still below the allowable seepage rate = 100 L/day/m.
Fig. 10 At y/H = 0.25; K = 0.240 g (SF = 2.399)

Fig. 11 At y/H = 0.5; K = 0.20 g (SF = 2.012)
Fig. 12 FS at y/H = 0.75; K = 0.182 g (SF = 1.781)
Fig. 13 FS at y/H = 1; K = 0.165 g (SF = 1.777)
Phreatic Line
The model also shows that the outlet of the phreatic line lies at the toe drain, which
is rock fill formation. This shows that the seepage exit from the dam through its
designated rock toe.
4 Conclusion
From the analysis it can be concluded as follows:

1. The boring on the Klego Dam along the crest confirms that the body of the dam
is uniform of clay protected with rip rap of stones and the existence of hard rock
at the downstream side of the embankment,
2. The safety of the Klego Dam is still in safe state even though analyzed using
the new earthquake map and under OBE and MDE earthquake loads,
3. The lowest factors of safety obtained are: 1.949 at downstream face under
rapid drawdown; 2.212 at the upstream face under OBE loadings; and 1.777
at upstream face under MDE loadings, which are surpassed the minimum
allowable FS of 1.5.
4. The seepage is 14.8 L/day/m’ which is still lower than the allowable of
seepage 199 L/day/m’.
5. The phreatic line also shows the exit falls in the safe side at the toe drain.
Acknowledgements The authors express their gratitude to the Pemali Juana River Basin Authority
for the data used in the study.
References
1. Chen Y, Lin P (2018) The total risk analysis of large dams under flood hazards. Water 10(2)
2. Ministry of Housing and Regional Infrastructures No. 360/2004 on: Stability Analysis of
Embankment type Dam under earthquake Condition. Jakarta (2004)
3. Hernández UM, Nava-Tristán OE, Liu XL, Sandoval EM, Hach F (2008) Seismic guidelines for
earth and rock fill dams. In: The 14th World Conference on Earthquake Engineering October
12–17, Beijing, China
4. Maaty AS, Rothenburg L, Lelievre B (1985) Dynamic Analysis of stability of Earth dams. In:
Proceedings of ISSMG: international society for soil mechanics and geotechnical engineering
5. Chhatre MV, Muralidhar B (2010) Dynamic analysis of matatila Earthed dam: a case study. In:
Proceeding of Indian geotechnical conference, 16–18 Dec
6. Paul DK (2004) Assessment of damage and rehabilitation of small and medium Earth dams.
In: 13th world conference on earthquake engineering, Vancouver, BC, Canada, 1–6 Aug 2004,
Paper No. 2160
7. Rashidi M, Mohsen Haeri S (2017) Evaluation of behaviors of earth and rockfill dams during
construction and initial impounding using instrumentation data and numerical modeling. J Rock
Mech Geotech Eng 9:709–725
Impact of Climate Change on Dam Safety
Heri Sulistiyono, Ery Setiawan, and Humairo Saidah
Abstract Dams are built to overcome surface water problems. These problems
include excess water in the rainy seasons and scarcity of water in the dry seasons.
As the occurrence of climate change has been recognized by people around the
world, engineers have to be able to consider its impact on the safety of dams. This
study proposes a procedure for studying the impact of climate change on the safety of
gated spillway dams. The procedure consists of a statistical consistency test, selection
of global climate change variables, modelling of local climate change, calculating
flood discharge due to climate change, and evaluation of dam safety. This study was
conducted at the Batujai Dam, Central Lombok Regency. Global climate change
data from 1998 to 2100 taken from the Intergovernmental Panel on Climate Change
(IPCC) and maximum daily rainfall data from 1998 to 2020 from the Pengadang
Station were used to demonstrate the application of the proposed procedure. The
results show that the proposed procedure can be used to evaluate the impact of
climate change on dam safety. It is also known that climate change has a significant
effect on the tendency of increasing flood discharge entering the Batujai Reservoir.
The four spillway gates: P1 , P2 , P3 , and P4 have to be opened 1.0, 1.2, 1.2, and 1.0 m,
respectively to anticipate the 1000-year return period of a flood event due to climate
change, at a rate of 1833.54 m3 /s.
Keywords Climate change · Maximum daily rainfall · Gated spillway dams ·

Safety of dam
H. Sulistiyono (B) · E. Setiawan · H. Saidah

Department of Civil Engineering, University of Mataram, Mataram, Indonesia
E. Setiawan
H. Saidah
158 H. Sulistiyono et al.
1 Introduction
Dams are built for many benefits. However, there are potential dangers if floods
overtop the dam embankments. The dam will break and a flash flood will occur
downstream. In the Dam Design guidelines, embankment heights should be designed
based on at least a 1000-year return period of flooding. Since the consensus on climate
change in Kyoto in 1992, engineers must evaluate the impact of climate change on
dam safety.
Global climate change predictions have been simulated by agencies around the
world under the coordination of the IPCC. The resolution of the global simulation is
still very rough, however, and cannot be used directly for local designs. The results of
global climate change simulation have to be downscaled following the local historical
data to obtain local scaled climate change data that can be used in local designs and
evaluations [1].
This study proposes the procedure for dam safety evaluation based on the impact
of climate change, to develop a local rainstorm model based on climate change, to
calculate the 1000-year return period of flooding due to climate change, and to obtain
the opening height of the spillway gates required to secure the dam.
2 Literature Review
The problem of the increase in water demand caused by the increase in population in
the Tasikmalaya Regency can be solved through the construction of a water reservoir,
namely the Manonjaya Dam [2]. The Manonjaya Dam is an earth-fill type dam with
a vertical core on the Citanduy River with a catchment area of 590,422 km2 . The
purpose of the Manonjaya Dam is to meet the needs for domestic water and irrigation
in the Tasikmalaya Regency area.
The problem of flooding in the rainy season and the lack of clean water during the
dry season in Ponorogo Regency, East Java, can be overcome through the construction
of the Bendo Dam [3]. A 1000-year return period of a flood of 676.37 m3 /s was used
to design the dam height. The design life of the dam is 50 years.
Climate change refers to permanent changes in the statistical characteristics of
climate factors such as solar radiation, temperature, wind, humidity, rainfall, air
pressure, and so on, over a long period such as tens of years to millions of years
[4–6]. Currently, the world is experiencing global climate change [7]. The impacts
of climate change in Indonesia include an increase in the average air temperature
from January 1971 to December 2006 of 0.5 °C; an increase in the maximum air
temperature of 0.7 °C; a decrease in the minimum air temperature of about 1.2 °C;
change in the maximum rainy season from January–March to October–December;
and the occurrence of drought in some areas. In addition, many small islands such as
Lombok, Sumbawa, Flores, Sumba, Timor, Solor, and Alor will be more vulnerable
to climate change than the large islands [8].
Impact of Climate Change on Dam Safety 159
3 Methodology
The study procedure consists of the following five strategic steps: Preparation, Stan-
dard Statistical Screenings, Climate Change Modelling, Hydrological Frequency
Analysis, and Hydraulic Analysis to check for overtopping. The procedure is shown
in Fig. 1.
Fig. 1 The procedure of the study

3.1 Step: 1 Preparation
In this step, both historical local rainfall data and global climate change data are
collected. The range of historical local rainfall data (HLRD) has to be within the
range of the global climate change data (GCCD). The period of HLRD has to be at
least 20 years to ensure statistical validity in the calibration and verification processes.
HLRD is collected from all rainfall stations around the catchment area of the reservoir.
The GCCD is the result of global climate models (GCM). In the Intergovernmental
Panel on Climate Change (IPCC), GCCD of various GCMs from various countries are
stored. One has to register before downloading the data. The GCCD cannot be applied
directly for local designs because the resolution is very coarse. The GCM variable
needs to be downscaled to get better-scaled variables that can be used for local
designs. Downscaling models can be grouped into the three following approaches:
dynamic models, statistical models, and change factor methods [1, 9]. Regression
models are proposed in this study procedure for the climate change downscaling
model.
3.2 Step: 2 Standard Statistical Screening
Statistical screening tests are used to ensure all the data are ready. The term ‘ready’
refers to data that are consistently collected in the expected condition. The Rescaled
Adjusted Partial Sums (RAPS) method [10] is applied in this procedure for the
standard statistical screening test. As RAPS has been frequently explained in many
hydrological references, this method is not described further in this paper.
3.3 Step: 3 Climate Change Modelling
This study applies the three most suitable GCM variables as independent variables
to simulate local climate variables. The selection of suitable GCM variables is made
using common statistical analysis: correlation coefficients (CC) [11].
In this study, the best multiple linear regression is selected and used to downscale
the GCM variable into local rainfall variables.
3.4 Step: 4 Hydrological Frequency Analysis
In this study, the frequency analysis is used to obtain the estimated return period of
rainstorms due to climate change. Then, the flood return period is calculated based
on the return period of rainstorms. The Nakayasu Synthetic Unit Hydrograph is used
to simulate the flood [10]. The Unit Hydrograph as shown in Fig. 2 is calculated
using Eq. (1) to Eq. (11):
A R0
QP = (1)
3.6(0.3t P + T0.3 )
where: Qp = peak flow (m3 /s), A = catchment area (km2 ), Ro = unit rainfall (mm),
tp = time lag from the initial rain to the peak flow (hours), T0.3 = the time it takes for
the discharge to decrease, from peak discharge to 30% of peak discharge (hours).
In Fig. 2, Tp and T0.3 are calculated using Eqs. (2) and (3):
Tp = tg + 0.8 tr (2)
T0.3 = α · tg (3)
where tg is calculated based on Eq. (4) or Eq. (5), and tr is calculated using Eq. (6):
tg = 0.21L0.7 for L < 15 km (4)
tg = 0.40 + 0.058 L for L > 15 km (5)
tr = (0.5 ∼ 1.0)tg (6)
where tr = the effective duration of rain.

The rising curve for the range of 0 ≤ t ≤ Tp
Fig. 2 The illustration of Nakayasu synthetic unit hydrograph

2.4
t
Qa = Q P (7)
TP
The declining curves for the range of

(1) Qd > 0.3 QP or TP ≤ t < T0.3

t − TP
Q d = Q P 0.3 (8)
T0.3
(2) 0.3 QP > Qd > 0.32 QP or T0.3 ≤ t < 1.5 TP :

(t − TP ) + 0.5T0.3
Q d = Q P 0.3 (9)
1.5T0.3
(3) 0.32 QP > Qd or t ≥ 1.5 T0.3

(t − TP ) + 1.5T0.3
Q d = Q P 0.3 (10)
2T0.3
The relationship between the shape of the drainage area and T0.3 can be stated as
a constant of α:
0.47( AL)0.25
α= (11)
tg
where: Qa = runoff before reaching peak discharge (m3 /s), Qd = runoff after
reaching peak discharge (m3 /s), t = time (h), L = length of river channel (km), tg =
concentration time (h), and α = constant.
3.5 Step 5: Hydraulic Analysis for Overtopping
The flood routing through the spillway aims to simulate the capacity of the spillway
concerning changes in the inflows of the dam. Equations (12) and (13) were used in
the simulation. The simulation must consider the upstream and downstream water
conditions.

Qs = μ · as · bs · 2gh 1 (12)
where: Qs = discharge (m3 /s), μ = discharge coefficient (given: 0.8), as = height of

the opening (m), bs = width of opening (m), g = gravitational acceleration (9.8 m/s2 ),
and h1 = hydraulic energy at the opening (m).
S1 O1 S2 O2
I+ − = + (13)
t 2 t 2
where: I = Inflow (m3 /s), S1 = Storage at time 1, S2 = Storage at time 2, t =

routing interval (h), O1 = Outflow at time 1 (m3 /s), O2 = Outflow at time 2 (m3 /s).
4 Case Study
To demonstrate the procedure, this study was applied in the Batujai Dam in Penujak
Village, Central Lombok Regency, West Nusa Tenggara Province, Indonesia, as
shown in Fig. 3.
4.1 Step: 1 Preparation
In this study, 23 years of maximum daily rainfall data from 1998 to 2020 from
Pengadang Rain Stations were used as local historical data and Global Climate
Change data from 1998 to 2100 were downloaded from the IPCC website. The
data are: X1 = mean 2 m wind speed (m/s), X2 = evaporation (mm/day), X3 =
precipitation (mm/day), X4 = screen (2 m) temperature (°C), X5 = screen spec.
humidity (kg/kg), X6 = sea level pressure (hPa), X7 = surface temperature (°C), X8
= solar flux at surface (W/m2 ), X9 = surface pressure (hPa).
Fig. 3 Study location

4.2 Step: 2 Standard Statistical Screening
From the calculations of the RAPS test of the annual rainfall data from the Pengadang
Station, it is found that these are consistent.
The next step is to select the three most suitable GCM variables, which are air
temperature (X4), specific air humidity (X5), and air temperature at the Earth’s
surface (X7). The three GCM variables are then considered as independent variables
in the multiple linear regression model for modelling local rainfall due to climate
change.
4.3 Step: 3 Climate Change Modelling
The multiple linear regression model is expressed as Y = 0.453 + 0.44 X4 + 0.253

X5 − 0.81 X7. The proportion of the variance for a dependent variable that can be
explained by the variables is quite high, being approximately 80.7%.
The local rainfall data due to climate change are generated using the regression
model. The results are shown in Fig. 4.
Figure 4 shows that the predicted mean local maximum daily rainfall in the area
of study has an increasing trend. The next step is to look for the 1000-year return
period of rainfall and the 1000-year return period of flooding.
Fig. 4 Predicted local maximum daily rainfall due to climate change

Table 1 Results of flood

Simulation: Routing output for Batujai Reservoir
routing simulation
Q-inflow 1833.5 m3 /s
Q-outflow 875.24 m3 /s
Reduction—Q 52.26 % 4 × @ 11 m
Opening (m) Gate #2 Gate #3 Gate #4
Gate #1 1.20 1.20 1.00
4.4 Step: 4 Hydrological Frequency Analysis
It is found that the statistical distribution of the data is a 2-parameter exponential.

The amount of the 1000-year return period of a rainstorm is 797.2 mm. This amount
is used to simulate the inflow of the Batujai Dam using the Nakayasu Hydrograph.
From the simulation, it is obtained that the peak of the 1000-year return period of
flooding that enters the Batujai reservoir is 1833.538 m3 /s.
4.5 Step 5: Hydraulic Analysis for Overtopping
In this step, flood routing simulation through a gated spillway is conducted to obtain
whether the 1000-year return period of flooding overtops the dam or not, and how to
arrange the gates of the spillway to avoid overtopping. The results of the simulation
using Eqs. (1)–(13) are presented in Table 1.
Table 1 shows that when the amount of the 1000-year period of flooding of 1833.54
m3 /s comes into the reservoir, the four gates have to be opened by 1.0, 1.2, 1.2, and
1.0 m respectively to avoid overtopping of the dam.
5 Conclusion
Based on the results of the analysis, the following conclusions can be drawn:
(1) The proposed procedure can be reasonably applied for the evaluation of dam
safety based on the impact of climate change;
(2) The study successfully developed a local rainstorm model affected by climate
change;
(3) The amount of the 1000-year return period of flooding due to climate change
is 1833.54 m3 /s; and
(4) The opening heights of the four spillway gates required to secure the dam are
1.0, 1.2, 1.2, and 1.0 m respectively.
References
1. Wilby RL, Charles SP, Zorita E, Timbal B, Whetton P, Mearns LO (2004) Guidelines for use
of climate scenarios developed from statistical downscaling methods. Supporting Material of
the IPCC. http://www.ipcc-data.org/guidelines/dgm_no2v1_09_2004.pdf
2. Shofa K, Hidayat M, Wahyuni S, dan Darsono S (2013) Perencanaan Bendungan Manonjaya
Kabupaten Tasikmalaya, Jawa Barat. Jurnal Karya Teknik Sipil 2(3):189–204
3. Busiri A, Fajar R, Sangkawati S, dan Budieny H (2016) Perencanaan Bendungan Bendo
Ponorogo. Jurnal Karya Teknik Sipil 5(2):159–169
4. IPCC (2001) Climate Change 2001. The scientific basis; impacts, adaptation, and vulnerability;
mitigation; and model evaluation. The Third Assessment Report (TAR) of the Intergovern-
mental Panel on Climate Change (IPCC). Cambridge University Press, Cambridge, UK and
New York, USA
5. IPCC (2007) The synthesis report and the physical science basis—summary for policymakers.
IPCC Plenary XXVII, Valencia, Spain
6. Pryor SC (2009) Understanding climate change: climate variability, predictability, and change
in the midwestern United States. Indiana University Press, Bloomington
7. Gleick PH (1997) Water planning and management under climate change. J Am Water Works
Assoc
8. Hilman M, Hatta GM, Hilman D, Gunawan A, Suryani Y, Irwani S, Setiawan G, Von Luepke
H, Hertz T, Panjiwibowo C, Pasaribu K, Suroso SA, Hadi TW, Sofian I, Latief H, Abdurahman
O, Halil, Julianto H, Riawan E, Muhammad A, Anggraini L (2010) A synthesis report: risk
and adaptation assessment to climate change in Lombok Island, West Nusa Tenggara Province.
Ministry of Environment of the Republic of Indonesia
9. Sulistiyono H, Lye LM (2011) Predicted regional hydrologic and climatic variables under
climate change scenarios using statistical downscaling techniques for future water resource
studies in Lombok, Indonesia. In: The 64th Canadian Water Resources Association (CWRA),
Canada
10. Harto S (1993) Analisa Hidrologi. Jakarta, Gramedia Pustaka Utama
11. Prak JL (2011) An empirical study on the reliability of correlation coefficients. Br J Psychol
21(4). https://doi.org/10.1111/j.2044-8295.1931.tb00602.x
Assessment, Protection, Maintenance,
Repair, and Retrofitting of Buildings
and Infrastructures
Investigating Materials
for Refurbishment Strategies of Heritage
Buildings: A Case Study of Soesman
Kantoor, Semarang
Ferry Hermawan , Didi Wibowo Tjokro Winoto, Ismiyati ,

Bambang Purwanggono , and Robby Soetanto
Abstract The management of heritage buildings presents complex challenges for

local authorities regarding repair methods, the selection of materials and post-repair
impacts. Restoring the function of heritage buildings is often still constrained by
commercialisation and the problem of maintaining sociocultural values. This raises
an important question, what approach strategy is most suitable for maintaining the
characteristics of heritage building in the long term? Here, we present a case study
with a materials laboratory test strategy for a heritage building in Semarang. The
parties involved include local authorities, architects, heritage building experts and
academics. Three technical assessments are conducted, namely, materials analysis
using X-ray fluorescence spectroscopy and brick compression and paint adhesion
tests. This study provides an overview of the strategic approach from the aspects of
measuring, interpreting and providing the best solutions for the material elements of
heritage buildings in the future. The results promote improving heritage buildings on
the basis of stakeholder mapping and building materials improvement strategies. This
research contributes to local authorities in arranging heritage buildings with tech-
nical considerations in line with the characteristics of the building and development
programs that consider local knowledge.
Keywords Strategy · Refurbishment · Heritage
F. Hermawan (B) · D. W. T. Winoto · Ismiyati · B. Purwanggono

Diponegoro University, Semarang 50275, Indonesia
B. Purwanggono
R. Soetanto
Loughborough University, Loughborough, LE 113TU, UK
170 F. Hermawan et al.
1 Introduction
1.1 Background
The Old Town area of Semarang (Kota Lama Semarang) has high historical value
with many buildings from the Dutch colonial era. Some of these heritage buildings
are listed as cultural heritage buildings according to Law No. 11 of 2020 on Cultural
Heritage [1]. Soesman Kantoor is one of these buildings and is located in Kota Lama
Semarang. According to law, buildings that have existed for at least five decades
can be classified as cultural heritage buildings [1]. In addition to this criterion, these
buildings are expected to have essential historical, scientific or cultural value.
In the modern era, the attentiveness of cultural heritage buildings is often
neglected. The European Charter of Architecture Heritage 1975 Article 6 states
that architecture’s cultural heritage is being unsettled in Indonesia and around the
world [2]. In mid-January and April 2011, two cultural heritage buildings collapsed
in Kepodang, Kota Lama Semarang. This was caused by the buildings’ lifetime
and natural factors, such as flooding, seawater intrusion and extreme weather [3].
As reviewed through financial, information and technological aspects, each city’s
evolution will always be the same from time to time. Cultural heritage buildings
have unique features, the significant elements of a culture that reflect an area’s iden-
tity. Cultural values grow and live within society and have their own uniqueness that
makes each stand apart. If these values are maintained, each region will have an
identity that differentiates it from others, thus representing its unique charm [4].
The refurbishment and improvement of building function are essential for the
aim that cultural heritage can function as a new attraction for tourism and bring an
area back to life. The repair of cultural heritage buildings is different from modern
buildings. Professionals must handle cultural heritage buildings to avoid irreversible
damage. The repair of existing structures must be carried out attentively and by not
changing or spoiling any characteristics or details, although changes might be needed
[5].
Research on the correct method for cultural heritage building repairs is essential
because incorrect action could cause even more damage. The cultural values might
be lost if damage to a single component of a cultural heritage building occurs. This
research leads to the important question of what is the best method for repairing such
buildings?
1.2 Research Method
The refurbishment of cultural heritage buildings in Semarang must be carried out by

paying particular attention to the materials and procedures. Laboratory tests can be
used to identify the material specification of existing building components. Labora-
tory test results are used as a benchmark to obtain a better specification of building
Investigating Materials for Refurbishment Strategies … 171
Fig. 1 Laboratory and material test activities
repair or at least at the same grade as the required criteria. The analysis referred to
various standards. There are four tested cultural heritage building components, i.e.,
bricks, mortar, wood and paint.
The source of quantitative data came from the results of the laboratory tests. The
tests were conducted on existing building material to obtain information regarding
cultural heritage building characteristics. The tests were conducted on materials
collected from an existing building, Soesman Kantoor.
In the case study building, there are four primary testing materials. i.e., bricks,
mortar, wood and paint. The sample collection and testing were conducted by refer-
ring to the American Society for Testing and Materials (ASTM). The flow of materials
testing activity from materials collection to laboratory tests and the output can be
seen in Fig. 1.
2 Results and Discussion
2.1 Materials Test of Soesman Kantoor’s Building
a. Brick Compressive Strength Test Results

The compressive strength of brick samples taken from Soesman Kantoor was
tested. The compressive strength test aims to acquire the required minimum
specifications of brick compressive strength to replace the new bricks during
the reparation. The selected bricks should have a greater or equal compressive
strength to avoid even greater damage to the existing building. The compressive
strength test used automatic cement concrete compression machines with a
3000 kN capacity. The compressive strength test results were obtained from the
load-receive brick cross-section area divided by the maximum load from the
compression machine. The compressive strength test resulted in the compressive
strength of each brick [6].
Older bricks are generally larger than modern ones. The dimensions of the
bricks in Soesman Kantoor are 25 cm × 12 cm × 5 cm for length, width and
height, respectively. Bricks with the same dimensions can preserve the building’s
aesthetics. In addition to dimensions, colour and texture are also important
factors to consider [7]. The bricks in Soesman Kantoor are of a rectangular
prism shape, with a reddish-orange colour and a smooth surface.
b. Mortar Composition Test Results
Mortar was tested to reveal the composition of the aggregate. The test used
a petrographic method to find the chemical content inside the components of
the mortar [8, 9]. Petrographic mortar with microscopic and stereoscopic irra-
diation principles that involved X-ray and electron microscopy was used. The
petrographic test could be tested with wavelength dispersive X-ray fluorescence
(WDXRF) spectroscopy.
WDXRF spectroscopy is an analytical technique used to detect chemical
elements in solid and liquid materials. X-ray fluorescence (XRF) is a “sec-
ondary” characteristic X-ray emission (or fluorescence) of materials excited
through high-energy X- or gamma rays. WDXRF is known for its high accuracy,
precision, and reliability.
The mortar test sample was taken from the indoor wall of Soesman Kantoor.
The minimum mass of samples tested was 10 g. The mortar composition test
was conducted with a petrographic examination utilising WDXRF spectroscopy
to analyse the metal contents in solid and liquid materials, both metal elemental
and metal oxide. Generally, lime mortar consists of several primary chemical
compositions, including CaO, MgO, SiO2 , Fe2 O3 , Al2 O3 and CO2 [10].
c. Wood Specific Gravity Test Results
The testing of wood samples was completed using a wood specific gravity test.
This test aims to discover the minimum wood specific gravity that is to be used in
the building component repair of the existing cultural heritage building. Wood
specific gravity is generally proportional to wood strength or its mechanical
properties [11]. The greater the wood specific gravity, the stronger the wood.
Wood specific gravity compares the oven dry wood weight at 105 °C and the
weight of water with a volume equal to that of the wood. The specifications of
new materials must be better or equal to avoid further damage.
After being weighed, the wood sample was dried in a laboratory bench oven
and with digital control to eliminate moisture from the wood samples. The
final wood sample gravity was obtained from the weighing after being dried in
an oven under 103 ± 2 °C for 24 hours The results of wood specific gravity
were obtained from the wood sample dry weight divided by moisture level
and volume. The mean specific gravity from the five wood samples was 0.571
gr/cm3 .
d. Paint Adhesive Power Test Results
The paint adhesive power test was directly conducted on the Soesman Kantoor
exterior wall utilising a cross cut tape test method. The test was completed on
three walls in different locations of Soesman Kantoor. The part of the wall to be
tested must be smooth and clean. It was then cut longitudinally and circularly,
each with six lines of a 2 cm length and a 2 mm gap between each line. We
observed the test area surface on the cutting line area. Sticky tape with a 1-
inch width was then applied until the whole area was closed and the tape was
flattened with fingers or additional tape. The tape was taken off with a near 180°
direction, with constant speed and at once. The condition of the paint surface
was observed [12]. In location 1 (old wall), the test result showed that the paint’s
grade was zero because more than 65% of the paint surface was peeled. For the
side wall (location 2), the test result showed a grade 2 paint indicated from the
peeled paint at the side of the cutting line and some parts of the square area.
The peeled paint area was ~15–35%. For the third location (front wall), the test
result showed a grade 5 paint, as seen from the zero-peeled paint on the test area
where the cutting edge also remained intact.
2.2 Refurbishment Strategy Formulation for Soesman

Kantoor
Brick Repair
The repair was completed by first determining the correct replacement materials. The
criterion for alternative brick materials was materials with a minimum compressive
strength of 100.63 kg/cm2 . Brick wall repair that included an area of more than
0.19–0.27 m2 or more than 20% of an existing wall, or exceeding 2.4–3.0 m length,
as a required consideration to meet the physical strength criteria of the brick. It
needs to meet the physical strength criteria to level the building movements due to
the load distribution and expansion and shrinkage of the existing and new bricks.
The installation of new bricks with different specifications might cause cracks and
fracturing of the bricks.
Another aspect to consider is the appearance of the bricks, which includes their
dimensions, colour and texture. The old brick dimensions are different from the
size of the contemporary brick. It is suggested to ask the brick manufacturer for
the specific brick size to make it more economical. The most crucial must-match
dimension is for the brick parts to be exposed [7]. For the other unexposed parts, a
variation in size is allowed as long as it does not obstruct the repair. The dimensions
of bricks known as the government’s model are 26 × 12.4 × 52 cm.
Contemporary bricks with the same colour and texture, but larger than the existing
bricks, could be used by cutting it with a wet saw to the desired size. This is an
economical way for the small needs of bricks. Using materials other than bricks is
not recommended due to the building’s visual aspect. Repair work on new brick
components must be marked to identify the components that have been repaired.
Mortar Repair
Mortar repair on the cultural heritage building wall was completed by stripping the
old mortar layer, a process known as repointing [11]. Repointing increases weather
resistance, reduces water seepage, maintains structural strength and appearance, and
extends the building’s lifetime. Field inspection was essential to point the damaged
mortar. If the previous mortar repair was not matched to the original mortar, it needed
to be stripped. Original mortar identification is crucial to obtain a suitable new mortar.
The colour and texture of the original mortar had to be investigated visually, while
the mortar composition was observed in the laboratory.
After getting the correct mortar composition, a test must be carried out to compare
to the original mortar, in terms of colour, texture and the finishing of the surface.
This assessment was carried out once the mortar hardened. The new mortar trial
assessment must represent the conditions and surface material to be peeled and
have proper lighting and the same environmental condition. The mortar repair must
use compositions similar to the existing conditions so it can sufficiently stick. The
composition was acquired from the laboratory test results with WDXRF to find the
ratio of the correct mortar constituent components. The most dominant component
of mortar test results in the laboratory were calcium oxide (CaO), which represents
lime material with ~85.2%, ferric oxide (Fe2 O3 )/clay with ~32.4%, silica (SiO2 )/sand
with ~16.4% and aluminium oxide (Al2 O3 )/clay with ~3.96%. The composition of the
mortar mixture according to the existing test results were lime:clay:sand with a ratio
of 8:3:1. The laboratory test results show the composition of the old building mortar,
but it is still in the percentage of mass per molecule that significantly involves oxygen
and causes the mass to exceed 100%. The mass per element is acquired through the
calculation that involves atom mass and molecular weight.
The simple lime mortar process was completed by burning the lime in a furnace
to form quicklime. The lime was then mixed with water to form slaked lime, which
forms a hydrated lime putty or lime powder. The powder was then mixed with sand
and water to form mortar. The type of lime mortar that is known as non-hydraulic
has a prolonged setting time through the reaction with carbon dioxide in the air. The
setting speed can be increased by using limestone, which is burned in a furnace,
to form a hydraulic lime that breaks down if it comes into contact with water. The
application of mortar to the gap of the bricks must be compact and should not leave
any cavities. The application must consist of double layers, where each is compacted
beforehand.
Wood Repair
The wood component repair was carried out by determining the material with a
specific gravity that was better or at least equivalent to the existing wood component.
The wood specific gravity that must be met is 0.571 gr/cm3 or equivalent to wood
strength grade level II, according to SNI 7973:2013 [12].
Structural reinforcement in cultural heritage buildings was carried out by replacing

wood materials with the same quality. There are 17 various types of wood in Indonesia
with various specific gravities that can be used as replacements in heritage build-
ings, including timber (Lagerstroemia speciosa), teak tree (Tectona grandis), rubber
(Hevea brasiliensis) and cinnamon (Cinnamomum purrectum). Knowledge of the
various types of wood would reduce the dependency on a specific type of wood,
so that other types of wood that have not been used could increase. Wood material
repair for construction elements could use alternative wood types, but it is better to
use wood with a better or at least equivalent grade.
In addition to structural strengthening, damages caused by termites must also
be replaced by replacing wood materials of the same quality. To avoid termites,
maintenance is essential by spraying with anti-termite liquid and traditional cleaning
using cloves, tobacco and banana fronds with the ratio of 1:1:1.
Differently coloured wood components could be repaired by means of camouflage,
i.e., change the shape and materials to make it look like the original. Maintenance that
could be carried out includes monitoring climatological conditions with an extreme
thermometer [13].
Paint Repair
Repainting cultural heritage buildings requires special paint that allows water vapor
to move freely through walls. Painting with a mixture of cement and regular sand or
cement plaster will only cause cracks and will not bond well with the old cement-
sand mixture that sticks to the wall. Plaster made of regular sand cement and cement
plaster would be too stiff and hard, thereby potentially causing cracks on the wall.
Paints must be breathable, but it could not be found easily in paint products in the
modern era because modern buildings tend to seek waterproof paints to prevent rain
from getting into buildings. Coloured mortar could be acquired by adding pigments
into the mortar mixture. This method can only be achieved if the desired colour can
be achieved using aggregates that can provide that colour. The breathable paint was
made from a lime-based mixture. Lime could also provide a white colour if it is the
desired one. The paint composition used referred to the BOW analysis. The mixture
for a 10 m2 paint was one part Tras Muria, one part limestone and three parts sand.
The pigment added into the mortar must be smooth to be mixed evenly, but
it should not cause a reaction that can damage the mortar. This kind of pigment
could be achieved through pigments with a metal-oxide nature, i.e., lime (CaO). The
pigment addition must be carried out in moderation to maintain the composition and
mortar performance. The suggested pigment addition is a maximum of 10% of the
total mixture.
The advantage of using lime is its flexibility and its resistance to shrinkage
cracking from wall movements. The cracking that occurs due to the structural move-
ment can occasionally “self-heal” with water circulation. The coloured mortar also
does not need to worry about adhesion because it is homogeneous with the mortar.
The disadvantage of lime is that it requires a long time to harden. It could turn soft and
crumble if continuously exposed to water. The summary of refurbishment strategies
as presented in Table 1.
Table 1 Summary of refurbishment methods

Component and characteristics Visual Repair method
Bricks Long rectangular prism brick, Repair using bricks with a
Dimensions flat and no-cracked surface, minimum compressive strength
25 cm × 12 cm × 5 cm and angled corners. Dark of 105.03 kg/cm2 . The new
Colour orange bricks or with as close brick dimension must be made
Dark orange as possible colour to the as close as possible to the old
Compressive Strength existing bricks if it is exposed one, i.e., 25 cm × 12 cm ×
5 cm
The colour and texture of the
bricks are adjusted to the
existing one to maintain the
aesthetic value, i.e., reddish
orange with a flat surface
Alternative material is not
recommended because it might
damage the appearance of the
new brick wall
Mortar The colour is dominantly dark Making mortar with a
Colour white due to the mixture of composition significantly
Composition lime and clay paste similar to the existing mortar,
i.e., the mixture of
lime:clay:sand with the ratio of
8:3:1. Application
Wood The wood dimension depends The structural wood repair
Dimension on the repaired components must meet the conditions of
Colour The colour of the wood is minimum wood specific
Specific gravity adjusted to the original gravity of 0.571 gr/cm3 . The
condition type of wood may vary as long
as the specific gravity is
sufficient. Joinery could be
done if the desired wood
dimension is unavailable, by
paying attention on the
technical guidelines of SNI
7973, 2013
Non-structural component
repair
Paint White paint or adjusted to the Repair using pigment-coloured
Composition colour of the original paint mortar. The pigment used
Waterproof paint is not could come from one of the
allowed aggregates with the desired
Grade 5 is the expected colour. Lime aggregate works
adhesive power, in which no as an adhesive and pro-vide
paint is peeled during the tape white colour to the paint. The
test adhesive power in paint is not a
problem because the paint
itself is the mortar’s
homogeneous part
3 Conclusions
It is strategic to conduct a comprehensive investigation before repairing cultural

heritage buildings to acquire thorough information regarding the buildings. The
laboratory test on Soesman Kantoor building components is a way to obtain the
characteristics of each material. The testing was carried out on the elements with the
most damaged materials, i.e., bricks, mortar, wood and paint.
The differences in materials and manufacturing methods of two time periods
show the technology development that becomes more efficient each day. However,
the material authenticity and work methods must be maintained as the original; thus,
modern materials and manufacturing methods cannot be carelessly applied.
There is a wide range of cultural heritage building components, so it could not be
generalised. The repair must be done per case because each issue needs a different
solution and could not be equated to other building components. The repair guidelines
of a cultural heritage building with Soesman Kantoor as the case study focuses on
brick, mortar, wood and paint components. All components could be repaired by
following the minimum criteria for replacement materials. If a material cannot be
repaired as well as the original, it can be replaced as long as it is not invasive and
reversible.
Acknowledgements Thankful to research group of Heritage from Laboratory of Construction

Management, Department of Civil Engineering, Diponegoro University and support funding from
Directorate General Empowering Research and Development Ministry of Research and Tech-
nology, National Research Council and Innovation-Budgetary Year 2020, Scheme of Master Thesis
Research Contract Number 225-52/UN7.6.1/PP/2020.
References
1. Regulation No.11 regarding Cultural Heritage. Jakarta. President Republic of Indonesia (2010)
2. ICOMOS (1975) European charter of the architecture heritage 1975 Art. 6
3. Utami P (2013) Bangunan Tua di Kota Lama Semarang Roboh. Kompas, 13 Januari
4. UNESCO (2016) The HUL guidebook managing heritage in dynamic and constantly changing
urban environments
5. Rasmussen TV (2014) Model for refurbishment of heritage buildings. ICVEST 2014 building
for a changing world, 2
6. ASTM C62-17 (2017) Standard specification for building brick (Solid masonry units made
from clay or shale). ASTM International, West Conshohocken, PA
7. The Brick Industry Association (2011) Brick brief: brick replacement, Virginia, USA
(gobrick.com). Accessed Oct 2020
8. ASTM C141/C141M-14 (2014) Standard specification for hydrated hydraulic lime for
structural purposes. ASTM International, West Conshohocken, PA
9. ASTM D2395-07ae1 (2007) Standard test methods for specific gravity of wood and wood
based materials. ASTM International, West Conshohocken, PA
10. ASTM D3359-09e2 (2009) Standard test methods for measuring adhesion by tape test. ASTM
International, West Conshohocken, PA
11. ASTM E2260-03(2012)e1 (2012) Standard guide for repointing (tuckpointing) historic
masonry. ASTM International, West Conshohocken, PA
12. SNI 7973:2013 (2013) Spesifikasi Desain untuk Konstruksi Kayu, BSN, Jakarta
13. Ardian M (2018) Manual Pemeliharaan Bangunan Menara, Masjid dan Makam Sunan Kudus,
Kudus. (docplayer.info). Accessed Aug 2020
Self-monitoring and Localization
of Crack of Concrete Beam with Fibers
and Carbon Black Subjected to Bending
Genjin Liu and Yining Ding
Abstract In order to endow concrete an ability of sensing cracks, conductive mate-

rials (carbon fiber, steel fiber and carbon black) are added into concrete. The effect
of the conductive materials on the post crack performance of concrete subjected to
bending is studied. In order to localize the crack, multiple sections of a concrete
beam under bending are monitored. The influence of the conductive materials on the
self-sensing property of the crack opening displacement (COD) of fiber reinforced
concrete was investigated. The results illustrated that there is a linear relationship
between COD and FCR. The self-localization of the flexural cracks will be realized
based on the fractional change in resistance of concrete.
Keywords Self-localization of crack · Self-sensing of concrete crack · Fiber ·

Fractional change in resistance
1 Introduction
Rehabilitation and maintenance in civil engineering is not less important than plan-
ning and designing. The excessive load, natural hazard, lack of construction proce-
dures and material degradation are causes of infrastructure defects. It is well-known,
reinforced concrete works with cracks in the service stage. Concrete structure may
lose its serviceability if excessive cracking happens.
Rehabilitation and maintenance help to extend the service life of RC. Structure
repairing and strengthening belong to this topic. Localizing the crack and sensing
the crack widening of concrete are prerequisite of the rehabilitation of RC. Some
literatures have studied the strain sensing ability of short carbon fiber (CF) cementi-
tious materials; However, the monitoring ability range is limited in the elastic period
before cracking because the short carbon fibers are incapable to bridge the macro
G. Liu · Y. Ding (B)

State Key Laboratory of Coastal and Offshore Engineering, Dalian University of Technology,
Dalian 116024, China
180 G. Liu and Y. Ding
crack [1–3]. In addition, carbon fibers are brittle material, which hardly improve the
toughness of cementitious matrix obviously. One way to enhance the crack-bridge
ability and toughness of cementitious materials is to add macro steel fibers (SF) into
the matrix [4–6].
The self-sensing ability of cementitious materials with coarse aggregate are less
focused. The experimental investigations on concrete reinforced with carbon fiber
(CF), carbon black (CB) and macro steel fiber (SF) for crack sensing. The combi-
nation of two distinct fibers (SF, CF) as admixtures in concrete are effective for
enhancing the electric conductivity due to the formation of a continuous conducting
path way [3]. In addition, the carbon black with its high conductivity, low cost and fine
filler effect can be used as ideal admixture for conductive concrete before cracking
[7]. In the composite, CB exhibits the short-range contacts, whereas CF and SF
exhibit a long-range conductive bridging effect [8].
This paper mainly focuses on the post-crack self-sensing behavior of deflection
softening and deflection hardening concrete beams added with CF, CB and SF. The
results provide compelling evidence that a strong relationship exists between resis-
tivity and the load bearing capacity change and crack propagation. Furthermore, how
to localize of concrete crack is still an open problem, and the localization of crack is
a prerequisite for accurately self-sensing in crack widening of concrete. The purpose
of this article also develops a method for sensing the crack location and development
of concrete under bending.
2 Experimental Program
2.1 Materials and Mixture Design
The base mix design of the concrete is listed in Table 1. The CB, CF and SF content is
between 0–1, 0–2 and 40–60 kg/m3 . The diameter of CB is smaller than 90 nm. The
diameter of CF is 12 µm and length is 6 mm. The diameter of macro SF is 0.55 mm
and the length is 35 mm. Methylcellulose (0.4%wt) and defoamer (0.19%vol) are
added for CF reinforced specimens. The content of the conductive materials for the
concrete samples are listed in Table 2.
Table 1 Mix design (kg/m3 )

Cement CEM I Fly ash Water Fine aggregate Coarse aggregate Superplasticizer
42.5 (0–5 mm) (5–10 mm)
390 155 272.5 848 822 5.5
Self-monitoring and Localization of Crack … 181
Table 2 Comparison of
Specimen SF (kg/m3 ) CB (kg/m3 ) CF (kg/m3 )
content of conductive
admixtures SF40 40 – –
SF60 60 – –
SF40B1 40 1 –
SF60B1 60 1 –
SF40CF2B1 40 1 2
SF60CF2B1 60 1 2
2.2 Test Set-up
The concrete mixtures with conductive materials are produced in a laboratory mixer.
Specimens designed for bending test are beams with the size of 100 mm × 100 mm ×
400 mm. Four-probe method is adopted for electric resistance measurement, outer
electrical contacts A and D (130 mm apart) are designed for passing current while
the inner contacts B and C are prepared for measuring of voltage. Voltmeter V1 and
V2 are designed to measure the resistance of the pure bending part of the beam and
fixed resistor Rf, respectively.
The Relationships between FCR (Fractional change in resistance) and COD
(crack opening displacement) are discovered and established by the experimental
investigation. Details of the beams and electrical measuring system are shown in
Fig. 1a.
2.3 Testing Methods
The test set up are shown in Fig. 1b. Based on ASTM C1609, the experiment tests a
simply supported beam under third-point loading [9]. The deformation rate of mid-
span is 0.2 mm/min. LVDTs are used to measure the deflection of mid-span. An
extensometer is attached at the mid-span of the beam to measure the COD during
the test.
50 100 100 100 50

(a)
15 30 15 100 15 30 15
A B Conductive adhesive type C D
V1
Rf
AC power
V2
Load
(b)
Specimen
Rigid Baffle
Steel Bracket LVDT
Conductive adhesive type Conductive adhesive type
Extensometer
50 100 100 100 50
Fig. 1 Sample configuration for measuring the resistivity of the specimen, a probes; b schematic
of the experimental set-up
3.1 Crack Sensing
Figure 2a–f display the relationship between FCR and COD on the tension side of
the beams. From Fig. 2 it can be observed that the relationships between FCR and
COD for beams with 40 kg/m3 SF (series 40 specimens) are different to that of beams
with 60 kg/m3 SF (series 60 specimens).
For series 40 specimens, a monotone linearity increasing relationship between
FCR and COD can be found. In order to analysis the effect on the self-sensing
performance of different specimens, the gauge factor (GF), defined as the fractional
change in resistance per unit crack opening displacement, is introduced to quantify
the slop of the FCR–COD relationships. The GF values can be acquired from the
20 12
(a) SF40
Load
FCR
Linear fitting of FCR
10
(b)
15
8
Load (kN)
FCR (%)
10 1.08 6
1
4
5
2
0
0 1 2 3 4 5 6
COD (mm)
(b)25 SF60 Load

12
FCR
10
20
FCR (%)
Load (kN)
15
6
Zone I Zone II 1.74
10 1
4
5
2
1.49
1
0 0
0 1 2 3 4 5 6
COD (mm)
(c) 20
Load
16
SF40B1 FCR
18 Linear fitting of FCR 14
16
12
14
10
12
Load (kN)
FCR (%)
10 8
8
6
6
4
4 2.29
1 2
2
0
0 1 2 3 4 5 6
COD (mm)
Fig. 2 Relationship between load and COD, FCR and COD of a SF40, b SF60, c SF40B1, d
SF60B1, e SF40CF2B1, f SF60CF2B1 [10]
(d) 25
Load
10
SF60B1 FCR
20 8
Load (kN) 15 6
FCR (%)
Zone I Zone II
10 1.70 4
1
5 2
0.47
1
0 0
0 1 2 3 4 5 6
COD (mm)
20
(e) Load
16
SF40CF2B1 FCR
Linear fitting of FCR 14
15 12
FCR (%)
Load (kN)
10
10
1.49
8
1
6
5 4
0 0
0 1 2 3 4 5 6
COD (mm)
(f) 25 18
SF60CF2B1 Load
FCR
16
20
14
12
Load (kN)
15
FCR (%)
Zone I Zone II 10
2.72
8
10 1
6
4
5
1.55 2
1 1
0 0
0 1 2 3 4 5 6
COD (mm)
Fig. 2 (continued)
slope of the regression line fitted to the data from FCR–COD curve for the beams
(Fig. 2a, c, e). However, for series 60 specimens, bi-linearity increasing relationships
between FCR and COD are observed instead (see in Fig. 2b, d, f). The slope of FCR
and COD curve changes when the load reaches to the peak load. GF1 is the slope of
the regression line fitted to those data from FCR–COD curve in zone I, corresponding
to multiple cracks initial and propagation period before the peak load. And GF2 is
the slope of the regression line fitted to those data from FCR–COD curve in zone II,
in this region crack localization happened after post peak load.
The GF index are exhibited in Fig. 2, the GF varies from 0.47 to 2.72 for the spec-
imens. And it can be concluded that, GF2 is larger than GF1 for series 60 specimens.
Using specimen SF60B1 as a representative specimen, GF2 (1.70) is 3.6 times GF1 .
It means that, crack sensing ability of the specimen is better in crack localization
period than that in multiple cracks initial and propagation period.
• Compared with beam SF40, the GF of SF40B1 increased by 112%. And GF
of SF60B1 is 68.4% less than that of SF60 in zone I; however, the GF values
are almost the same with each other in zone II for the two specimen. It can be
concluded that 1 kg/m3 of CB will enhance the GF of concrete beam reinforced
with 40 kg/m3 steel fibers, but for beams with 60 kg/m3 steel fibers if added with
1 kg/m3 of CB the GF will reduce in zone I.
• Compared with SF40B1, the GF of SF40B1CF2 decreased by 34.9%. Whereas,
GF of SF60CF2B1 is 229.8 and 60% higher than that of SF60B1 in zone I and
zone II, respectively. It indicates that no obvious effect on the sensitivity with the
addition of 2 kg/m3 CF.
• SF40B1 doped with 40 kg/m3 SF and 1 kg/m3 CB possess the highest GF when
compared with the other beams of series 40 specimens.
• SF60CF2B1 doped with 60 kg/m3 SF, 2 kg/m3 CF, and 1 kg/m3 CB possess the
highest GF when compared with the other beams of series 60 specimens. In addi-
tion, specimen with triphasic conductive materials (SF60CF2B1) provides more
stable GF in both zone during the test than specimen with diphasic conductive
materials (SF 60B1).
3.2 Crack Location
In order to realize the crack region location, concrete beam with 40 kg/m3 steel fiber
and 1.5 kg/m3 carbon black is produced. Monitoring zone of the specimen is divided
into two parts, namely section AB (in red colour) and section BC (in blue colour),
see in Fig. 3.
Figure 4 exhibits the load-time-FCR curves of the SF40B1.5 under flexural load.
From Fig. 4, it can be seen that the load-time curve of the specimen increases up to
the peak load corresponding to the cracking point. Followed that, a great drop down
of loading (30% of the peak load) is observed during crack happening.
Fig. 3 Cracking pattern of SF40B1.5
18 125
16
100
14 FCR1
12 75
Load (kN)
FCR (%)
10
50
8
6 25
4
0
2 FCR2
0 -25
0 300 600 900 1200 1500
Time (s)
Fig. 4 Load-time-FCR curve of specimen SF40B1.5
According to Fig. 4, FCR of each section of the specimen almost equals to zero
before crack. In addition, FCR-Time curve of section AB (red curve) increase rapidly
after crack (at 426th seconds), however, the FCR-Time curve of section BC (blue
curve) decreased slightly (up to 10%) after crack.
The re-distribution of the potential of the specimen leads to the difference of FCR
curves for the cracked section (section AB) and the un-cracked section (section BC).
Part of the conductive paths can be cut off by the crack appeared in section AB,
meanwhile, the electrical current can only flow through the un-cracked concrete area
and the way bridged by the steel fibers.
A large beam with size of 120 mm × 200 mm × 1400 mm is also made attached
with different distance of inner electrodes. The distance of the inner electrodes are
200 mm (A1 –A2 ), 400 mm (B1 –B2 ), 500 mm (C1 –C2 ), 600 mm (D1 –D2 ) and 800 mm
(E1 –E2 ). Figure 5 illustrates the schematic diagram of electrodes. Figure 6 displays
the electrodes of the smart concrete beam with 20 kg/m3 and 1 kg/m3 carbon black.
The FCR-time-load curves of the smart concrete beam under bending is shown in
Fig. 7. From Fig. 7, it can be seen that the value of FCR for all the sections (besides
section A1–A2) increase rapidly after the crack appears. However, section A1–A2
Fig. 5 Schematic diagram of electrodes
Fig. 6 Electrodes of the smart concrete beam
Fig. 7 FCR-time-load curves of the smart concrete beam

shows a slightly reduction of FCR. Thus, it can be judged that the crack is outside
the section A1–A2 and inside the section B1–B2. From the experiment, a crack is
obeserved in section A2–B2. That implies that the crack location method is effective
to a certain extent.
4 Conclusion
(1) Self-sensing of concrete crack can be achieved by adding conductive materials

like SF/CB/CF into concrete. As the load bearing capacity declines rapidly at
the first cracking point, the FCR of all beams shows a clear increasing trend.
(2) A linearly increasing of FCR–COD correlation can be found for series 40
specimens. And, a bi-linearly increasing of FCR–COD are observed, and GF2
in zone II is larger than GF1 in zone I.
(3) The self-localization of the flexural cracks can be realized based on the
fractional change in resistance of concrete.
References
1. Ding Y, Chen Z, Han Z et al (2013) Nano-carbon black and carbon fiber as conductive materials
for the diagnosing of the damage of concrete beam. Constr Build Mater 43(Jun):233–241
2. Ding Y, Huang Y, Zhang Y et al (2015) Self-monitoring of freeze–thaw damage using triphasic
electric conductive concrete. Constr Build Mater 101:440–446
3. Ding Y, Han Z, Zhang Y et al (2016) Concrete with triphasic conductive materials for self-
monitoring of cracking development subjected to flexure. Compos Struct 138:184–191
4. Banthia N, Nandakumar N (2003) Crack growth resistance of hybrid fiber reinforced cement
composites. Cement Concr Compos 25(1):3–9
5. Banthia N, Sappakittipakorn M (2007) Toughness enhancement in steel fiber reinforced
concrete through fiber hybridization. Cem Concr Res 37(9):1366–1372
6. Buratti N, Mazzotti C, Savoia M (2011) Post-cracking behaviour of steel and macro-synthetic
fibre-reinforced concretes. Constr Build Mater 25(5):2713–2722
7. Wen S, Chung D (2007) Partial replacement of carbon fiber by carbon black in multifunctional
cement–matrix composites. Carbon 45(3):505–513
8. Wu S, Mo L, Shui Z et al (2005) Investigation of the conductivity of asphalt concrete containing
conductive fillers. Carbon 43(7):1358–1363
9. ASTM C 1609 (2012) Standard test method for flexural performance of fiber-reinforced
concrete (Using beam with third-point loading)
10. Ding Y, Liu G, Hussain A et al (2019) Effect of steel fiber and carbon black on the self-sensing
ability of concrete cracks under bending. Constr Build Mater 207:630–639
Seismic Performance Analysis
of Multi-story Buildings with Addition
of Bracing Based on SNI 1726: 2019
(Case Study: Airlangga University
Parking Building)
Krisnamurti , Willy Kriswardhana , and Achmad Wahyu Ramadiyan
Abstract There are various alternatives to improve the performance of a structure

in resisting lateral forces. One of them is the addition of types X, V, Z, and inverted V
braces or stiffeners to the elements of the portal structure. The research aims to study
the effect of X and V’s bracing types on the behavior of special moment-resisting
frame systems of building using the response spectrum analysis method. For analysis
purposes, 3 models of the frame were used, namely without bracing, with type X
braced frame, and type V braced frame. The three models can be used to fulfill the
requirement of service limit performance, ultimate limit performance, and control
of the P-delta effect. The biggest drift between floors occurs in modeling without
bracing. Significant changes occur in modeling with bracing type X and type V with
a drift reduction of 51.5 and 63.8%. The result shows that the use of bracing can
increase the stiffness and strength of the structure. The optimal bracing performance
in this study is bracing with type X.
Keywords Moment-resisting frame systems · Bracing · Response spectrum

analysis
1 Introduction
Indonesia has a high potential for earthquakes, one of which is in East Java. There
are two sources of earthquakes in the province of East Java, namely a plate collision
in the south of East Java (crack of the Indo-Australian plate) and an active fault in
Java (Baribis Kendeng Fold Thrust Zone) [1], both sources threaten most areas in
East Java and potentially a tsunami. The active fault passes through several cities,
one of which is Surabaya [2]. The large earthquake potential has a direct impact
on the buildings in the city, especially high-rise buildings. The little resistance to
Krisnamurti (B) · W. Kriswardhana · A. W. Ramadiyan

Department of Civil Engineering, University of Jember, Jalan Kalimantan 37 Kampus Bumi
Tegalboto, Jember 68121, Indonesia
190 Krisnamurti et al.
earthquake loads can cause the building to sway and even collapse and endanger
human safety.
The design and analysis of earthquake-resistant building structures in Indonesia
are guided by the latest regulation, namely SNI 1726-2019 concerning “Earthquake
Resistance Planning System for Buildings and Non-Buildings.” SNI 1726-2019
contains several methods for analyzing earthquake resistance in a building, namely
response spectrum analysis methods. This analysis method used to determine the
behavior and seismic forces acting on the structure [3] dan to calculate the displace-
ment and drift of the building [4]. Several structural systems can be used as earth-
quake resistance, as shear wall, braced frame, and moment-resisting frame system.
The use of shear walls and bracing in the same building have the same perfor-
mance level [5]. It stated that adding bracing with the right location can significantly
improve the structure’s performance [6]. In general, bracing is placed diagonal using
various configurations in the structure’s portal. The stiffness increased efficiently
using bracing because a diagonally mounted brace can only withstand axial forces
while receiving horizontal shear forces. The shape of concentric bracing can be the
V
Z (diagonal), X, V, and inverted V ( ) which can be seen in Fig. 1 [7].
In the comparison of bracing with the types X, V, Z, and inverted V, it is known
that the optimum structural performance was obtained from bracing type X with
the smallest displacement value compared to other types of bracing [8]. Meanwhile,
the type V bracing placed on both building axes (x and y axes) will increase the
performance of the structure and there is a decrease in displacement compared to
structures without bracing [9].
Airlangga University Parking Building is a 10-story building with a ± 36,340 m2
and a height of 30,8 m. The dimension of building is 112,8 × 32,3 m. The type of soil
in the location is medium soil. The building is planned to contain 414 motorbikes on
the first floor and 124 cars on the 2nd to 10th floors. With ten floors and large external
loads, the building that can accommodate up to 1000 cars in its planning does not
use shear walls and is guided by the SNI 1726:2012 regulations. This problem is
the basis for this research to review the suitability of planning against regulations
of SNI 1726-2019 and to add bracing as a substitute for shear walls. Based on the
explanation above, the aim of this research is to analyze the seismic performance of
the structure using the response spectrum analysis method in the parking building of
Airlangga University using the braced frame variation.
Fig. 1 Various forms of concentric braces [7]

Seismic Performance Analysis of Multi-story Buildings … 191
2 Research Methodology
The structural data using a 10-story parking building, which is for 3D modeling.
The structural data consist of floor plans, beams, columns, and plates along with the
quality of concrete and reinforcing steel, as well as cross-sections and longitudinal
sections to determine the height of each floor. The structural model used in this study
is a 10-story parking building, which consists of 3 models. Model 1 is an existing
building structure. Models 2 and 3 are a modification of model 1. Bracing is used as
an element to resist lateral forces (earthquake forces). Model 2 is a model with type
X bracing, while model 3 is a model with type V bracing. In practice, the bracing
elements in this structure must be prepared as precast elements to avoid difficulties
in the casting process (Fig. 2).
The materials quality are:
– the beam and slab concrete (f ’c) = 25 MPa.

– the column and bracing concrete (f ’c) = 25 MPa.
– the reinforcing steel (f y) = 240 MPa (BJTP 24).
the reinforcing steel ( f y) = 240 MPa(BJTP 24).

= 400 MPa(BJTD 40).
The 3D model of the building structure is in the Fig. 3, and Fig. 4. Dead load
(DL), additional dead load (SDL), live load (LL), earthquake load for X direction
(EX), and earthquake load for Y direction (EY) use as action load. Life and death
loads refer to PPIUG 1987 (regarding building loading) and SNI 1727: 2013. The
earthquake load refers to SNI 1726: 2019, and The Surabaya earthquake area is using
with moderate ground conditions. Earthquake loading calculations using response
spectrum analysis methods.
Fig. 2 Airlangga University parking building plan campus B

Fig. 3 3D building
modeling without bracing
Fig. 4 a Front View of Type X Brace Placement in Building. b Side View of Type X Braces in the
Building. c Front View of Type V Brace Placement in Building. d Side View of Type V Braces in
the Building
Story drift is a determinant of a building’s service limit performance. To determine

the amount of story drift, a structural analysis calculation tool is used. The allowable
limit for the change in the deviation between levels must not exceed the following
equation:
S < 0, 03/R × h (1)
where S = storey drift displacement (mm), R = modification coefficient, and h

= storey height. From Eq. (1) it is obtained the service limit performance value
of 12.75 mm. Figure 5 and Fig. 6 shows the graph S value of each model and
the direction of the earthquake, Figs. 5 and 6 show that the three models that have
been made have met the service limit performance controls. The highest S value is
experienced by building models without bracing while building models with bracing
type X experience the lowest S value. The ultimate limit performance is regulated
in SNI 1726: 2019 article 7.12.1 [10], with the deviation limit value between floors
determined by the following equation:
Storey drift limit (limit ) = 0, 015hsx (2)
Fig. 5 Comparison diagram of service limit performance on X-direction earthquake
Fig. 6 Comparison diagram of service limit performance on Y-direction earthquake

Storey drift maximum (x ) = S × Cd (3)
x < limit (4)
where hsx = height of the level below the x-level; S = storey drift displacement
(mm); Cd = Deflection Factor. From Eq. (2) it is obtained the Ultimate Limit Perfor-
mance value of 55 mm. Figure 8 and Fig. 9 show the graph x value of each model
and the direction of the earthquake.
Figure 7 and Fig. 8 shows that the three models have met the Ultimate Limit Perfor-
mance controls. The highest x value is experienced by building models without
bracing while with bracing type X experience the lowest x value. Control of P-
delta effects performance refers to the SNI 1726: 2019 clause 7.8.7, to determines
Fig. 7 Comparison diagram of ultimate limit performance on X-direction earthquake
Fig. 8 Comparison diagram of ultimate limit performance on Y-direction earthquake

11
10
9
8
7
6
Floor
5
4
3
2
1
0
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
θ
θX θY θ Max
Fig. 9 Control diagram of P-Delta effect of buildings without bracing in the Surabaya zone
earthquake with x and y directions load
the stability coefficient (θ):
Px Ie
θ= (5)
Vx hzx Cd
where Px = total load above level x, (kN); = storey drift displacement (mm); Ie
= primary factor quake; Vx = seismic shear force at the x and x-1 levels (kN); hzx
= height x level (mm); Cd = deflection magnification factor. The value of stability
coefficient (θ) should not exceed θmax . The value of θmax determined by:
0.5
θmax = (6)
βC d
From Eq. (6), the θmax value is 0.09091.

Figures 9, 10, and 11 shown that the three models have θx and θy values below
the maximum theta, which means that the p-delta effect that occurs in the building is
in a safe condition. Based on the control and structural performance analysis that has
been carried out, it is found that the parking building of Airlangga University, which
has been planned, is safe against earthquake loads for the Surabaya earthquake zone.
However, there is a significant difference in the value story drift between the building
model using and without bracing. Figures 12 and 13 show a comparison diagram
story drift of the parking building, which uses bracing and without bracing.
Figure 12 and Fig. 13 shown that the story drifts value of a building model
with bracing has decreased by more than 50% compared to the story drift of a
building model without bracing. This shows that bracing has a major effect on the
performance of the building structure in resisting earthquake forces. Previous studies
also mentioned the same statement. The performance of bracing structures is better
compared to structures without bracing [6]. The picture above shows the structure’s
optimal performance value is a structure with the addition of X-type bracing when the
11
10
9
8
7
6
Floor
5
4
3
2
1
0
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
θ
Teta X Teta Y Teta Max
Fig. 10 Control diagram of P-Delta effect of buildings with bracing type X in the Surabaya zone
11
10
9
8
7
Floor
6
5
4
3
2
1
0
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
θ
Teta X Teta Y Teta Max
Fig. 11 Control diagram of P-Delta effect of buildings with bracing type V in the Surabaya zone
value of “deviation between floors” is used as a reference. Previous studies obtained

the same results. The structural model with type X bracing has the best performance
compared to other types of bracing models [8].
Fig. 12 Comparison diagram of the total deviation of buildings with and without braces for
earthquake loads in the X direction of the Surabaya zone
Fig. 13 Comparison diagram of the total deviation of buildings with and without braces for
earthquake loads in the Y direction of the Surabaya zone
4 Conclusion
(1) The story drift of the Airlangga University parking building has decreased signif-
icantly after being given bracing. The value story drift of the building model without
bracing is 16,269 mm, but after bracing is used, it has decreased to 10,714 mm for
the V type brace and 8,156 mm for the X type brace; (2) Based on the performance
control, the building model structure with bracing and without bracing has met the
requirements for the structure’s performance limits. The optimal type of bracing in the
Airlangga University parking building is type X bracing. The increase in building
performance of type X bracing reached 68.6% for service limit performance and
73.6% for ultimate limit performance, while type V bracing increased 62.11% for
service limit performance and 67.3% for ultimate limit performance.
References
1. Tim PSGN (2017) Peta Sumber Dan Bahaya Gempa Indonesia Tahun 2017. PuslitbangPP,
Balitbang Kemen PUPR, Bandung 40393
2. Kuncoro H, Kartini GAJ, Meilano I, Susilo S (2019) Identification of fault mechanism in
Eastern part of Java Island using Continuous GNSS data from 2010 to 2016. In: Seminar
Nasional Geomatika, vol 3. Bandung, pp 805–812
3. Rendra R, Kurniawandy A, Djauhari Z (2015) Kinerja Struktur Akibat Beban Gempa Dengan
Metode Respon Spektrum dan Time History (Studi Kasus: Hotel SKA Pekanbaru). JOM
FTEKNIK 2(2):1–15
4. Afrida I, Wahyuningtyas WT, Krisnamurti (2020) Analisis Ketahanan Gedung Apartemen
Surabaya dengan Menggunakan Metode Respon Spektrum, Berkala Sainstek 8(4):132–139
5. Hutahaean SG, Aswandy (2016) Kajian Pemakaian Shear Wall dan Bracing pada Gedung
Bertingkat, Reka Racana Jurnal Online Institut Teknologi Nasional 2(4):100–111
6. Haryono S, Arumningsih D, Purnamawanti D (2015) Penggunaan Struktur Bresing Konsentrik
Tipe X Untuk Perbaikan Kinerja Struktur Gedung Bertingkat Terhadap Beban Lateral Akibat
Gempa, Jurnal Teknik Sipil dan Arsitektur 16(20)
7. Smith BS, Coull A (1991) Tall building structures: analysis and design. Willey, Inc., New York
8. Aryandi D, Herbudiman B (2017) Pengaruh Bentuk Bracing terhadap Kinerja Seismik Struktur
Beton Bertulang. Reka Racana Jurnal Online Institut Teknologi Nasional 3(1):1–11
9. Repadi JA, Sunaryati J, Thamrin R (2016) Analisis Kinerja Struktur Beton Bertulang Dengan
Variasi Penempatan Bracing Inverted V. Jurnal Rekayasa Sipil 12(2):32–39
10. SNI 1726:2019 (2019) Tata cara perencanaan ketahanan gempa untuk struktur bangunan
gedung dan non gedung. Jakarta
Various Strut—Macro Modelings
for Infilled Frame Analysis
Isyana Ratna Hapsari, Marwahyudi, Edy Purwanto, Senot Sangadji,

and Stefanus Adi Kristiawan
Abstract Equivalent diagonal strut is a simple macro model used to represent the
behavior of masonry walls as an infilled frame structure when exposed to lateral
loads. The model has been shown capable to produce the response of an infilled frame
structure that is comparable to the experiment results. Several important parameters
need to be considered when initiating diagonal strut modeling. The selection of strut
type related to the number of struts, i.e. single, double, or triple struts will greatly
affect the behavior of infill walls. In addition, the residual area after crack (Am2)
is a very important parameter to be defined properly when modeling the strut. This
study shows that the double strut (DS) and triple strut (TS) give response of infilled
frame structure close to testing, with an error is around 9–35% when the response is
assessed based on peak load, displacement at peak load, and initial stiffness of the
structure. For DS model, further improvement can be made by assigning the Am2 to
around 55 and 78%.
Keywords Macro modeling of masonry · Diagonal strut · Number of struts ·

Residual area of strut
I. R. Hapsari (B) · Marwahyudi

Civil Engineering Department, Faculty of Engineering, Universitas Sebelas Maret,
Marwahyudi
E. Purwanto · S. A. Kristiawan
SMARTCrete Research Group, Department Faculty of Engineering, Universitas Sebelas Maret,
Civil Engineering, Surakarta, Indonesia
S. A. Kristiawan
S. Sangadji
SMARTQuake Research Group, Civil Engineering Department, Faculty of Engineering,
200 I. R. Hapsari et al.
1 Introduction
Research on masonry infill frame structures has been carried out for a long time.
This is not surprising since masonry infill frame is a common structure for residential
buildings in many countries. Previous studies have shown that masonry infill walls
contribute to the increase of stiffness and strength of structures [1]. Therefore, the
contribution of infill walls should be considered in the modeling to describe the
structure’s response accurately when exposed to earthquake loads. However, the
actual behavior of infill is very complex. Non-standard characteristics of masonry in
each region due to workmanship contributes an additional complexity to the behavior
of this infill wall. Furthermore, the bonding between bricks and mortar also affects
the behavior of the masonry structure.
According to Crisafulli [2] there is a grouping of parameters that can affect the
response of the masonry infilled frame:
1. Masonry unit: dimension, mechanical properties, type of brick, type of brick
surface character
2. Mortar joints: mechanical properties, lime content, the thickness of the species,
if any reinforcement is used
3. The structural frame: mechanical properties of the concrete or steel used,
dimension, the amount of longitudinal and transverse reinforcement used.
4. General i.e. type of excitation loading, the number of floors, the number of
bays, the presence of openings or not, the position of the openings, and the
construction technique.
In previous studies, several macro models have been introduced to describe the
infill panels. One of such models is the equivalent diagonal strut. The type of diagonal
strut its self has transformed; starting from what was introduced by Crisafulli et al.
[3] in the form of a single strut. The single strut has the advantage of the simplicity
of mathematical formulas but it is still capable to provide sufficient estimation of the
stiffness value. However, the single strut is unable to predict the shape of the deflection
that occurs in the column. According to Kumar [4], a single strut gives an overestimate
result of the displacement of the column. In addition, this model produces a bending
moment and shear forces that are very low or unrealistic when compared to the strut
model quantification through FEM Software. Meanwhile, the double strut model
is more capable of illustrating the similarity of the model crack pattern to the test
object. El-Dakhakhni [5] and Tanganelli et al. [6] stated that triple struts produce the
most rational bending moment values if compared to other strut models. In addition,
the principle that the walls do not immediately fail simultaneously can capture more
accurately when it is estimated by using triple strut modeling.
The purpose of this paper is to identify parameters of the equivalent diagonal
compressive strut, which will determine the suitability of the model to represent the
behavior of the infilled frame structure. Three types of diagonal compressive strut
models will be studied i.e. single strut (SS), double strut (DS), and triple strut (TS).
The test results of the infilled frame by Kakaletsis et al. [7] will be used as a validation
Various Strut—Macro Modelings for Infilled Frame Analysis 201
for the numerical simulation. Numerical modeling of the infilled frame with the diag-
onal compressive strut approach is carried out using the SeismoStruct 2020 [8]. Struts
are built-in SeismoStruct by defining 32 parameters covering the geometric, mechan-
ical, and empirical parameters. Modeling in this software considers the behavior of
the fiber-based cross-section about the uniaxial stress-strain relationship. The results
of this study can be used as a reference for simple analysis of masonry infilled frame
yet keeping the accuracy of the numerical model in predicting the behavior of the
actual structure.
2 Description of Model Structure
2.1 Geometry and Materials
Several data to build a numerical model in this study will use the existing two-
dimensional portal model researched by Kakaletsis [7]. There are two types of portals
in Kakaletsis’s study i.e the open frame reinforced concrete (OF) model and the
masonry infilled reinforced concrete frame (IF) model. The OF modeling and vali-
dation are needed to ensure that OF behavior is as expected, so that in the subsequent
modeling i.e infilled frame (IF), we can quantify the contribution of infill walls to the
behavior of the structure. The constitutive model of concrete follows that of Mander
[9] for both confined and unconfined concrete sections. As for the steel material,
the model of Menegotto-Pinto [10] is selected. The boundary conditions of a model
are defined as fixed support. These structural models can be described as shown in
Fig. 1. The geometric parameters and mechanical properties of the model structures
can be seen in Tables 1 and 2. Especially for geometric dimensions using a model
with a size of 1/3 of the prototype.
Columns use the longitudinal reinforcement 6D100 and stirrup reinforcement
D80-200. Whereas for beams use the longitudinal reinforcement 6D80 and stirrup
Fig. 1 3D modeling of the open frame (left)—infilled frame (right)

Table 1 The geometry of

Model parameters Dimension (mm)
structure elements
Column section 150 × 150
Column height 1150
Beam section 100 × 200
Beam length 1350
Wall thickness 60
Table 2 Mechanical
Mechanical properties Values
properties of the model
Compressive strength of concrete (Mpa) 28.51
Specific weight of concrete (kg/m3 ) 2200
Modulus of elasticity—concrete (Mpa) 23692
Diagonal compressive strength of masonry (Mpa) 5.11
reinforcement D80-200. Based on additional information through numerical valida-

tion by Sukrawa [11] using strut and shell modeling, he showed that the modulus of
elasticity of the masonry (Em) obtained from the experiment i.e. 670.3 MPa is too
small when compared with the test results of other studies. For this reason, Sukrawa
used the value Em = 2810 MPa, referring to the equation presented by FEMA 356
[12], i.e. Em = 550 fcm. This modulus of elasticity will be applied to the current
study.
2.2 The Simplified Infilled Frames—Macro Model
There are 2 (two) types of concepts for modeling the masonry infilled frame. The
first is a micro model where the infill structure is divided into many elements in
detail using the discretization method to take into account the local effects when
subjected to earthquake loads. This complex micro model requires large computeri-
zation efforts and time from the modeling process to the analysis stage. The second
type of infilled frame modeling is the macro model, a simplification of the masonry
model based on a physical understanding of the infill behavior in order to approx-
imate the actual behavior of the infilled frame when subjected to loads. The main
objectives of making this macro model are to reduce the complexity and the need for
very large computation.
Crisafulli’s research related to infill walls proved that when the infilled frame is
subjected to increased lateral loads, some parts of the walls or connection between
masonry and frames suffered cracks. If the lateral load continues to increase then
part of the wall will separate (gap) from the frame. After the existence of the gap,
the stress at the corner of the tensile area decreases while those near the corner of
the compressive area increase significantly. The stress distribution that occurs after
separation produces a diagonal stress trajectory. Based on these results, Crissafuli

proposed a compressive diagonal strut to simplify the model of the masonry infill
panel. Crisafulli and Crowley et al. [13] stated that diagonal compressive single strut
can represent the behavior of infilled frame. However, further, development has been
proposed to improve the prediction of the behavior by introducing double strut and
triple strut [3, 5].
In this current study, 3 (three) types of diagonal struts for masonry infill panels
modeling will be discussed. The principle differences underlying the three types of
diagonal strut models can be seen in Fig. 2 [3]. Which is the hz value is the vertical
distance between the nodes on the strut, and Am1 is the strut area 1 or initial strut. The
model parameter of the infilled frame based on the equivalent diagonal compressive
strut in SeismoStruct 2020 uses 3 main criteria. The first criterion is the geometric
parameter related to the dimensions of the infilled frame to be modeled, according
to Fig. 3 below (a) illustrate the real wall model, (b) masonry modeling built by
the geometric parameters, and (c) strut approach for macro modeling as the infilled
frame in SeismoStruct 2020. The second criterion is mechanical parameters, which
are parameters related to the material that makes up the infilled frame. And the third
criterion is the empirical parameters, which are the parameters needed to represent
the behavior of infilled frame due to cyclic loading.
The general parameters related to the 3 (three) criteria mentioned above are
presented in Table 3. The parameters used for modeling the diagonal strut refers
to the values and formulas of the previous research on infilled frame [2, 14–16].
Fig. 2 Macro modeling of masonry infilled using various strut types [3]
Fig. 3 a Masonry visual. b Geometrical parameters of masonry infilled for strut modeling. c Macro
modeling of masonry infilled in SeimoStruct 2020
Table 3 Infilled parameters, equations, and references

Infilled panel model parameters Unit Used value Equation Ref
1. Mechanical properties
a. Tensile Strength, ft MPa 0 [14]
b. Bond shear strength, τ 0 MPa 0.08 [7]
c. Coefficient of friction, μ 0.77 [7]
d. Maximum shear stress, τ max MPa 1 [14]
e. Strain at max stress, (εm ) MPa 0.0012 [14]
f. Ultimate strain, (εu ) MPa 0.024 εu = 20 × εm (1) [14]
g. Closing strain, (εcl ) MPa 0.003 [14]
h. Specific Weight, W N/mm3 1.7E-005
2. Empirical Properties
a. Starting Unloading Stiffness 1.7 [14]
Factor, (γ un )
b. Strain Reloading Factor, (α re ) 0.2 [14]
c. Strain Inflection Factor, (α rh ) 0.7 [14]
d. Complete Unloading Strain 2 [14]
Factor, (β a )
e. Stress Inflection Factor, (β ch ) 0.9 [14]
f. Zero Stress Stiffness Factor, 1 [14]
(γ plu )
g. Reloading Stiffness Factor, 1.1 [14]
(γ plr )
h. Plastic Unloading Stiffness 3 [14]
Factor, ex1
i. Repeated Cycle Strain Factor, 1 [14]
ex2
j.Reduction Shear Factor, (α s ) MPa 1.43 [2]
k. Out-of-plane Failure Drift % 1 [15]
(continued)
Table 3 (continued)
Infilled panel model parameters Unit Used value Equation Ref
l. Proportion of stiffness assigned % 70 [14]
to shear, γ s
3. Geometrical Properties

E m tw sin(2θ )
a. Dimensionless relative λh = h 4
4E c Ic h w ….(2) [2]
stiffness, λh
b. Strut Area Reduction Strain, 0.0006 [16]
(ε1 )
c. Residual Strut Area Strain, (ε2 ) 0.001 [16]
0.5×column width
d. Horizontal Offsets, x oi % Varies xoi = width o f in f ill panel (3) [14]
0.5×beam depth
e. Vertical Offsets, yoi % Varies yoi = height o f in f ill panel (4) [14]
3 Numerical Simulation
3.1 Open Frame Model for Validation
The cyclic loads employed in this Kakaletsis study were replaced by lateral loads
which were increased incrementally per 10 kN until the maximum load was reached.
Before conducting a parametric study on the modeling of infilled wall structures
using a diagonal compressive strut, we validated the open frame structure modeling
first. The open frame modeling and analysis process is assisted by the SeismoStruct
2020. The numerical modeling is validated by the results of Kakaletsis [7]. The
resulting capacity curve (see Fig. 4) can provide an overview of the fitness of the
model. It can be concluded that the result of the open frame modeling in SeismoStruct
is comparable to the experiment result.
Fig. 4 Capacity curve open 50

frame between experiment
and numerical modeling 40
Base Shear (KN)
30
OF_KAKALETSIS
20
OF_SEISMOSTRUCT
10
0
0 5 10 15 20 25 30
Displacement (mm)
3.2 Numerical Model with Various Types of Struts
In this chapter, the numerical study is carried out to compare the results of the infill
panel testing conducted by Kakaletsis with several types of infill panel models. The
following stages of modeling are conducted in this study: the first is modeling the
masonry infilled frame using 3 types of diagonal struts, the second is assigning
variations in the elastic modulus of masonry (Em), and the third is defining various
residual strut area (Am2). From these stages of numerical modeling, it can show
the most suitable diagonal compressive strut model in simulating experimental test
results.
In the first stage of modeling, the fundamental parameter that distinguishes the
three strut types is strut area 1 (Am1). The formula for getting the value of strut area
1 (Am1) in single strut (SS) model is:
Am 1single = bw1 .tw (5)
Where t w is the masonry thickness and bw1 is the width dimension of the diagonal
strut, as recommended by FEMA 356 as follows:
bw1 = 0, 175(λh )−0,4 dw (6)
The definition of d w is the length of the diagonal strut being modeled. For dimen-
sionless relative stiffness, λh is a parameter that expresses the relative stiffness of
the infill panel to the frame structure. This parameter is sufficient to describe the
elasticity behavior of the infill panel. Meanwhile, to get the area of the strut area 1
(Am1) in the double strut (DS) and triple strut (TS) model are as follows:
bw1.tw
Am1double = (7)
2
bw1.tw
Am1tri ple = (8)
4
The definition of the hz value is the vertical distance between the nodes on the
strut in double strut (DS) modeling with the following formula:

h z = 1 3.z (9)
Meanwhile, the hz value in triple strut (TS) modeling uses the following formula:

h z = 1 2.z (10)
where z is the actual contact length value determined using the formula (11).
Meanwhile, the actual contact length (z) in single strut modeling is defined as 0.
π
z= (11)
2λ
The capacity of curves obtained from numerical modeling of masonry infilled
frame (IF) structure using various strut models is shown in Fig. 5. This curve is
constructed by first setting the displacement control value = 30 mm, and iteration
is carried out in 50 steps. From Table 4 and Fig. 5, it can be concluded that double
strut (IF_DS) and triple strut (IF_TS) give the closest to the structural response
of the experimental results. Variances of response values in IF_DS and IF_TS are
relatively smaller than the single strut model (IF_SS). This can be seen through the
quantification of model errors based on the peak load (IF_DS = 11%, IF_TS = 16%)
and initial stiffness (IF_DS = 35%, IF_TS = 9%).
In the second study, the mechanical properties of the double strut i.e the
elastic modulus of masonry (Em) will be adjusted by the results of the Kakaletsis
(670.3 MPa). The double strut model with this Em value is assigned as
IF_DS_EM670. The resulting capacity curve (see Fig. 6) shows that the Em value
contributes to the added strength of the overall structure. The IF_DS_ EM670 model
can withstand a maximum lateral load of 95 kN, an increase of 8% from the initial
140
130
120
110
100
90
Base Shear (KN)
80
70
60
50
40 IF_EXP_KAKALETSIS
30 IF_SS
20 IF_DS
10 IF_TS
0
0 5 10 15 20 25 30
Displacement (mm)
Fig. 5 Capacity curve infilled frame between experiment and single, double, triple strut
Table 4 Parameters of capacity curves determined by various struts models

Parameters Models Model Error
quantification IF_EXP_KAKALETSIS IF_SS IF_DS IF_TS IF_SS IF_DS IF_TS
(%) (%) (%)
Peak load (kN) 79 130.96 88.08 66.70 66 11 16
Displacement at 8 9 10.2 10.8 13 28 35
peak load (mm)
Initial 19.38 39.92 26.22 21.12 106 35 9
stiffness (kN/mm)
Fig. 6 Capacity curve 100

infilled frame between 90
experiment, double strut 80
(DS) and double strut with 70
Base Shear (KN)

Em adjustment (DS_EM670) 60
50
40
IF_EXP_KAKALETSIS
30
IF_DS
20
IF_DS_EM670
10
0
0 5 10 15 20 25 30
Displacement (mm)
DS modeling which uses an Em value of 2810 MPa. The smaller Em produces a

greater value of the strut width (bw1 ), because the value of the strut width referring
to formula (6). Hence, this strut width (bw1 ) value will lead to the wider area of
strut 1 (Am1), according to the calculation in formula (7). This is the reason why the
smaller Em will influence the strength of infilled frame structure.
Table 5 shows the parameters of capacity curves determined using a double strut
model with various Em values. IF_DS_EM670 produces a larger error value, where
the peak load and displacement at peak load indicate error values of 20% and 80%,
respectively, in comparison to the experimental result. Thus, it is reasonable to use
Em estimated from FEMA 356 (2810 MPa) for constructing the diagonal strut model.
The geometric parameter is one of the influencing parameters to the behavior of
infilled frame against the lateral loads according to Tanganelli et al. [6]. Thus, in this
third stage of the study, the contribution of the residual area of the strut (Am2) to
the infilled frame behavior will be explored. Crisafulli [2] and Smyrou [14] stated
that the area of the residual strut (Am2) occurs due to cyclic or increased lateral
loads, causing the contact length between the frame and the infill panel to decrease.
This will automatically affect the area after the panel has cracked. Despite the lack
of information about the practical value to estimate this parameter, Crisafulli and
Smyrou used the experimental results report from Decanini [17] which suggests that
the strut equivalent width is reduced by 20–50% due to the cracking of the infill
panel.
Table 5 Parameters of capacity curves determined from double strut model with various Em
Parameters Models Model error
quantification IF_EXP_KAKALETSIS IF_DS IF_DS_ EM670 IF_DS IF_DS_
(%) EM670
(%)
Peak load (kN) 79 88.08 94.71 11 20
Displacement 8 10.20 14.4 28 80
at peak load (mm)
Initial 19.38 26.22 19.6 35 1
stiffness (kN/mm)
Fig. 7 Variation of the strut Am

area related to the alteration
of axial strain value
Am1
Am2
ɛ1 ɛ2 ɛ
According to Meslem [16], when the elastic limit of the infill panel is exceeded
due to the cracks the assumption used between the area of strut 1 and 2 has linear
variation as a function of axial strain (see Fig. 7). And after doing the parametric
study, Meslem chooses the residual area of strut (Am2) = 55%.
Meanwhile, according to Al-Chaar [18], a reduction factor used to quantify the
strut area is 0.7 and 0.4 for moderate and heavy damage infill panel, respectively.
Kumar [4] argues that Am2 has varied values in the range of 60 and 80%, and based
on his parametric study, Am2 = 70% was able to give results that were almost alike
experimental. Wang et. al. [19] and Totoev [20] give Am2 value about 40–50% for
their proposed 2 infill wall models.
In this current study, we estimated the residual value of the strut area (Am2) based
on the stiffness degradation data of the test results from Kakaletsis. The resulting
stiffness degradation is shown in Figure 8. This graph can show the value of stiffness
in elastic conditions of 19.38 kN/mm corresponds to the displacement of 2.58 mm
where the first crack occurs. The comparison of the stiffness at the first crack with
the initial stiffness is equal to 78%. Meanwhile, the residual strength of the infill
panel occurs at a displacement of 22.5 mm (see Fig. 9) corresponding to a stiffness
of 2.40 kN/mm (see Fig. 8). At this residual strength, the stiffness ratio remains
10%. For detailed information regarding the behavior of infill walls in an elastic
state, see Fig. 9. Through several literature studies and consideration of the value of
30.00
Stiffness (kN/mm)
20.00 19.38
IF_EXP_KAKALETSIS
10.00
2.40
0.00
0.00 5.00 10.00 15.00 20.00 25.00 30.00
Displacement (mm)
Fig. 8 Stiffness degradation of Kakaletsis’s infilled frame experiment

90
80 79 78 77
75 75
70 70 70
Base Shear (kN)

60 60 62
50 54 51.5 52.553.5
50
40 40
30 30 IF_EXP_KAKALETSIS
20 20
10 10
0 0
0 5 10 15 20 25 30
Displacement (mm)
Fig. 9 Plastic—elastic condition for identifying first crack and residual strength condition of infilled
frame
stiffness degradation, we propose the value of residual strut area (Am2) in the range
of 44–78%. This range of Am2 will be used to modify the IF_DS (see Table 6).
A comparison of the resulting capacity curves between the double strut model
with various Am2 values can be seen in Fig. 10. Because at this modeling phase
only the Am2 variable is adjusted, the character and initial stiffness values have
a similarity. From this reason, we do not need to show this comparison of values.
The graph shows IF_DS_AM2_55% gives the most similarity values of peak load
achievement. And IF_DS_AM2_78% shows the closest values of stiffness at peak
load (see Table 7).
Table 6 Type of double strut model with proposed residual strut area (Am2)
Infilled frame model Am2 (%) Reason for proposing
IF_DS 70 Initial setting
IF_DS_AM2_78% 78 First crack condition
IF_DS_ AM2_55% 55 Estimation
IF_DS_ AM2_44% 44 Mean values between first crack—residual condition
100
90
80
70
Base Shear (KN)
60
50
IF_EXP_KAKALETSIS
40 IF_DS
30 IF_DS_AM2_78%
20 IF_DS_AM2_55%
10
0
0 5 10 15 20 25 30
Displacement (mm)
Fig. 10 Capacity curve infilled frame between experiment, double strut (DS) and double strut with
Am2 variation (DS_AM2_78%, DS_ AM2_55% and DS_ AM2_44%)
Table 7 Parameters of capacity curve determined from double strut model with various Am2
Parameters Models Model error
quantification IF_EXP_KAKALETSIS IF_ IF_DS_AM2_78% IF_DS_AM2_55% IF_DS_AM2_44% IF_ IF_DS_AM2_78% IF_DS_AM2_55% IF_DS_AM2_44%
DS DS (%) (%) (%)
(%)
Peak Load 79 88.08 92.86 79.14 72.60 11 18 0 8
(kN)
Displacement 8 10.20 9.60 10.20 10.80 28 20 28 35
at Peak Load
(mm)
Various Strut—Macro Modelings for Infilled Frame Analysis
Stiffness 9.88 8.64 9.67 7.76 6.72 13 2 21 32

at Peak Load
(kN/mm)
211
Based on the numerical study, it is shown that the smaller the Am2 value will
greatly affect the decrease in the strength and stiffness of the infilled frame. From
this analysis, we can conclude that the determination of the residual strut area after
cracked (Am2) is very important in initiating infill panel modeling as a diagonal
compressive strut. Based on experiments conducted by Kakaletsis, it is stated that
nonlinear behavior is initiated by infill cracking. These cracks start as a 45° diagonal
line and then join with several horizontal shear cracks. The failure mode of the
numerical simulation shows that the masonry loses its contribution when the masonry
stiffness decreases about 40% remaining from the initial condition.
4 Summary and Conclusions
Three types of diagonal struts i.e. single strut (SS), double strut (DS), and triple
strut (TS) have been used in this study to represent the influence of masonry infilled
panel on the behavior of frame structure. It is shown that the double strut (DS) and
triple strut (TS) models have the capability to estimate the structural response of
masonry infilled frame structures. The corresponding model errors for DS and TS
are 11 and 16% when assessed on the basis of peak load, and 35 and 9% based on
initial stiffness.
The modulus elasticity masonry (Em) influences the overall behavior of the struc-
ture. A smaller value of Em increases the strut width (bw1), thereby increasing the
strength of the structure as a whole. IF_DS_EM670 with Em of 670.3 MPa can
withstand a maximum lateral load of 95 kN, which represents an increase of 8%
compared to IF_DS model with Em of 2810 MPa. However, the IF_DS_EM670 has
bigger (more deviant) model errors i.e. 20% and 80% when assessed based on its
peak load and displacement at peak load, respectively.
The Am2 has been determined in this study based on the stiffness degradation.
The proposed Am2 is in the range of 44–78%, representing a range of conditions
from initial stiffness to the stiffness at residual strength. IF_ DS_ AM2_55% model
with Am2 equal to 55% provides the best estimate of the infilled frame behavior with
model errors 0% when viewed from the peak load achievement. Whereas IF_ DS_
AM2_78% model with Am2 = 78% can show the value of stiffness at peak load
with model error in 2%.
References
1. Stavridis A, Koutromanos I, Shing PS (2012) Shake-table tests of a three-story reinforced

concrete frame with masonry infill walls. Earthq Eng Struct Dyn
2. Crisafulli FJ (1997) Seismic behaviour of reinforced concrete structures with masonry infills.
In: Thesis, University of Canterbury
3. Crisafulli FJ, Carr AJ (2007) Proposed macro-model for the analysis of infilled frame structures.
Bull New Zeal Soc Earthq Eng 40(2)
4. Kumar M, Khalid F, Ahmad N (2018) Macro-modeling of weak masonry infilled reinforced

concrete frame for seismic analysis. NED Univ J Res
5. El-Dakhakhni WW, Elgaaly M, Hamid AA (2003) Three-strut model for concrete masonry-
infilled steel frames. J Struct Eng 129(2):177–185. https://doi.org/10.1061/(asce)0733-944
5(2003)129:2(177)
6. Tanganelli M, Rotunno T, Viti S (2017) On the modelling of infilled RC frames through strut
models. Cogent Eng 4(1). https://doi.org/10.1080/23311916.2017.1371578
7. Kakaletsis DJ, Karayannis CG (2009) Experimental investigation of infilled reinforced concrete
frames with openings. ACI Struct J (106-S14):132–141
8. SeismoSoft (2020) Seismostruct-a computer program for static and dynamic nonlinear analysis
of framed structures, [Online]. Available: http//www.seismosoft.com
9. Mander JB, Priestley MJN, Park R (1988) Theoretical stress-strain model for confined concrete.
ASCE J Struct Eng 114(8):1804–1826
10. Menegotto M, Pinto PE (1973) Method of analysis for cyclically loaded R.C. plane frames
including changes in geometry and non-elastic behaviour of elements under combined normal
force and bending
11. Sukrawa M (2015) Earthquake response of RC infilled frame with wall openings in low-rise
hotel buildings. Procedia Eng 125:933–939. https://doi.org/10.1016/j.proeng.2015.11.118
12. FEMA356 (2000) Prestandard and commentary for the seismic rehabilitation of buildings. vol
36, no 17
13. Crowley H, Pinho R (2006) Simplified equations for estimating the period of vibration of
existing building. In: First European conference on earthquake engineering and seismology,
no. 1122
14. Smyrou E (2006) Implementation and experimental verification of models for nonlinear
analysis of masonry infilled R.C. frames. In: Dissertation, University of Pavia, Italy, pp 1–108
15. Da Porto F, Guidi G, Dalla Benetta M, Verlato N (2013) Combined in-plane/out-of-plane
experimental behaviour of reinforced and strengthened infill masonry walls. In: 12th Canadian
masonry symposium, pp 1–11
16. Meslem A, D’Ayala D (2012) Toward worldwide guidelines for the development of analytical
vulnerability functions and fragility curves at regional level. In: Proceedings 15th WCEE
17. Decanini LD, Fantin GE (1986) Modelos simplificados de la manposteria incluida en portieos,
Caracteristicas de rigidez y resistencia lateral en estado limite (in Spanish). Jornadas Argentinas
Ing. Estructural 2:817–836
18. Al-chaar G (2002) Evaluating strength and stiffness of unreinforced masonry infill structures.
ERDC/CERL TR-02-1
19. Wang Z, Totoev YZ, Page A, Sher W, Lin K (2015) Numerical simulation of earthquake
response of multi-storey steel frame with SIM infill panels. Adv Struct Eng Mech
20. Totoev YZ, Williamson D, Wang Z (2014) Vibrations of multi-storey RC frame with SIM
panels; numerical simulation. In: Procedings 9th international mason conference
Strengthening of Non-engineered
Building Beam-Column Joint to Increase
Seismic Performance with Variation
of Steel Plate Width
Edy Purwanto, P. Amarta Adri, S. A. Kristiawan, Senot Sangadji,
and S. Halwan Alfisa
Abstract Most of the residential buildings were built without following the require-
ments specified in the design Code of earthquake-resistant building. Such non-
engineering buildings (NEB) are susceptible to seismic hazard. Beam-column joint
failure is a typical example of structural failure in NEB induced by seismic load.
Strengthening of the beam-column joint by mean of external steel plate is expected
to increase the seismic performance of NEB. This study aims to find the most effective
plate width as a strengthening material for the beam-column joint. ATENA software
will be used for the investigation. The structural model is portal structure without
strengthening as a control, namely the PU-200 model and with an external steel plate,
namely PUP-200 with various width of plates. From the results of the analysis, it is
found that strengthening joint with a plate width of 50 mm has better performance
than plate widths of 100 mm and 150 mm. The increase performance is quantified
in term of increase of load by 14.60%, reduction of displacement by 18.43% and
reduction of crack width by 23.65%.
Keywords Strengthening · Steel plate · Non-engineered building
E. Purwanto (B) · P. A. Adri · S. A. Kristiawan · S. Sangadji · S. H. Alfisa

Civil Engineering Department, Universitas Sebelas Maret, Surakarta, Indonesia
P. A. Adri
S. A. Kristiawan
S. Sangadji
S. H. Alfisa
216 E. Purwanto et al.
1 Introduction
Indonesia is located on top of three major tectonic plates in the world, namely the
Eurasian, Australian and Pacific plates which produce frequent earthquakes due to
their relative movement. Several types of hazard may be associated with the earth-
quake, however, ground shaking with high intensity may damage structures like
building and bridges.
Observation from the several past earthquake event exhibit dire situation where
most of non-engineered residential buildings were severely damaged and collapse,
even in the moderate strength of earthquake. This non engineered buildings (NEB),
typically built by and for the middle to lower economic class, were built sponta-
neously and informally without any input or supervision from a qualified engineer.
Often built with poor workmanship whose rely only on past experience, these typical
non-engineered buildings have not met engineering criteria for earthquake resistant
structure [1]. Quite large number of this typical structures demonstrate poor perfor-
mance during earthquake shaking resulted in people having no place to live and many
other secondary losses; such as business interruption.
Complete collapse of structure often occurred as the structure have unable to
withstand additional lateral seismic load, exhibit large enough horizontal deformation
resulting in the detachment of elements in the NEB due to, for example, failure of
beam-column joints, as well as the removal of roof frames and building walls.
Generally, failure of beam-column joint may take one of the following modes: joint
shear failure without beam reinforcement yielding (J), joint shear failure with beam
reinforcement yielding (BJ), beam flexural failure (BF), column flexural failure (CF),
beam reinforcement pull-out failure (P), and beam reinforcement 90 hook anchorage
failure (A) [2].
Several factors involved in the beam-column joint failure including properties of
concrete and reinforcement, dimension of beam and column, ratio of longitudinal
reinforcement, anchorage of reinforcement, the presence of joint shear reinforcement,
spacing of shear reinforcement, etc. [2–6].
The poor performance of beam-column joint in NEB to resist seismic load can
be related to the lacks implementation of engineering principle specifically in the
design and construction of this part. Observation in the field shows these insufficient
engineering principles implementation include the use of low concrete strength, the
use of plain reinforcement instead of deformed bar, no shear reinforcement within the
joint, not enough anchorage development due to insufficient reinforcement detailing,
considerable spacing between shear reinforcement in the column and beam, etc.
It is noted that the degradation of joint due to seismic load will reduce the rigidity
of the joint. Consequently, the load transfer between elements connected by the joint
becomes less effective. Hence, the capacity of structure to resist lateral load decreases
[7, 8].
Improving the seismic performance of existing NEB becomes urgent to reduce
the vulnerability of such building in the earthquake-prone area. Truong et al. investi-
gated various retrofitting technique for enhancing beam-column joint performance.
Strengthening of Non-Engineered Building Beam-Column Joint … 217
They showed head re-bar anchoring, CFRP wrapping, haunch, and steel jacketing
could partially enhance the strength and deformation capacity of the beam–column
joints. Several investigators [9–11] used carbon-fiber-reinforced-plastic (CFRP) for
reinforcing beam-column joints to increase the ability to accept seismic loads so that
it has better load capacity, deformation capacity, and ductility. Meanwhile Thiru-
malini [12] used glass-fiber reinforced-plastics (GFRP) and Rahmi [13] used steel
plates and props to strengthen the beam-column joint in order to increase the seismic
performance of building.
This study proposes the use of steel plates to externally strengthen the beam-
column joints of NEB. The L-shape steel plates were glued and bolted in the external
side of the beam column joint. The main consideration of using this strengthening
material and method is that the installation is easy and the cost is affordable to the
owner of the buildings.
The efficacy of this material and method for such strengthening will be numerically
examined using ATENA 3D v.5 (Advanced Tool for Engineering Nonlinear Analysis)
software. The parameters investigated to quantify performance include load capacity,
crack pattern, ductility and energy dissipation. There are 4 variations of the test object
model, namely PU-200 as the control model and PUP-200 as the reinforcement test
model with plate width variations of 5, 10 and 15 cm.
2 Test Object Model
The test object model to be studied is a simple building construction that represents
non-engineered buildings widely used by the community, especially those that are
prone to earthquake loads. The quality of the concrete, steel and reinforcements is
in accordance with the conditions in the field. Furthermore, a nonlinear analysis
was performed using ATENA software. The concrete used is f’c 17 MPa, the yield
strength of reinforcing steel fy 400 MPa with the reinforcement lay out as in Figs. 1
and 2.
There are two models of test objects, namely PU-200 as a control, where the
test object model has column dimensions 150 × 150 mm with the number of main
Fig. 1 Test object model PU-200 (control)

Fig. 2 Test object model PUP-200 with variations in steel plate width
reinforcement 4-D10 with stirrup D6-200 mm, while the beam has dimensions of 150
× 200 mm with the number of bending reinforcement 4-D10 with stirrup D6-200 mm.
Meanwhile, the second model is retrofitted with L-shape steel plates. This model,
denoted PUP-200 has the same specifications as PU-200 but with the addition of
reinforcing steel plates on the beam-column joint with variations in plate widths
of 50, 100 and 150 mm. PU-200 and PUP-200 modeling can be seen in Figs. 1
and 2. In numerical testing with ATENA 3D v5, the concrete material is modeled
as 3D Nonlinear Cementitious 2, reinforcing steel as bilateral reinforcement, and
reinforcing steel plates as 3D Bilinear Steel Von Mises. The global element size of
the macro elements is 2.5 cm. The supports at both ends of the specimen column are
restrained in the direction of the X, Y and Z axes. Each macro elements has a perfect
connection. The beam-column connection is made as a rigid connection because the
reinforcement of the beam and column is fully joined into the joint.
3 Test Result and Analysis
3.1 Load Displacement Relationship
ATENA software was used to numerically analyze the model with the Newton
Raphson method was used as solution parameters. Load capacity, displacement,
stress, and crack width were obtained from the analysis as shown in Figs. 3, 4 and 5.
In this study, the forces inputted into the load-cases have been determined to be
5 × 10–4 MN or 50 kg to the –X direction at the top of the beam column linkages,
while the point of monitoring for the displacement of the model specimen is at the
other end. The number of load steps for analysis in this study is determined to be
a maximum of 50 steps and will stop when the frame model is not strong enough
to withstand lateral loads or collapses. The relationship between lateral loads and
displacement is shown in Fig. 6.
Figure 3 shows that the PU-200 specimen has the largest cracks in the joint area
that this will trigger a total failure while the PUP-200 specimen with reinforcement
PU-200 PUP-200-050
PUP-200-100 PUP-200-150
Fig. 3 Crack pattern in the test object model PU-200, PUP-200-050, PUP-200-100, PUP-200-150
plates with a width of 50, 100 and 150 mm has the largest cracks in the area of beam,
hence that the collapse pattern will be partial in the beam section only.
Figure 4 shows that the PU-200 steel reinforcement will yield in the column
around the joint, while the PUP-200 test object with reinforcement plates with a
width of 50, 100 and 150 mm, yielding of reinforcement will occur first on the beam.
This can be inferred that the retrofit model demonstrate a better performance as the
brittle failure pattern will be change in more acceptable ductile failure pattern.
Figure 5 shows that the retrofitted model PUP-200 with variations in plate widths
of 50, 100 and 150 mm shows reduction in concrete stress compared to non-retrofitted
model PU-200. This is in agreement with the failure pattern in Fig. 3 and stress
reinforcement in Fig. 4.
The main parameters that can be deduced from Fig. 6 are presented in Table 1.
Based on Table 1, the strengthening of the non-engineered building model with a
plate width of 50 mm (PUP-200-050) will increase by 14.60%, while for plate widths
of 100 mm (PUP-200-100) and 150 mm (PUP-200-150), it increases by 9.91% and
8.05%, respectively. Hence that the reinforcement of a L-shape steel plate with a
width of 5 cm provides the greatest value of increase. 50 mm wide L-shape steel
plate is also the most effective retrofitting material at reducing the displacement of
the test object model, which is 18.43%. Meanwhile the reinforcement of 100 mm and
150 mm wide steel plates has a greater displacement than the model, respectively
PU-200 PUP-200-050
PUP-200-100 PUP-200-150
Fig. 4 Reinforcement stresses in the test object model PU-200, PUP-200-050, PUP-200-100, PUP-
200-150
14.44% and 15.90%. 50 mm wide L-shape steel plate is also the most effective
retrofitting material at reducing the crack width, which was 23.65%, while for the
reinforcement of 100 mm and 150 mm wide steel plates had a crack width that
was greater than the control specimen model, respectively equal to 28.08% and
17.49%. Therefore, 50 mm wide L-shape steel plate provides the best performance
for strengthening the beam column joint.
PU-200 PUP-200-050
PUP-200-100 PUP-200-150
Fig. 5 Concrete stress on the test object model PU-200, PUP-200-050, PUP-200-100, PUP-200-
150
Fig. 6 Graph of the 1200

load–displacement
relationship on the test object 1000
model 800
Load (kg)
PUP 200
600
PUP-200-050
400
PUP-200-100
200
PUP-200-150
0
0 20 40 60 80
Displacement (mm)
Table 1 Recapitulation of test result

Test object model Load Displacement Crack
Maximum Increase Maximum Increase Width Increase
(Kg) % (mm) % (mm) %
PU-200 915.30 – 52.96 – 4.06 –
PUP-200-050 1049.00 14.61 43.20 −18.43 3.10 −23.65
PUP-200-100 1006.00 9.91 60.61 14.44 5.20 28.08
PUP-200-150 989.00 8.05 61.38 15.90 4.77 17.49
Table 2 The ductility value of test object model

Test object model Displacement (mm) Ductility Increase
Yield Maximum %
PU-200 14.61 52.96 3.63 –

PUP-200-050 15.14 43.20 2.85 −21.30
PUP-200-100 11.88 60.61 5.10 40.73
PUP-200-150 11.23 61.38 5.47 50.18
3.2 Ductility
Ductility of the frame can be calculated from Fig. 3 by obtaining the displacement
value at ultimate condition (u) and yield condition (y). Where y is determined
from the displacement at 75% of the peak lateral load. From these data, the ductility
(µ) of the test object model can be calculated using the formula:
u
μ = (1)
y
where µ is ductility, u is displacement at maximum condition and y is

displacement at yield condition.
Table 2 shows that the increase in the ductility value of the steel plate reinforcement
model with a width of 50 mm, 100 mm and 150 mm respectively is −21.30%, 40.73%
and 50.18%, so that the largest ductility value will be achieved in the test specimen
model with reinforcing a steel plate width of 150 mm.
3.3 Energy Dissipation
Based on the graph in Fig. 3, the calculation of the energy dissipation in each model
of the test object is carried out by calculating the area under the graph. The results
of the energy dissipation calculation are presented in Table 3.
Table 3 The energy

Test object model Energy dissipation Increase
dissipation value of the test
object model (kg.mm) %
PU-200 39.647,05 –
PUP-200-050 34.954,45 −11,84
PUP-200-100 51.268,69 29,31
PUP-200-150 51.583,00 30,11
From Table 3 above shows that the energy dissipation in the steel plate reinforce-
ment model with a width of 50 mm, 100 mm and 150 mm is −11.84%, 29.31% and
30.11% respectively. Hence that the greatest energy dissipation will be achieved in
the test object model with the reinforcement of a steel plate width of 150 mm.
Based on the analysis, it can be recapitulated that retrofitting the beam column joint
of the test object with 50 mm L-shaped steel plate demonstrates the most effective
seismic performance while also cost effective as it used the least retrofitting material.
This strengthening technique will dissipate more earthquake energy by performing
more ductile controlled failure pattern and reduce lateral displacement within certain
allowable displacement requirements.
4 Conclusion
From the test results of the PU-200 and PUP 200 test object models with plate
variations, it can be concluded that the increase in the load on the strengthening of
the beam column joint with plate width variations of 50 mm, 100 mm and 150 mm
is respectively 14.60%, 9.91%. and 8.05% so that the strengthening of 50 mm wide
steel plate has the greatest value of increase. For a review of the displacement, the
strengthening of 50 mm wide steel plates will effectively reduce the displacement by
18.43% while for the widths of 100 mm and 150 mm still has a displacement value
that is greater than the test object model before strengthening, respectively 14.44%
and 15.90%. If viewed from the reduction in crack width, the strengthening of a steel
plate with a width of 50 mm will also be able to reduce the crack width by 23.65%,
while for the crack width in the test specimen model with reinforcement of a steel
plate with a width of 100 mm and 150 mm, it is greater than with the test object model
before strengthening, each increased by 28.08% and 17.49%. When viewed from the
ductility and energy dissipation values in the steel plate reinforcement model with a
width of 150 mm, it has a higher value than the strengthening for steel plates with a
width of 100 mm and 50 mm.
In the event of earthquake, the ability to carry additional lateral seismic load
and perform the smallest possible displacement is important factors of structural
performance. It will prevent the collapse of the building. NEB does not have these
characteristics as it has not been designed properly. It has been demonstrated numeri-
cally that strengthening the beam column joint with 50 mm wide L-shaped steel plate
provide the most effective seismic performance by reducing lateral displacement and
increase ductility. Applying this simple strengthening technique will help commu-
nity to more resilient toward earthquake hazard by making these typical vulnerable
non-engineered buildings earthquake resistant.
References
1. Boen T (2001) Earthquake resistant design of non-engineered buildings. In: Indonesia1

Earthquake Resistant Design of Non-Engineered Buildings In Indonesia 1
2. Park S, Mosalam KM (2012) Parameters for shear strength prediction of exterior beam-column
joints without transverse reinforcement. Eng Struct 36:198–209. https://doi.org/10.1016/j.eng
struct.2011.11.017
3. Alaee P, Li B (2017) High-strength concrete exterior beam-column joints with high-yield
strength steel reinforcements. Eng Struct 145:305–321. https://doi.org/10.1016/j.engstruct.
2017.05.024
4. Faleschini F, Hofer L, Zanini MA, dalla Benetta M, Pellegrino C (2017) Experimental behavior
of beam-column joints made with EAF concrete under cyclic loading. Eng Struct 139:81–95.
https://doi.org/10.1016/j.engstruct.2017.02.038
5. Ricci P, De Risi MT, Verderame GM, Manfredi G (2016) Experimental tests of unreinforced
exterior beam-column joints with plain bars. Eng Struct 118:178–194. https://doi.org/10.1016/
j.engstruct.2016.03.033
6. Barbhuiya S, Choudhury AM (2015) A study on the size effect of RC beam-column connections
under cyclic loading. Eng Struct 95:1–7. https://doi.org/10.1016/j.engstruct.2015.03.052
7. Unal M, Burak B (2013) Development and analytical verification of an inelastic reinforced
concrete joint model. Eng Struct 52:284–294. https://doi.org/10.1016/j.engstruct.2013.02.032
8. Shafaei J, Zareian MS, Hosseini A, Marefat MS (2014) Effects of joint flexibility on lateral
response of reinforced concrete frames. Eng Struct 81:412–431. https://doi.org/10.1016/j.eng
struct.2014.09.046
9. Deng Y et al (2021) Experimental and analytical investigation on flexural behaviour of RC
beams strengthened with NSM CFRP prestressed concrete prisms. Compos Struct 257:113385.
https://doi.org/10.1016/j.compstruct.2020.113385
10. Hadi MNS, Tran TM (2014) Retrofitting nonseismically detailed exterior beam–column joints
using concrete covers together with CFRP jacket. Constr Build Mater 63:161–173. https://doi.
org/10.1016/j.conbuildmat.2014.04.019
11. Esmaeeli E et al (2015) Retrofitting of interior RC beam–column joints using CFRP strength-
ened SHCC: cast-in-place solution. Compos Struct 122:456–467. https://doi.org/10.1016/j.
compstruct.2014.12.012
12. Thirumalini S, Shanmugapriya, Atchaya (2021) Experimental and numerical investigation of
short GFRP-reinforced steel tubular short columns. Mater Today Proc. https://doi.org/10.1016/
j.matpr.2020.11.103
13. Rahmi Y, Saputra A, Siswosukarto S (2017) Numerical modelling of interior RC beam-
column joints for non-engineered buildings strengthened using steel plates. In: MATEC Web
Conference, vol 138. https://doi.org/10.1051/matecconf/201713802007
A Proposed Method of FRP Anchorage
for FRP Confined Over-Reinforced
Concrete Beam
Nuroji, Ay Lie Han , Sri Tudjono, Lena Tri Lestari, and Tiara Murtisari
Abstract Fiber Reinforced Polymer (FRP) is a material for strengthening,

retrofitting or confining concrete elements. Research on the FRP performance demon-
strated that debonding was one of the major failure modes, reducing the ultimate
capacity of the composite action between the FRP and concrete. Debonding between
FRP and concrete is mostly characterized in the interfacial transition zone (ITZ).
Increasing the bond performance in the ITZ is conducted by applying an anchorage
system to prevent premature bond loss. The aim of this paper is to propose three
concepts of FRP anchorage systems and evaluate their usefulness and theoret-
ical effectiveness. For evaluation, an over-reinforced flexural member is externally
strengthened using carbon fiber wraps, the compression area of this member is
confined using a u-shape configuration. The three types of FRP anchorages intro-
duced in this study are the spike, insertion, and stitch anchor. It is expected that the
anchors will increase the bond performance between the concrete and the CFRP,
preventing premature failure and increasing the concrete strength and deformation
behavior due to the confinement effect. At further stages, full size elements will be
tested to prove this hypothesis, and to evaluate which type of the three anchors is the
most effective.
Keywords Fiber-reinforced polymer · Anchorage · Confinement
Nuroji (B) · A. L. Han · S. Tudjono · L. T. Lestari · T. Murtisari

Universitas Diponegoro, Semarang 50275, Indonesia
A. L. Han
S. Tudjono
L. T. Lestari
T. Murtisari
226 Nuroji et al.
1 Introduction
Improving and strengthening of existing structural elements in the field are often
needed as a result of additional building loads in the form of live, dead, or earthquake
loads. The use of Fiber Reinforced Polymers (FRP) for the external reinforcement
of structural elements is one method that is being researched extensively, since the
FRP has a very high tensile strength exceeding the tensile strength of commonly
reinforcing steel [1]. Further, the composite action between the FRP and concrete
provides an excellent compatibility behavior in transferring both stress and strain
between the two materials. FRP can be used to increase the shear [2–6], the bending
and the concrete strength due to the confinement effect [7–15] of a concrete element,
leading in an overall enhancement of the member. One of the problems in related
prior studies was that the beam failed before the FRP reached its maximum strength
due to debonding [16]. Previous studies have found that FRP sheets are debonding
by average of 50% of their tensile capacity. Therefore, additional anchorage systems
are required to improve the bond behavior in the interfacial transition zone (ITZ)
[17–22]. Anchorage system for externally bounded FRP aims to: (I) prevent or delay
interfacial crack opening; (II) increase the total available interfacial shear stress
transfer; and (III) to provide a stress mechanism where no bond length is available
beyond the critical section [18, 23]. The effectiveness of CFRP wraps can be improved
by applying multiple layers of wraps [24], but this method will not increase the bond
capacity between the wraps and the concrete.
The effective strain of the FRP reinforcement must be limited by the debonding
strain, Efd, to avoid cracking due to debonding failure (Eq. 1) [25].
ε f d = 0.41 f c n E f t ≤ 0.9ε f u (1)
where f c , n, Ef, tf, and Efu are concrete compression strength, number of FRP
layers, elastic modulus tensile of FRP, thickness of FRP for each layer, and ultimate
strain of FRP, respectively. Several studies have shown that anchorage FRP is viable
method to delay or prevent FRP debonding to concrete [19, 23, 26, 27].
In the research conducted by [3, 16, 28], experimental testing was carried out on
over-reinforced beams that were reinforced more with external FRP reinforcement,
but debonding occurred in the stress area due to the beam bending response followed
by the shear strain between the concrete and FRP. The members were utilized with
u-wraps in the compression zone of the member. U-wraps are commonly used for
shear reinforcement of flexural elements [3, 29–32].
A Proposed Method of FRP Anchorage … 227
2 Concept and Method
2.1 Specimen and Anchorage Details
Figure 1 shows a CFRP wrap-confined beam without anchorage. Three types of FRP
anchorage used in this study are shown in Fig. 2.
Over-reinforced beams are used to obtain a wide compressive area of the beam.
CFRP Wrap is installed in the stressed area of the beam and a U-shape is used to
consider that the FRP cannot be wrapped completely due to the presence of a slab at
the existing structure.
In this study, three types of FRP anchor are introduced, they are spike, insertion,
and stitch. Anchors will improve the strength of the far end fibers of the FRP, and
directly prevent the shear-bond strains to increase [2, 33–37]. The spike anchor is
made of CFRP string which is partially hooked into the concrete cover and the rest is
fan-shaped spread over the surface of CFRP wrap (Fig. 2a). This anchor model uses
the model proposed by researchers and proven highly effective [19, 20, 38, 39] with
slight adjustments to the beam dimensions used in this study. Another method of
anchoring is an insertion (Fig. 2b), the CFRP wrap edge is inserted into the concrete
cover opening and filled with epoxy. The opening groove is curved to avoid the CFRP
wrap being torn at the outside corner of the opening due to a stress concentration. In
the stitch anchor, CFRP string [4, 40] is used which is formed as shown in Fig. 2c.
Fig. 1 CFRP wrap-confined beam without anchor

228 Nuroji et al.
Fig. 2 Variety of FRP anchorage a Spike b Insertion c Stitch
The concrete is perforated using a drill then the anchor stitch is inserted and glued
using epoxy.
2.2 Experimental Test Set-Up
The test will be carried out when the concrete reaches 28 days with a 2-point-load
system (Fig. 3). The two-point loading system is used so that at middle span of
the beam, pure bending occurs without any shear intervention. The FRP is installed
200 mm over the load point to avoid excessive stresses on the ends of the CFRP
wrap. The compression strain gauge is installed horizontally and transversally for
both concrete and FRP to assess the confinement. Strain gauge for steel is only
installed on the tensile reinforcement due to the impossibility to be attached on
small diameter compression reinforcement. In order to obtain beam deflection and
curvature, one LVDT is installed at the middle of the beam span and two LVDT with
a yoke for each other are installed next to it.
Fig. 3 Setup of testing

230 Nuroji et al.
3 Discussion
The three types of FRP anchors are expected to be able to overcome debonding
by providing confinement to the compressed concrete region of the beam. In over-
reinforced beams, failure will occur due to the maximum compressive strain of the
concrete before the reinforcing steel yields. As a result of FRP anchor, U-shaped
FRP can hold the compressive stress after maximum compressive strain of concrete
is reached so that the ultimate load increases until the tensile reinforcement yields.
FRP anchorage has a role to anticipate U-shaped FRP debonding phenomenon which
is caused by shear stress along the interface of U-shaped FRP and concrete. For spike
and stitch anchor, the anchor space/distance is important to make sure the anchors
work fine and do not lead stress concentration around the anchor hook. By comparing
the three types of FRP anchors, it can be seen which types of anchoring are more
effective to use in beam reinforcement using FRP.
References
1. Rasheed HA (2014) Strengthening design of reinforced concrete with FRP

2. Bae S-W, Belarbi A (2013) Behavior of various anchorage systems used for shear strengthening
of concrete structures with externally bonded FRP sheets. J Bridg Eng 18(9):837–847. https://
doi.org/10.1061/(asce)be.1943-5592.0000420
3. Tee HH, Al-Sanjery K, Chiang JCL (2018) Behaviour of over-reinforced concrete beams with
double helix and double square confinements related to ultimate bending and shear strength. J
Phys Sci 29. https://doi.org/10.21315/jps2018.29.s2.7
4. Eslami A, Moghavem A, Shayegh HR, Ronagh HR (2020) Effect of FRP stitching anchors
on ductile performance of shear-deficient RC beams retrofitted using FRP U-wraps. Structures
23:407–414. https://doi.org/10.1016/j.istruc.2019.11.007
5. Lavorato D, Bergami AV, Fiorentino G, Fiore A, Santini S, Nuti C (2018) Experimental tests
on existing RC beams strengthened in flexure and retrofitted for shear by C-FRP in presence of
negative moments. Int J Adv Struct Eng 10(3):211–232. https://doi.org/10.1007/s40091-018-
0193-1
6. Mahrenholtz P, Cho JY, Park JM, Eligehausen R (2018) Characterization of shear strength of
FRP anchors. In: MATEC web conference. vol 199. pp 4–8. https://doi.org/10.1051/matecc
onf/201819909008
7. Chaallal O, Hassan M, Shahawy M (2003) Confinement model for axially loaded short
rectangular columns strengthened with fiber-reinforced polymer wrapping. ACI Struct J
100(2):215–221. https://doi.org/10.14359/12485
8. Fam A, Qie FS, Rizkalla S (2004) Concrete-filled steel tubes subjected to axial compression
and lateral cyclic loads. J Struct Eng 130(4):631–640. https://doi.org/10.1061/(asce)0733-944
5(2004)130:4(631)
9. Madas P, Elnashai AS (1992) A new passive confinement model for the analysis of concrete
structures subjected to cyclic and transient dynamic loading. Earthq Eng Struct Dyn 21(5):409–
431. https://doi.org/10.1002/eqe.4290210503
10. Priastiwi YA, Imran I, Nuroji Hidayat A (2014) Behavior of ductile beam with addition confine-
ment in compression zone. Proc Eng 95:132–138, Elsevier. https://doi.org/10.1016/j.proeng.
2014.12.172
11. Priastiwi YA, Imran (2015) Nuroji: the effect of different shapes of confinement in compression
zone on beam’s ductility subjected to monotonic loading. Proc Eng 125:918–924, Elsevier
https://doi.org/10.1016/j.proeng.2015.11.098
12. Tee HH, Al-Sanjery K, Chiang JCL (2018) Behaviour of over-reinforced concrete beams with
double helix and double square confinements related to ultimate bending and shear strength. J
Phys Sci 29:77–98. https://doi.org/10.21315/jps2018.29.s2.7
13. Nguyen DH, Hong WK (2019) Part I: the analytical model predicting post-yield behavior
of concrete-encased steel beams considering various confinement effects by transverse
reinforcements and steels. Materials (Basel) 12(14). https://doi.org/10.3390/ma12142302
14. Aylie H, Antonius Okiyarta AW (2015) Experimental study of steel-fiber reinforced concrete
beams with confinement. Proc Eng 125, Elsevier. https://doi.org/10.1016/j.proeng.2015.11.158
15. Nematzadeh M, Fazli S (2020) The effect of active and passive confining pressure on compres-
sive behavior of STCC and CFST. Adv Concr Constr 9(2):161–171. https://doi.org/10.12989/
acc.2020.9.2.161
16. Suprapto H, Tudjono S, Susilorini RMR (2020) Study of the role of CFRP shear on increased
bending capacity of reinforced concrete beams. In: Engineering, information and agricultural
technology in the global digital revolution, pp 134–138
17. Aljaafreh T, Beneberu E, Yazdani N (2018) Anchorage effect on flexural fiber reinforced
polymer (FRP) laminate strengthening of lightweight concrete beams. J Eng Archit 6(1):14–25.
https://doi.org/10.15640/jea.v5n2a2
18. Jumaah R, Kalfat R, Al-Mahaidi R, Abdouka K (2017) Anchorage systems used in FRP
strengthening of concrete members. In: High tech concrete: where technology and engineering
meet—proceedings of the 2017 fib symposium, June, pp 877–886. https://doi.org/10.1007/
978-3-319-59471-2_102
19. Villanueva Llauradó P, Ibell T, Fernández Gómez J, González Ramos FJ (2017) Pull-out and
shear-strength models for FRP spike anchors. Compos Part B Eng 116:239–252. https://doi.
org/10.1016/j.compositesb.2017.02.029
20. del Rey Castillo E, Kanitkar R, Smith ST, Griffith MC, Ingham JM (2019) Design approach
for FRP spike anchors in FRP-strengthened RC structures. Compos Struct 214:23–33. https://
doi.org/10.1016/j.compstruct.2019.01.100
21. del Rey Castillo E, Griffith M, Ingham J (2018) Straight FRP anchors exhibiting fiber rupture
failure mode. Compos Struct 207:612–624. https://doi.org/10.1016/j.compstruct.2018.09.073
22. Li W, Liu W, Yang X, Xing F (2020) Experimental study on FRP-to-concrete bonded joints
with FRP sheet anchor system. Adv Mater Sci Eng. https://doi.org/10.1155/2020/2514313
23. Grelle SV, Sneed LH (2013) Review of anchorage systems for externally bonded FRP laminates.
Int J Concr Struct Mater 7(1):17–33. https://doi.org/10.1007/s40069-013-0029-0
24. Khoso AR, Akhund MA, Meghwar SL, Hussain F (2018) Effect of multiple layers of carbon
fiber reinforced polymer on flexural strength of reinforced concrete. 1–11. https://doi.org/10.
20944/preprints201807.0368.v1
25. American Concrete Institute and ACI Committee 440: Guide for the design and construction
of externally bonded FRP systems for strengthening concrete structures (2017). American
Concrete Institute
26. Orton SL, Jirsa JO, Bayrak O (2008) Design considerations of carbon fiber anchors. J Compos
Constr 12(6):608–616. https://doi.org/10.1061/(asce)1090-0268(2008)12:6(608)
27. del Rey Castillo E, Dizhur D, Griffith M, Ingham J (2019) Strengthening RC structures using
FRP spike anchors in combination with EBR systems. Compos Struct 209:668–685. https://
doi.org/10.1016/j.compstruct.2018.10.093
28. Mohamed HA (2018) Improvement in the ductility of over-reinforced NSC and HSC beams by
confining the compression zone. Structures 16:129–136. https://doi.org/10.1016/j.istruc.2018.
09.005
29. Lee J, Lopez MM (2016) Characterization of FRP Uwrap anchors for externally bonded FRP-
reinforced concrete elements: an experimental study. J Compos Constr 20(4). https://doi.org/
10.1061/(asce)cc.1943-5614.0000642
232 Nuroji et al.
30. Kim YJ, Bhiri M (2020) Grid U-wrap anchorage for reinforced concrete beams strength-
ened with carbon fiber-reinforced polymer sheets. ACI Struct J 117(1):3–16. https://doi.org/
10.14359/51716772
31. Huang L, Zhang C, Yan L, Kasal B (2018) Flexural behavior of U-shape FRP profile-RC
composite beams with inner GFRP tube confinement at concrete compression zone. Compos
Struct 184:674–687. https://doi.org/10.1016/j.compstruct.2017.10.029
32. Ali AM, Tarkhan MA (2015) Experimental investigation of confining the compression zone in
over-reinforced beams. Int J Eng Sci Res Technol 4(11):611–617
33. Cortez Flores IA, Fernández Gómez J, Villanueva Llauradó P (2019) Influence of multiple
anchor arrangement in the behaviour of FRP-to-concrete anchored joints. Compos Struct
230:111528. https://doi.org/10.1016/j.compstruct.2019.111528
34. Singh A, del Rey Castillo E, Ingham J (2019) FRP-to-FRP bond characterization and force-
based bond length model. Compos Struct 210:724–734. https://doi.org/10.1016/j.compstruct.
2018.12.005
35. Del Rey Enrique C, Michael G, Jason I (2016) Fibre rupture tensile capacity of FRP straight
anchors
36. Aules W, Saeed YM, Rad FN (2020) A novel anchorage system for strengthening
slender RC columns with externally bonded CFRP composite sheets. Constr Build Mater
245:118423.https://doi.org/10.1016/j.conbuildmat.2020.118423
37. Zhang HW, Smith ST, Kim SJ (2012) Optimisation of carbon and glass FRP anchor design.
Constr Build Mater 32:1–12. https://doi.org/10.1016/j.conbuildmat.2010.11.100
38. del Rey Castillo E, Dizhur D, Griffith M, Ingham J (2019) Experimental testing and design
model for bent FRP anchors exhibiting fiber rupture failure mode. Compos Struct 210:618–627.
https://doi.org/10.1016/j.compstruct.2018.11.091
39. Sun W, Liu S, Zhang C (2020) An effective improvement for enhancing the strength and
feasibility of FRP spike anchors. Compos Struct 247:112449. https://doi.org/10.1016/j.compst
ruct.2020.112449
40. Saeed YM, Aules WA, Rad FN, Raad AM (2020) Tensile behavior of FRP anchors made from
CFRP ropes epoxy-bonded to uncracked concrete for flexural strengthening of RC columns.
Case Stud Constr Mater 13:e00435. https://doi.org/10.1016/j.cscm.2020.e00435
Experimental Investigation on the Shear
Behavior of Patched RC Beams Without
Web Reinforcements: Efficacy
of Patching Position with Respect
to the Shear Span
Adji P. Abrian, Stefanus A. Kristiawan, Halwan A. Saifullah,

P. Muhammad Rafi, R. Muhammad Hafizh, Andreas M. Simanjuntak,
and D. Abel Bismo
Abstract Spalling on concrete covers is a type of degradation that commonly

appeared in reinforced concrete structures. This form of degradation could occur
at the web of shear-span due to, for example, corrosion of reinforcements within the
shear span. The degradation will eventually affect the shear strength of that particular
section, resulting in un-conservative shear capacity. Patch repair may be an option
for restoring the strength of the damaged RC beam. The shear behavior of patched
RC beams without web reinforcements is investigated in this study, with a particular
focus on determining the effectiveness of patching position with respect to the shear
span. Unsaturated Polyester Resin (UPR)-mortar was used in this investigation as
a patch repair material. UPR-mortar was applied to patch RC beam web damage at
0.25 and 0.5 of the shear span (measured from the loading point). The RC patched
beams accompanied by some normal beams were loaded until failure in the labo-
ratory. The results indicate that patched RC beams exhibit a higher shear capacity
than the original undamaged RC beam. Furthermore, UPR-mortar is more effective
when applied at 0.25 of the shear span rather than 0.5 of the shear span. This is
demonstrated by the evidence that after an initial diagonal crack was formed, the RC
beam with patching at 0.25 of the shear span still carries a substantial load before the
diagonal crack propagates, causing the beam to fail in shear. Whereas in RC beam
with patch repair position at 0.5 of the shear span, the RC beam immediately fails
in shear once the significant diagonal crack was formed. However, in NC beams the
shear failure occurred earlier than in patched RC beams.
A. P. Abrian · S. A. Kristiawan · H. A. Saifullah (B) · P. M. Rafi · R. M. Hafizh ·

A. M. Simanjuntak · D. A. Bismo
Department of Civil Engineering, Faculty of Engineering, Universitas Sebelas Maret, Surakarta,
Indonesia
S. A. Kristiawan
A. M. Simanjuntak
234 A. P. Abrian et al.
Keywords Crack · Patch repair · RC beam · Shear · UPR-mortar
1 Introduction
A reduction in the shear capacity of RC beams can be expected if there is damage

to that particular beams. One type of damage often found in RC beam elements is
spalling of concrete covers. Spalling is a type of damage that occurs in the concrete
cover which can be triggered by, for example, corrosion of reinforcements. The
damage raises a concern about the structural component’s safety, especially when it
occurs within the shear span of the beam. If spalling is not repaired, it will decrease
the shear strength of the respected RC beams, which eventually resulting in un-
conservative shear capacity. Shear failure is more dangerous than flexural failure
because it typically occurs unexpectedly [1–4]. According to the literature review,
the flexural strengthening performance of the RC beam has been fully resolved, but
its mechanism is not as complicated as shear [5–7]. For this reason, any attempts to
repair a deficient shear capacity of RC beam are encouraged.
Diagonal cracks must occur in the beam before failure in all cases of shear failure
mode. Several factors that influence the shear failure of RC beam without web rein-
forcements are concrete properties, longitudinal reinforcement properties, longitu-
dinal reinforcement ratio, shear span-to-depth ratio, and beam size [8]. For RC slender
beams with a ratio between shear span a to effective depth d is between 2.5 and 6.5,
it has been determined that three components contribute to the shear transfer mech-
anism, i.e., un-cracked concrete in the flexural compression zone, V cy ; the vertical
component of the aggregate interlock, V ay ; and dowel action, V d [9, 10]. An attempt
has been carried out by Kristiawan et al. [11] to recover the shear capacity of the
damaged RC beam using UPR-mortar. The results show that after an initial diagonal
crack was formed, the patched RC beam can still withstand a greater load before
the beam fails in shear. The shear force capacity is increased by 14% compared to
the normal RC beam. Interestingly, the diagonal cracks did not penetrate the UPR
mortar. As a result, the UPR mortar remains un-cracked when the repaired RC beam
fails in shear [11]. This could be the reason why by using UPR-mortar, the shear
capacity of the patched RC beam can be increased.
The position of patching can affect the efficacy of the repair. This is because the
effective contribution of UPR-mortar to interfering with the propagation of diagonal
cracks is location-dependent. The current study examines the influence of UPR-
mortar repaired positions on the shear performance of RC beam. In this inquiry, two
repaired places are chosen, which are 0.25 and 0.50 of the shear span measured from
the loading point.
Experimental Investigation on the Shear Behavior … 235
2.1 Detail Specimen
Three different types of RC beam specimens with a shear span to effective depth
ratio of 3.0 were prepared and tested. The first beam specimen, NC (see Fig. 1a),
was used as a control beam, which was not patched with UPR-Mortar. While the
remaining two beams were patched RC beams, PR-1 (see Fig. 1b) and PR-II (see
Fig. 1c), with the repair position at 0.25 and 0.50 of the shear span, respectively.
The beams were made of ordinary concrete with a compressive strength of 22.5 MPa
and tensile strength of 2.5 MPa at the loading test’s age (63 days). On the other
(a) NC
(b) PR-I (Patch repair position at 0.25 of the beam shear span)
(c) PR-II (Patch repair position at 0.5 of the beam shear span)
Fig. 1 Details of reinforcement and cross sections (units in mm)

hand, UPR-mortar used had a compressive strength of 67 MPa and tensile strength
of 5.09 Mpa (3 days). A relatively high tensile reinforcement (with a yield strength
of 399 MPa) ratio was used to achieve a bending capacity of at least 40% greater
than the shear capacity ensuring the shear failure mode. The detail of RC beams can
be seen in Fig. 1.
2.2 Testing of the RC Beams
The loading test was carried out at the Structural Engineering Laboratory, Univer-
sitas Sebelas Maret. Figure 2 shows the setup of beam testing. A hydraulic jack with
a capacity of 500 kN was installed on the loading frame for imposing load. The
load cell, which was located in direct contact with the hydraulic jack, was used to
measure the applied load. During the loading test, the RC beam deflection and crack
development were monitored. A linear vertical displacement transducer (LVDT) was
positioned at midspan to measure the maximum deflection of RC beam. Demount-
able mechanical strain gauges (DEMEC) with a gauge length of 200 mm were also
attached to the RC beam’s surface to track the diagonal crack development. Addi-
tionally, strain gauges were installed on the tension reinforcement at a location that
Fig. 2 The setup of beam testing

coincides with the beam’s midspan. The development of crack was marked at each
load increment of 5 kN.
3 Behavior of Test Specimens
3.1 Failure Mode
For all RC beams, the maximum longitudinal reinforcement strains at the ultimate
load (923 µ at the maximum recorded in the NC beam) did not exceed the reinforce-
ment bar’s yield strain of approximately 1995 µ. This indicated that shear failure
modes occurred without longitudinal reinforcement yielding. The failure mode of
all beam specimens is typical of diagonal tension (see Figs. 3, 4 and 5). After the
initial diagonal cracking formation, which indicated the beginning of shear cracks,
P = 30 kN
(a) NC, initial flexural cracking

P = 86 kN P = 86 kN (max)
(b) NC, significant diagonal cracking (c) NC, ultimate load
Fig. 3 Crack patterns of the NC beam
P = 30 kN
(a) PR-I, initial flexural cracking

P = 110 kN P = 120.4 kN (max)
(b) R-I, significant diagonal cracking (c) PR-I, ultimate load
Fig. 4 Crack patterns of the PR-I beam

P = 35 kN
(a) PR-II, initial flexural cracking

P = 95 kN P = 101 kN (max)
(b) PR-II, significant diagonal cracking (c) PR-II, ultimate load
Fig. 5 Crack patterns of the PR-II beam
the patched RC beams resisted more shear force accompanied by further develop-
ment of shear cracks. The diagonal crack penetrating the compression zone near the
loading points causes shear failure.
3.2 Crack Pattern
Figures 3, 4 and 5 shows the crack patterns exhibited in the beam specimens. Gener-
ally, initial flexural cracks were located around the mid-span where the flexural stress
at maximum. Flexural cracks propagated toward the loading point as the applied load
increased. When the load was sufficient to induce diagonal tension stress greater than
the concrete’s tensile strength, diagonal cracks will appear. When the applied load
was increased until a certain level, more flexural cracks were formed, and the previous
flexural cracks extend in length and width. Once the significant diagonal cracks had
formed, a further increase in load would cause the beams would fail immediately.
In the NC beam case, a continuous propagation of the diagonal crack is observed
from the lower surface of the concrete toward the loading point. Distinct phenomena
were observed in the PR-I dan PR-II cases. The presence of UPR-mortar could resist
the diagonal crack propagation. Obviously, this is due to the high tensile strength
of the UPR-mortar. The resistance against the propagation of diagonal crack by the
patching material causing an improvement in the shear resistance of patched RC
beams. The patching location also affects the extent of the propagation of the crack
pattern. The crack pattern on the PR-I beam with patching location at 0.25 of the
shear span gives a more significant effect in resisting the diagonal crack than the
patching location at 0.5 of the shear span. This may be due to the contribution of
UPR-mortar to preserve the compression zone at a high load level when this repair
material close to the loading point (the PR-I beam). The intensity of cracks at this
compression zone of PR-I prior to shear failure indicates the above argument (see
Fig. 4).
3.3 Load–Deflection Response
Figure 6 shows the relationship between the load and midspan deflection. The applied
load-midspan deflection curves could only be recorded up to the peak load under the
load control system’s implementation. The beam’s flexural cracking capacity was
exceeded at a load of approximately 30–35 kN, resulting in the stiffness reduction of
the RC beam. The amount of stiffness reduction appears to be different for the three
types of beam, with the NC beam having the most significant decrease. It indicates
the influence of patching repair becomes more pronounced after the first flexural
cracking occurrence.
The parameters describing the characteristic of load–deflection curves are summa-
rized in Table 1. The ultimate load for the NC, PR-I, and PR-II beams were 86, 120.4
and 101 kN, respectively. Although the corresponding deflections had almost the
same value, the patched beams exhibited a higher ultimate load (40 and 17% for
130
120 Ultimate load
110 19%
100 40%
90 17%
Load (kN)
80
70
60
50
40
30 NC
20 PR I
10 flexural cracking load
PR II
0
0 1 2 3 4 5 6 7 8
Midspan deflection (mm)
Fig. 6 Load and mispan deflection of RC beams
Table 1 Flexural cracking load, ultimate load, and deflection at the ultimate load of RC beams
Specimen Flexural cracking load (kN) Ultimate load Deflection at the ultimate load
(kN) (mm)
NC 30 86 5.75
PR-I 30 120.4 5.69
PR-II 35 101 5.84
PR-I and PR-II, respectively) than the normal beam. The efficacy of patching posi-
tion with respect to the shear span is also apparent. The PR-I beam with a patching
position at 0.25 of the shear span shows the largest shear capacity. The effectiveness
of UPR-mortar to reduce the intensity of cracks at similar load levels (see Figs. 3,
4 and 5) could be the reason for such behavior. This phenomenon has also been
identified in the flexural behavior of patched RC beams [12].
3.4 Diagonal Crack
The development of diagonal cracks was measured by the DEMEC gauge every load
increment of 5 kN. Figure 7 shows the relationship between the load and diagonal
crack width. In the present study, a significant diagonal cracking load was identified
by a sudden increase in the diagonal crack’s width. All RC beams indicated almost
the same significant inclined crack width. However, the patched beams exceeded the
NC beam in terms of significant diagonal cracking load, ultimate load, and maximum
diagonal crack width, as shown in Table 2. Additionally, the patching position affects
the resistance of RC beams to the applied loads. This is demonstrated by the more
130
120
110 19% 40%
100
17%
90
Load (kN)
80
70
60 Significant diagonal cracking load
Ultimate load
50
40
30 NC
20 PR I
10 PR II
0
0 1 2 3 4 5 6 7
Crack width (mm)
Fig. 7 Beam specimens diagonal crack width
Table 2 Significant diagonal cracking load, ultimate load, and maximum diagonal crack width of
RC beams
Specimen Significant diagonal cracking load Ultimate Maximum diagonal crack width
(kN) load (kN) (mm)
NC 80 86 4.1
PR-I 115.9 120.4 5.4
PR-II 95 101 4.8
excellent shear resistance and the higher corresponding maximum diagonal crack
width in the PR-I case compared to the PR-II.
The presence of patching can influence the shear transfer mechanism. The high
tensile strength of the patch repair material effectively prevents the diagonal crack
from propagating. Consequently, the beam holds a sufficient compression zone to
provide a considerable V cy . Furthermore, in the diagonally cracked beam, the un-
cracked UPR-mortar provides slip resistance, promoting better aggregates interlocks
on the surface of diagonal crack and improving V ay .
4 Conclusion
The failure mode of all beam specimens is typical of diagonal tension. UPR-mortar
can be applied to repair the damaged RC beam and increase its shear capacity. The
un-cracked feature of UPR-mortar affects the shear transfer mechanism by arresting
diagonal crack propagation and providing slip resistance. The position of UPR-
mortar with respect to the shear span influences the efficacy of the repair material to
recover the shear resistance. The closer the UPR-mortar to the loading point will be
more effective in preserving the compression zone, thus increasing the shear strength.
Acknowledgements This research has been funded by Sebelas Maret University through PNBP-
UNS Grant No. 452/UN27.21/PN/2020. The specimens were built and tested at the Structural
Laboratory of the Department of Civil Engineering, Sebelas Maret University, with the help of
the laboratory technicians and undergraduate students (Atep Nuryana, Alfian Muhammad, Fendy
Prayugo, Inggrid Monalita, Ummi Sarah, and Elisabeth Nathasya).
References
1. Joint ACI-ASCE Committee 445 (2009) Recent approaches to shear design of structural
concrete : ACI manual of concrete practice (ACI 445R-99)
2. Jayaprakash J, Abdul Samad AA, Anvar Abbasovich A, Abang Ali AA (2008) Shear capacity
of precracked and non-precracked reinforced concrete shear beams with externally bonded
bi-directional CFRP strips. Constr Build Mater 22(6):1148–1165
3. Ali N, Samad AAA, Mohamad N, Jayaprakash J (2015) Shear behavior of pre-cracked
continuous beam repaired using externally bonded CFRP strips. Proc Eng 53:129–144
4. Khalifa A, Nanni A (2000) Improving shear capacity of existing RC T- section beams using
CFRP composites. Cement Concr Compos J 22(2):165–174
5. Saadatmanesh H, Ehsani RM (1991) RC beams strengthened with GFRP plates I: experimental
study. Struct Eng J 117(11):3417–3433
6. GangaRao HVS, Vijay PV (1998) Bending behavior of concrete beams wrapped with carbon
fabrics. Struct Eng J 124(1):3–10
7. El-Mihilmy Mahmoud T, Joseph WT (2000) Analysis of reinforced concrete beams strength-
ened with FRP laminates. Struct Eng J 126(6):684–691
8. Thamrin R, Tanjung J, Aryanti R, Nur OF, Devinus A (2016) Shear strength of reinforced
concrete T-beams without stirrups. J Eng Sci Technol 11(4):548–562
9. Hassoun MN, Al-Manaseer A (2015) In: Structural concrete: theory and desain. 6th edn.
10. Wright JK, MacGregor JG (2011) In: Reinforced concrete: mechanics and design. 6th edn.
Prentice Hall
11. Kristiawan SA, Agus S, Senot S, Hapsara BW (2017) Shear failure of patched reinforced
concrete beam without web reinforcements. Key Eng Mater 737 KEM:441–447
12. Kristiawan SA, Supriyadi A, Sangadji S, Santosa D (2017) Cracking behavior and its effect on
the deflection of patched-reinforced concrete beam under flexural loading. In: MATEC web of
conferences. vol 138
Secondary AE Analysis of Pre-corroded
Concrete Beam
Ahmad Zaki and Zainah Ibrahim
Abstract Acoustic emission technique of non-destructive testing (NDT) method is

the more successful method of corrosion assessment of reinforced concrete (RC)
structures for the evaluation of pre-corroded concrete beam specimens under load
testing. The tests were used for small scales of pre-corroded concrete beams subject to
cyclic loading before the first visible cracks for secondary AE source. The secondary
AE source caused by the cyclic load at the maximum load level is half of the estimated
first visible cracks (i.e., 4 kN). To study the impact of significant damage due to
corrosion before the first visible cracking occurs, the AE parameters included the
cumulative signal strength (CSS) and Ib-value are proposed. The results showed that
the high CSS ratio (more than 69%) is related to the damage growth of the beam
specimens. The Ib-value tends to decrease from loading cycle 2 to cycle 3, indicates
that the occurrence of damage.
Keywords AE · NDT · Corrosion · Assessment · Secondary AE
1 Introduction
Steel corrosion of reinforced concrete (RC) structures is a global problem that causes
the deterioration of RC structures. Corrosion damage typically needs sufficient
repairs and maintenance. Moreover, the Non-destructive testing (NDT) approach
is practically a necessity in this study to rehabilitate the corroded RC structures
efficiently and cost-effectively. Over the past few years, the acoustic emission (AE)
A. Zaki (B)
Department of Civil Engineering, Faculty of Engineering, Universitas Muhammadiyah
Yogyakarta, Bantul 55183, Daerah Istimewa Yogyakarta, Indonesia
Z. Ibrahim
Department of Civil Engineering, Faculty of Engineering, University of Malaya,
50603 Kuala Lumpur, Malaysia
244 A. Zaki and Z. Ibrahim
technique more useful techniques for structural health monitoring and damage assess-
ment of the corroded RC structures [1–7]. The AE technique can be operated without
disturbing any progressions associated with the RC structures [8–11]. The conver-
sion of transient elastic waves (ultrasonic frequency range) from a confined source
within a material into electrical signals via connected piezoelectric sensors is referred
to as AE technique [12–15]. The sources of AE are deformation processes such as
crack growth, corrosion, and other material degradation. Localised energy release
gives rise to elastic waves that are detected by sensors placed on the concrete surface
[16, 17].
The AE technique successfully analyses the energy signals produced by the
concrete cracking induced by the corrosion of steel reinforcement by the AE sensors
mounted on the concrete surface. The AE technique has been widely used for corro-
sion, especially the monitoring of corroded RC structure. AE parameters, including
AE hits, energy, amplitude, signal strength, etc. were successfully used to describe
the corrosion processes for reinforcement of RC structures [5, 6, 18–24]. The AE
technique enables damage localization of steel corrosion and for detecting the micro-
cracking and the macro-cracking [21, 25–33]. However, there is a limited study of
the impact of the significant damage due to steel corrosion before visible cracking
occurs. In this study, the AE technique is the proposed technique to monitor the
pre-corroded concrete specimens subjected to cyclic loading before the first visible
cracking occurs (at early stage). Early warning could be given to allow for repair
work before the concrete structure is seriously damaged. The CSS and Ib-value of AE
data have been carried out to characterize the corrosion damage of the pre-corroded
specimens.
2 Materials and Methods
2.1 Fabrication of Concrete Specimens
The deformed steel reinforcement of sizes 12 mm in compliance to BS 4449 [34]

was used as tension reinforcement. The pre-corroded specimens were prepared using
corroded steel reinforcement before concrete casting The steel reinforcements were
corroded before concrete casting. The steel reinforcements were immersed in a 5%
sodium chloride (NaCl) solution in a plastic container. A direct current (DC) power
supply was used as the electrolysis. As the anode, the copper wire connected the steel
reinforcement to the positive terminal—meanwhile, the steel bar connected to the
negative terminal as the cathode. Therefore, after the corrosion process, to remove
the corrosion product (rust) of the steel reinforcements were cleaned in accordance
to ASTM G1-03 [35]. The corroded steel reinforcement weighed to assess the mass
loss. Three pre-corroded specimens with a dimension of 100 × 100 × 500 mm3 were
fabricated. The details of the pre-corroded specimen are shown in Fig. 1.
Secondary AE Analysis of Pre-Corroded Concrete Beam 245
Fig. 1 The details of the beam specimen (in mm)
The beam specimens with 28 MPa of compressive strength were designed in

accordance to BS 8110–1 [36]. The reinforced concrete properties are 2192.91 kg/m3
of dry density, 392,95 kg/m3 of cement content, 0,5429 of water to cement ratio,
67 mm of slump value, and 200.000 MPa of modulus elasticity of reinforcement.
The specimens with steel reinforcements of 0% (S0), 9.06% (S9), and 27.56% (S28)
of corrosion levels were prepared for pre-corroded beam specimens. The corrosion
level (%) of the beam specimen could be measured as the mass loss of the steel
reinforcement before and after accelerating corrosion [37, 38]. The bottom cover
was 26 mm from the concrete surface. Ordinary Portland cement (OPC), sand from
the river, and 10 mm of crushed type granite were used to prepare the specimens.
2.2 Data Acquisition
The specimens were subjected to three-point load testing, as shown in Fig. 2. The
flexural load testing using 100 kN of Universal Testing Machine (INSTRON Satec
Series). Each specimen was supported at both ends with distance between each
support to the loading point was 200 mm. The data acquisitions of the specimens
were conducted using the AE technique throughout the cyclic load testing. A PCI-
2 AE System of MITRAS Group, Inc. was connected to the six AE sensors. The
sensors were mounted using wax to the specimen surface as the coupling agent.
After testing the specimens with the AE technique, the AE data will be analyzed for
AE parameters analysis.
Fig. 2 Data acquisitions

using INSTRON Satec series
of 100 kN and AE technique
Fig. 3 Cyclic load testing for secondary AE
The source types of AE monitoring of the specimens can be classified as primary

AE and secondary AE [39]. The primary AE is related to crack growth and fracture
mechanic parameters, while in secondary AE comprises all other sources, including
the micro-fracture. The secondary AE sources induced by the cyclic load testing with
a maximum load level is half of the estimated first visible cracks, that is 4 kN, as
shown in Fig. 3. The loading cycle load is intended to study the effect of significant
damage before the first visible cracking occurs. Furthermore, the number of cycles
was chosen based on many references, such as Paul and Adam [40].
Table 1 presents a summary of the secondary AE parameters of the pre-corroded

beam specimens along with the three different cycles (i.e., Cycles 1, 2, and 3) of
cyclic load testing. The loading rate of secondary AE is 0.1 mm/min. The secondary
AE parameters are duration, amplitude, average frequency (AF), hits, and counts.
3.1 Cumulative Signal Strength (CSS)
Figure 4 shows the curve of signal strength and load with time for all the three beam
specimens under three cycles. Figure 5 shows the curve of CSS with time for all the
three beam specimens. In general, the CSS rate decreases from cycle 1 to cycle 3.
It means that the AE activities decrease due to the limitation of AE activities during
uncracked beam as the beam specimen was subjected up to 4 kN load, at which it is
50% of the theoretical first visible flexural cracking load at middle span of the beam
Table 1 Secondary AE parameters

Beam Cycle Duration (× Amplitude AF Hits % Of total Counts
specimens 1000 µs) (kHz) hits
S0 1 23.572 29.58 14.08 8992 52.87 2,986,932
2 24.848 23.23 7.26 5148 30.27 986,366
3 24.023 22.61 6.77 2868 16.86 532,277
S9 1 24.033 34.28 19.25 14,667 61.59 8,640,423
2 16.408 23.27 5.55 5595 23.50 531,306
3 14.821 23.70 5.91 3551 14.91 311,610
S28 1 38.696 29.49 22.50 4837 35.94 6,012,840
2 33.668 23.54 11.23 4182 31.07 1,879,292
3 30.228 22.71 9.09 4440 32.99 1,440,196
Note Duration: Time between signal start and signal end, Amplitude: Largest voltage peak in the
waveforms, Average Frequency (AF): Nature of source event, Hits: Detection of an AE signal,
Counts: Number of times AE signal crosses threshold
specimen. Therefore, to evaluate the damage of beam specimens, the CSS ratio is
proposed.
A study done by Paul and Adam [33] proposed an evaluation criterion based on
the cumulative signal strength (CSS) ratio of RC beams that were strengthened with
carbon fiber-reinforced polymer (CFRP). It is defined as the peak CSS at the end of
the reload hold period divided by the peak CSS at the end of the initial load hold
period. The study suggested that a peak CSS ratio between 30 and 50% was an
appropriate value to select as an indication of significant damage. Table 2 shows the
CSS ratio for the specimens. The suggested ratio of CSS ratio based on the study
done by Paul and Adam [33] is not suitable in this case due to the CSS ratios obtained
for the specimens that are not absolutely within the range of 30–50% for significant
damage due to steel corrosion. It can be attributed to the sensitivity of individual
sensors, attachment methods of sensors, concrete material, and steel corrosion of
pre-corroded specimens of this study that are different from the study by Paul and
Adam [33].
However, from the results obtained, by comparing with the S0 beam specimen,
it can be deduced that S9 and S28 beam specimens with higher corrosion levels
experienced very significant damage. Due to the specimen having a much higher
CSS ratio compared to the S0 beam specimen, especially for the CSS ratio obtained
using cycle 2 and 3 values, which are more than 60%. This result is supported by
the findings from the mechanical loading test, where these specimens failed at the
lowest peak load for S28 beam specimens. On the other hand, the result for S9 beam
specimens, their first CSS ratios were the lowest with less than 18%, but their second
CSS ratio was very significant, with a percentage of more than 69%. The high value
of CSS ratio is related to the damage growth of the specimens [33]. The lowest CSS
ratio might imply the steel corrosion of the specimen does not affect the damage
growth of the beam specimens at the first CSS ratio.
(a) S0 beam specimen
(b) S9 beam specimen
Fig. 4 Signal strength (pVs) and load versus time (s)

(c) S28 beam specimen
Fig. 4 (continued)
3.2 Ib-Value
Figure 6 shows the Ib-value for the specimens, i.e., S0, S9, and S28 beam specimens.
The beam specimens experienced steel corrosion has a trend different from S0 beam
specimen. For the S0 specimen, the trend established that the Ib-value tends to
increase from cycle 1 to cycle 3 of loading. However, for S9 and S28 beam specimens,
the Ib-value tends to decrease from cycle 2 to cycle 3 of loading in terms of increasing
in corrosion level. This result signifies the occurrence of damage from cycle 2 to
cycle 3 of loading. It indicates that the occurrence of damage is generated due to
steel corrosion (loss in the cross-section) of the beam specimens. The cross-section
of the S28 beam specimen is lower than S9 beam specimen.
Fig. 5 Cumulative signal strength (pVs) against time (s) of the beam specimens
Fig. 5 (continued)
Table 2 CSS ratio

Beam specimens Cycle CSS CSS ratio 1 (%) CSS ratio 2 (%)
S0 Cycle 1 5.75 × 108
Cycle 2 38.09
Cycle 3 1.13 × 108
S9 Cycle 1 69.96
Cycle 2 1.04 × 108
Cycle 3 7.25 × 107
S28 Cycle 1 5.43 × 108 47.33 89.49
Cycle 2 2.57 × 108
Cycle 3 2.30 × 108
Fig. 6 Ib-value against time graph of the beam specimens

Fig. 6 (continued)
4 Conclusion
This experiment deliberates the mechanical behaviour of the RC beam specimens

with pre-corroded steel reinforcement, together with the AE monitoring results for
secondary AE signals analysis. The results obtained in this experiment showed the
great potential of the AE technique to be used in the corrosion evaluation of the
beam specimens. The AE parameters used in this experiment provide some useful
information and trend that could be illustrated on the corrosion level experienced
by the corrosion beam specimens, as discussed in this experiment. However, these
findings do not apply to a real corroded structure that is subjected to water and
chloride attack. In addition, there is only one factor contributing to the loss of concrete
strength, i.e., flexural strength, shear strength, load-bearing capacity etc., which is
the decrement in the cross-section of the steel reinforcement of the pre-corrosion
beam specimens. Thereafter, based on the limitation gained in the experiment, the
next experiment was utilized for post-corrosion beam specimens.
References
1. Malhotra VM, Carino NJ (1991) Handbook on nondestructive testing of concrete. CRC Press,
Florida
2. Ohtsu M, Yuyama S (2001) Recommended practice for in-situ monitoring of concrete structures
by acoustic emission. J Acoust Emiss 184–190
3. Carpinteri A, Lacidogna G, Niccolini G (2011) Damage analysis of reinforced concrete
buildings by the acoustic emission technique. Struct Control Health Monit 660–673
4. Calabrese L, Campanella G, Proverbio E (2013) Identification of corrosion mechanisms by
univariate and multivariate statistical analysis during long term acoustic emission monitoring
on a pre-stressed concrete beam. Corros Sci 161–171
5. Zaki A, Tan JY, Chai HK, Aggelis DG (2015) Assessment of corrosion damage using acoustic
emission technique under load testing. In: The 7th Asia Pacific young researchers and graduates
symposium, University of Malaya, Kuala Lumpur
6. Zaki A, Chai HK, Behnia A, Aggelis DG, Tan JY, Ibrahim Z (2017) Monitoring fracture of steel
corroded reinforced concrete members under flexure by acoustic emission technique. Constr
Build Mater 609–618
7. Zaki A, Chai HK, Aggelis DG, Alver N (2015) Non-destructive evaluation for corrosion
monitoring in concrete: a review and capability of acoustic emission technique. Sensors
19069–19101
8. Ohtsu M (1987) Mathematical theory of acustic emission and moment tensor solution. J Soc
Mater Sci 1025–1031
9. Colombo S, Main IG, Forde MC (2003) Assessing damage of reinforced concrete beam using
“b-value” analysis of acoustic emission signals. J Mater Civ Eng 280–286
10. Grosse CU, Reinhardt HW, Finck F (2003) Signal-based acoustic emission techniques in civil
engineering. J Mater Civ Eng 274–279
11. Ohno K, Ohtsu M (2010) Crack classification in concrete based on acoustic emission. Constr
12. Beattie AG (1983) Acoustic emission, principles and instrumentation. J Acoust Emission 34
13. Shiotani T (2006) Evaluation of long-term stability for rock slope by means of acoustic emission
technique. NDT & E International 217–228
14. Alvarez MG, Lapitz P, Ruzzante J (2012) Analysis of acoustic emission signals generated from
SCC propagation. Corrosion Sci 5–9
15. Lapitz P, Ruzzante J, Alvarez MG (2007) AE response of A-brass during stress corrosion crack
propagation. Corros Sci 3812–3825
16. Wells D (1970) An acoustic apparatus to record emissions from concrete under strain. Nuclear
Eng Design 9
17. Luo X, Haya H, Inaba T, Shiotani T, YN.: Experimental study on evaluation of breakage in
foundations using train-induced acoustic emission. In: Structural engineering world congress,
T9–1
18. Yoon DJ, Weiss WJ, Shah SP (2000) Assessing damage in corroded reinforced concrete using
acoustic emission. J Eng Mech 273–283
19. Ohtsu M, Tomoda Y (2007) Acoustic emission techniques for rebar corrosion in reinforced
concrete. In: Advances in construction materials, Springer, Switzerland, pp 615–621
20. Kawasaki Y, Tomoda Y, Ohtsu M (2010) AE monitoring of corrosion process in cyclic wet–dry
test. Constr Build Mater 2353–2357
21. Ohtsu M, Tomoda Y (2007) Corrosion process in reinforced concrete Identified by acoustic
emission. Mater Trans 1184–1189
22. Ohtsu M, Tomoda Y (2008) Phenomenological model of corrosion process in reinforced
concrete identified by acoustic emission. ACI Mater J 194–199
23. Kawasaki Y, Wasada S, Okamoto T, Izuno K (2014) Evaluation for RC specimen damaged
from rebar corrosion by acoustic emission technique. Constr Build Mater 157–164
24. Zaki A, Ibrahim Z (2021) Corrosion assessment of pre-corrosion concrete specimens using
acoustic emission technique. J Eng Technol Sci 210211
25. Leelalerkiet V, Shimizu T, Tomoda Y, Ohtsu M (2005) Estimation of corrosion in reinforced
concrete by electrochemical techniques and acoustic emission. J Adv Concr Tech 137–147
26. Aggelis DG, Soulioti DV, Sapouridis N, Barkoula NM, Paipetis AS, Matikas TE (2011)
Acoustic emission characterization of the fracture process in fibre reinforced concrete. Constr
27. Behnia A, Chai HK, Yorikawa M, Momoki S, Terazawa M, Shiotani T (2014) Integrated non-
destructive assessment of concrete structures under flexure by acoustic emission and travel
time tomography. Constr Build Mater 202–215
28. Kawasaki Y, Wakuda T, Kobarai T, Ohtsu M (2013) Corrosion mechanisms in reinforced
concrete by acoustic emission. Constr Build Mater 1240–1247
29. Behnia A, Chai HK, Shiotani T (2014) Advanced structural health monitoring of concrete
structures with the aid of acoustic emission. Constr Build Mater 282–302
30. ElBatanouny MK, Ziehl PH, Larosche A, Mangual J, Matta F, Nanni A (2014) Acoustic
emission monitoring for assessment of prestressed concrete beams. Constr Build Mater 46–53
31. Behnia A, Chai HK, GhasemiGol M, Sepehrinezhad A, Mousa AA (2019) Advanced damage
detection technique by integration of unsupervised clustering into acoustic emission. Eng Fract
Mech, 212–227
32. Guzman C, Torres D, Hucailuk C, Filipussi D (2013) Analysis of the acoustic emission in a
reinforced concrete beam using a four points bending test. In: International congress of science
and technology of metallurgy and materials. Elsevier Procedia, Amsterdam, pp 148–154
33. Aggelis DG, Tsangouri E, Van Hemelrijck D (2015) Influence of propagation distance on
cracking and debonding acoustic emissions in externally reinforced concrete beams. Meccanica
1167–1175
34. BS (2005) BS 4449: Specification Steel for the reinforcement of concrete. Weldable reinforcing
steel. Bar, coil and decoiled product. British Standards Institution, London
35. ASTM (2003) ASTM G1–03: Standard practice for preparing, cleaning, and evaluating
corrosion test specimens, ASTM International, West Conshohocken, PA
36. BS (1997) BS 8110–1: Structural use of concrete. Code of practice for design and construction,
British Standards Institution, London
37. Apostolopoulos CA, Michalopoulos D (2007) Impact of corrosion on mass loss, fatigue and
hardness of BSt500s steel. J Mater Eng Perform 63–67
38. Majhi S, Mukherjee A, George NV, Uy B (2019) Corrosion detection in steel bar: a time-
frequency approach. NDT E Int 102150
39. Holford KM (2000) Acoustic emission–basic principles and future directions. Strain 51–54
40. Paul Z, Adam R (2006) Evaluation of strengthened RC beams: cyclic load test and acoustic
emission methods. ACI Struct J 832–841
Cable Force Prediction Technique Using
Subspace and Effective Vibration Length
Method
Muhammad Ibnu Syamsi , Hao-Lin Wang, and Chung-Yue Wang
Abstract Inevitably, cable structure is often used as a structural part of either a

middle or long-span bridge, so its existing tension force is essential to be eval-
uated and monitored periodically. This research aims to demonstrate the use of
covariance type of Stochastic Subspace Identification (SSI-COV) to identify the
dynamical system properties of bridge’s cable and estimate the cable tension.
The concept of effective vibration length proposed by previous researchers is
implemented in this research to overcome the unknown rotational stiffness of the
cable’s supporting system. During the measurement, multiple-synchronized vibra-
tion measuring devices are deployed on a cable to get the corresponding eigenvector
and its effective length. The cable force of an existing bridge is identified using two
different approaches, the least-square approximation and the two-frequency method.
It is found that the deviation between those two mentioned approaches is very small.
The overall process evidence that this non-destructive testing procedure is practicable
to estimate the internal force of a cable in real structures.
Keywords Cable force · Subspace method · Effective vibration length
1 Introduction
Assessing the existing condition of infrastructure is a need to maintain its perfor-

mance and operational condition. Non-destructive testing (NDT) is the most widely
used technique to be used and developed among some measurement methods. As a
part of NDT, the vibration-based approach has been successfully applied in many
fields to detect the dynamic tension force of a stayed cable of structure [1–4]. Conse-
quently, a signal processing technique is required to extract the dynamical properties
M. I. Syamsi (B)
Universitas Muhammadiyah Yogyakarta, Bantul 55183, Indonesia
M. I. Syamsi · H.-L. Wang · C.-Y. Wang
National Central University, Taoyuan 32001, Taiwan
258 M. I. Syamsi et al.
of a system. Operational Modal Analysis (OMA) is the most efficient method to

reveal the system properties because of its advantage, which does not require any
knowledge of exiting input. Compared to the traditional vibration test, this technique
is more feasible for a large structure. One of the OMA methods developed in the
time-domain approach is Stochastic Subspace Identification (SSI).
SSI method starts its popularity since a paper published by Van Overschee and De
Moor [5] along with the book [6]. A well-reviewed paper reported by Brincker and
Andersen [7] helps engineers better understand the difficulties of the theory behind
this technique. Besides, SSI has been proved to be the most powerful output-only
modal analysis technique in many applications, including civil engineering [8, 9].
Many researchers have also performed this technique in a bridge’s structure. Maga-
lhaes, Cunha [10] proposed an automated subspace identification by combining this
method with machine learning and applied it in a long span arch bridge. Further
application of the previous researcher’s automated subspace technique is demon-
strated by Ubertini, Gentile [11]. Moreover, the comparison with robust OMA in the
frequency domain, Frequency Domain Decomposition (FDD), shows a promising
result.
Meanwhile, many researchers have developed the vibration-based approach of a
cable to assess the tension. The most traditional approach uses the string theory, where
the relationship is expressed in terms of natural frequency and its corresponding mode
order. However, the application of this theory is very limited and mostly unsuitable
to be implemented in the real cable structure. Thus, the modern cable formula is
developed from the concept of axially-loaded Euler–Bernoulli beam theory [12].
Some researchers reported the application of the formula for long-term cable force
monitoring [13, 14]. Since the cable’s end support system in the real case lies between
hinged and clamped, then the application of the formula became challenging. The
concept of effective vibration length is then proposed to overcome the uncertainty
end-restrain of cable and demonstrated in the real stayed cable of bridges [2, 15–17].
Besides, an alternative determination for cable force even with the absence of flexural
rigidity using a combination of two frequencies is introduced by Cheng [3].
It is known that vibration-based is one of the non-destructive testing (NDT)
methods requiring signal processing knowledge. Studies in taking advantage of the
vibration-based approach with the robust SSI technique to measure and monitor the
bridge cables have also been done by previous researchers [1, 18]. However, the
covariance-driven type of SSI has not been widely applied in determining effective
vibrational length. Besides, two different approaches in predicting cable forces are
presented, which is the least-square approximation and the two-frequency methods in
which effective cable lengths are adopted in those two. Therefore, this research aims
to demonstrate the mentioned methods in a real arch-cable railway bridge. Besides,
a challenge of using the subspace method for determining the effective length is also
discussed.
Cable Force Prediction Technique Using Subspace … 259
2 Subspace Method
The SSI method is started by considering the state-space form of the second-order
differential equation of a system. It is assumed that the input process and measurement
process follows the gaussian distribution to represent the noisy signal in the real
condition. Therefore, the discrete-time stochastic state-space model of a system is
expressed as
x(k + 1) = Ax(k) + w(k) (1)
y(k) = C x(k) + v(k) (2)
where A ∈ Rn×n describe the discrete-time system matrix, x(k) ∈ Rn is the state
vector in time step k, C ∈ Rl×n stands for the allocation matrix output, y ∈ Rl
is output measurement, while both w ∈ Rn and v ∈ Rl are the noise input and
measurement noise, which both are white noise, zero-mean.
There are two types of SSI: SSI-Data, which directly works on the raw time series
data and SSI-COV requires preprocessing the raw response data into the covariance
matrices [19]. The post-processing data in SSI-COV let this type of SSI runs faster
than the SSI-DATA [10]. Once the covariance about the time lag i are obtained, then
rearrange them into a Toeplitz matrix, T 1|i , as shown in Eq. (3)
⎡ ⎤
Ri Ri−1 ... R1
⎢ Ri+1 Ri ... R2 ⎥
⎢ ⎥
T 1|i =⎢ . .. .. .. ⎥ (3)
⎣ .. . . . ⎦
R2i−1 R2i−2 . . . Ri
Factorize this huge matrix by taking the SVD and gain the system matrix by
involving the Moore–Penrose pseudo-inverse as explained in Van Overschee and De
Moor [6]. Once the system matrix, A, is identified, the eigenvalue decomposition
is performed to attain the dynamical system properties through the following well
known complex system poles formula.
A = −1 (4)
μ j = eλ j (5)

λ j = −ζ j ω j + iω j 1 − ζ j2 (6)
where is a diagonal matrix containing eigenvalues μ j , stands for the eigen-

vector, λ j and μ j are the continuous and discrete system poles, respectively, ω j is
the circular natural frequency, and ζ j represents the system damping ratio. The mode
shape, , can be obtained can be through Eq. (7).
=C (7)
A stabilization diagram is performed to distinguish between the physical system

properties among the noisy signal by varying the order of system poles under specific
stable criteria reported in Peeters, Van der Auweraer [20].
3 Straight Cable Theory and Effective Vibration Length
Some successful vibration-based technique to attain the force of cable has reported
[1, 2, 14, 18, 21]. The analytical formula of cable tension force derived from a simply
supported axially-loaded Euler–Bernoulli beam is shown in Eq. (8). This equation
is one of the most often equation to be used for estimating the cable force [4].
4m L 2 f n2 n2π 2 E I
T = − (8)
n2 L2
where T stand for the force, L is the cable length, m is the linear mass, E I is
the bending stiffness, and f n is the frequency of mode n. Without removing the
cable’s bending stiffness effect, Cheng [3] propose a direct tension force formula
shown in Eq. (9). In this equation, flexural rigidity is implicitly replaced through the
combination of two-mode frequencies, mode p and q thus, the cable force estimation
can be performed with the absence of bending stiffness information. The equivalent
two-mode frequency is then denoted as f c .

q 4 f p2 − p 4 f q2
T = 4m L 2

= 4m L 2 f c2 (9)
p2 q 2 q 2 − p2
However, in a real case, cable’s support system commonly involves special

support, which neither belongs to hinged nor fixed. Hence, the application of the
formula leads to some error due to the complexity of rotational stiffness. Chen [15],
ef f
Wu [17] and Chen [2] proposed the concept of effective vibration length (L n ) to
be applied in any cable end-restraint condition. The corresponding vibration length
is obtained through the mode shape of a simply-supported case in which expressed
in the form of the sine wave function. Therefore, by obtaining the cable’s effective
vibration length and replacing the actual length, the tension force can be identified
through the same relationship.
The cable mode shape can be categorized into two groups according to their
movement about the mid-axis, symmetrical mode shapes (n = 1, 3, 5, . . .) which
are belong to the cosine function and unsymmetrical ones (n = 2, 4, 6, . . .), which
belong to the sinusoidal form. And the mode shape vector for r measured location
can be formed as shown in Eq. (10) where the n = n 1 , n 2 , . . . , n m are the m major
modes [15, 17].
⎧ ⎫ ⎧ ⎫
⎪
⎪ φ1n ⎪⎪ ⎪
⎪ cosin nπLx1 ⎪
⎪
⎪
⎨ φ2n ⎬⎪ ⎪
⎨ cosin nπ x2 ⎪
⎬
L
φn = .. ⎪ = an ⎪ .. , n = n1, n2, . . . , nm (10)
⎪
⎪ . ⎪ ⎪ . ⎪
⎪
⎪
⎩ ⎪
⎭ ⎪
⎩ ⎪
⎭
φr n cosin nπLxr
where

cos(·) if n = odd
cosin(·) = (11)
sin(·) if n = even
4 Implementation
A steel-arch railway bridge located in Taiwan is taken as the object in this research to
illustrate the theories as mentioned in the previous section, shown in Fig. 1a. A total
of 16 cables on each side support the bridge deck. According to the basic information
of the cable, it is recorded that the cable diameter is 9.5 cm consisting of 19 strands in
Fig. 1 a Bridge A b velocity-meter deployment c measurement signal

Fig. 2 Sensor deployment configuration
which the area is 140 mm2 per each and covered by HDPE material. By considering
all section and material properties, it is estimated that the linear mass of the cable is
23.26 kg/m. This mass information is required for detecting the cable force through
the vibration-based formula.
Five vibration transducers are used to measure the vibration of a cable deployed
on a different position along the longitudinal axis, as shown in Fig. 1b. The first
sensor position is 1.5 m from the bridge deck inline to the cable’s local axis, while
the gap between each of them are set to be equal distance, which is 1 m. The first
sensor location is different from others to accommodate the uncertainty boundary
condition. The sensors placement configuration can be observed in Fig. 2a. Each
sensor must be adjusted such that they have the same alignment about the horizontal
axis.
Through five synchronized vibration sensors, the vibration signals of a cable are
recorded in the same period. The sampling frequency is set to 1 kHz, and the data
capturing duration is about 5–10 min under ambient vibration. Onsite random signals
from ambient vibration excitation in a certain short period are shown in Fig. 1c. By
taking an example of the cable around the middle span, named B2W11, the vibration
responses of the inspection are shown in Fig. 2b.
The vibration responses obtained through the five sensors are used to perform
the modal analysis using SSI. Since the preprocessing of the data required in the
covariance-driven SSI type, the correlation between signals is computed and arranged
in the Toeplitz matrix. In this process, the vibration data is scaled down become
100 Hz, and the time lag is set to be 40 s. This scaling down step is performed to
save the memory during the computation without losing the important information.
Once the Toeplitz matrix is established, SVD is taken to factorize the system and
obtained the system matrix. The classical modal analysis of the system poles is then
performed, and the stabilization diagram can be established, as shown in Fig. 3.
The stabilization diagram is constructed under the stable criteria

f < 1%,
ζ < 5%, and

MAC < 2% [20]. Any poles satisfying all three parameters are
denoted as stable poles, while the poles that are not stable in all criteria are named
the new pole. There are also obtain poles that just stable in frequency and damping
ratio, frequency and MAC, or even just the frequency. The physical dynamic system
poles can generally be figured out by observing the stable poles’ vertical line.
As shown in Fig. 3, under the 50 Hz frequency range, there are many stable lines
detected, which indicated by a green-circle mark. It is shown in Fig. 4, 15 modes
are extracted through SSI, but not all the detected modes belong to the cable modes.
There are many possibilities that the modes obtained are mixed with other elements of
the bridge. Since the vibration-based approach requires the cable frequency, thus it is
Fig. 3 Stabilization diagram
Fig. 4 Extracted mode shapes

Fig. 5 Effective vibration of the first three cable mode
needed to select the right frequency which belongs to the cable. The selection of cable
frequency can be executed by matching detected modes to the cable modes’ natural
behaviour. The fitting curve is then taken to the extracted mode shape following the
trigonometric function shown in Eq. (10) within the range of real cable length. Cable
effective lengths are identified by observing the zero-crossing distance between the
maximum and minimum.
As shown in Fig. 5, three cable frequencies are detected along with the effec-
tive vibration modes. It is seen that there is a gap between the end-zero-crossing to
the corresponding end-boundary support, so the effective length is shorter than the
real length (L = 12.77 m). This indicates that the supporting system is not hinged,
i.e. the rotational stiffness exists at both ends. The effective lengths of these three
modes are obtained by calculating the distance of their corresponding end-zero-
crossing. Through these processes, it is found that the frequency of mode 1, mode
2, and mode 3 are 5.81, 11.98, and 18.89 Hz, respectively; and their corresponding
effective vibration length are 10.64, 10.77, 10.65 m. With the obtained mode frequen-
ef f
cies ( f n ) and their corresponding effective lengths (L n ), the remaining unknown
remains the tension force (T ) and flexural rigidity (E I ). Three selected modes are
then substituted into Eq. (8), resulting in three linear functions with T and E I as the
unknowns. The optimal values of these unknowns can be solved through the least
squared approximation [2]. Finally, the tension force (T ) obtained is 351.093 kN and
the bending stiffness (E I ) is 89.18 kNm2.
Besides using least-squares approximation, tension force can also be attained
through the two-frequency approach introduced by Cheng [3]. But since it is found
that the cable effective length shorter than its actual length, so it is not a hinged cable
system. As a result, the application of Eq. (9) cannot be directly adopted. It requires
computing not only the equivalent frequency of two modes but also its corresponding
equivalent vibration length. After computing these two equivalencies, the force of
cable B2W11 obtained using the two-mode frequency method is 351.457 kN.
Overall in the final step, two different approaches to find out the tension forces
are used. Even though the force values produced through these two methods have a
gap of about 0.1% to each other, but this small deviation is acceptable.
Meanwhile, the minimum breaking strength for one strand is 260.7 kN. If the
maximum allowable force is assumed to be 85% of the breaking strength, the allow-
able cable force is 4210.3 kN. Consequently, it is computed that the measured cable
force is only 8.3% of the allowable one; in other words, the cable is safe.
5 Conclusion
This research presents a cable force estimation approach applied on a real steel arch
bridge through a vibration-based approach. Five sensors are mount at a different
location along the cable axis to get the vibration responses. Random signals are
obtained under the ambient environment, and SSI-COV is used to extract the cable
dynamical properties. Three cable natural frequencies ( f n ) are obtained along with
their effective vibration modes, where f 1 , f 2 , and f 3 are 5.81, 11.98, and 18.89 Hz,
respectively, and their corresponding effective vibration lengths are 10.64, 10.77,
10.65 m. Two different approaches to reveal the force are utilized, the first uses
the least-squares approximation, and the second is the two-frequency approach. The
force obtained using the first method is 351.093 kN, while the second one is 351.457
kN, so the gap between these two results is 0.1%. Compared to the allowable force,
the measured cable force indicates that it is still under a safe level. Besides, it is also
attained from the first method that the bending stiffness of cable B2W11 is 89.18
kNm2.
Overall, three main concerns can be pointed out from this study. First of all, the
use of SSI helps in extracting the mode frequency of the system under operational
conditions, and cable modes selection is required as the additional post-procedure
since the extracted modes are mixed with other bridge elements. Secondly, the appli-
cation of effective cable length can successfully reveal the cable force through two
different approaches with a slight deviation. Lastly, the cable internal force disclosure
is performed under uncertain end-restraint condition, indicating that this procedure is
a proper method to estimate an existing stayed cable internal force on a real structure.
References
1. Haji Agha Mohammad Zarbaf SE et al. (2017) Stay force estimation in cable-stayed bridges
using stochastic subspace identification methods. J Bridge Eng 22(9):04017055
2. Chen C-C et al (2016) Tension determination of stay cable or external tendon with complicated
constraints using multiple vibration measurements. Measurement 86:182–195
3. Yu C-P, Cheng C-C, Chiang C-H (2011) Alternative determination of cable forces using flexural
theory of axially loaded member. In: Non-destructive characterization for composite materials,
aerospace engineering, civil infrastructure, and homeland security. International Society for
Optics and Photonics, pp 79832R
4. Kim BH, Park T (2007) Estimation of cable tension force using the frequency-based system
identification method. J Sound Vibr 304:660–676
5. Van Overschee P, De Moor B (1996) Continuous-time frequency domain subspace system

identification. Signal Process 52(2):179–194
6. Van Overschee P, De Moor B (1996) Subspace identification for linear systems: theory—
Implementation—applications. Kluwer Academic Publishers, London
7. Brincker R, Andersen P (2006) Understanding stochastic subspace identification. In: Proceed-
ings of the 24th IMAC, St. Louis 126
8. Döhler M, Andersen P, Mevel L (2011) Data merging for multi-setup operational modal analysis
with data-driven SSI. In: Conference proceedings of the society for experimental mechanics
series. Springer, New York, NY, pp 443–452
9. Giraldo DF et al (2009) Modal identification through ambient vibration: comparative study. J
Eng Mech 135(8):759–770
10. Magalhaes F, Cunha A, Caetano E (2009) Online automatic identification of the modal
parameters of a long span arch bridge. Mech Syst Signal Process 23(2):316–329
11. Ubertini F, Gentile C, Materazzi AL (2013) Automated modal identification in operational
conditions and its application to bridges. Eng Struct 46:264–278
12. Humar J (1990) Dynamics of structures. Prentice-Hall, Englewood Cliffs, New Jersey
13. Sim S-H et al (2013) A wireless smart sensor network for automated monitoring of cable
tension. Smart Mater Struct 23(2):025006
14. Kim S-W et al (2013) Vision-based monitoring system for evaluating cable tensile forces on a
cable-stayed bridge. Struct Health Monit 12(5–6):440–456
15. Chen C-C et al (2013) Determination of stay cable force based on effective vibration length
accurately estimated from multiple measurements. Smart Struct Syst 11(4):411–433
16. Wu W et al (2012) Determination of stay cable force based on multiple vibration measurements
to consider the effects of unsymmetrical boundary constraints. In: 6th European workshop on
structural health monitoring. NDT. net Bad Breisig, Dresden, Germany, pp 1049–1056
17. Wu W et al (2010) Determination of stay cable force based on multiple vibration measurements
to consider the effects of uncertain boundary constraints. In: Proceedings of 5th European
workshop on structural health monitoring. Naples, Italy, pp 282–287
18. Geier R, De Roeck G, Petz J (2005) Cable force determination for the Danube channel bridge
in Vienna. Struct Eng Int 15(3):181–185
19. Peeters B (2000) System identification and damage detection in civil engeneering
20. Peeters B et al (2007) Operational modal analysis for estimating the dynamic properties of a
stadium structure during a football game. Shock Vib 14(4):283–303
21. Zui H, Shinke T, Namita Y (1996) Practical formulas for estimation of cable tension by vibration
method. J Struct Eng 122(6):651–656
Implementation of Microbially Induce
Calcite Precipitation (MICP) by Bacillus
Subtilis and Adding Sand in Repairing
Shear Strength Parameters of Peat
Firman Syarif and Dian M. Setiawan
Abstract These days, the utilization of microbially induced calcite precipitation

(MICP) has gotten mainstream as a ground improvement strategy for sandy soil. In
any case, this strategy was not all that much applied to peat soil. This research focused
on bacterial calcium carbonate precipitation from Bacillus Subtilis and its impact on
shear strength parameter of peat soil. An uncommon injection framework was ready
for actuating bacterial solution for the examples. The bacterial arrangement provided
to the examples by gravity for 7 days in explicit molds intended for this work. From
this exploration. The most noteworthy compressive strength esteem is found in the
expansion of 15% cementation arrangement of 0.42 kg/cm2 of the first soil test +
5% sand, which has a free compressive strength worth of 0.30 kg/cm2 .
Keywords Bacillus subtilis · Organic soil · Peat soil · Shear strength
1 Introduction
In the previous 18,000 years there was an arrangement of peat soil in southeast
Asia. While Indonesia’s peat conformed to 5000–8000 years prior. This peat soil is
framed in territories that have for quite some time been lowered in water. Indonesia
has 20 million hectares of peatland which is number four after Canada, Russia, and
America in the broadest class of peat land on the planet. The circulation of peatlands
in Indonesia is in four huge islands, in particular Sumatra 35%, Kalimantan 32%,
Sulawesi 3%, and Papua 30%. On the island of Sumatra, peatlands are typically
found in the marshes along the east coast with a space of 7.2 million hectares. On
F. Syarif
Universitas Islam Riau, Pekanbaru 28284, Riau, Indonesia
D. M. Setiawan (B)
Universitas Muhammadiyah Yogyakarta, Bantul 55183, Yogyakarta, Indonesia
Texas A&M University, College Station, TX 77843, USA
268 F. Syarif and D. M. Setiawan
Sumatra Island, the Province with the most extensive peatland is Riau Province with
a space of ±4.04 million Ha or 56.1% of the absolute peatland territory in Sumatra
[1–5].
Peat soil is a problem that often occurs in the construction world. Peat soils have
a low bearing capacity and can cause large land settlement during loading, Peat
soil is very soft soil (very soft soil) with a very low bearing capacity and is easily
compressible if the load is working on it [6]. There is a fundamental problem that
arises due to the loading on the peat soil layer, namely the higher the organic content,
the lower the bearing capacity and shear strength and the greater the compression [7].
Peat soil compression is very different from clay soil, where the compression that
occurs on peat soil is a long process of compression. The amount of compression
due to the consolidation of a soil layer is very dependent on the size of the load
given and the thickness of the soil layer that is loaded [8]. So that in this study, the
addition of bacillus subtilis bacteria to peat soil using bio-grouting technique method
to determine the effect of the addition of these bacteria on the compressive strength
of peat soil.
This repair technique uses the biogrouting method with Bacillus subtilis bacteria.
Biogrouting is a soil stabilization technique involving microorganisms that are
induced by calcium carbonate (CaCO3 ) precipitation. Precipitation of calcium
carbonate acts as a crystallizer between cells that stimulates the cementation process
between the soil grains. Biogrouting also means injection of adhesive in soil or rock
that passes water with the aim of closing pores in the soil, biogrouting also functions
as a penetration or penetration and compacting or compaction in the soil [9].
Bacillus are gram-positive rod-shaped bacteria with the optimal temperature for
growth between 25–35 °C. Although bacillus was considered strictly aerobic, it was
discovered later that they could live anaerobically under the prescribed conditions.
Bacillus subtilis has a relatively different physiology from other non-pathogenic
bacteria, which is relatively easy to genetically manipulate and easy to breed so that
it can be developed on an industrial scale [10].
Aside water, cementitious materials are the most eaten up substances on Earth.
Along these lines, they should be concentrated, to show adequate execution, worried
about the presentation of a couple of works trying to improve the materials’ mechan-
ical properties. Besides, in the point of view of formative difficulties on improving
their characteristics and durability, increases of minute creatures, including the class
Bacillus, are being read for break filling, consequently, growing compressive char-
acteristics, through calcium carbonate (CaCO3 ) precipitation. Consequently, leading
this interaction is moreover called biocementation, subsequently, it should be seen
that organisms exist for billions of years, and countless their biotechnological appli-
cations are not yet settled. Furthermore, Albeit’ various experts are enthusiastic
about the experimentation of bioprecipitation of CaCO3 , albeit the whole strategy
isn’t exceptionally clear and portrayed. In the meantime, microorganisms are fit
for hurrying calcium carbonate creation through a couple of frameworks, the phys-
iology and innate contemplations required are puzzling and difficult to grasp. A
couple of microorganisms with the ability to make minerals were used in fixing
limestone tourist spots and furthermore in filling pores and parts in concrete and
Implementation of Microbially Induce Calcite Precipitation … 269
Fig. 1 Process of calcite precipitation process on MICP technique [14]
different cementitious materials, fit for improving mechanical properties and sturdi-
ness. Among the microorganisms used in the investigation on biocementation, the
living being B. subtilis sticks out, in view of their capacity to convey calcite, a translu-
cent sort of CaCO3 . This cycle is sped up within the sight of a medium with a calcium
source [11].
An enzyme-reagent mixed solution from bacteria (i.e., purified urease and CaCl2 -
urea), created the hastened calcite after the substance response, is infused into the
dirt. The hastened calcite may give spans between the grains of sand, confining their
development, and thus, improving the solidness and the strength of the dirt [12–
14]. A schematic of the entire interaction recorded above, and grouting instrument
expected are represented in Fig. 1.
2 Material and Method
2.1 Material
Peat Soil and Sand. The soil sample (see Fig. 2) that the researchers will use in this
research is peat soil whose condition has been disturbed, where the soil is taken at a
depth of +50 cm from the topsoil surface using a hoe, then the soil is transported to
the laboratory to be dredged by drying it in the open sun. Then the soil is sifted to
pass sieve no 4. The sand used by researchers for this research is white sand that has
been sieved until it passes using sieve number 40.
Fig. 2 Peat soil and sand
Fig. 3 Reagents include Bacillus Subtilis (a), CO (NH2 ) urea (b), and CaCL2 , and the microor-
ganism (c)
Reagent. Reagent used in this research are CO (NH2 ) urea, CaCL2 , and the
microorganism, Bacillus Subtilis as a grouting material seen in Fig. 3.
2.2 Method
Preliminary Testing. The initial research process was carried out by a preliminary
test which was a test to obtain the physical properties of the soil. This test aims to
check the soil conditions to match the ground conditions in the field. Water content
testing (ASTM D 2216-98) was carried out to obtain the percentage of water weight
against the dry soil weight. Examination of specific weight (ASTM D 854-02) test
was carried out to obtain the specific weight of the soil which is the ratio of soil
weight to water weight. Soil mechanical properties test was carried out with the
Proctor test (ASTM D 698) to obtain the optimum moisture content and maximum
density, then the data was used as a comparison to the soil density in the test.
Main Testing. This test is carried out after the preparation of the cementation solution
and the preliminary testing has been completed. This testing process begins with
mixing the cementation solution in which bacillus subtillis is already present in the
peat soil that has been added with sand. The cementation solution and water are
adjusted to the optimum amount of water content, then mixed with peat soil that has
been mixed with sand to test its Free Compressive Strength which is the main test
of this study. The amount of sand added to the compressive strength sample is 5%
of the weight of the soil. The amount of cementation solution mixed in the sample
of this specimen has 6 variations, namely sample 1 = 0%, sample 2 = 5%, sample
3 = 10%, sample 4 = 15%, sample 5 = 20%, sample 6 = 25%. The percentage
of the cementation solution is the optimum moisture content of the test for the free
compressive strength of the peat soil seen in Figs. 4 and 5.
Fig. 4 Reagent manufacturing process (a) and equipment (b)

The UCS is perhaps the most ordinarily utilized stone designing boundaries for its
mass arrangement, or strength assurance. Also, the mean worth and its changeability
are frequently expected to dependably address material property, despite the fact
that it once in a while addresses the flawless structures. Besides, the test essentially
records the breakdown of burden during uniaxial stacking of a tube-shaped example,
and many recommended this appraisal as a list, as opposed to just an interesting
designing boundary. Subsequently, it is an intermediary for rock strength, subject to
the stacking rate, example size, and other various components. Moreover, it is not
equivalent to the Hoek–Brown strength basis boundary σci [11].
Compaction testing is carried out to obtain the maximum dry bulk weight (γdmax )
of the original soil of 0.467 g/cm3 and the optimum moisture content (OMC) of the
original soil of 157%, can be seen in Fig. 6.
Fig. 5 UCS test
MOISTURE - DENSITY CURVE

0.60
0.55
Bulk Dencity Gr/cm 3
0.50
0.45
0.40
0.35
130 140 150 160 170 180 190
Water Content(%)
Fig. 6 Proctor standard result

Based on the tests that have been carried out, the physical properties of the peat
soil in Siak Regency, Riau Province, were obtained in Table 1.
In this research, the free compressive strength test is carried out on the disturbed
sample soil or in a state that has been mixed with other materials. However, for the
test object that has been given a mixture of cementation solution, it must be stored
for 14 days. This is to measure the ability of each sample of the specimen to free
compressive strength.
The free compressive strength test value on the original soil sample +0% sand and
+0% bacteria, then continued with the original soil test object +5% sand and with
those that have been mixed the addition of cementation solution with variations in the
percentage of the addition of 0, 5, 10, 15, 20, and 25% can be seen from Table 2 and
Fig. 7.
From Fig. 7, there is an increase in the strength of the free compressive strength of
the peat soil from the sample of the original soil sample +5% sand with that which
has been mixed with bacterial cementation solution with variations in the addition
of 0, 5, 10, 15, 20, and 25%. In this case, with the addition of bacteria, the bearing
capacity of the peat soil occurs.
Table 1 Physical properties of the peat soil in Siak Regency

No Description Standard Value Unit
1 Specific gravity (GS) 1.2–1.95 1.30 –
2 Water content (%) 200–900% 407.5 %
3 γdmax – 0.467 g/cm3
4 Optimum water content (%) – 157 %
Table 2 UCS test results

No Strain (%) Compressive strength (kg/cm2 )
0% 5% 10% 15% 20% 25%
1 0.00 0.00 0.00 0.00 0.00 0.00 0.00
2 0.175 0.06 0.06 0.06 0.03 0.06 0.07
3 0.35 0.09 0.10 0.10 0.06 0.09 0.10
4 0.70 0.16 0.16 0.16 0.12 0.14 0.16
5 1.05 0.22 0.22 0.22 0.19 0.19 0.23
6 1.40 0.29 0.27 0.26 0.23 0.24 0.29
7 1.75 0.30 0.30 0.30 0.29 0.30 0.33
8 2.10 0.29 0.31 0.31 0.36 0.36 0.33
9 2.45 0.28 0.28 0.41 0.41 0.28
10 2.80 0.42 0.38
11 3.15 0.40
Then from Fig. 7, the highest value of each sample of the test object is also
obtained, namely the value of qUmax (kg/cm2 ) and can be seen in Table 3.
The results of Table 4 on the graph of the relationship between the value of the
free compressive strength (qUmax ) of the sample of the original soil sample +5%
sand with variations in the solution of bacterial cementation, there is an increase
and decrease in the percentage of 5, 10, 15, 20 and 25%. The value of the free
compressive strength of the original soil test object +5% sand is 0.30 kg/cm2 , while
with the addition of variations in bacteria 5% = 0.31 kg/cm2 , 10% = 0.31 kg/cm2 ,
15% = 0.42 kg/cm2 , 20% = 41 kg/cm2 , and 25% = 0.33 kg/cm2 . The highest value
of free compressive strength is found in the addition of 15% cementation solution
of 0.42 kg /cm2 of the original soil sample + 5% sand, which has a free compressive
strength value of 0.30 kg/cm2 .
From the results of the free compressive strength test sample of peat soil using the
UCS tool with the addition of variations in the addition of +5% sand and bacteria 0,
5, 10, 15, 20, and 25% there is an increase and decrease in the strength of the peat
Compresive Strength Graph

45
40
35 + 5% Sand + 0%
30 Bacteria
Stress kg/cm2
+ 5% Sand + 5%
25 Bacteria
+ 5% Sand + 10%
20
Bacteria
15 + 5% Sand + 15%
Bacteria
10 + 5% Sand + 20%
5 Bacteria
+ 5% Sand + 25%
0 Bacteria
75
0.1
Strain %
Fig. 7 Curve for UCS test
Table 3 qUmax value of each

Adding reagent (%) Strain (%) qUmax (kg/cm2 )
sample specimen
0 2.10 0.30
5 2.45 0.31
10 2.45 0.31
15 3.15 0.42
20 2.80 0.41
25 2.45 0.33
Table 4 Comparison of the value of free compressive strength of this study of bio-grouting
stabilization with other stabilizers
No 10% calcium carbonate + Sand and bio-grouting Due to preloading [18]
8% Matos [17]
Load qUmax (kg/cm2 ) Percentage qUmax (kg/cm2 ) Time qUmax (kg/cm2 )
(kg) (%) (kPa
1 day)
1 1 1.1 0 0.30 10 0.09
2 2 1.6 5 0.31 20 0.11
3 3 1.8 10 0.31 10 0.12
4 – – 15 0.42 20 0.14
5 – – 20 0.41 – –
6 – – 25 0.33 – –
soil. This is due to the addition of the subtilis bacteria cementation solution which
can make the pores of the peat soil particles closed or filled with bacteria so that the
peat soil does not easily decline. In [15, 16] the indication of this result is there are
adding some amount of water on the reagent therefore the sample with reagent has a
large amount of water and effect on the compression strength. There are also some
investigations should be conducted to analyze this effect i.e., water effect, pH, and
amount of carbonate. Organic soil had many amounts of water content and this effect
should be analyzed in another research.
From the comparison of Table 4, the maximum free compressive strength (qUmax )
value is the highest on peat soil with stabilization of 10% lime + 8% matos while
the lowest value (qUmax ) is due to preloading, and it turns out that in the addition
of sand and bio-grouting with variations in the addition of bacteria 0, 5, 10, 15%,
20, and 25% there was an increase in the value of the free compressive strength of
the peat soil. From the UCS test, the sample with 10% concentrate of reagent had
a compression strength qu = 2.93 kg/cm2 and the sample without the reagent had a
compression strength qu = 4.97 kg/cm2 .
4 Conclusion
From the results of the research on the compressive strength of peat soil using a
mixture of sand and bio-grouting techniques with the help of the bacterium Bacillus
subtilis, the results obtained in accordance with the research objectives are as follows:
the relationship between the compressive strength (qUmax ) of the original soil sample
+5% sand with variations in the concentration of the bacterial cementation solution
at a percentage of 0, 5, 10, 15, 20, and 25%. The value of the free compressive
strength of the original soil test object +5% sand is 0.30 kg/cm2 , while with the
addition of variations in bacteria 5% = 0.31 kg/cm2 , 10% = 0.31 kg/cm2 , 15% =
0.42 kg/cm2 , 20% = 41 kg/cm2 , and 25% = 0.33 kg/cm2 . The highest compressive
strength value is found in the addition of 15% cementation solution of 0.42 kg /cm2
of the original soil sample +5% sand, which has a free compressive strength value
of 0.30 kg/cm2 .
References
1. Mochtar NE (1998) Koefesien tekanan tanah ke samping at rest (ko) tanah gambut berserat
serta pengaruh overconsolidation ratio (OCR) terhadap harga Ko. Jurnal Teknik Sipil ITB 5(4)
2. Mochtar NE, Yulianto FE (2017) Behaviour change in peat stabilized with fly ash and lime
caco3 due to water infiltration. ARPN J Eng Appl Sci 12(17):4967–4972
3. Moeyadi H, Kazemian S, Huat BBK (2013) Shear strength parameters of improve peat by
chemical stabilizer. Geotech Geol Eng 31:1089–1106
4. Wibowo P, Suyatno N (1998) An overview of indonesian wetlands sites. In: Wetlands
International—Indonesia programme (WI-IP)
5. Yulianto FE (2017) Perilaku tanah gambut berserat permasalahan dan solusinya. In: Konferensi
Nasional Teknik Sipil dan Infrastruktur – I Jurusan Teknik Sipil, Universitas Jember
6. Parlan (2016) Pengaruh jumlah plat helical terhadap daya dukung pondasi tiang helical pada
tanah gambut. Jurnal Online Mahasiswa Fakultas Teknik Universitas Riau 3(2)
7. Wardana IGN, Wardata I (2010) Korelasi strain rate dengan kadar organik pada test konsolidasi
metode constant rate of strain. Jurnal Ilmiah Teknik Sipil 14(1):43–56
8. Fatnanta (2014) Pengaruh kadar lempung dan kadar air pada sisi basah terhadap nilai cbr pada
tanah lempung kepasiran (sandy clay). Jurnal Online Mahasiswa Fakultas Teknik Universitas
Riau 1(2)
9. Pangesti DR (2005) Pedoman grouting untuk bendungan. Depertemen Pekerjaan Umum,
Derektoral Jendral Sumber Daya Air, Derektoral Sungai, Danau dan Waduk, Jakarta
10. Soesanto L (2008) Pengantar pengendalian hayati penyakit tanaman, PT Raja Grafindo Persada,
Jakarta
11. Barkouki TH, Martinez BC, Mortensen BM, Weathers TS, De Jong JD, Ginn TR, Spycher
NF, Smith RW, Fujita Y (2011) Forward and inverse bio-geochemical modeling of microbially
induced calcite precipitation in half-meter column experiments. Transp Porous Med 90(23):23–
39
12. Neupane D, Yasuhara H, Kinoshita N, Unno T (2013) Applicability of enzymatic calcium
carbonate precipitation as a soil strengthening technique. J Geotech Geoenviron Eng
139(12):2201–2211
13. Putra H, Yasuhara H, Kinoshita N, Hirata A (2017) Optimization of enzyme-mediated calcite
precipitation as a soil-improvement technique: The effect of aragonite and gypsum on the
mechanical properties of treated sand. Curr Comput-Aided Drug Des 7(2):59
14. Yasuhara H, Putra H, Kinoshita N (2016) Enzyme-mediated carbonate precipitation (EMCP)
for soil improvement: the effect of magnesium substitution. In: The 2016 world congress on
advances in civil, environmental, and materials research (ACEM16), Jeju Island, Korea (2016)
15. Syarif F, Hardianto F, Mahardika G (2020) Penerapan teknik biocementation oleh bacillus
subtilis dan pengaruhnya terhadap permeabilitas pada tanah organik. Jurnal Saintis 20(1):47–52
16. Syarif F, Hardianto F, Mahardika G (2019) Microbially induce calcite precipitation as bio
grouting by bacillus subtilis on its shear strength parameter effects on organic soil (peat) from
Siak Regency Riau Province Indonesia. In: The 7th International Conference of Euro Asia Civil
Engineering Forum (EACEF) 2019—Green Engineering for Infrastructure and Safety against
Hazards, The Universitat Stuttgart (2019).
17. Prabowo A, Fauziah M (2018) Pengaruh stabilitas tanah menggunakan kapur dan matos
terhadap kuat geser dan konsolidasi tanah gambut. Universitas Islam Indonesia, Prosiding
Sidang Program Studi Teknik Sipil (SPSTS) FTSP
18. Waruwu A (2013) Korelasi nilai kuat tekan dan cbr tanah lempung yang distabilisasi dengan
abu batu dan semen. Jurnal Rancang Teknik Sipil 2(2):99–107
Sensitivity Analysis on the Effect
of Reinforcement Materials Addition
for Soil Stabilization
Ida Agustin Nomleni, Raden Harya Dananjaya, and Yusep Muslih Purwana
Abstract Sensitivity analysis is a widely used method to analyze uncertainty in the

model. The sensitivity is calculated as the ratio of the output changes and parameters
changes. This study was performed using a sensitivity index for sensitivity analysis
to determine the effect of reinforcement material addition on clay soil. In this study,
54 samples of soil with additional material have been prepared for the unconfined
compression test. The reinforcement materials used in this test are cement, tire waste
strips, and water-cement ration for the hydration process. The unconfined compres-
sion test results were then analyzed for sensitivity analysis input. Based on this study,
it can be concluded that the addition of reinforcement material, including cement,
and tire waste strips, and water-cement ratio, can increase the unconfined compres-
sion strength of high plasticity clay. The maximum unconfined compressive strength
value is 577.10 kN/m2 with 3% of tire waste strips, 20% cement, and 35% water-
cement ratio. The cement proportion is the most sensitive variable for unconfined
compression strength from the sensitivity analysis. The sensitivity analysis results
can help to become the baseline to help design improvement using tire waste strips
on high plasticity clay.
Keywords Sensitivity analysis · Sensitivity index · Unconfined compression test
1 Introduction
Soil plays a vital role in civil construction because all buildings stand on soil. High
plasticity clay is a problematic soil because it has a low bearing capacity, so it needs
stabilization to increase its bearing capacity. Several studies have been performed to
stabilize high plasticity clay, for example, using a mixture of tire waste, cement, and
I. A. Nomleni (B)
Department of Civil Engineering, National Central University, Taoyuan, Taiwan
R. H. Dananjaya · Y. M. Purwana
Department of Civil Engineering, Universitas Sebelas Maret, Surakarta, Indonesia
278 I. A. Nomleni et al.
clay. Research to improve clay properties has been conducted, with the general aim
is to improve the engineering properties of clay, especially to increase soil bearing
capacity and reduce soil compressibility. Some soil improvement methods used are
mechanical improvements by compacting, hydraulic improvements with dewatering,
and physical and chemical improvements.
The increasing number of motorized vehicles in Indonesia has led to increased
vehicle tire waste. Based on data reported by the Indonesian Central Bureau of Statis-
tics, there was an increase in motorized vehicles from 2014 until 2017 [1–4]. The
increasing number of vehicles has led to an increase in vehicle tire waste. Vehicle tires
waste in large quantities can cause various health problems when piled up or burned.
Therefore, the reuse of tire waste is needed to reduce environmental degradation. One
effective way of utilizing tire waste is by using it as a mixture in geotechnical construc-
tion materials. Many researchers worldwide have implemented vehicle tire waste. It
because tires developed rapidly throughout the world [5].
Research on low plasticity clay based on the Unified Soil Classification System
(USCS) was performed [6]. In this study, crumb rubber and cement addition in clay
mixture were investigated. Crumb rubber was cut into 0.8–2 mm. The study used
crumb rubber of 2.5, 5, 7, 0.5, 10% of the soil’s dry weight. It shows that the compres-
sive strength and tensile strength increased in crumb rubber and cement addition. [7]
conducted a study of soil reinforced by tire waste cut into fiber and cement. The study
was performed on Perth sandy soil subjected to unconfined compressive strength and
California bearing ratio test. From this study, it can be concluded that the addition of
cement and or the addition of waste tires in Perth sandy soil can increase the value
of the California bearing ratio and the unconfined compressive strength.
Iman et al. [8] stated that the model’s sensitivity could be calculated as the ratio
of the model’s output and parameters changes. There are two reasons that cause
problems in the sensitivity analysis study. The first problem arises when the model’s
response to parameter change is non-linear. In that case, the result will depend on
how the initial value is chosen because small initial values produce slight variations,
and larger initial values produce more significant variations. On the other hand, if
the initial parameter is located near the upper or lower limit of a valid range of
parameters, it can cause unacceptable values lying outside the range. Iman et al. [9]
suggested several reasons for using sensitivity analysis, namely: (a) to determine
parameters that require additional research to strengthen the theoretical basis so that
it can reduce the uncertainty of output, (b) to determine parameters that do not affect
the model so that it can be removed from the final model, (c) to determine parameters
which give a significant influence on output, (d) to determine parameters which have
a significant correlation with output, and (e) to know the change of output when input
is changed.
Identification and representation of uncertainty are essential components in
applying a model. One important approach often used to identify model uncertainty
is the sensitivity analysis method. Many sensitivity analysis methods can be used,
i.e., (a) the differential sensitivity analysis method, which can be analyzed based on
sensitivity coefficient, (b) one-time sensitivity analysis or local sensitivity analysis,
(c) factorial design method, which can be calculated by selecting several samples
Sensitivity Analysis on the Effect of Reinforcement … 279
from each parameter and analyzing the model for all sample combinations, (d) sensi-
tivity index which can be used to determine the sensitivity parameters by calculating
the percentage difference in output, (e) important factors, (f) subjective sensitive
analysis.
In [10], the sensitivity index method was used in the sensitivity analysis. This
method is a simple method used to determine sensitivity parameters. The sensitivity
index in the model is calculated based on the output % difference when the input
parameter values vary from minimum to maximum. Lenhart et al. [11] used the ∂y/∂x
differential approach in determining the sensitivity index value. Using the differential
approach shows that changes in output parameters are affected by changes in input
parameters. The sensitivity index for this calculation’s input can be calculated with
Eqs. 1 and 2 [11].
y2 − y1
I = (1)
2x
(y2 − y1 )/y0
I= (2)
2x/y0
In this research, sensitivity analysis was conducted using the sensitivity index
method. The sensitivity index was obtained from the clay mixtures subjected to
unconfined compression test (Fig. 1).
Fig. 1 Relationship between an output variable y and a parameter x proposed by Lenhart (2002)
Primary data from unconfined compression test results have been used in this
research. The unconfined compressive strength test was conducted on high plasticity
clay. The reinforcement materials used in this test are cement, tire waste strips, and
water-cement ration for the hydration process. Tire waste has been cut in a 1 mm ×
1 mm × 30 mm dimension and mixed with soil, cement, and water in various weight
conditions. The weight of tire mixtures varied from 0–7.5% of dry weight of the soil
mixture, i.e. 0, 1.5, 3, 4.5, 6, and 7.5%. The cement’s weight in the mixture is 10, 15,
and 20% of the soil’s dry weight. The water-cement ratios used in this study are 25,
35, and 45% of the cement’s dry weight. Due to variations of the mixture composi-
tions, a wide range of the data is obtained. The data obtained from the unconfined
compression test then calculated and analyzed using sensitivity analysis. The sensi-
tivity analysis used in this study is based on the hypothesis of a linear relationship
between output and input parameters.
3.1 Testing Material
A preliminary test aims to determine the soil properties. In this study, soil was taken
from Ploso village, Mojogedang, Karanganyar, at a depth of 0 to 0.5 m below the
ground level. This soil’s specific gravity is 2.66, with a liquid limit of 57.94% and a
plastic limit of 29.31%. According to the Unified Soil Classification System (USCS),
this soil can be classified as CH (inorganic clay or high plasticity clay). The summary
of the preliminary test results can be seen in Table 1.
Table 1 Soil properties

Soil properties Notation Value
Specific gravity Gs 2.66
Liquid limit (%) LL 57.94
Plastic limit (%) PL 29.31
Plasticity index PI 28.64
Dry density (g/cm3 ) γd opt 1.404
Moisture content (%) wopt 29
Pass No. 200 sieve (%) 90.52
USCS classification CH
3.2 Unconfined Compression Test Results
This research was conducted under Indonesian National Standard (SNI) 6887:2012,
Standard Test Methods for Compressive Strength of Molded Soil–Cement Cylin-
ders [12]. Figures 2 and 3 show that shredded tire waste strips addition can affect
the unconfined compressive strength value. The value of unconfined compression
strength without the addition of tire waste is 173.55 kN/m2 . At the addition of 1.5%
tire waste strips, the value of unconfined compressive strength is increased up to
Fig. 2 Tire waste strips with

a dimension of 1 mm ×
1 mm × 30 mm
(a) 0% (b) 1.5% (c) 3%
(d) 4.5% (e) 6% (f) 7.5%
Fig. 3 Unconfined compression test results for soil sample with reinforcement material addition;
10% of cement and the water-cement ratio of 25%
250
0%TW,10%C,25%WCR
200 1.5%TW,10%C,25%WC
3%TW,10%C,25%WCR
4.5%TW,10%C,25%WC
)
150
6%TW,10%C,25%WCR
100 7.5%TW,10%C,25%WC
0%TW, 0%C, 0%WCR

50
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
ε (%)
Fig. 4 Relationship between compression strength value and strain on the soil sample with
reinforcement material addition
179.80 kN/m2 . When the addition of tire waste is 3%, the unconfined compression
strength reaches its maximum value. The value of unconfined compressive strength
decreases by adding 4.5, 6, and 7.5% of tire waste (Fig. 4).
3.3 Sensitivity Analysis
The sensitivity analysis can be done if the model has two or more dependent variables.
In this analysis, the water-cement ratio of 45 and 20% cement is the dependent
variable. Dependent variables are needed to get the simple 3D statistical model for
predicting the unconfined compression strength of the model.
3.3.1 The Optimum Value of Reinforcement Material Addition
The first step of sensitivity analysis is determining the value of the addition of the
optimum tire waste strips percentage for adding a water-cement ratio of 45% and
20% cement using 3D statistical models. Based on the 3D statistical model, we can
get the percentage of tire waste strips that needs to add to the unconfined compression
test sample. Based on this analysis, the percentage of tire waste strips can be found
with Eq. 3. The maximum percentage of tire waste addition is 4.05%.
Z(X, Y) = 17, 9419−1, 7260 Y + 0, 0529 Y2 −0, 1617 X

+ 0, 0048 XY + 0, 0012 X2 (3)
where:
X= water cement ratio (%) when X value is 25 ≤ X ≤ 45.

Y= cement (%) when Y value is 10 ≤ Y ≤ 20.
Z= tire waste strips percentage (%).
3.3.2 Sensitivity Analysis for Additional Reinforcement Material
The second step of sensitivity analysis is determining the sensitivity index for the
addition of 3% tire waste for the addition of a water-cement ratio of 45% and 20%
cement using 3D statistical models. The water-cement ratio, cement, and sensitivity
index can be plotted with a 3D statistical model.
The sensitivity index value can be obtained using Fig. 5 and or Eq. 4. The index
sensitivity for the addition of 3% tire waste strips for adding a water-cement ratio of
45 and 20% cement is 29.49.
Z(X, Y) = −22, 8984−8, 3648 Y + 0, 3147 Y2 + 4, 7260 X

+ 0, 0360 XY − 0, 0747 X2 (4)
where:
Fig. 5 3D statistical model of optimum value of reinforcement material addition


Z= sensitivity index.
3.3.3 Unconfined Compression Strength
The third step of sensitivity analysis determines the optimum unconfined compressive
strength for the addition of 3% tire waste strips for the addition of a water-cement
ratio of 45 and 20% cement using 3D statistical models. The model can also be
predicted using the following equation or Fig. 6. The predicted unconfined compres-
sion strength is 430 kN/m2 can be obtained with the proposed equation as follow
(Fig. 7):
qu (X, Y) = −708, 7262 − 19, 4120Y + 1, 1322Y2 + 58, 2342X

+ 0, 2456XY − 0, 8729X2 (5)
where:

qu = predicted unconfined compression strength of model.
Fig. 6 3D statistical model of sensitivity index

Fig. 7 3D statistical model for predicting the value of unconfined compression strength
3.3.4 Unconfined Compression Test Calculation in the Polynomial

Model
The unconfined compressive strength values in the model can be predicted using
Eq. (8). The unconfined compression strength value in the predicted model is
399.03 kN/m2 , while the unconfined compression strength value in the laboratory test
is 412.40 kN/m2 .
qu
IS = (6)
x
qu = x × IS (7)

qu = qopt − (xopt −xi × IS) (8)
where:
qu = unconfined compression value in the model (kN/m2 ).

qopt = the value of unconfined compression strength with the reinforcement
material addition (kN/m2 )
x opt = independent variable at optimum conditions (the percentage of tire waste
strips, cement, and the addition of water at optimum conditions)
xi = the independent variable reviewed (percentage of the tire waste strips,
cement, and water)
IS = sensitivity index (kN/m2 ).
4 Conclusions
Based on this research, it can be concluded that the addition of reinforcement material,
including cement, and tire waste strips, and water-cement ratio, can increase the
unconfined compression strength of high plasticity clay. The maximum unconfined
compressive strength value is 577.10 kN/m2 with 3% of tire waste strips, 20% cement,
and 35% water-cement ratio. The cement proportion is the most sensitive variable
for unconfined compression strength from the sensitivity analysis. The sensitivity
analysis results can help to become the baseline to help design improvement using
tire waste strips on high plasticity clay.
References
1. Badan Pusat Statistik Indonesia (2014) Statistik Indonesia 2014. Badan Pusat Statistik ISSN:
0126-2912
0126-2912
0126-2912
0126-2912
5. Humphrey DN (1999) Civil engineering application of tire shreds. Presented at the tire industry
conference, Hilton Head, South Carolina
6. Yadav JS, Tiwari SK (2017) Assessment of geotechnical properties of uncemented/cemented
clayey soil incorporated with waste crumb rubber. J Mater Eng Struct 37–50
7. Bazazorde SM (2018) UCS and CBR behaviour of Perth sandy soil reinforced with waste tyre
fibres and cement. Retrieved from https://ro.ecu.edu.au/theses/2076
8. Iman RL, Helton JC (1998) An investigation of uncertainty and sensitivity analysis techniques
for computer models. Risk Anal 8:71–90
9. Iman RL, Helton JC, Campbell JE (1981) An approach to sensitivity analysis of computer
models: part I-introduction, input variable selection and preliminary assessment. J Qual Technol
13:174–183
10. Hamby DM (1994) A review of techniques for parameter sensitivity analysis of environmental
models. Environ Monit Assess 32:135–154
11. Lenhart T, Eckhardt K, Fohrer N, Frede H-G (2002) Comparison of two different approaches
of sensitivity analysis. Phys Chem Earth 27:645–654
12. Badan Standarisasi Nasional. SNI 6887:2012: Metode uji kuat tekan silinder campuran tanah-
semen. Badan Standarisasi Nasional, Jakarta (2012)
Numerical Simulation of Slope Stability
for Soil Embankment Reinforced
with Inclined Bamboo Piles
Ngudiyono and Tri Sulistyowati
Abstract Slope stability is an important parameter when designing for the embank-
ment. Poor-quality and degradation properties of soil material became triggers slope
failure. Bamboo culm (Dendrocalamus asper) provided a diameter of 6–15 cm and
high tensile strength parallel to the grain, and it was suitable for slope stabilizing.
Slope stabilization by bamboo pile is a passive technique in which soil displace-
ments require to activate pile forces. In this manuscript, two-dimensional (2D) model
slope reinforcement with bamboo pile with the shear strength reduction (SSR) finite
element method (FEM) by using ABAQUS SE software has simulated. The bamboo
piles with 5 m length and 6 cm, 10 cm, 15 cm for diameter were placed two in the edge,
one in the middle, and inserted perpendicular to slope plane embankment with the
inclination of 33.69°. The model of the slope was analyzed, its influence on bamboo
piles soil interaction by evaluating the safety factor (FOS). The result has shown
that reinforced embankment slope with inclined bamboo piles had increased slope
stability, the safety of factor (FOS) increase to 1.515. The shear strength reduction
(SSR) finite element method (FEM) can be used as an alternative for slope stability
analysis.
Keywords Slope stability · SSR · FEM · Reinforcement · Bamboo pile
1 Introduction
Slope stability is an important parameter when designing embankment roads, high-

ways, dams, and other civil engineering structures. The problem of poor-quality soil
material and degradation properties of soil material became sometimes triggers slope
failure. Several slope reinforcement techniques widely applied are geosynthetic, soil
Ngudiyono (B) · T. Sulistyowati

T. Sulistyowati
288 Ngudiyono and T. Sulistyowati
nailing, concrete or steel mini-pile, concrete block, etc. The technologies that require
skilled labor are relatively expensive.
Bamboo is a natural material widely available globally, especially in Asia, and
has been widely used as a traditional and modern building material. Bamboo also
green building material, renewable, sustainable, typically be harvested in less than
three years to four years, and relatively inexpensive. Bamboo culm (Dendrocalamus
asper) provided with diameter 6–15 cm, tensile strength parallel to grain 228 MPa,
flexural strength 134.97 MPa, and modulus of elasticity 12,888.48 MPa [1, 2] that
make it suitable for slope stabilizing. Bamboo piles inserted in a soil mass can provide
stability by adding the reactive portion to the resistant forces contrary to the ground.
Slope stabilization by bamboo pile is a passive technique in which soil displacements
require to activate pile forces [3, 4].
Using bamboo piles as material for reinforcement of slope sandy soil was carried
out by Munawiret al. [5, 6]; the result of the experiment and numerical by using
PLAXIS showed that pile reinforcement applied on the slope had improved the
stability of the slope. It had indicated by the significant increase of safety factor and
bearing capacity. Wahhab and Ramadhan were used bamboo as soil reinforcement to
landslide in Kali Wungu–Boja road, Darupono Village, Kendal District, two varieties
of bamboos with a diameter of 20 and 15 cm were analyzed. The safety factor was
analyzed by using PLAXIS and SLOPE/W software. The result showed that of safety
factor was give the illustration that bamboo could use as an alternative material to
slope reinforcement [7]. Using the Dendrocalamus Giganteus species bamboo-pile
for slope stability using the finite element method (FEM) analysis of bamboo-pile
was developed PLAXIS software for 2D and 3D. The results presented that the
capabilities of the bamboo piles are potential as an alternative for slope stability [8].
Generally, there are two methods for slope stability analysis, including limit equi-
librium methods (LE), such as Ordinary, Bishop’s, Janbu’s, Spencer’s, Morgenstern-
Price, and numerical analysis methods such as finite element method (FEM). In the
LE method for slope stability analysis, the critical slip surface has to be estimated.
However, in FEM analysis, no potential slip surface was estimated in advance, and
the stress–strain correlation in the slope can also be considered [8]. In this study, two-
dimensional (2D) analyses evaluated bamboo piles using ABAQUS Student Edition
(SE) software to simulate the slope model. The software is a personal finite element
analysis tool for solving limited size problems (1000 nodes for Abaqus/Standard
and Abaqus/Explicit. This numerical study aims to analyze the performance of the
slope reinforced with inclined bamboo piles as an alternative material used in slope
reinforcement. The variation of diameter bamboo piles was modeled to evaluate the
influence of inclined bamboo piles on the safety factor (FOS).
Numerical Simulation of Slope Stability for Soil Embankment … 289
2 Shear Strength Reduction (SSR) Finite Element Method

(FEM)
Nowadays, shear strength reduction (SSR) has increasingly been used with the finite
element method (FEM) for slope stability analyses. Previous researchers have widely
used this method to analyze the slope stability of several cases [8–13].
In the SSR finite element method elastic–plastic strength is assumed for slope
materials. The factor of safety (FOS) is applied to reduce the strength of soil until
failure occurs. For Mohr–Coulomb material model shear strength reduced by a factor
of safety can be determined from the equation below.
τ c tan ∅
= (1)
FOS FOS FOS
Equation (1) can be re-written as,
τ
= c∗ + tan ∅∗ (2)
FS
where
c
c∗ = (3)
FOS

tan ∅
∅∗ = ar ctan (4)
FOS
Equations (3) and (4) are reduced Mohr–Coulomb shear strength parameters, and
these values can be input into the ABAQUS SE software and then analyzed. There are
several types of criteria for determining the failure slope model: numerical iteration
non-convergence, suddenly change displacement, plastic strain, or equivalent plastic
strain [8–13].
In the computation process, the field variables change following increment step
time. The process of reducing material strength be realized by adjusting the mate-
rial parameters to change with field variables. Since the incremental time change
is automatic and manual modification is no longer required, the strength is also
automatically decreased. Eventually, according to specific failure criteria, a suitable
increment step time is known. Then the safety of the factor and the critical sliding
surface of the slope can be found.
Fig. 1 The geometry of the slope model
3 Numerical Simulation
In the section describing step by step finite element analysis for slope model embank-
ment reinforced with inclined bamboo piles using ABAQUS SE software, including
model geometry, material properties, meshing, boundary condition, and loading as
follows:
3.1 Model Geometry
According to the previous study [14], the geometry of the homogeneous embankment
slope model was typically 20 m in height and 35 m long at the base with an inclination
of 33.69°. The bamboo pile with 5 m length and 6, 10, 15 cm for diameter were placed
two in the edge, one in the middle, and inserted perpendicular to the slope model
plane. The geometry slope reinforced with bamboo piles is shown in Fig. 1.
3.2 Material Properties
The soil material properties of slope based previous study [14], the elastic–plastic
Mohr–Coulomb failure criterion is used in the constitutive model of the soil in the
analyses. Because clay assumed to be undrained, so the modulus of elasticity (E),
Poisson’s ratio (v), cohesion (c), friction angle (ϕ), and unit weight (γ) are 200 MPa,
0.25, 10°, 20 kPa, and 20 kN/m3 respectively. In contrast the dilation angle (ψ)
assumed zero because of the saturated clay. Meanwhile, the properties of the bamboo
piles species Dendrocalamus asper are 5 m for length and 6, 10, 15 cm for diameter
with the modulus of elasticity (E) 12,000 MPa and Poisson’s ratio (v) 0.2.
Fig. 2 Meshing and boundary condition of the slope model
3.3 Meshing, Boundary Condition, and Loading
Because the ABAQUS SE has having limitations, the 2D FEM technique has been
adopted to solve the slope model in this study. There are 1185 element type CPE3
(3-node linear plane strain triangle) used for soil and type B21 (2-node linear beam
in a plane) for the bamboo pile. The bamboo piles are assumed as an embedded
elements beam in a set of solid elements. The boundary conditions (BC) on both
sides are assigned to be roller which deformation on x-direction, and the fixed is
used for the bottom. The meshing and BC of the model are shown in Fig. 2. Only
the gravity load was applied in the slope model.
4.1 The Safety of Factor (FOS)
The slope stability analyses based on SSR using ABAQUS SE software, the factor
of safety (FOS) has evaluated base criteria non-convergence of the numerical itera-
tion, suddenly change the dis-placement node control on the top slope model. The
displacement of the versus FOS has shown in Fig. 3. It illustrated that the displace-
ment changes suddenly at FOS = 1.293 for clay embankment slope unreinforced.
Meanwhile, the clay embankment slope reinforced inclined bamboo piles with diam-
eters of 6, 10, 15 cm increase 1.503, 1.512, 1.515, respectively. The main principle
of reinforcement in slopes is to provide stabilizing forces. When bamboo piles are
inserted in the slope, additional pullout resistances are provided, so this structure
changes safety of factor. The simulation results also indicate that the safety factor
depends on the diameter of bamboo piles. With the larger diameter of bamboo piles,
the pullout resistance was increased.
Fig. 3 Displacement versus FOS of the slope model
Fig. 4 The equivalent plastic strain (PEMAG) slope model unreinforced
4.2 The Equivalent Plastic Strain (PEMAG)
The bamboo piles are inserted in the slope, and it also changes the stress–strain and
failure mechanism. Figure 4 until Fig. 7 shows the equivalent plastic strain (PEMAG)
at increment 43rd for unreinforced and 158th for reinforced slope. In the figures, the
contour indicates the location of the potential failure of the slope. The different
contours shown in the figures indicate the location where the slope failure started
(Figs. 5, 6 and 7).
Fig. 5 The equivalent plastic strain (PEMAG) slope model reinforced with inclined bamboo piles
diameter 6 cm
diameter 10 cm
diameter 15 cm
5 Conclusion
According to the 2D FEM simulation of soil embankment slope reinforced with

inclined bamboo piles with the shear strength reduction (SSR) using ABAQUS SE
software, the factor of safety (FOS), and the failure mechanism of the slope were
obtained as follow:
1. The factor of safety (FOS) of the slope depends on the diameter of bamboo
piles. The clay embankment slope reinforced with inclined bamboo piles with
diameter 6, 10, 15 cm increase 1.503, 1.512, 1.515, respectively.
2. When the bamboo piles were inserted in the slope, it also changes the stress–
strain and failure mechanism. The equivalent plastic strain (PEMAG) contour
indicates the location of the potential failure of the slope.
3. The shear strength reduction (SSR) finite element method (FEM) can be used
as an alternative for slope stability analysis of soil embankment slope model
reinforced with inclined bamboo piles.
References
1. Morisco (1999) Rekayasa Bambu, Nafiri Offsett, Yogyakarta

2. Irawati IS, Saputra A (2012) Analisis Statistik Sifat Mekanik Bambu Petung. Proceeding
Simposium Sinar Bambu I, Yogyakarta
3. Kumar P, Chakrabortty P (2018) A simple analytical model for bamboo reinforced slope using
ordinary method of slices. Ind Geotech Conf
4. Davila AMJ (2019) Numerical Analysis of Bamboo Piles for Slope Stability, Geotechnical
Engineering in the XXI Century. Pontificia Universidade Católica do Rio de PUC-Rio. Brazil
5. Munawir A, Dewi M, Zaika Y, Soehardjono A (2013) Bearing capacity on slope modeling with
composite bamboo pile reinforcement. Int J Eng Adv Technol (IJEAT) 2(5)
6. Munawir A, Dewi M, Zaika Y, Soehardjono A (2013) Safety factor of continuous footing on
slope modeling with composite bamboo pile reinforcement. EJGE 18
7. Wahhab MA, Ramadhan MR (2017) Analysis of slope stability using bamboo as an alterna-
tive slope reinforcement (Case Study In Kaliwungu—Boja Street, Darupono Village, Kendal
District). In: Proceedings of international conference: problem, solution and development of
coastal and delta areas. Semarang, Indonesia
8. Ho IH (2014) Parametric studies of slope stability analyses using three-dimensional finite
element technique: geometric effect. J GeoEng 1:33–43
9. Dawson EM, Roth WH, Drescher A (1999) Slope stability analysis by strength reduction.
Geotechnique 49(6):835–840
10. Griffiths DV, Lane PA (1999) Slope stability analysis by finite elements. Geotechnique
49(3):387–403
11. Yang X, Yang G, Yu T (2012) Comparison of strength reduction method for slope stability
analysis based on ABAQUS FEM and FLAC. Appl Mech Mater 170–173:918–922
12. Dai w, Jiang P, Ding J, Fu B (2017) The influence of strength reduction method on slope
stability under different instability criteria. Int Conf Arch Eng New Mater
13. Peng X, Li J, Li Y, Jiang Y (2020) Slope instability judgment criteria in FEM based on strength
reduction method. IOP Conf Ser: Earth Environ Sci 571
14. Cai F, Ugai K (2000) Numerical analysis of the stability of a slope reinforced with piles. Soil
Found 40(1):73–84
Analysis of Community Satisfaction
Level on the Road Rehabilitation
and Reconstruction Project (Learn
from Palu Disasters Area)
Andri Irfan Rifai, Eko Prasetyo, and Rhismono
Abstract The rehabilitation and reconstruction project stages after the Palu disaster
were almost completed. The massive earthquake disaster followed by the tsunami
and the liquefaction required a relatively complex pre-during-past rehabilitation and
reconstruction series. These stages will certainly be considered successful if the
community affected by the disaster can feel performance satisfied. This research
was conducted in the project area of rehabilitation and reconstruction of roads and
bridges in Palu, Sigi, and Donggala. The research method was carried out through
a questioner distributed to all communities around the project site. The analysis is
carried out using a community profile and importance-performance analysis. The
results showed that the community got satisfaction above the average (almost 4.00).
The community is satisfied with the road’s condition and the bridge’s current state
compared to the past. Meanwhile, essential indicators that must be improved are
information and socialization and street lighting.
Keywords Community Satisfaction · Liquefaction · Rehabilitation ·

Reconstruction
1 Introduction
The earthquake that occurred in Palu—Indonesia, on September 28, 2018, caused

severe damage. Until now, this incident still leaves quite a trauma. The earthquake that
occurred has quite complete phenomena, including the movement of faults, tsunamis,
landslides, and the devastating liquefaction period. Several incidents that took place
simultaneously have paralyzed various community activities [1]. In particular, this
liquefaction phenomenon has received attention from the world. The mudflow event
A. I. Rifai (B)
Universitas Internasional Batam, Batam, Indonesia
E. Prasetyo · Rhismono
Directorate General Highway, Batam, Indonesia
298 A. I. Rifai et al.
during the liquefaction had devastated infrastructure and housing on a massive scale
and simultaneously during the incident.
Palu City and its surroundings based on topographic, geological, and seismolog-
ical conditions can suffer damage due to earthquakes, including secondary disasters
(tsunami, liquefaction, and cliff landslides). A similar earthquake with a magnitude
of 7.6 on the Richter scale occurred in Palu in 1938 with many victims [2]. Several
records of earthquake events in the last 100 years can be used as a lesson for all
stakeholders to be better prepared for earthquakes.
Apart from being a natural occurrence that is destructive and has a devastating
impact on human life and property loss, Earthquakes is also a natural laboratory on
a full scale. This can provide many benefits; among them is to study earthquakes’
characteristics to mitigate disasters in the future. Civil engineering experts have a
fundamental relationship to earthquake disasters. Most of the Indonesia’s territory
is in earthquake-prone areas. Of course, the Palu earthquake that occurred should
be used as a valuable lesson for practitioners, especially civil engineering, because
it provides a complete picture of natural disasters. Apart from having to evaluate
preparing a mitigation plan, we must immediately evaluate the rehabilitation and
reconstruction that has been carried out within the past three years. Valuable for the
continuation of humans to build an economy and a proper life order [3].
Evaluation of the rehabilitation and reconstruction process is needed to ensure that
development direction is on the right track. One of the measuring tools that can be
used is the community satisfaction level [4]. This can be used as a measure of accuracy
in carrying out trauma healing. The community’s existence and involvement can
boost the success of the preparation, implementation, and implementation processes
in implementing rehabilitation and reconstruction.
One method that is widely used to measure community satisfaction is the Impor-
tance Performance Analysis (IPA). The scientific method has proven to be a widely
applicable tool that is relatively easy to manage and interpret. Thus, it is widely
used among researchers and managers in various industries [5]. However, despite its
widespread, this IPA model still has to be correctly ascertained about its method and
implementation. The purpose of this study is to determine community satisfaction
with rehabilitation and reconstruction using the IPA method.
2 Literature Review
Palu is an active earthquake area where according to the USGS Seismicity Map, it
has an epicenter depth of less than 150 km [6]. The earthquake in the Palu region is
included in the transform zone type, which is an earthquake caused by sliding two
tectonic plates parallel to each other, but in opposite directions. The two of them
neither gave each other nor supported each other. The earthquakes in this zone are
generally shallow crustal earthquakes caused by the Palu-Koro Fault and the Matano
Fault [7].
Analysis of Community Satisfaction Level on the Road … 299
2.1 Disasters Vulnerability
Palu City is the capital city of Central Sulawesi Province below the equator with an
altitude of 0–700 m above sea level. Palu’s vulnerability to disasters is primarily the
result of the Palu-Koro fault. This fault is the main fault on the island of Sulawesi
and is classified as an active fault extending from central Sulawesi to the Karimata
Strait [8]. The fault starts from the boundary of Sulawesi waters with the Makassar
Strait to the bay of Bone. This fault crosses the city of Palu and reaches the Koro
River area. Morphologically, the Palu-Koro Fault is a left-lateral slip [9].
The Palu-Koro fault is very active with a shift in the range of 33 mm/year. Because
of its large dimensions, it is also called the Palu-Koro fault zone or system. This land
fault with a length of 250 km is the second-longest fault in Indonesia, after the large-
Sumatra fault [10]. Initially, this fault was called the Fossa Sarassina fault, and then
it was called the Palu-koro fault, Lariang in the Koro River segment (Koro Valley).
These geological and geophysical scientists familiar with the Palu-Koro fault agree
that the fault is active [11].
2.2 Rehabilitation and Reconstruction
The scope of implementation of post-disaster rehabilitation and reconstruction is

carried out through several activities. Among them are improving the environment
in the disaster area, repairing public infrastructure and facilities, helping for commu-
nity housing repairs, social-psychological recovery, health services, reconciliation
and conflict resolution, social, economic, and cultural recovery, restoring security
and order, restoring government functions, and restoration of public service func-
tions. The main objective of rehabilitation and reconstruction is to restore disturbed
community activities and public services [12].
In post-disaster management, especially in the handling of reconstruction, a proper
reconstruction process is needed. Based on sound planning, so that it is right on target
and is also orderly in the use of funds and can increase community resilience to the
threat of future disasters [13]. A good post-disaster reconstruction process must
recover the community’s condition, both physically, mentally, socially, and econom-
ically, and reduce vulnerability to disasters, not exacerbate existing vulnerability
conditions that lead to disasters.
For the reconstruction and rehabilitation process to run well, a particular standard
is needed to manage disaster management. The government (central and regional) and
non-governmental organizations and members of the public can carry out the rehabil-
itation and reconstruction process with planned, on time, on quality and budget, and
under the target [14]. Every rehabilitation and reconstruction program must prepare a
detailed technical plan, covering aspects of volume, network system, stages of work,
the cost, the technical requirements for the implementation, and the actors who can
do it.
Each rehabilitation and reconstruction program must meet specific achievement

indicators, mainly so that each component of public infrastructure and facilities can
adequately support the resumption of the communities’ social and economic life
in the disaster area. One of the critical indicators that must be considered is the
disaster’s social-psychological recovery to return to normal functions [15]. This can
be achieved by activating community elements so that they can return to normal
social functions.
2.3 Community Satisfaction
The main target of rehabilitation and reconstruction is the repair and restoration of all
aspects of public services by normalizing all aspects of governance and community
life in post-disaster areas [16]. The ability to provide essential services to the commu-
nity cannot be adequately achieved without proper planning. The shift of authority
from the central government to regional government is often not accompanied by
the transfer of responsibility for services and protection to the public. As a result,
regional government responses tend to be slow and often centrally dependent when
disasters occur. This situation becomes even more complicated if the disaster affects
more than one area.
As part of overall disaster management, the rehabilitation phase’s implementation
must be linked to other stages. In this sense, the rehabilitation phase activities relate to
the pre-disaster and emergency stages and the reconstruction phase. The relationship
and coordination between these stages dramatically determine the effectiveness and
efficiency of disaster management. Therefore, disaster management staging should
not be positioned as a goal but to achieve efficiency and effectiveness of disaster
management as a whole [17]. Based on this definition, synchronization and coor-
dination should be the keywords for disaster management that various parties must
implement. The planning, preparation, implementation of post-rehabilitation and
reconstruction processes are a series of interrelated activities.
In the implementation of rehabilitation and reconstruction, it is necessary to
mobilize resources, including human resources, equipment, materials, and funds,
by considering the available local resources. Human resources who understand and
have professional skills are indispensable in all post-disaster rehabilitation processes
and activities. Resources in the form of equipment, materials, and funds are provided
and ready to be allocated to support the rehabilitation process [18]. Recovery activ-
ities involving the community are carried out by helping the community to revive
and reactivate social, economic, and cultural activities. By actively involving the
community, it is hoped that the rehabilitation and reconstruction activities carried
out will get a high level of satisfaction.
A community satisfaction survey is needed in connection with some of the above,
especially the community satisfaction level. This is a comprehensive measurement
of the level of community satisfaction with the quality of rehabilitation and recon-
struction services provided by public service providers [19]. The survey needs to
Table 1 Community
Code Level Weight
satisfaction level
VD Very dissatisfied 1
NS Not satisfied 2
QS Quite satisfied 3
S Satisfied 4
VS Very satisfied 5
be carried out to know each indicator’s weaknesses or deficiencies in public service

providers and determine the performance of the rehabilitation and reconstruction
implementation being carried out.
3 Research Method
The research method used in this research is descriptive research. This research is
used to answer questions about what or how an event or phenomenon. The intensive
field observation of the phenomenon under study is required. This research was
conducted on the community and stakeholders in the post-disaster area of Palu.
Respondents in the community satisfaction survey were 125 respondents, but those
who returned the questionnaire and filled it in completely were 110 respondents.
3.1 Survey Instrument
For demonstration purposes, community and stakeholder data around the rehabili-
tation and reconstruction sites were collected using a structured questionnaire. The
questionnaire consists of two parts. In the first part, respondents are asked to provide
information about the respondent’s characteristics and information. The second part
contains questions about the level of importance and respondent satisfaction with
each attribute of the rehabilitation and reconstruction activity. The main attributes
of implementing rehabilitation and reconstruction are taken from existing literature
and measured by existing and tested scales adapted from several previous studies.
The assessment for the satisfaction level on the guideline using the Likert Scale
method can be seen in Table 1.
3.2 Data Analysis
The application IPA begins with the identification of attributes that are relevant to
the observed choice situation. Attribute lists can be developed using mean, median,
Fig. 1 Importance
performance analysis
or ranking measures. The aggregated importance scores and performance attributes

are classified into high or low categories; then, each attribute is assigned to one of the
four predefined performance interest quadrants by pairing the two rank sets. IPA has
been widely applied to evaluate importance and performance value in the market,
identify opportunities for improvement, and guide strategic planning efforts. In IPA,
service attributes are plotted in a two-dimensional matrix based on each attribute’s
importance and performance. The average or median importance and performance of
all attributes divide the matrix into four quadrants [20]. The priority for improvement
is then summarized based on the matrix points’ location, which can be seen in Fig. 1.
Quadrant 1 (high importance and high performance) has the management scheme
for this quadrant is continuing to work well. Quadrant 2 (low importance and high
performance) have a management scheme for this quadrant that is likely to overdo.
Quadrant 3 (low importance and low performance): having a management scheme
for this quadrant is a low priority. Whereas Quadrant 4 (high importance and low
performance) has a management scheme for this quadrant is concentrated here.
4 Data Analysis and Result
The policy for implementing rehabilitation and reconstruction in Palu is the central
government’s responsibility or local governments affected by the disaster. In carrying
out rehabilitation and reconstruction activities in the strategic and national assets, it
is directly carried out by the central government. One of them is the rehabilitation
and reconstruction of national roads and other strategic roads carried out by the
Directorate General of Highway-Indonesia, which loan from the World Bank. The
series of stages of the rehabilitation and reconstruction process can be seen in Fig. 2.
Fig. 2 Pre-during-post rehabilitation and reconstruction stage
4.1 Damage Description
The damage that occurred in Palu after the disaster varied greatly but could be
classified into several types of damage. Damage due to fault motion, fractures, and
earthquake shocks caused by fault motion is an offset where the left side moves north,
and the right side shifts south. The length of the most significant shear on the right
side is about 4 m, while the left side shifts to the north along 3 m. This shift is visible
on the map visible on googles map. Of course, buildings that are traversed by faults
will suffer significant damage and soil fractures where fractures can be the impact
of the movement of faults (or reactivated faults) with smaller offsets. Earthquake
shocks are in the form of vibrations both horizontally and vertically. In Palu City, in
general, the impact of damage due to shocks was not too much, except for buildings
of low quality.
Furthermore, the damage caused by the tsunami is due to inundation (submerged
buildings) and tsunami currents (the speed or force that acts to push or pull the
building). The impact of current velocity is mainly the scouring of the subgrade.
If it is loose sand, the erosion rate is very high. Generally, buildings with shallow
foundations fail because the scour reaches the foundation’s base, and the buildings
are relatively light, so they are easily carried away by the flow of water. Another
damage is due to the tsunami and at the same time carrying debris to cars and ships,
so collisions with these objects often result in heavy damage.
Also, there is damage due to massive liquidation with high functionality. There
are 4–5 locations that are prominent and wide, namely in Balaroa, Petobo, Jono Oge,
Lolu village (also in Jono Oge), and Sibalaya. Although some spots also occur in the
sand boil, they are not prominent and are not recorded.
Avalanche damage on several hills. Many locations on the hill suffer from local
landslides, and a notable one is in the village of Poi near South Sibalaya. This
landslide caused damage and threatened the people below it because the debris in
that location was loose and ready to slide due to the rainfall that occurred (it could
cause flash floods).
Damage due to liquefaction-induced landslides and landslides on the seabed.
Avalanches of this kind are induced by liquefaction. The landslides in Balaroa and
Sibalaya were a phenomenon of liquefaction-induced landslides. It is possible that
the submarine landslides that occurred in Palu Bay which caused the tsunami impact
had the exact mechanism as in Sibalaya.
Table 2 Sample profile

Measure Items Frequency Percentage (%)
Gender Male 82 74.55
Female 28 25.45
Age <21 2 1.82
21–25 18 16.36
26–30 15 13.64
31–35 16 14.55
36–40 44 40.00
>41 15 13.64
Education Diploma degree or 23 20.91
lower
Bachelor’s degree 74 67.27
Master’s degree 13 11.82
Profession Government 32 29.09
employees
Private employees 23 20.91
Entrepreneur 42 38.18
Unemployment 3 2.73
Student 10 9.09
4.2 Sample Profile
Based on the results of the questionnaire, the respondent’s profile can be seen in
Table 2. Among the respondents, 74.55% were men, and 25.45% were women.
Meanwhile, based on age, the respondents were dominated by the 36–40-year-old
group, with most of them being undergraduate. Furthermore, based on the type of
work, entrepreneurs occupy the most positions.
The questionnaire data obtained from 110 respondents have been tested for instru-
ment reliability using the Cronbach’s alpha value. The value of the reliability test
results can be used if it exceeds the alpha value of 0.600. Based on the reliability test
results, the value of Cronbach’s alpha for reliability importance statistics is 0.944
and for reliability performance statistics is 0.909.
4.3 Importance Performance Analysis
IPA is carried out to determine the importance of each service quality attribute based
on user perceptions of the importance of service quality. The analysis to determine
the level of importance of service quality is carried out in 3 stages: pre-stage, during-
stage, and post-stage.
Based on the questionnaire results regarding the importance level and performance
attributes of the rehabilitation and reconstruction implementation, descriptive statis-
tics are obtained on the data from all respondents in the study area, which can be
seen in Table 3. The I is importance, P is performance, and G is the gap between
importance and performance.
Based on the results of a survey of respondents in the rehabilitation and recon-
struction after the Palu Disasters area, the average value of the rehabilitation and
reconstruction series’s satisfaction level was obtained. From these values, it is known
that the indicator which has the highest level of importance according to the respon-
dents is the quality road and bridge (4.34). Meanwhile, the indicator with the highest
level of satisfaction is the current state of the road, and the bridge is compared to the
past (3.99).
In the pre-reconstruction and rehabilitation stages, all performance indicators
show an average number of 3.81. This shows that the community is quite satisfied with
the pre reconstruction and rehabilitation performance. Likewise, the average perfor-
mance during reconstruction and rehabilitation with an average performance of 3.81.
Meanwhile, post-reconstruction and rehabilitation showed a higher number, namely
3.88. This shows that the community feels the overall performance satisfaction stages
are above average.
The gap value is the ratio between importance value and performance value. When
viewed from the value gap between importance and performance values, all attributes
have a minus value, which means that no indicator can meet the community’s expec-
tations. The highest gap value is the current availability of street/environment lighting
(−0.56). The performance indicators are considered to be still far below the value
expected by the community. The indicator of the access road to residence current
state is compared to the past, including attributes with a high enough importance and
its performance is also relatively high. So, it can be concluded that respondents saw
changes in access road to residence conditions better than before.
The IPA matrix of performance rehabilitation and reconstruction indicators is
formed with an average value of importance as the X-axis and the average value
of the performance level as the Y-axis, according to Table 3. Based on the matrix,
indicators whose importance values are above the Y-axis are included in the high
importance category (included in quadrants I and II). In Fig. 3, we can see the IPA
matrix for each stage.
Based on the IPA matrix in Fig. 3, it can be seen that in the pre-reconstruction and
rehabilitation stages. There are three indicators in quadrant I (keep up the good works)
1 indicator in quadrant II, three indicators in quadrant III, and one indicator in quad-
rant IV. From the questioners’ results, the indicator in quadrant IV (concrete here)
should get more attention, namely information and socialization about reconstruction
and rehabilitation. While at the stage during rehabilitation and reconstruction, there
are no indicators that require special attention. Furthermore, there is no indicator in
quadrant IV at the post-reconstruction and rehabilitation stage, but the indicator of
the current street/environment lighting availability is quite interesting. It can be seen
that the community expects more on the availability of street lighting.
Table 3 The average level of community satisfaction

ID Satisfaction indicator I P G
A. Before the rehabilitation and reconstruction
A1 Information and socialization 4.11 3.71 −0.40
A2 The time the reconstruction program began 4.06 3.85 −0.21
A3 Road and bridge damage identification 4.22 3.85 −0.37
process
A4 Participation in the reconstruction and 4.07 3.75 −0.32
rehabilitation process
A5 Collaboration between local communities in 4.07 3.78 −0.29
reconstruction and rehabilitation
A6 The wishes of the people are fulfilled 4.10 3.90 −0.20
A7 Easy administration/disbursement process 3.84 3.62 −0.22
A8 The role of government in the 4.30 3.99 −0.31
reconstruction process
Mean A 4.10 3.81 −0.29
B. During the rehabilitation and reconstruction
B1 The Role of the Facilitator in the 4.05 3.78 −0.27
reconstruction and rehabilitation process
B2 Labor availability 4.08 3.77 −0.31
B3 Work experience and skills 4.23 3.86 −0.37
B4 Availability of material for reconstruction 4.28 3.85 −0.43
and rehabilitation
B5 Quality material available for reconstruction 4.25 3.88 −0.37
and rehabilitation
B6 Quality of road and bridge 4.34 3.92 −0.42
B7 Community participation in the 3.93 3.62 −0.31
reconstruction and rehabilitation
Mean B 4.17 3.81 −0.35
C. After the rehabilitation and reconstruction
C1 With the results of existing assistance 3.98 3.69 −0.29
C2 The current state of the road and bridge is 4.17 3.99 −0.18
compared to the past
C3 The road and bridge become more 4.25 3.94 −0.31
earthquake-resistant
C4 The comfort of road and bridge compared to 4.15 3.94 −0.21
before
C5 The quality of the road and bridge now 4.17 3.97 −0.20
compared to before
C6 The road and bridge was had been as a 4.22 3.96 −0.26
community wish
(continued)
Table 3 (continued)
ID Satisfaction indicator I P G
C7 Satisfaction with the current design 4.15 3.95 −0.20
C8 The access road to residence compared to 4.11 3.95 −0.16
before the reconstruction and rehabilitation
C9 Current availability of street/environment 4.11 3.55 −0.56
lighting
Mean C 4.15 3.88 −0.26
Grand mean 4.14 3.84 −0.30
Fig. 3 IPA matrixes for each stage
5 Conclusion
From the data analysis that has been carried out, it can be concluded that the commu-
nity gets satisfaction above the average for all stages. Consecutively starting from
the pre-during-post reconstruction and rehabilitation, the average performance satis-
faction level obtained was 3.81; 3.81; and 3.88, with an overall average of 3.84. The
satisfaction performance almost met the satisfied category (almost 4.00). An inter-
esting indicator is that the community is satisfied with the access road to residence
current state compared to the past. This is indicated by the minimal gap between
importance and performance. This study also found several important things that
must be improved: information and socialization and street/environment lighting.
This is consistent with several references that state that socialization is an important
stage in rehabilitation and reconstruction.
Acknowledgements The authors are grateful to the editor and reviewers for constructive comments
on the earlier version of the paper, the Rehabilitation and Reconstruction Project, which supported
this research. We would also liketo thank those in Universitas Internasional Batam, Indonesia, who
assisted in this study.
References
1. Yulianto E, Utari P, Satyawan IA (2020) Communication technology support in disaster-prone

areas: Case study of earthquake, tsunami and liquefaction in Palu, Indonesia. Int. J Disaster
Risk Redu 45:101457
2. Rajindra R, Suardi I, Sabara Z, Pushpalal D, Samad MA, Yani A, Umam R (2019) Diver-
sity, resilience, and tragedy: three disasters in Palu of Indonesia. Int J Innovat Creat Change
5(2):1592–1607
3. Liu H, Zhang D, Wei Q, Guo Z (2017) Comparison study on two post-earthquake rehabilitation
and reconstruction modes in China. Int J Disaster Risk Reduct 23:109–118
4. Wijegunarathna EE, Wedawatta G, Prasanna LJ, Ingirige B (2018) Long-term satisfaction of
resettled communities: an assessment of physical performance of post-disaster housing. Proced
Eng 212:1147–1154
5. Hosseini SY, Bideh AZ (2014) A data mining approach for segmentation-based importance-
performance analysis (SOM–BPNN–IPA): a new framework for developing customer retention
strategies. Serv Bus 8(2):295–312
6. Rusydi HE (2017) Vulnerability zoning of earthquake disaster of Palu. Int J Sci Appl Sci Conf
Ser 1(2), 137–143
7. Jaya A, Nishikawa O, Jumadil S (2019) Distribution and morphology of the surface ruptures
of the 2018 Donggala-Palu earthquake, Central Sulawesi, Indonesia. Earth, Planets Space
71(1):1–13
8. Rifai AI, Hendra H, Prasetyo E (2020) Data mining applied for liquefaction mapping and
prediction learn from Palu Earthquakes. Civil Eng Arch 507–514
9. Patria A, Putra PS (2020) Development of the Palu-Koro Fault in NW Palu Valley Indonesia.
Geosci Lett 7(1):1–11
10. Natawidjaja DH (2018) Updating active fault maps and sliprates along the Sumatran Fault
Zone, Indonesia. IOP Conf Ser: Earth Environ Sci 118(1):012001
11. Massinai MF, Ariansyah MR, Trisartika R, Massinai MA (2020) Relative present-day motion
on Palu-Koro Fault. Earth Environ Sci 575(1):012185
12. Moreno J, Shaw D (2019) Community resilience to power outages after disaster: a case study
of the 2010 Chile earthquake and tsunami. Int J Disaster Risk Reduct 34:448–458
13. Lee DW (2020) The expertise of public officials and collaborative disaster management. Int J
Disaster Risk Reduction 50:101711
14. Tagliacozzo S (2018) Government agency communication during post-disaster reconstruction:
Insights from the Christchurch earthquakes recovery. Nat Hazard Rev 19(2):04018001
15. Asadzadeh A, Kötter T, Salehi P, Birkmann J (2017) Operationalizing a concept: The systematic
review of composite indicator building for measuring community disaster resilience. Int J
Disaster Risk Reduction 25:147–162
16. Nemeth DG, Kuriansky J, Onishi Y (2021) The 3 Es of psychosocial recovery after Disaster.
The Intersection of Trauma and Disaster Behavioral Health Springer, Cham. pp 297–316
17. Yong Z, Zhuang L, Liu Y, Deng X, Xu D (2020) Differences in the disaster-preparedness
behaviors of the general public and professionals: evidence from Sichuan Province, China. Int
J Environ Res Public Health 17(14):5254
18. Santha SD (2018) Social interfaces in disaster situations: analyzing rehabilitation and recovery
processes among the fisherfolk of Tamil Nadu after the Tsunami in India. In The Asian Tsunami
and post-disaster aid. Springer, Singapore, pp 65–78
19. Sofyan M (2019) Community satisfaction of the Urban flood control system improvement
project (UFCSI). Ilomata Int J Soc Sci 1(1):29–34
20. Chang RD, Zuo J, Soebarto V, Zhao ZY, Zillante G, Gan XL (2017) Discovering the transition
pathways toward sustainability for construction enterprises: Importance-performance analysis.
Journal of construction engineering and management, 04017013
Seepage and Piping Control of Earth Fill
Dam
(Case Study of Pidekso Dam Indonesia)
Muhammad Zainal Arifin, Yusep Muslih Purwana,

Abstract The Pidekso Dam is one of the major dams in the Central Java, situated
in Wonogiri Indonesia. The stability analysis of the dam is required due to this dam
is planned for national water and food security program covering the irrigation area
of 15 × 106 m2 , increasing the planting intensity from 20 × 106 m2 to 36 × 106 m2
and as a source of water around 300 L/s for Wonogiri, Sukoharjo, Solo and their
surroundings. The control is needed for the safety of the dam from seepage and
piping. Excessive seepage can cause piping and lead to dam failure. Seepage control
is carried out by 3 alternatives, extended curtain grouting, cut off wall and upstream
blanket respectively. Each alternative is divided into 4 models: (1) original length
(L), (2) 2/3L, (3) 1/3L and (4) 2/3L with doubles in the upstream and downstream
of the dam. The result indicates that, the installation curtain grouting with model 1
is the most effective alternative to reduce seepage up to 67.82% of dams.
Keywords Cut off wall · Curtain grouting · Seepage control · Upstream blanket
1 Introduction
The Pidekso Dam is one of the large dam in the Central Java area, located in Pidekso
Village, Giriwoyo District, Wonogiri Regency. The map of the study location can
be seen in Fig. 1. The stability analysis of the dam is required due to this dam is
planned for national water and food security program covering the irrigation area of
15 × 106 m2 , increasing the planting intensity from 133% (20 × 106 m2 ) to 240%
(36 × 106 m2 ) and as a source of water 300 L/second for Wonogiri, Sukoharjo, Solo
and their surroundings [1].
M. Z. Arifin · Y. M. Purwana (B) · R. H. Dananjaya (B)

Department of Civil Engineering, Universitas Sebelas Maret, Surakarta, Indonesia
M. Z. Arifin
312 M. Z. Arifin et al.
Fig. 1 Study location map
Dam stability studies need to be carried out in the construction and planning of
seepage and piping. Excessive seepage can cause piping and lead to dam failure.
For example, the dam failure that occurred in the Lempake-Indonesia dam and the
Brumadinho-Brasil dam was due to the dam’s condition which had seepage due
to piping resulting in flooding which caused material and non-material loss and
casualties [2, 3].
Embankment dams are constructed across rivers and valleys to retain water for
flood protection, water storage and energy generation purposes. Embankment dams
consist of homogeneous dams (constructed using one type of soil) or zonal dams
(consisting of more than one soil type with high hydraulic permeability different)
[4]. Seepage is a flow that continuously flows from upstream to downstream. This
water flow constitutes the flow from the reservoir water through the material which
passes the water (permeable), either through the body of the dam or the foundation
[5]. Piping is a kind of movement or erosion of fine particles in soil along inter-
space channel between skeleton grains, by the action of seepage [6]. Costa (1985)
reported that approximately 34% of dam failures were caused by overtopping, 30%
by foundation defects, and 28% by piping [7]. If seepage is considered as a poten-
tial problem, countermeasures—such as filters, drains, clay blankets and flatter side
slopes–introduced at the design stage can reduce any risks to a minimum [8].
Seeing the condition of the study area and the many objectives of the Pidekso
Dam development, it is necessary to carry out detailed technical planning. Among
other things, regarding the geological condition of the dam foundation, foundation
repair, seepage capacity, the possibility of piping, and the stability of the dam body in
various conditions. So that through this research entitled “seepage and piping control
of earth fill dam”. So that a dam that is safe against seepage and the piping can be
planned.
Seepage and Piping Control of Earth Fill Dam … 313
1.1 Geotechnical Site and Construction Material
The body of the Pidekso Dam is a embankment dam with an upright core in the
centre, a dam length of 387 m stretching from the Southeast to the Southwest, and
a maximum height of 32.00 m from the riverbed or 44.00 m from the bottom of
the excavation. The elevation of the top of the dam is designed at El.189.00 m
with a freeboard height from the normal water level (El.185.0 m) [9]. Image of the
cross-section of the dam can be seen in Fig. 2.
The embankment material used for the Pidekso Dam consists of a watertight,
random, filter, transition and rip-rap core which is divided into zones as seen in
Table 1.
2 Methodology
Seepage analysis that indicates the occurrence of piping, is determined based on the
safety factor for the piping as follows [10]:
I cr
SF = (1)
I cal
Elv (m)
200
190
180
Upstream Downstream
170
160
150 Type of embankment material:

1. Waterproof core
140 2. Filter
3. Transition
130
4. Random land
5. Rip-rap
Fig. 2 Geometry section of the Pidekso Dam
Table 1 Dam material input data

Material Earthfill Randomfill Rockfill Filter Transition Foundation
Classification ML SC GM SM-SC SM-SC CH
Permeability (m/s) 10–7 5.86 × 10–6 – 10–5 10–6 5 × 10–6
where:
Icr = Critical hydraulic gradient (dimensionless).
Ical = Output gradient from seepage analysis results (dimensionless).
SF = Safety factor for symptoms piping > 4 (dimensionless).
Gs − 1
I cr = (2)
1+e
where:
Gs = specific gravity.
e = porosity number.
The Laplace equation is the basis of mathematical equations for several models
and methods used in seepage analysis and is used as a basis for modelling seepage
flow in the case of steady flow conditions [11].
∂ 2μ ∂ 2μ ∂ 2μ
+ + 2 = 0. (3)
∂x2 ∂γ 2 ∂z
Fortunately, now that computers and numerical software tools are so readily avail-
able, it is no long necessary to rely on graphical solutions to the Laplacian equation.
With a software tool like SEEP/W these types of situations can be readily consid-
ered in a seepage analysis [12]. Seepage analysis was carried out on 3 alternative
installations of extended seepage controls as follows:
1. Alternative 1: installations of extended seepage control with cut off wall.
2. Alternative 2: installations of extended seepage control withurtain grouting.
3. Alternative 3: installations of extended seepage control with upstream blanket.
From each alternative is divided into 4 models: (1) original length (L), (2) 2/3L,
(3) 1/3L and (4) 2/3L with doubles in the upstream and downstream of the dam. This
analysis is calculated against normal water level conditions, a more detailed sketch
can be seen Fig. 3.
Fig. 3 a Extended cut off wall. b Extended curtain grouting. c Extended upstream blanket
The seepage discharge obtained from the modelling results with Seep/W was
analyzed against the allowable seepage discharge. The permissible seepage discharge
is less than 0.05% mean annual run-off [13]. Safety factor against piping analyzed
based on Eqs. (3) and (4), while the hydraulic slope value of the flow and foundation
material obtained from modelling and data from soil investigation and testing.
3.1 Without Extended Seepage Control
Seepage analysis in conditions without extended seepage control is carried out to

determine the seepage discharge and piping. Seepage discharge analysis results and
exit gradient by using the Seep/W program presented on Fig. 4. From Fig. 4, got exit
gradient in the zone around the base of the core zone is 0.15. The critical hydraulic
(a)
5.59 x 10-5 m3/sec/m
0.15
(b)
Fig. 4 a Dam sketches without extended seepage control. b Seepage and exit gradient analysis
magnitude of this embankment material is equal to 0.948, so the size of the Safety
Factor is 6.32, which is greater than allowed. So it can be concluded that the dam
safe against the dangers of piping.
The amount of discharge per time per length of the dam results from the analysis
as amounting to 5.59 × 10–5 m3 /s/m. With the length of the dam section of 250 m, the
amount of seepage discharge for the entire length of the dam is: 5.59 × 10–5 × 250
= 1.4 × 10–2 m3 /s or 1206.60 m3 /day. This discharge is smaller from 0.05% of the
effective volume of the dam per day, which is 8535 m3 /day, so it can be concluded
that the dam without extended seepage control safe against volume seepage.
3.2 Curtain Grouting
Seepage analysis with extended seepage control of curtain grouting was carried out
on the located at the base of the dam core. The determination of the curtain grouting
depth model is based on empirical calculation according to Eq. 4 which is then
tabulated in Table 2 [14].
D = 1/3h + c (4)
Where:
D = depth of the curtain grouting.
H = height of the dam, EL.189 - EL.153 = 36 m.
C = coefficient that varies between 8 to 20, the value is 18.
The magnitude of the seepage discharge and the piping hazard is analyzed
according to the appropriate model Table 2 using the Seep/W program which can be
seen in Fig. 5 which produces the amount of seepage discharge and exit gradient.
The seepage discharge value is then calculated based on the seepage permit discharge
and the exit gradient value then calculated using Eq. (3). These results are more fully
tabulated in Table 3.
Based on Table 3, Overall, the seepage discharge for the dam with extended
seepage control of curtain grouting at the base of the foundation is still less than
the allowable seepage permit discharge and safe against piping hazards. A complete
comparison of the grouting curtain length and the seepage discharge can be seen in
Fig. 5.
Table 2 Length model of curtain grouting extended seepage control

Model Information Length (m)
Model 1 Length (3/3) of empirical calculation of Eq. 4 30
(a) 5.59 x 10-5 m3/sec/m
46 0.15
(b) 1.80 x 10-5 m3/sec/m
30
0.15
(c) 3.22 x 10-5 m3/sec/m
30
0.15
Fig. 5 Exit gradient analysis of a upstream blanket, D = 46.75 m. b Curtain grouting, D = 30 m.

c Cut of wall, D = 30 m
Table 3 Result of curtain grouting extended seepage control

Model Length (m) Seepage (m3 /day) SF Piping Seepage reduction (%)
Model 1 30 388.3032 6.32 67.82
Model 2 20 472.4784 6.32 60.84
Model 3 10 647.5032 6.32 46.34
3.3 Cut off Wall
Seepage analysis with extended cut off wall seepage control was carried out on the cut
off wall depth model which is located at the base of the dam core. The determination
Table 4 Result of cut of wall extended seepage control

Model Length (m) Seepage (m3 /day) SF Piping Seepage reduction (%)
Model 1 30 694.764 6.32 42.42
Model 2 20 985.3272 6.32 18.34
Model 3 10 1143.396 6.32 5.24
Model 4 20 801.2088 6.32 33.60
800
Seepage (m3/day)
600
400
y = -12.96x + 761.96
200 R² = 0.9607
0
0 5 10 15 20 25 30 35
Length of Curtain Grouting (m)
Fig. 6 Variation of curtain grouting verses seepage
of the cut off wall depth model is based on the empirical calculation of the cut off
wall depth according to Eq. 4 which is then tabulated in Table 2.
tabulated inward Table 4.
Based on Table 4, Overall, the seepage discharge for the dam with extended
seepage control of cut off wall at the base of the foundation is still less than the
allowable seepage discharge and safe against piping hazards. A complete comparison
of the cut off wall length and the seepage discharge can be seen in Fig. 6.
3.4 Upstream Blanket
Seepage analysis with extended upstream blanket seepage control is carried out on
the upstream blanket depth model which is located at the base of the dam core.
Determination of the upstream blanket depth model based on empirical calculation
of the upstream blanket depth according to Eq. 5 according to SNI 8065: 2016 which
is then tabulated in Table 5 [15].
√
2
x= (5)
Kb
Z b.K f.Z f
Where:
Zb = thickness of the waterproof base (m).
= 1 m.
Kb = K coefficient of waterproof base.
= 1 × 10–7 (m/s).
Kf = K coefficient of foundation (m/s).
= 1 × 10–5 (m/s).
Zf = depth of the water graduated layer on the foundation (m) = 9.93.
tabulated in Table 6.
Table 5 Length model of upstream blanket extended seepage control

Model Information Length (m)
Model 1 Length (3/3) of empirical calculation of Eq. 5 46.75
Model 4 Length (2/3) of empirical calculation of Eq. 5 with double points 31.17
upstream and downstream
Table 6 Result of upstream blanket extended seepage control

Model Length (m) Seepage (m3 /day) SF piping Seepage reduction (%)
Model 1 46.75 1109.5704 6.32 8.04
Model 2 31.17 1136.7216 6.32 5.79
Model 3 15.58 1170.2664 6.32 3.01
Model 4 31.17 1153.2888 6.32 4.42
Seepage (m3/day) 1500
1000
y = -22.432x + 1389.8
500 R² = 0.9717
0
0 5 10 15 20 25 30 35
Length of Cut of Wall (m)
Fig. 7 Variation of cut off wall verses seepage
1180
Seepage (m3/day)
1160
1140
y = -1.9476x + 1199.5
1120 R² = 0.9963
1100
0.00 10.00 20.00 30.00 40.00 50.00
Length of Upstream Blanket (m)
Fig. 8 Variation of upstream blanket verses seepage
Based on Table 6, Overall, the seepage discharge quantity for the dam with
extended seepage control of upstream blanket at the base of the foundation is still
less than the allowable seepage discharge quantity and safe against piping hazards.
A complete comparison of the upstream blanket length and the seepage discharge
can be seen in Fig. 7.
3.5 Comparison of Analysis Results
The following is a recapitulation of the results of various extended seepage control

installations models Curtain Grouting, Cut Off Wall and Upstream Blanket with a
variety of models Length which can be seen in Fig. 9.
1400
1200
Seepage (m3/day) 1000
800
600
400
200
0
Cut off Wall Curtain Grouting Upstream
Blanket
Model 1 Model 2 Model 3 Model 4
Model 1 original length (L)

Model 2 2/3L
Model 3 1/3L
Model 4 2/3L with doubles in the upstream and downstream of the dam
Fig. 9 Comparison of seepage discharge results from the installations of extended seepage control
4 Conclusion
Based on the results of the stability dam to the amount of seepage and safety against
piping, the following conclusions can be made:
1. The stability of the Pidekso Dam without and with the installations of extended
seepage controls to the safety of the seepage magnitude is safe, which is less
than the allowable seepage discharge and safe against of the piping.
2. The longer the addition of curtain grouting, cut off walls and upstream blankets
on the foundation of the dam, so the smaller the seepage discharge from the
dam
3. Of the several alternatives extended seepage control, then alternative curtain
grouting 30 m is the most effective alternative to reduce seepage up to 67.82%
of dams.
References
1. Engineering Consultant (2011) Technical Specifications Report, PT. Mettana, Wonogiri

2. Nyoman N (2014) Hydraulic analysis of flood due to dam failure (Case Study of Lempake
Dam, Samarinda City). Thesis, Gadjah Mada University. Yogyakarta.S
3. Raman (2019) An investigation of the Brumadinho Dam Break with HEC RAS simulation.
Journal on arXiv:1911.05219v1, New Jersey. USA
4. Molla E (2019) Seepage through homogeneous earth dams provided with a vertical sheet pile
and formed on impervious foundation. Ain Shams Eng J 10:529–539
5. Kanchanha (2015) Adequacy of seepage analysis in core section of the earthen dam with
different mix proportions. In: International conference on water resources, coastal and ocean
engineering (ICWRCOE 2015), Aquatic Procedia, vol 4. India, pp 868–875
6. Yongbiao (2012) Prediction methods to determine stability of dam if there is piping. In: Inter-
national conference on mechanical, industrial, and manufacturing engineering, IERI Procedia,
vol 1. China, pp 131–137
7. Sheng-shui (2019) Numerical modeling of earthen dam breach due to piping failure. Water
Science and Engineering 12(3):169e178
8. Andreaa (2015) Unsaturated Slope Stability and Seepage Analysis of a Dam. Sustain Sol
Energy Environ Energy Procedia 85:93–98
9. Planning Consultant (2011) Final Report of the FS Pidekso Dam Review in Wonogiri Regency.
PT. Virama Karya, Wonogiri
10. Peng S (2020) Inverse analysis of laboratory data and observations for evaluation of backward
erosion piping process. J Rock Mech Geotech Eng 12:1080e1092
11. Shivakumar (2015) Seepage and stability analyses of earth dam using finite element method.
In: Proceedings of international conference on water resources, coastal and ocean engineering
(ICWRCOE 2015), Aquatic Procedia, vol 4. India, pp 876–883
12. GEO-SLOPE International Ltd (2012) Seepage Modeling with SEEP/W. In: An engineering
methodology July 2012 Edition. Canada
13. National Standardization Agency (2002) Procedure for dam body design. RSNI T-01–2002.
Jakarta
14. Water Resources and Construction Education and Training Center (2005) Guidelines for dam
grouting. Ministry of Public Works, Jakarta
15. National Standardization Agency (1992) Methods of analysis and control method of seepage
water for urugan-type dam. SNI 8065: 2016. Jakarta
Maintenance Management of Buildings
and Infrastructures
State-of-the-Art of Artificial Intelligence
Methods in Structural Health Monitoring
I. G. E. A. Putra
Abstract Development of technology nowadays occur in every aspect of life,

including in computational technology. Artificial Intelligence (AI) over the past years
became trending topic in field related to computation as it is mimic the way human
think. This is led into effective problem-solving execution for a complex problem.
Solution based on AI method is proven to be time-efficient compare with traditional
ways and require significantly less resource to conduct experiment. AI method is
able to perform a decision making based on series of given data with relatively short
time, less error and efficient in overall computation. The objective of this paper is to
investigate what is the most popular AI method and with the intended use in Struc-
tural Health Monitoring (SHM) by referring to reliable publications over the past
decade. This paper shows the Neural Network is the most popular AI method in
terms of SHM with 38% from total research over the past decade which is used as a
performance evaluation.
Keywords Structural health monitoring · Artificial intelligence · Machine

learning · Neural network · Pattern recognition
1 Introduction
Structural health monitoring in structural engineering can be defined as a process of

performing damage detection which involves observation of structural system over
time. The structural system health is determined based on the analysis result of period-
ically measured dynamics response. It is also includes extraction of damage-sensitive
feature and feature statistical analysis [1]. Damage that occurs on the structure can
be detected either through parametric method (fundamental frequency, mode shape
stiffness and damping) or non-parametric method [2]. In assessing the parametric
performance of existing buildings, ASCE 41-17 provides three evaluation proce-
dures; Tier 1: Screening, Tier 2: Deficiency-Based Evaluation and Retrofit and Tier
I. G. E. A. Putra (B)
Department of Civil Engineering, Universitas Tarumanagara, Jl. S. Parman No. 1, Jakarta 11440,
Indonesia
326 I. G. E. A. Putra
3: Systematic Evaluation and Retrofit [3]. It is expected that the use of artificial intel-
ligence methods in the field of structural health monitoring would make the work
easier in making efficient decision [4].
Research on artificial intelligence has been conducted since the term “Artificial
Intelligence” was founded at a meeting at Dartmouth College in 1956. Artificial intel-
ligence consists of several disciplines such as computer science, cybernetics, infor-
mation theory, psychology, linguistics and neuropsychology. The main purpose of
artificial intelligence concept is to imitate the way of the human brain executing intel-
ligence functions for a certain thing [5]. Recently, there has been numerous research
on artificial intelligence in the world of structural health monitoring. Sun et al. [6]
conducted a study on bridge structural health monitoring using neural networks and
extreme machine learning. Neural networks are also used in predicting the effective-
ness of the transfer length of precast concrete by considering the strands characteris-
tics, concrete properties, the detailed geometry and the manufacturing methods [7].
Chun et al. [8] used the supervised machine learning to evaluate damage by measuring
the velocity of the structure at several points. Evaluation of steel connections is also
performed using a neural network to investigate the behavior of innovative C-shaped
shear connectors or tilted angle connections [9]. Structural evaluation which involves
a combination of artificial intelligence methods with commonly used methods. In
this case the Monte Carlo method also occurs in the investigation of possible bridge
failure due to the fatigue failure in overloaded traffic [10]. Hung et al. [11] estimated
the ultimate load factor on a nonlinear inelastic steel truss with the help of a deep
learning algorithm. While Liu et al. [12] developed a deep learning algorithm to auto-
matically detected damage after a disaster on a concentrically braced frame structure
with seismic loading.
This paper focuses on investigation of artificial intelligence method used in the
field of structural health monitoring for the past decade. A general description
regarding each artificial intelligence method is provided along its application in
the structural health monitoring.
2 Research Method
In this research, several papers are used as references. They have been checked for
the conformity to the research purpose. The papers that conform then are being inves-
tigated with the purposes, the artificial intelligence method used and the mechanism
action in structural health monitoring. The reference used in this research has been
reviewed and approved by the publisher. Because it has been peer-reviewed, the
quality of the reference is proven. Reference articles are obtained from the reliable
sources. The literature review process in this study is as follows:
• From academic sites on the internet, a search was made for Candidate papers. The
sites used must have proven quality such as Scopus, Wiley Online Library, ASCE
Library, Sage, Springer and Elsevier.
State-of-the-Art of Artificial Intelligence Methods … 327
• Several keywords used for search candidate paper were “Structural health monitor-
ing”, “Artificial intelligence structural engineering”, “Artificial intelligence struc-
tural health monitoring”, “Machine learning structural health monitoring”, “AI
methods in structural health monitoring”, “Deep learning structural health moni-
toring” dan “Neural network structural health monitoring”. Papers that will be
used as data are those published in the period 2010–2020.
In the process of assessing the contents of the artificial intelligence method, the
quantitative and qualitative methods are used. By using quantitative data analysis
methods, the popularity level of the artificial intelligence method in terms of struc-
tural health monitoring will be obtained. By using qualitative methods, the common
utilization of AI methods in SHM will be obtained.
In this study, more than 256 samples of data were obtained from journals that had
been indexed by Scopus and had proven quality, such as Scopus, Wiley Online
Library, ASCE Library, Sage, Springer and Elsevier and 101 papers were selected.
Samples selected from 2010 to 2020 obtained 38% using the Neural Network method,
27% Machine Learning, 17% Support Vector Machine, 8% Pattern Recognition, 5%
Principal Component Analysis and the other 5% are Bayesian Network (see Fig. 1).
Application of AI Methods in SHM
Neural Network
Support Vector Machine

27%
38%
Pattern Recognition
5% Principal Component Analysis
5%
Bayesian Network
8%
17%
Machine Learning
Fig. 1 Application of AI methods in SHM

Application of AI Methods in SHM
20%
Performance Evaluation
45%
Damage Detection
Damage Prediction
35%
Fig. 2 Common usage of AI methods
The data above is then checked for contents to determine the purpose of using
the artificial intelligence method. Generally categorized as; performance evaluation,
damage detection and damage prediction. The data obtained are 45% for performance
evaluation, 35% for damage detection and 20% for predicting damage (see Fig. 2).
3.1 Application of Artificial Intelligence Methods
Neural Network
Posenato et al. [13] implement Neural Network method to interpret anomalous
behaviour of Riccolo viaduct located on Swizz. The result of analysis then compared
to other statistical methods in which neural network superior toward others. An inte-
grated multi-objective harmony search with artificial neural networks also proposed
to minimize the run time the analysis in obtaining a cost optimal of a post-tensioned
concrete box girder road bridge [14]. Several bridges performance evaluation also
conducted using neural network methods [15–18]. Siddique et al. [19] performs a
comparative performance evaluation of a model to predict the 28 days compressive
strength of self-compacting concrete using neural network method while Saridemir
et al. [20] evaluate its tensile strength. Dantas et al. [21] develop a neural network
algorithm predict the compressive strength of concrete specimen at the age of 3, 7, 28
and 91 days which consist of demolition waste material. A concrete dam also identi-
fied to be analyzed using neural network method in order to evaluate its performance
under thermal effect [22].
An early damage detection of International Bridge, a cable stayed bridge which

connects Portugal and Spain, was performed using neural network [23]. Norman
et al. [24] identify the performance of Creek Bridge in Canada, a long span bridge,
and develop genetic algorithm to detect anomaly which often related with damage
(cracks) on a certain part of the structure [25–27]. Concepcion II et al. [28] develop a
thermal-vibration approach using hybrid soft computing using neural network algo-
rithm to a reinforced concrete bridges experimental model to detect anomaly during
its critical condition. A damage detection action also performed for a composite
cantilever beam using signal which obtained from an optical fiber sensor supple-
mented using back propagation neural network [29, 30]. Ng et al. [31] perform a
damage detection on an experimental model of 5 story steel structure acted as a
benchmark structure using pattern recognition supplemented by Bayesian designed
artificial neural network, which successfully identify its damage.
Mondoro et al. [32] develop an optimal risk-based management strategy of bridges
in coastal region to predict required repair action to address the deterioration perfor-
mance under traffic load and retrofit action by developing a neural network algorithm.
By this study it is expected to minimize the potential failure during the occurrence
of hurricanes. A framework of an autoencoder-based was developed to learning a
mapping between the characteristic of vibrations and damages characteristics using
Artificial Neural Network and Deep Learning method. Common usage of Neural
Network can be seen on Fig. 3.
Yan et al. [33] develop a support vector machine to predict the elastic modulus of
normal and high strength concrete with the introduction of insensitive loss function
based on learning theory. Similar research also developed to investigate the compres-
sive strength of high-performance concrete using time-weighted fuzzy support vector
machine inference model [34] and genetic weighted pyramid operation tree [35].
Fig. 3 Common usage of

neural network Neural Network
26% Performance
Evaluation
49% Damage
Detection
Damage
Prediction
25%
Support vector machine also used in diagnosis the damage [36] located in Wuhan
[37]. Santos et al. [38] present a damage detection action by performing feature
extraction, data normalization and statistical classification using a support vector
machine method for Bridge in Portugal. A combination of support vector machine
with the likes of pattern recognition method performed to investigate an experimental
model for bridges [39, 40]. Radhika et al. [41] conduct a research of damage detection
caused by cyclone by combining support vector machine method and pattern recog-
nition technique. Several damage predictions also performed such as harnessing the
structural data to recovery of a random missing vibration responses time-history by
Yang et al. [42]. In bridge structure, calculation of damage in cable stayed bridge by
analyzing its natural frequency and comparing the result of support vector machine
analysis in the numerical example which shows its efficiency in the process [43].
Concrete dams also use support vector machine to predict the damage possibility.
SVM benefited the user who’s working on a high dimensional space, in specifying
a decision function it uses a subset of training data and it is versatile since it is able
to use a different kernel function and make customization out of it. Common usage
of support vector machine can be seen on Fig. 4.
Pattern Recognition
Abdollahzadeh et al. [44] use a pattern recognition technique to investigate the
performance of beam column joint due to behavior of extreme non-elastic response,
constant changes in roughness and resistance by using two models of the structure.
Several simulation models also investigated using the same method such as image-
based damage detection [45] and structural behavior based on learning algorithm.
Lin et al. [46] perform a damage detection of a prestressed concrete using pattern
recognition technique. Elwood et al. [47] also conduct similar research about damage
detection in concrete structures using simulation model. Binary data is utilized to
identify damage using an image-based pattern recognition which generated from
self-powered sensor networks. A three stories bookshelves structure is analyzed
Performance
27% Evaluation
36%
Damage
Detection
Damage
Prediction
37%
Fig. 4 Common usage of support vector machine

using experimental model to identify its damage by utilizing its frequency response
spectrum [48]. Common usage of pattern recognition can be seen on Fig. 5.
Principal Component Analysis
Datteo et al. [49] present a study of structural health monitoring of G. Meazza stadium
in the form of simulation model to detect any possibilities of damage in the future by
applying principal component analysis. Damage prediction using principal compo-
nent analysis also conducted by Balsamo et al. [50] in a simulation model using small
training dataset. Reddy et al. [51] using principal component analysis to detect crack
occurs in wind turbine blade based on data of wind turbine in Sweden in 1997–2005.
Principal component analysis was investigated on its usage purpose as perfor-
mance evaluation method on structural health monitoring which often used together
as a subset to other Artificial Intelligence method. Common usage of principal
component analysis can be seen on Fig. 6.
Pattern Recognition
17%
Performance
33% Evaluation
Damage Detection
Damage Prediction
50%
Fig. 5 Common usage of pattern recognition
Principal Component Analysis

Performance
Evaluation
27%
37% Damage Detection
36% Damage Prediction
Fig. 6 Common usage of principal component analysis

Bayesian Network
Performance
Evaluation
29% 28%
Damage Detection
Damage
43%
Prediction
Fig. 7 Common usage of Bayesian network
Bayesian Network
Bayesian network method used as fault diagnosis on a rotary machinery in an indus-
trial application since this kind of particular method has become an interesting topic
for academia throughout the world [52]. A study of cyclone damage detection using
satellite image and processed using Bayesian network. An optimization technique
using Bayesian network as a subset of artificial network is conducted in order to
obtain an optimal damage detection result. A methodology to diagnosis system level
using Bayesian network was presented by Bartram et al. [53] in the presence of
heterogeneous information. Common usage of Bayesian network can be seen on
Fig. 7.
Machine Learning
Machine learning application in terms of performance evaluation was the most
popular application among the others. An observation of prestressed concrete in
terms of its initial yield and onset failure in bonded posttension concrete beams
using machine learning method [54, 55]. A prediction of elastic modulus of high-
performance concrete is conducted while Nedushan et al. [56] investigate similar
object supplemented by ANFIS and optimal nonlinear regression models. A steel
end plate was investigated using machine learning technique in order to obtain the
most optimum type of connection [57]. Tamura et al. [58] use machine learning in
his research to evaluate the optimal placement of steel brace on a steel frame. Several
simulations of machine learning models was conducted to evaluate the performance
of smart city concept [59], building performance prediction [60] and beam column
joint in reinforced concrete structure [61].
Damage detection using machine learning technique in a composite material was
presented by Sarkar et al. [62]. Machine learning in combination with several other
artificial-based method to investigate the damage that occurs in a cable stayed bridge
structure is. Several models based on machine learning framework was developed to
detect damage in bridge using one-class machine learning [63], detect damage in a
bridge under operational used and environmental variability [64] and detect damage
with the most efficient energy in structure with delayed signal [65].
Machine Learning
16%
Performance
Evaluation
Damage Detection
26% 58%
Damage Prediction
Fig. 8 Common usage of machine learning
In predicting damage that may occur, a machine learning framework in combina-

tion of Bayesian network is presented to evaluate Church in Portugal [66]. Zheng et al.
[67] investigate the damage possibility by a collision in pier using probabilistic-based
classification in machine learning. An experimental model to investigate damage
location using comparison of sparse representation and Fourier discriminant method
was presented by Wang et al. [68] and estimation of punching shear of FRP reinforced
concrete slabs using hybrid machine learning presented by Vu et al. [69]. Common
usage of machine learning can be seen on Fig. 8.
3.2 Artificial Intelligence Method as a Performance

Evaluation, Damage Detection and Damage Prediction
in SHM
Performance evaluation is defined as an ability of a certain AI algorithm in measuring

how well is an object serve its function. Machine learning is mostly utilized as a
performance evaluation (58%) followed by Neural Network (49%). Machine learning
consist of component of task (T), performance measure (P) and improving from
dataset by experience (E). The use machine learning in several research to evaluate
the optimal placement of steel brace on a steel frame by extracting the characteris-
tics using support vector machine to determine classes and binary decision tree are
conducted. Characteristics are incorporated with maximum value of additional stress
in existing beams and column and also the maximum interstory drift angle. Research
regarding evaluation of anomalous behavior of Riccolo viaduct located on Swizz by
using Neural Network and using several functions such as Auto regressive moving
average (ARMA), Discrete Wavelet Transform (DWT), Robust Regression Analysis
(RRA) and several other in defining the algorithm for dot product is conducted.
Damage detection is defined as an ability of a certain AI algorithm in detecting

flaws of a system or object during a sudden change due to either natural disaster
or human error. Pattern recognition and Bayesian network are mostly used as a
damage detection by 50% and 43% respectively. A damage detection of a prestressed
concrete is performed using pattern recognition technique by developing a discrete-
time stochastic state space model and a standard form of matrix. These parameters
then investigated using a damage index and the result is verified using several study
cases. If the result from damage index is considerably acceptable then it can be
concluded that the model is proven to be able to detect damage.
Damage prediction is defined as an ability of a certain AI algorithm in predicting
a potential flaw of a certain system or object before any disaster or other dangerous
event occurs. All the mentioned AI method above is shown to have a rather small
portion in terms of predicting damage compare to performance and damage detection.
Support vector machine has the relative highest percentage by 36% in its population.
A support vector machine is developed to predict the elastic modulus of normal
and high strength concrete with the introduction of insensitive loss function based
on learning theory. The SVM is developed based on learning theory and regression
technique toward loss function. The result then compared with the experimental
result which shown SVM perform well in predicting elastic modulus of concrete.
4 Conclusion
In this research, it shows that the neural network is the most popular artificial
intelligence method in the past decade in terms of structural health monitoring.
Neural network mostly used for both performance evaluation and damage detec-
tion. Machine learning is the second most popular artificial intelligence method and
just like neural network its intended use is mostly for either performance evaluation
or damage detection. Third most popular artificial intelligence method is support
vector machine and it is used as performance evaluation as well as damage detection.
Other artificial intelligence methods that have almost equal proportion are pattern
recognition, principal component analysis and Bayesian network with each majority
intended used as a performance evaluation and damage detection. In the utilization
of structural health monitoring, artificial intelligence is often used as a performance
evaluation tool followed by damage detection and the least is for damage prediction.
References
1. Farrar CR, Worden K, Wiley J (2013) Structural health monitoring a machine learning
perspective. Library of Congress Cataloging-in-Publication Data
2. Rafiei MH, Adeli H (2017) A novel machine learning-based algorithm to detect damage in
high-rise building structures. Struct Des Tall Spec Build 26(18):1–11. https://doi.org/10.1002/
tal.1400
3. Wijaya U, Putra IGEA, Tavio (2020) Assessment and evaluation procedure of existing industrial
building structure based on ASCE 41–17. J Phys Conf Ser 1477(5). https://doi.org/10.1088/
1742-6596/1477/5/052045
4. Smarsly K, Dragos K, Wiggenbrock J (2016) Machine learning techniques for structural health
monitoring. In: 8th Eur Work Struct Heal Monit. EWSHM 2016, vol 2, pp 1522–1531
5. Modha DS, Ananthanarayanan R, Esser SK, Ndirango A, Sherbondy AJ, Singh R (2011)
Cognitive computing. Commun ACM 54(8):62–71. https://doi.org/10.1145/1978542.1978559
6. Sun L, Shang Z, Xia Y, Bhowmick S, Nagarajaiah S (2020) Review of bridge structural health
monitoring aided by big data and artificial intelligence: from condition assessment to damage
detection. J Struct Eng (United States) 146(5) https://doi.org/10.1061/(ASCE)ST.1943-541X.
0002535
7. Alhassan MA, Ababneh AN, Betoush NA (2020) Innovative model for accurate prediction of
the transfer length of prestressing strands based on artificial neural networks: Case study. Case
Stud Constr Mater 12:e00312. https://doi.org/10.1016/j.cscm.2019.e00312
8. Chun P-j, Yamane T, Izumi S, Kuramoto N (2020) Development of a machine learning-
based damage identification method using multi-point simultaneous acceleration measurement
results. Sensors 20(10):2780. https://doi.org/10.3390/s20102780
9. Shariati M et al (2020) A novel approach to predict shear strength of tilted angle connectors
using artificial intelligence techniques. Eng Comput. https://doi.org/10.1007/s00366-019-009
30-x
10. Yan W, Deng L, Zhang F, Li T, Li S (2019) Probabilistic machine learning approach to bridge
fatigue failure analysis due to vehicular overloading. Eng Struct 193(May):91–99. https://doi.
org/10.1016/j.engstruct.2019.05.028
11. Hung TV, Viet VQ, Van Thuat D (2019) A deep learning-based procedure for estimation of
ultimate load carrying of steel trusses using advanced analysis. J Sci Technol Civ Eng NUCE
13(3):113–123. https://doi.org/10.31814/stce.nuce2019-13(3)-11
12. Liu H, Zhang Y (2019) Deep learning-based brace damage detection for concentrically braced
frame structures under seismic loadings. Adv Struct Eng 22(16):3473–3486. https://doi.org/
10.1177/1369433219859389
13. Posenato D, Kripakaran P, Inaudi D, Smith IFC (2010) Methodologies for model-free data
interpretation of civil engineering structures. Comput Struct 88(7–8):467–482. https://doi.org/
10.1016/j.compstruc.2010.01.001
14. García-Segura T, Yepes V, Frangopol DM (2017) Multi-objective design of post-tensioned
concrete road bridges using artificial neural networks. Struct Multidiscip Optim 56(1):139–150.
https://doi.org/10.1007/s00158-017-1653-0
15. Bao Y, Chen Z, Wei S, Xu Y, Tang Z, Li H (2019) The state of the art of data science and engi-
neering in structural health monitoring. Engineering 5(2):234–242. https://doi.org/10.1016/j.
eng.2018.11.027
16. Shrestha YR, Ben-menahem SM (2019) Organizational decision-making structures in the age
of artificial intelligence, pp 1–18. https://doi.org/10.1177/0008125619862257
17. Sri C et al (2018) Structural damage identification based on autoencoder neural networks and
deep learning. Eng Struct 172(April):13–28. https://doi.org/10.1016/j.engstruct.2018.05.109
18. Cury A, Crémona C, Diday E (2010) Application of symbolic data analysis for structural
modification assessment. Eng Struct 32(3):762–775. https://doi.org/10.1016/j.engstruct.2009.
12.004
19. Siddique R, Aggarwal P, Aggarwal Y (2011) Prediction of compressive strength of self-
compacting concrete containing bottom ash using artificial neural networks. Adv Eng Softw
42(10):780–786. https://doi.org/10.1016/j.advengsoft.2011.05.016
20. Yan K, Xu H, Shen G, Liu P (2013) Prediction of splitting tensile strength from cylinder
compressive strength of concrete by support vector machine. Adv Mater Sci Eng. https://doi.
org/10.1155/2013/597257
21. Dantas ATA, Batista Leite M, De Jesus Nagahama K (2013) Prediction of compressive strength
of concrete containing construction and demolition waste using artificial neural networks.
22. Kang F, Li J, Zhao S, Wang Y (2019) Structural health monitoring of concrete dams using
long-term air temperature for thermal effect simulation. Eng Struct 180: 642–653. https://doi.
org/10.1016/j.engstruct.2018.11.065
23. Santos JP, Cremona C, Orcesi AD, Silveira P (2017) Early damage detection based on pattern
recognition and data fusion. J Struct Eng (United States) 143(2):1–11. https://doi.org/10.1061/
(ASCE)ST.1943-541X.0001643
24. Noman AS, Deeba F, Bagchi A (2013) Health monitoring of structures using statistical pattern
recognition techniques. J Perform Constr Facil 27(5):575–584. https://doi.org/10.1061/(ASC
E)CF.1943-5509.0000346
25. Gui G, Pan H, Lin Z, Li Y, Yuan Z (2017) Data-driven support vector machine with optimization
techniques for structural health monitoring and damage detection. KSCE J Civ Eng 21(2):523–
534. https://doi.org/10.1007/s12205-017-1518-5
26. Dorafshan S, Thomas RJ, Maguire M (2018) Data in brief SDNET2018: an annotated image
dataset for non-contact concrete crack detection using deep convolutional neural networks.
Data Br 21:1664–1668. https://doi.org/10.1016/j.dib.2018.11.015
27. Heo G, Kim C, Lee C, Hur J, Seo S (2017) A damage assessment technique based on a revised
Statistical Pattern-Recognition Technique (SPRT). KSCE J Civ Eng 21(3):882–888. https://
doi.org/10.1007/s12205-016-0764-2
28. Ii RSC, Ilagan LC (2019) Application of hybrid soft computing for classification of reinforced
concrete bridge structural health based on thermal-vibration intelligent system parameters. In:
2019 IEEE 15th Int Colloq Signal Process its Appl, March, pp 207–212
29. Mallik N, Wali AS, Kuri N (2016) Damage location identification through neural network
learning from optical fiber signal for structural health monitoring. In: ACM international
conference proceeding series, pp 157–161. https://doi.org/10.1145/3036932.3036937
30. Senniappan V, Subramanian J, Papageorgiou EI, Mohan S (2017) Application of fuzzy cognitive
maps for crack categorization in columns of reinforced concrete structures. Neural Comput
Appl 28:107–117. https://doi.org/10.1007/s00521-016-2313-9
31. Ng CT (2014) Application of Bayesian-designed artificial neural networks in phase ii structural
health monitoring benchmark studies. Aust J Struct Eng 15(1):27–36. https://doi.org/10.7158/
S12-042.2014.15.1
32. Mondoro A, Frangopol DM, Soliman M (2017) Optimal risk-based management of coastal
bridges vulnerable to hurricanes. J Infrastruct Syst 23(3):1–12. https://doi.org/10.1061/(ASC
E)IS.1943-555X.0000346
33. Yan K, Shi C (2010) Prediction of elastic modulus of normal and high strength concrete by
support vector machine. Constr Build Mater 24(8):1479–1485. https://doi.org/10.1016/j.con
buildmat.2010.01.006
34. Cheng MY, Chou JS, Roy AFV, Wu YW (2012) High-performance concrete compressive
strength prediction using time-weighted evolutionary fuzzy support vector machines inference
model. Autom Constr 28:106–115. https://doi.org/10.1016/j.autcon.2012.07.004
35. Cheng MY, Firdausi PM, Prayogo D (2014) High-performance concrete compressive strength
prediction using Genetic Weighted Pyramid Operation Tree (GWPOT). Eng Appl Artif Intell
29:104–113. https://doi.org/10.1016/j.engappai.2013.11.014
36. Çevik A, Kurtoğlu AE, Bilgehan M, Gülşan ME, Albegmprli HM (2015) Support vector
machines in structural engineering: a review. J Civ Eng Manag 21(3):261–281. https://doi.
org/10.3846/13923730.2015.1005021
37. Tan D, Qu W, Zhang J, Wei G, Liu J (2012) Damage diagnosis of cable of large span cable-stayed
bridge based on the support vector machine. Appl Mech Mater 190–191:958–961. https://doi.
org/10.4028/www.scientific.net/AMM.190-191.958
38. Santos JP, Crémona C, Orcesi AD, Silveira P (2013) Multivariate statistical analysis for early
damage detection. Eng Struct 56:273–285. https://doi.org/10.1016/j.engstruct.2013.05.022
39. McGetrick PJ, Kim CW (2014) An indirect bridge inspection method incorporating a wavelet-
based damage indicator and pattern recognition. In: Proceedings of international conference
structural dynamics, EURODYN, vol 2014 January, no June 2014, pp 2605–2612
40. Bandara RP, Chan THT, Thambiratnam DP (2014) Frequency response function based damage
identification using principal component analysis and pattern recognition technique. Eng Struct
66:116–128. https://doi.org/10.1016/j.engstruct.2014.01.044
41. Radhika S, Tamura Y, Matsui M (2015) Cyclone damage detection on building structures from
pre- and post-satellite images using wavelet based pattern recognition. J Wind Eng Ind Aerodyn
136:23–33. https://doi.org/10.1016/j.jweia.2014.10.018
42. Yang Y, Sun P, Nagarajaiah S, Bachilo SM, Weisman RB (2017) Full-field, high-spatial-
resolution detection of local structural damage from low-resolution random strain field
measurements. J Sound Vib 399:75–85. https://doi.org/10.1016/j.jsv.2017.03.016
43. Kang F, Li J, Dai J (2019) Advances in engineering software prediction of long-term temper-
ature effect in structural health monitoring of concrete dams using support vector machines
with Jaya optimizer and salp swarm algorithms. Adv Eng Softw 131:60–76. https://doi.org/10.
1016/j.advengsoft.2019.03.003
44. Abdollahzadeh G, Shabanian SM (2018) Experimental and numerical analysis of beam to
column joints in steel structures. Front Struct Civ Eng 12(4):642–661. https://doi.org/10.1007/
s11709-017-0457-z
45. Salehi H, Das S, Chakrabartty S, Biswas S, Burgueño R (2018) Structural damage identification
using image-based pattern recognition on event-based binary data generated from self-powered
sensor networks. Struct Control Heal Monit 25(4):1–21. https://doi.org/10.1002/stc.2135
46. Lin YQ, Ren WX, Fang SE (2011) Structural damage detection based on stochastic subspace
identification and statistical pattern recognition: II. Experimental validation under varying
temperature. Smart Mater Struct 20(11). https://doi.org/10.1088/0964-1726/20/11/115010
47. Elwood E, Corotis RB (2015) Application of fuzzy pattern recognition of seismic damage to
concrete structures. ASCE-ASME J Risk Uncertain Eng Syst Part A Civ Eng 1(4):1–12. https://
doi.org/10.1061/AJRUA6.0000831
48. Bandara RP, Chan TH, Thambiratnam DP (2014) Structural damage detection method using
frequency response functions. Struct Heal Monit 13(4):418–429. https://doi.org/10.1177/147
5921714522847
49. Datteo A, Lucà F, Busca G (2017) Statistical pattern recognition approach for long-time
monitoring of the G. Meazza stadium by means of AR models and PCA. Eng Struct
153(July):317–333. https://doi.org/10.1016/j.engstruct.2017.10.022
50. Balsamo L, Betti R, Beigi H (2014) A structural health monitoring strategy using cepstral
features. J Sound Vib 333(19):4526–4542. https://doi.org/10.1016/j.jsv.2014.04.062
51. Reddy A, Indragandhi V, Ravi L, Subramaniyaswamy V (2019) Detection of cracks and
damage in wind turbine blades using artificial intelligence-based image analytics. Measurement
147:106823. https://doi.org/10.1016/j.measurement.2019.07.051
52. Liu R, Yang B, Zio E, Chen X (2018) Artificial intelligence for fault diagnosis of rotating
machinery: a review. Mech Syst Signal Process 108:33–47. https://doi.org/10.1016/j.ymssp.
2018.02.016
53. Bartram GW, Mahadevan S (2012) System modeling for SHM using dynamic Bayesian
networks. AIAA Infotech Aerosp Conf Exhib 2012(June):1–12. https://doi.org/10.2514/6.
2012-2423
54. Salamone S, Veletzos MJ, Lanza Di Scalea F, Restrepo JI (2012) Detection of initial yield and
onset of failure in bonded posttensioned concrete beams. J Bridg Eng 17(6):966–974. https://
doi.org/10.1061/(ASCE)BE.1943-5592.0000311
55. Chou JS, Tsai CF, Pham AD, Lu YH (2014) Machine learning in concrete strength simulations:
multi-nation data analytics. Constr Build Mater 73:771–780. https://doi.org/10.1016/j.conbui
ldmat.2014.09.054
56. Ahmadi-Nedushan B (2012) Prediction of elastic modulus of normal and high strength concrete
using ANFIS and optimal nonlinear regression models. Constr Build Mater 36:665–673. https://
doi.org/10.1016/j.conbuildmat.2012.06.002
57. Greco L (2018) Machine learning and optimization techniques for steel connections, no. July,
pp 1–8
58. Tamura T, Ohsaki M, Takagi J (2018) Machine learning for combinatorial optimization of brace
placement of steel frames. Japan Archit Rev 1(4):419–430. https://doi.org/10.1002/2475-8876.
12059
59. Chin J, Callaghan V, Lam I (2017) Understanding and personalising smart city services using
machine learning, the Internet-of-Things and Big Data. In: IEEE international symposium Ind
Electron, pp 2050–2055. https://doi.org/10.1109/ISIE.2017.8001570
60. Geyer P, Singaravel S (2018) Component-based machine learning for performance prediction in
building design. Appl Energy 228(July):1439–1453. https://doi.org/10.1016/j.apenergy.2018.
07.011
61. Baker JW (2007) Measuring bias in structural response caused by ground motion scaling.
Pacific Conf Earthq Eng 056:1–6. https://doi.org/10.1002/eqe
62. Sarkar S, Reddy KK, Giering M, Gurvich MR (2016) Deep learning for structural health
monitoring: a damage characterization application. In: Proceedings annual conference Progn
Heal Manag Soc PHM, vol 2016 October, no. October, pp 176–182
63. Long J, Buyukozturk O (2014) Automated structural damage detection using one-class machine
learning. Conf Proc Soc Exp Mech Ser 4:117–128. https://doi.org/10.1007/978-3-319-04546-
7_14
64. Figueiredo E, Park G, Farrar CR, Worden K, Figueiras J (2011) Machine learning algo-
rithms for damage detection under operational and environmental variability. Struct Heal Monit
65. Salehi H, Das S, Biswas S, Burgueño R (2019) Data mining methodology employing artificial
intelligence and a probabilistic approach for energy-efficient structural health monitoring with
noisy and delayed signals. Expert Syst Appl 135:259–272. https://doi.org/10.1016/j.eswa.2019.
05.051
66. Ramos LF, Miranda T, Mishra M, Fernandes FM, Manning E (2015) A Bayesian approach for
NDT data fusion: the Saint Torcato church case study. Eng Struct 84:120–129. https://doi.org/
10.1016/j.engstruct.2014.11.015
67. Zheng W, Qian F (2017) Promptly assessing probability of barge–bridge collision damage of
piers through probabilistic-based classification of machine learning. J Civ Struct Heal Monit
7(1):57–78. https://doi.org/10.1007/s13349-017-0208-9
68. Wang Z et al (2013) Comparison of sparse representation and fourier discriminant methods:
damage location classification in indirect lab-scale bridge structural health monitoring. In: 2013
Bridg Your Passion with Your Prof - Proc 2013 Struct Congr, pp 436–446. https://doi.org/10.
1061/9780784412848.039
69. Vu DT, Hoang ND (2016) Punching shear capacity estimation of FRP-reinforced concrete slabs
using a hybrid machine learning approach. Struct Infrastruct Eng 12(9):1153–1161. https://doi.
org/10.1080/15732479.2015.1086386
Development of Preventive Maintenance
Guidelines for Simple-Classification
Government Buildings Based on Work
Breakdown Structure Within the DKI
Jakarta Provincial Government
Dyah Ayu Pangastuti and Yusuf Latief
Abstract DKI Jakarta, as the capital of the Republic of Indonesia, has the highest
population density in Indonesia, with a population of 16,334 people/km2 . Therefore,
community facilities and infrastructure within the DKI Jakarta Provincial Govern-
ment play an important role so that community service can be carried out properly.
As one of the main infrastructures, State Buildings must have building reliability as
stated in the technical requirements stipulated in Presidential Regulation Number 73
of 2016. Building maintenance is an activity to maintain the building’s reliability and
infrastructure, and facilities so that the building always functions properly. Based on
a survey conducted by the DKI Jakarta Provincial Office for The Creation of Works,
Spatial Planning and Land Use in 2019, the building assets of Province DKI Jakarta
are 9823. 60% of those buildings were classified as Simple-Building category. Due
to numerous building assets, building maintenance activities need to be supported by
an adequate maintenance guideline especially for the simple-classification building.
This study aims to develop preventive maintenance guidelines based on Work Break-
down Structure. The Work Breakdown Structure (WBS) will be used as the basis
for maintenance guidelines, including maintenance activities and frequency. The
methodology used in this research is study literature, survey and expert validation
for WBS for Simple-Building component and Preventive Maintenance Guidelines.
Keywords Simple-building maintenance · Maintenance guidelines · Preventive

maintenance
1 Introduction
Based on the DKI Jakarta Provincial Office for The Creation of Works, Spatial
Planning and Land Use survey results in 2019, the number of state-building assets in
the DKI Jakarta province are 9823 consisting “in need of repair” condition are 2994
buildings and “severely damaged” condition are 796 buildings. It can be concluded
D. A. Pangastuti · Y. Latief (B)

Department of Civil Engineering, Faculty of Engineering, University of Indonesia,
Depok 16424, Indonesia
340 D. A. Pangastuti and Y. Latief
that 38% of the state buildings in the DKI Jakarta Provincial Government are in
less reliable condition. Figure 1 shows that the building assets of the DKI Jakarta
Provincial Government with Simple-Buildings classification are more numerous than
buildings with non-Simple-Building classifications.
From interviews with the Head of DKI Jakarta Provincial Office for The Creation
of Works, Spatial Planning and Land Use, there were several obstacles in main-
taining state buildings DKI Jakarta Province. The technical incomprehension of those
responsible for maintenance activities in planning building maintenance activities
and other several factors that can be seen in Fig. 2.
Among the causes of building unreliability is the absence of an efficient system for
data recording, paper-based forms and a lack of technical personnel to monitor and
communicate information [1]. The influence of government policies, lack of clear
job descriptions, and lack of technical experts are common problems in building
maintenance [2]. Also, the lack of maintenance personnel, the lack of preventive
maintenance [3], and the lack of IT use, the lack of adequate maintenance procedures
are some of the main issues in the management of government assets [4].
For this reason, it is necessary to have a maintenance guidelines based on WBS
to improve maintenance performance [5, 6].
Fig. 1 DKI Jakarta provincial office for The creation of works, spatial planning and land use
building asset survey data 2019
Development of Preventive Maintenance Guidelines … 341
Fig. 2 Fishbone diagram about obstacles that occurs in building maintenance of state building
within DKI Jakarta Province
2 Literature Review
2.1 Simple-Building Classification
Simple-Building classification for State buildings according to Ministerial regulation

of Minister For Public Works and Human Settlements Number 22 in 2018 is described
in article 14 paragraph (2), which reads State Buildings with simple classification
as referred to in paragraph (1) letter a are buildings with simple technology and
specifications including:
1. Office buildings and other state buildings with a number of floors up to 2 (two)
floors;
2. Office buildings and other state buildings with an area of up to 500 m2 (five
hundred square meters); and
3. State houses include Type C, Type D, and Type E.
The technical specifications for simple classified State buildings are described
in attachment II Table 1, including building layout, building materials, building
structures, utilities, infrastructure & in buildings and safety facilities.
2.2 Work Breakdown Structure
Work Breakdown Structure (WBS) is a hierarchical decomposition oriented towards

the work that the project team will carry out to achieve project objectives and
create the required results [7]. In general, WBS can be defined as a result-oriented
grouping (deliverable) to project elements, which organize and determine the
project’s overall structure [8]. Furthermore, standard practice for work breakdown
structure explains that a more regularly made project describes individual work
components, determining the entire scope of work on a project. This activity is
called WBS [7].
WBS for this system is used for determines the building component so that no
component are overlooked for maintenance. It is also used as a basis for preparing
preventive maintenance guidelines for the system.
2.3 Preventive Maintenance
Building maintenance activities are classified into preventive maintenance, corrective

maintenance [9]. Periodic maintenance is a preventive measure. its activities include
regularly measuring, detecting, analyzing equipment problems to avoid failure [10].
The maintenance schedule is part of preventive maintenance. Efficiency and effec-
tiveness in scheduling are also determined by preventive maintenance planning [11].
Preventive maintenance activities are attached to the guidelines. The guidelines
will be containing: Preventive maintenance activities, frequency and interval each
activity for every component of the building.
Table 1 Problem formulations and research strategy

No Problem formulation Question used Strategy
1 How is the structure of a simple-building How Archive analysis, survey
WBS within the DKI Jakarta provincial
government?
2 How to develop maintenance guidelines How Archive analysis, survey
for simple buildings within the DKI
Jakarta Provincial government?
3 Research Method
3.1 Strategy and Process
After identifying the problems, some formulations of the problem were obtained,
which were then determined by the research strategy. The research strategy used in
this research is archive analysis (literature review and existing document for WBS
and maintenance guidelines) and surveys (validation and deep interview).
The research process consists of two stages analysis. The first analysis is a simple-
building WBS. The results of the validated WBS become basis for the identification
of maintenance activities and the schedule of maintenance activities which are then
validated. The second analysis is carried out to guide maintenance activities. Research
process flow shown in Fig. 3
There are several stages of a questionnaire to determine WBS, preventive main-
tenance activities and schedules. This study uses the Delphi method, which involves
expert respondent for every validation process: five experts in managerial positions
within 10–25 years of experience and seven expert technical staffs within 5–15 years
of experience in building maintenance activities.
Fig. 3 Research method flow

3.2 Research Variables
Based on the results of the literature study, it is found that the activity variables and
maintenance schedules form the maintenance guidelines. The list of activities and
maintenance schedules are compiled based on identifying the building components to
be maintained. This building component is defined by the type of work and the work
package as variables of the work breakdown structure. The resulting maintenance
guidelines are expected to improve the performance of building maintenance in
accordance with applicable regulations. Table 2 shows a list of the variables and sub
variables used in this research,
4 Result Discussion
The first stage analysis is validation of the types and packages of simple building
works. The questionnaire was compiled and obtained from 5 different references
representing each construction work. The questionnaire was given to 7 experts with
the answers “yes” or “no” to the types and packages of work deemed appropriate
for simple-buildings. Validation of acceptance was carried out with a minimum of
four “Yes” answers. Table 3 shows an example of expert data analysis, with the most
answers in quantity being used as the acceptance or denial judgment.
The validation results of the WBS study of Simple-Building education functions
are: Structural elements obtained 8 types of work and 16 work packages; Architec-
tural elements obtained 12 types of work and 78 work packages; Mechanical elements
obtained 4 types of work and 17 work packages; Electrical elements obtained 4 types
of work, and 15 work packages; Landscape and Housekeeping element obtained 3
types of work and 50 work packages. Total 55 Type or work and 176 work package
obtained. Example of WBS findings for electrical work are shown in Fig. 4.
4.2 Activity Maintenance and Schedule
After obtaining the WBS results for simple-buildings, then a literature study is
carried out to identify the activities and schedule of preventive maintenance activi-
ties. The activities and periodic schedule of preventive activities are then compiled
in a simple building maintenance guideline and validated using a questionnaire and
deep interviews with 5 experts with 10–15 years of experience in managerial posi-
tions. The guidelines consist of preventive maintenance activity and periodic schedule
for structural, architectural, mechanical, electrical, landscaping, and housekeeping
Table 2 Research variables and sub variables

Variables Sub variables References
X.1 Work X.1.1 Type of work X.1.1.1 Types of structural Muhammad
breakdown maintenance of work Naufal
structure simple buildings (2020)
X.1.1.2 Type of Luki Wijaya
architectural work (2020)
X.1.1.3 Type mechanical Yusuf
work Djonli
(2019)
X.1.1.4 Types of electrical Azhara
work Yudha
(2019)
X.1.1.5 Types of Rizkyana
landscaping and Fathoni
housekeeping work (2019)
X.1.2 Simple building X.1.2.1 Structural work Muhammad
maintenance work package Naufal
package (2020)
X.1.2.2 Architectural work Luki Wijaya
package (2020)
X.1.2.3 Mechanical work Yusuf
package Djonli
(2019)
X.1.2.4 Electrical work Azhara
package Yudha
(2019)
X.1.2.5 Landscape and Rizkyana
housekeeping work Fathoni
package (2019)
X.2 Preventive X.2.1 Preventive X.2.1.1 Structural Muhammad
maintenance maintenance preventive Naufal
guidelines activity maintenance (2020),
activities Anita
Handayani
(2020)
X.2.1.2 Architectural Luki Wijaya
preventive (2020),
maintenance Anita
activities Handayani
(2020)
(continued)
Table 2 (continued)
X.2.1.3 Mechanical Yusuf
preventive Djonli
maintenance (2019),
activities Anita
Handayani
(2020)
X.2.1.4 Electrical Azhara
preventive Yudha
maintenance (2019),
activities Anita
Handayani
(2020)
housekeeping Fathoni
Preventive (2019),
maintenance Anita
(2020)
X.2.2 Schedule/frequency X.2.2.1 Structural Muhammad
of preventive schedule/frequency Naufal
maintenance of Preventive (2020),
activities maintenance Anita
(2020)
X.2.2.2 Architectural Luki Wijaya
schedule/frequency (2020),
of Preventive Anita
maintenance Handayani
activities (2020)
X.2.2.3 Mechanical Yusuf
schedule/frequency Djonli
of Preventive (2019),
maintenance Anita
(2020)
X.2.2.4 Electrical Azhara
schedule/frequency Yudha
of Preventive (2019),
maintenance Anita
(2020)
(continued)
Table 2 (continued)
housekeeping Fathoni
Schedule/frequency (2019),
of preventive Anita
Maintenance Handayani
activities (2020)
Y.1 Building Y.1.1 Safety Law of the
Maintenance Y.1.2 Health Republic of
Performance Indonesia
Y.1.3 Convenience Number 28
Y.1.4 Easiness of (2002),
Regulation
of the
Minister of
Public
Works and
Public
Housing
Number 24
of 2010
works. There are 1049 maintenance activities for 176 work package in the guidelines.
Table 4 is an example of Tie Beam’s preventive maintenance guidelines.
From deep interviews with the experts it was found differences of opinion about
several activities or schedules. The differences of suggestions are collected in a list
then analyzed and further discussed. Table 5 is an example of Expert’s different
opinions on periodic schedule activity in structural work.
From the results of the analysis and discussion, it was agreed that several work
activities needed to be corrected, some if it was agreed to be eliminated. Examples
of the results of corrections to preventive maintenance work activities can be seen in
Table 6.
Total 6 activities were corrected, 4 of which were eliminated. The work that have
been eliminated are painting the ACP walls, painting the acoustic ceiling, coating
the acoustic ceiling and replacing the backup lamp battery for the emergency lights.
Most of the corrected activities are changes in schedule status to incidental because
there are conditions that require repairs, such as grouting the fractured part which was
previously scheduled every 2 years, being corrected to incidental activities because
its depending on the appearance of the crack. Further validation was carried out on
the maintenance guidelines obtaining 1045 maintenance activities with each activity
schedule.
Table 3 Validation analysis process example
348
WBS level 2 WBS level 3 WBS level 4 Does this type of work include WBS Level 4 Quantity Modus Status
Work type Work package simple-building maintenance activities?
Reference: Luki Wijaya Reference: Luki Wijaya (2020) Expert Yes No

(2020)
Code Family Code Type Code Package Design 1 2 3 4 5 6 7
alternatives
X.2 Architecture X.2.1 Parking and X.2.1.1 Parking Bicycle parking Yes Yes Yes No Yes Yes Yes 6 1 Yes Accept
Work pedestrian Motor parking Yes Yes Yes Yes Yes Yes Yes 7 0 Yes Accept
facilities
X.2.1.2 Sidewalk Sidewalk No No Yes No No No Yes 2 5 No Denied
X.2.2 Means of X.2.2.1 Smoking Smoking room No No No No No No No 0 7 No Denied
Smoking Room
X.2.3 Means of exit X.2.3.1 Exit Exit Yes No Yes No No No Yes 3 4 No Denied
X.2.4 Floor work X.2.4.1 Floor Floor h ardener Yes No Yes No No No Yes 3 4 No Denied
Ceramic floor Yes Yes Yes Yes Yes Yes Yes 7 0 Yes Accept
of the corridor
area
Ceramic floor Yes Yes Yes Yes Yes Yes Yes 7 0 Yes Accept
dry area
Tiled floor wet Yes Yes Yes Yes Yes Yes Yes 7 0 Yes Accept
area
Ceramic floor Yes Yes Yes Yes No Yes Yes 6 1 Yes Accept
fire ladder
Waterproof Yes Yes Yes No Yes Yes Yes 6 1 Yes Accept
coating
concrete floor
(continued)
D. A. Pangastuti and Y. Latief
Table 3 (continued)
WBS level 2 WBS level 3 WBS level 4 Does this type of work include WBS Level 4 Quantity Modus Status
Work type Work package simple-building maintenance activities?
Reference: Luki Wijaya Reference: Luki Wijaya (2020) Expert Yes No

(2020)
Code Family Code Type Code Package Design 1 2 3 4 5 6 7
alternatives
Waterproof Yes No No Yes No Yes Yes 4 3 Yes Accept
membrane
concrete floor
integral Yes No No Yes No Yes Yes 4 3 Yes Accept
waterproof
concrete floor
Granite floor Yes Yes No No No No Yes 3 4 No Denied
Parquet floor No No No No No Yes No 1 6 No Denied
Development of Preventive Maintenance Guidelines …
349
Fig. 4 WBS result for electrical work
As we can see in Table 7, the results of the validation of five experts obtained
five “Yes” answers. That means, the maintenance guidelines can be used as a simple
building maintenance guideline to improve maintenance performance. Suggestions
are given by the respective experts for the purposes of developing this preventive
maintenance guideline.
Table 4 Example of preventive maintenance activity and periodic schedule

Description
Structural elements contained in buildings or structures that use deep foundations or local shallow
foundations. This tie beam is located above the ground and on shallow local foundations such as
footplat foundations or deep foundations. This tie beam is the same as the beam except that it is located
in the lower structure
Tie beam
Description
Work family Structure
Work type Tie beam and sloof
Package Work Tie beam
Alternative desain Concrete tie beam
Activity Preventive maintenance actions Periodic schedule
Examination Visual observation of cracks in concrete Annual
Maintenance Use a chisel to tap the concrete around the crack to make sure that 2 years
none of the concrete has peeled off
Clean the chiseled surface 2 years
Grouting the crack 2 years
Repaint with emulsion paint or paint that is waterproof and acidic 2 years
on the surface
Table 5 Example of Expert’s different opinions on periodic schedule activity in structural work
Structural preventive maintenance guidelines Type of suggestion/correction
Type work Maintenance Frequency Expert Expert Expert 3 Expert Expert 5
activity (years) 1 2 4
(years) (years) (years)
Reinforced Destructive 3–5 10–15 3–5 Incidental 3–5 Incidental
concrete testing: core
columns, drill, and
reinforced compression test
concrete Non-destructive 3–5 10–15 3–5 Incidental 3–5 5 years
beams, testing: UPVT,
conventional infrared
concrete floor thermograph,
slabs, hammer test,
conventional loading test,
concrete covermeter test,
stairs, carbonation test,
conventional impact echo test
concrete
plate roofs
Steel column, Destructive 3–5 10–15 3–5 Eliminate 3–5 Incidental
steel beam testing:
radiography,
ultrasonic test,
and liquid
Penetrant test
Steel truss Non-destructive 3–5 10–15 3–5 3–5 3–5 5 years
roof testing
Table 6 Example of corrected preventive maintenance activity and schedule in electrical work
WBS level WBS level WBS level 5 Preventive Schedule/frequency Corrected
3 4 maintenance
Work type Work Work activity activity
package
Electrical LED light Treatment Replace the Incidental Eliminated
load back up
battery on the
emergency
lights
Solid Service Improve the Monthly Incidental
grounding resistance of
detention if the
resistance is
above the
standard
Lightning Maintenance Improve Monthly Incidental
rod grounding
resistance if it
is above
standard
Table 7 Preventive maintenance guidelines validation analysis

Is this guidelines for the maintenance of simple buildings for the state building could improve
maintenance performance in the DKI Jakarta Provincial environment?
Respondents Yes/No Suggestion
Expert 1 Yes According to the regulation of the Minister of Public Works and
observes the real/real conditions regarding daily maintenance and
maintenance (buildings)
Expert 2 Yes Generator work can be eliminated for school functions building
(no longer used in school buildings, replaced with Solar Panels
(future school construction will use Solar Panels)
Expert 3 Yes To be adjusted to the building object
Expert 4 Yes Can be developed for use by other buildings, especially in the city
hall complex
Expert 5 Yes Agree if the waterproofing coating in question is to use a special
waterproofing coating instead of re-coating it with the same
waterproofing material
5 Conclusion
1. Preventive guidelines can be prepared based on the identification of building

components using WBS based on relevant previous research and literature
studies.
2. Preventive maintenance guidelines cannot be equated for all types of building

classifications. In the same work package there can be different activities
or schedules for preventive maintenance activities. For example, in a simple
building, the structural test schedule is not necessary as often as a building with
a non-simple classification.
3. WBS and preventive maintenance guidelines can improve maintenance perfor-
mance, although not directly. Maintenance performance is very dependent on
the activeness and concern of the building manager for maintenance activities.
Further research can be developed to make this guideline implemented properly,
such as using web-based system for maintenance activity.
4. These activities can be developed for the preparation of the budget plan by
analyzing the budget requirements of WBS level 6, which are resources and
equipment.
References
1. Ismail Z-A (2014) Improving maintenance management practices for building facility. J Constr
Eng Proj Manag 4(3):21–32. https://doi.org/10.6106/jcepm.2014.4.3.021
2. Alshehri A, Motawa I, Ogunlana S (2015) The common problems facing the building main-
tenance departments. Int J Innov Manag Technol 6(3):234–237. https://doi.org/10.7763/ijimt.
2015.v6.608
3. Waziri BS, Vanduhe BA (2013) Evaluation of factors affecting residential building maintenance
in Nigeria: users ’ perspective. Civ Environ Res 3(8):19–25
4. Abdullah S, Razak AA, Hanafi MH, Salleh MN (2011) Managing government property assets:
the main issues from the malaysian perspective. J Techno-Social 3(1):35–52
5. Djonly Y (2019) Pengembangan Pedoman Pekerjaan Preventive Maintenance Komponen
Mekanikal Pada Bangunan Gedung Pemerintah Berbasis Work Breakdown Structure (WBS)
untuk Meningkatkan Kinerja Pemeliharaan dan Peawatan,” Universitas Indonesia
6. Wijaya L (2020) Pengembangan Pedoman Pekerjaan Preventive Maintenance Komponen
Arsitektur Pada Bangunan Gedung Pemerintah Berbasisi Work Breakdown Structure (WBS)
Untuk Meningkatkan Kinerja Pemeliharaan dna Perawatan. Universitas Indonesia
7. PMI (2021) Work Breakdown Structure Practice Standard | PMI. https://www.pmi.org/
pmbok-guide-standards/framework/practice-standard-work-breakdown-structures-2nd-edi
tion. Accessed 12 Apr 2021
8. Jung Y, Woo S (2004) Flexible work breakdown structure for integrated cost and schedule
control. J Constr Eng Manag 130(5):616–625. https://doi.org/10.1061/(asce)0733-9364(200
4)130:5(616)
9. Ruparathna R, Hewage K, Sadiq R (2017) Developing a level of service (LOS) index for
operational management of public buildings. Sustain Cities Soc 34:159–173. https://doi.org/
10.1016/j.scs.2017.06.015
10. Lee PC, Wang Y, Lo TP, Long D (2018) An integrated system framework of building
information modelling and geographical information system for utility tunnel maintenance
management. Tunn Undergr Sp Technol 79:263–273. https://doi.org/10.1016/j.tust.2018.
05.010
11. Basri EI, Razak IHA, Ab-Samat H, Kamaruddin S (2017) Preventive maintenance (PM) plan-
ning: a review. J Qual Maintenance Eng 23(2), 114–143. Emerald Group Publishing Ltd. https://
doi.org/10.1108/JQME-04-2016-0014
Development of Preventive Maintenance
Guidelines for Electrical Components
on Government Building Based on Work
Breakdown Structure
Azhara Yudha Pradipta, Yusuf Latief, and Rossy Armyn Machfudiyanto
Abstract This research aims to create preventive maintenance guidelines for elec-
trical components in buildings, especially government buildings, by breaking down
electrical components using work breakdown structures so that the identification
of electrical components is complete and well-structured. The methodology used
in this research uses archive analysis with validation by experts through a ques-
tionnaire based on the legislations and applicable research and previous research.
The results of this research are guidelines for preventive maintenance for electrical
components in government buildings that contain preventive maintenance steps for
each component at an alternative design level with a time interval for each compo-
nent. Preventive maintenance guidelines for electrical components in this building
are formed using work breakdown structure described from level 1 to level 4 with
their derivatives, namely alternative designs. From the results of the validation, there
are 22 (twenty-two) work packages and 45 (forty-five) alternative designs for elec-
trical components in buildings that need preventive maintenance with each activity,
procedure and preventive action and time intervals.
Keywords Government building · Guideline · Preventive maintenance · Work

breakdown structure
A. Y. Pradipta · Y. Latief (B) · R. A. Machfudiyanto (B)

Civil Engineering, Faculty of Engineering, Universitas Indonesia (UI), 16424 Depok, West Java,
Indonesia
R. A. Machfudiyanto
A. Y. Pradipta
356 A. Y. Pradipta et al.
1 Introduction
Minister of Finance, through letter S-841/MK.02/2014 dated December 16, 2014, on

the Postponement/Moratorium on Ministry/Institution Office Building Construction
asking Ministry/Institution to postpone the construction of new government office
buildings that began in 2015. Postponement/moratorium on new building construc-
tion is carried out to implement bureaucratic reform more effectively. Besides, it
is also to improve the quality of state spending in the 2015 fiscal year [1]. Along
with this, consequences arise in the operational phase of the building itself, namely
for the building manager in maintaining the reliability of the building so that it
is always worthy of function [2]. Furthermore, Wimala and Karim [3] stated that
building maintenance is still considered less important in developing countries so
that implementation priorities are not prioritized.
In Facility Management, two terms are known as Preventive Maintenance (PM)
and Corrective Maintenance (CM). Both are part of the operational phase after the
construction phase is completed and the difference is the time of implementation
where maintenance is carried out at certain time intervals to reduce the possibility
of malfunction while maintenance is carried out after a malfunction occurs [4, 5].
Electrical components in a building are a separate challenge to be described
because they are many. WBS or Work Breakdown Structure is the process of dividing
the results and project work into smaller ones so that it is easy to manage utilizing
decomposition techniques [6]. The purpose of WBS is to organize project elements,
known as project scope, all work is organized into smaller components so that the
project can be better managed. The cluster hierarchy of WBS provides details of
all levels, usually to the smallest level of resources consisting of materials, tools,
labor and also methods of implementation. The decomposition technique in WBS
can prevent the loss of an element in a component to be maintained. According to
U.S. Government Administration Office [7], WBS provides a clear picture of what
needs to be achieved and how the work will be done.
Without scheduling either routinely or periodically in building maintenance,
which can result in maintenance costs because preventive maintenance is asso-
ciated with costs to carry out maintenance/replacement (PM) activities and cost
savings needed to increase the costs required by the system (CM) [8]. Based on this,
research needs to be carried out on preventive maintenance of WBS-based electrical
components in government buildings.
There are many cases of fire whose main cause is a short circuit. The U.S. Fire
Administration noted that 119,000 fires were caused by malfunctioning electrical
components in non-residential buildings between 2003 and 2016 and ranked third as
the main cause of fires after cooking and intentional activities [9]. In Indonesia, cases
of fire in buildings that occurred in DKI Jakarta as a result of damage to electrical
components (electrical short circuit) as many as 494 events from a total of 692 events
in 2018 [10, 32].
Maintenance and operations are the phases passed after the initiation phase, the
planning phase and the construction phase are completed in a series of construc-
tion industry processes. Poor maintenance is associated with damage that occurs in
a building [11]. Many factors cause damage to electrical components in building
that have the impact as mentioned earlier. One of them is the poor performance of
maintenance of electrical components in buildings [12].
The maintenance and upkeep of a building become important because it plays a
role in maintaining the function of every component in the building and keeping the
building in its original condition [13, 14] so that user satisfaction can be achieved.
Lack of awareness regarding damage prevention and the mindset “If it hasn’t been
damaged, don’t fix it” [15] must be eliminated.
2 Research Objective
This research will produce a preventive maintenance guideline for electrical compo-
nents in buildings. In realizing it there are several stages so that the resulting
guidelines are optimal. These stages are:
1. Identify policies and regulations that are in force today and are relevant to
preventive maintenance
2. Identify the work package (level 4) electrical components of the building that
need to be prevented from maintenance
3. Identify the procedures and guidelines used in preventive maintenance work so
that a preventive maintenance guideline can be developed on a WBS-based
The above stages are carried out so that a guideline that is produced can cover
all electrical components in buildings, especially government buildings or in other
words comprehensive.
3 Literature Review
3.1 Government Buildings
Building is a physical form of the result of construction work which is integrated

with the place of domicile, partly or wholly on top and/or in land and/or water, which
functions as a place for humans to carry out their activities, both for residential or
residence place, religious activities, activities business, social, cultural and special
activities [2, 5, 16–18]. Furthermore, state buildings can simply be interpreted as
buildings owned by the state or government or in other words, state buildings are
buildings for official purposes that become/will become state property and are held
with funding sources that come from state budget funds, and/or region budget funds,
and/or other legitimate sources of financing [16, 18]. According to Wimala and
Tamin [13], maintaining government buildings is a necessity coupled with inadequate
policies and regulations related to the maintenance of buildings.
3.2 Preventive Maintenance
According to Beichelt and Fischer [19], there are two general maintenance policies,
namely (the age of replacement) carried out when damage occurs or the replacement
is done at a certain time and (minimal repair-replacement) replacement is always
done at when certain time. Furthermore, [19] also stated, minimal repair in practice
is not always possible to restore the function of a system after damage occurs. It can be
concluded that it is important to prevent damage to a system. Damage to a component
cannot be avoided but can be prevented by choosing the type of maintenance, namely
preventive maintenance [20–22] because corrective maintenance is considered less
effective in terms of frequent damage to a component that results in high maintenance
costs [23].
Preventive maintenance is an activity to minimize the damage done at certain
time intervals based on the experience or recommendations of the manufacturing
company (OEM Recommendation) [21]. It is important to note these time intervals
because too frequent time intervals will cause high maintenance costs [24]. Based
on Peng et al. [25], one of the activities that must be carried out before preventive
maintenance activities are inspection because inspection is a very effective activity
to measure the performance of a system. Furthermore Ref. [5], states that preventive
maintenance activities for building electrical components are inspection, service, and
maintenance.
WBS (Work Breakdown Structure) is a hierarchical decomposition of work-oriented

results that will be carried out by the project team to achieve project objectives and
create the required results [26]. WBS is created after the stage of determining the
scope is completed. The predetermined scope of work is structured in WBS to make
it easier to manage [27]. Every project activity, in this case, maintenance is important
to be planned as well as possible. WBS is an effective tool for managing projects at
the planning stage [28].
3.4 Guideline
Guidelines are defined as a framework, procedure, or process, or step by step used to

plan or decide on something [21]. In a preventive maintenance work guide, usually
presented several things, one of which is a table that lists maintenance activities
with a daily, weekly, monthly or yearly schedule. The purpose of a guideline is as a
reference for doing something that includes technical aspects and planning aspects
[29].
If reviewed according to policies and regulations in Indonesia, the Guidelines
are intended to meet the technical requirements of buildings [30]. Guidelines for
building maintenance are a reference to maintaining the reliability of buildings and
their infrastructure and facilities so that buildings are always worthy of function [5].
The research methodology used is divided into several stages, namely archive anal-
ysis to obtain data and information related to preventive maintenance including rele-
vant previous research, literature studies, and validation by experts in the field of
building maintenance. The entire methodology is intended to identify policies and
regulations that are applicable and relevant to preventive maintenance of buildings,
identify WBS electrical components of buildings consisting of level 1, level 2, level
3, level 4 including alternative designs and identify applicable procedures and guide-
lines related to preventive maintenance with the time interval required to carry out
each of its activities. After all the stages are carried out, a preventive maintenance
guideline for electrical components of the building can be made based on WBS that
has been identified previously, so that guidelines are obtained consisting of level 1 to
level 4 of electrical components along with alternative designs with each preventive
maintenance activity and intervals time for each component.
Validation carried out by experts in the field of maintenance includes validation
of applicable policies and regulations, validation of the WBS that has been made,
validation of the actions and time intervals of each electrical component and finally
validation of guidelines for preventive maintenance of electrical components of the
building as a whole.
5.1 Policies and Regulations
In Indonesia, the policies and legal regulations in the form of hierarchies are intercon-
nected with one another. The law is the highest source of law followed by govern-
ment regulations and ministerial regulations [31]. In connection with preventive
maintenance, the underlying policies and regulations are Law of the Republic of
Indonesia Number 28 of 2002 concerning Buildings, Government Regulation of the
Republic of Indonesia Number 36 of 2005 concerning Implementing Law of Republic
Indonesia Number 28 of 2002 concerning Building, Minister of Public Works Regu-
lation number 24/prt/m/2008 concerning guidelines for building maintenance and
Minister of Public Works Regulation number 16/prt/m/2010 concerning technical
guidelines for periodic inspections of buildings.
5.2 WBS for Electrical Components of Buildings
Based on the results of archival analysis and literature studies on policies and regu-
lations and previous research, WBS obtained for electrical components of buildings
consisting of level 1, namely the name of the project, level 2, namely the work
section, level 3, namely the type of work and level 4 in the form of work packages
with derivatives, namely alternative designs. Alternative design is not a level in WBS
because it cannot be measured [27] (see Fig. 1).
There are 4 (four) types of work (level 3) for WBS electrical components of
buildings that consist of power supply, distribution systems, electricity loads and
electronics. At level 4 (work packages) there are 22 (twenty-two) work pack-
ages consisting of transformers, UPS (Uninterruptible Power Supply), generators,
solar panels, MVMDP, LVMDP, electricity loads, lighting systems, lighting control
systems, switches, under floor duct/raised floor systems, cables, ground systems,
lighting rod systems, fire alarm systems and detectors, telephones, sound systems,
network systems, closed-circuit television (CCTV), signal amplifiers, building
automation systems (BAS), presence machines.
Derivatives from level 4 or work packages are alternative designs that are not
included in the WBS level but are still part of the WBS consisting of 45 (forty-five)
components which are described in full in Table 1.
Fig. 1 WBS for electrical components of buildings
Table 1 Alternative design

WBS level 4 (work package) Alternative design
Detail Detail
Transformer Dry transformer
Liquid transformer
UPS (uninterruptible power Local UPS (uninterruptible
supply) power supply)
Central UPS (uninterruptible
power supply)
Reserved power supply Generator
(continued)
Table 1 (continued)
Detail Detail
Gas generator
Renewable power supply Solar panel
Biodiesel generator
MVMDP Oil circuit breaker
Vacuum circuit breaker
LVMDP Miniature circuit breaker
Molded case circuit breaker
Cubicle Cubicle
Lighting system LED
HID
Fluorescent
Lighting control system Dimmer
Lighting censor
Time censor
Switch Electric socket
Switch
Under floor duct/Raised floor Under floor duct
system Raised floor system
Cable High voltage cable
Medium voltage cable
Low voltage cable
Grounding system Solid grounding
Lightning rod Lightning rod
Lightning rod–tight wire
Lightning rod—meshed cage
Fire alarm system and detector Smoke detector
Heat detector
Fire panel
Fire alarm
Telephone Telephone
Sound system Sound system
Network system Ethernet cable
Wi-Fi
CCTV TV
(continued)
Table 1 (continued)
Detail Detail
Wall display
Closed circuit television
(CCTV)
Signal amplifier Cellphone signal amplifier
Building automation system Censors
(BAS) Control panel
Presence machine Digital presence system
5.3 Preventive Maintenance Guidelines for Building

Electrical Components
Preventive maintenance activities in this guideline include inspection, service, and

maintenance [5, 25]. These activities are attached to every electrical component
included in the alternative design category. Each activity consists of preventive main-
tenance actions and procedures that have been validated by experts in the field of
maintenance. Every preventive maintenance actions and procedure for each electrical
component is equipped with a certain time interval. Generally, the time intervals for
preventive maintenance are daily, monthly, weekly and yearly [29]. The time interval
is also validated by experts in the field of maintenance.
For example, for liquid transformer components (Table 2), the procedures and
preventive measures are checking safety relays, checking bushings, checking termi-
nals, checking transformer stands, checking transformers’ physical conditions,
checking transformer temperatures, checking safety and measuring tools, checking
temperatures and transformer chamber air condition, check cable connections in
bushing terminals and grounding systems, clean the outside of the transformer, adjust
the temperature and condition of the transformer chamber air and replace equipment
and parts if damage occurs. All procedures and precautions are carried out at time
intervals every 6 (six) months except for replacements done at time intervals every
1 (one) year.
In the end, all phases that have been carried out are grouped into a comprehen-
sive format consisting of building electrical components, activities, procedures and
preventive measures of building electrical components together with the time inter-
vals of each of these components so that comprehensive guidelines for buildings are
produced as shown in Table 3.
Table 2 Preventive maintenance for liquid transformer

Component Activities Preventive actions and procedures Time interval
Liquid Transformer Inspection Check safety relays Biannually
Check bushing Biannually
Check the terminal Biannually
Check the transformer holder Biannually
Check the physical condition of the Biannually
transformer
Check the temperature of the transformer Biannually
Check security and measurement equipment Biannually
Check the temperature and air condition of Biannually
the transformer room
Check the cable connections on the bushing Biannually
terminal and the grounding system
Service Clean the outside of the transformer Biannually
Adjust the temperature and air condition of Biannually
Repair Replace equipment and spare parts if Annually
damage occurs
Table 3 Preventive maintenance guidelines

WBS Detail
WBS level 2 Work section Electrical
WBS level 3 Type of work Power supply
WBS level 4 Work package Transformer
Alt. design Liquid transformer
Activities Preventive actions and procedures Time interval
Biannually Annually
√
Inspection Check safety relays
√
Check bushing
√
Check the terminal
√
Check the transformer holder
√
Check the physical condition of the
transformer
√
Check the temperature of the transformer
√
Check security and measurement equipment
√
Check the temperature and air condition of
√
Check the cable connections on the bushing
terminal and the grounding system
√
Service Clean the outside of the transformer
(continued)
Table 3 (continued)
WBS Detail
WBS level 2 Work section Electrical
WBS level 3 Type of work Power supply
WBS level 4 Work package Transformer
Alt. design Liquid transformer
Activities Preventive actions and procedures Time interval
Biannually Annually
√
Adjust the temperature and air condition of
√
Repair Replace equipment and spare parts if
damage occurs
6 Conclusion
Some policies and regulations relating to government buildings, especially those

relating to maintenance, are sufficient to be used as a basis. In preventive main-
tenance, there are 4 (four) policies and regulations that apply as a reference. The
identification of electrical components obtained from the policies and regulations
can be compiled using WBS so that all electrical components in buildings can be
identified completely, thoroughly and well-structured. From the results of the valida-
tion, there are 22 (twenty-two) work packages and 45 (forty-five) alternative designs
for electrical components in buildings that need preventive maintenance with each
activity, procedure and preventive action and time intervals. The role of the WBS in
developing Preventive Maintenance Guidelines is very large. The results obtained
describe activities that are easy to understand and very neatly arranged. Each level in
the WBS has its own role according to the existing theory. The greater the level shows
the portion to break something down into more detail. Furthermore, Preventive Main-
tenance Guidelines can be developed again by identifying electrical components that
have not been listed according to technological advances.
Acknowledgements The authors would like to thank the members of the Maintenance Building
Team who have helped researchers to provide support in the form of both financial and moral support
in completing this research.
References
1. Direktorat Jenderal Anggaran Kementerian Keuangan (2015) Penundaan/Moratorium Pemban-

gunan Gedung Kantor Kementerian Negara/Lembaga, Januari 8. Dipetik September 24, 2019,
dari Direktorat Jenderal Anggaran Kementerian Keuangan. http://www.anggaran.depkeu.go.
id/dja/edef-konten-view.asp?id=1026
2. Pemerintah Indonesia (2002) Undang-Undang No. 28 Tahun 2002 Tentang Bangunan Gedung.
Lembaran Negara Republik Indonesia Tahun 2002, No. 134, Desember 16. Jakarta, DKI
Jakarta, Sekretariat Negara, Indonesia
3. Wimala M, Karim EKA, Adianto YLD, Setiawan TH, Roy AFV (2017) Pengembangan Sistem
Pelaporan Online Kerusakan Bangunan Asrama Kampus di Malaysia. Raka Racana: Jurnal
Online Institut Teknologi Nasional, Sept 2017
4. Chanter B, Swallow P (2007) Building maintenance management, 2nd edn. Blackwell
Publishing, Oxford
5. Pemerintah Indonesia (2008) Peraturan Menteri Pekerjaan Umum No. 24/PRT/M/2008 Tentang
Pedoman Pemeliharaan dan Perawatan Bangunan Gedung. Kementerian Pekerjaan Umum,
Jakarta, DKI Jakarta, Indonesia , Desember 30
6. Project Management Institute (2017) A guide to the project management body of knowledge
(PMBOK® Guide), 6th edn. Project Management Institute, Newtown Square
7. U.S. GAO (2009) GAO cost estimating and assessment guide: best practices for developing
and managing capital program costs. USGAO, Washington, District Of Columbia, USA
8. Moghaddam KS (2008) Preventive maintenance and replacement scheduling: models and
algorithms. Comput Ind Eng 1:85
9. U.S. Fire Administration (2016) Download fire data and data analysis tools. Dipetik Agustus
22, 2019, dari U.S. Fire Administration, Working for a fire-safe America: https://www.usfa.
fema.gov/data/statistics/order_download_data.html
10. Setiyo B (2014) Korsleting Listrik Penyebab Kebakaran Pada Rumah Tinggal Atau Gedung.
Edu Elektrika J 3(2).
11. Wardhana K, Hadipriono FC (2003) Analysis of recent bridge failures in the United States.
J Perform Constr Facil 17(3):144–150. https://doi.org/10.1061/(ASCE)0887-3828(2003)17:
3(144)
12. Ahzahar N, Karim NA, Hassan SH, Eman J (2011) A study of contribution factors to building
failures and defects in construction industry. Procedia Eng 20:249–255. https://doi.org/10.
1016/j.proeng.2011.11.162
13. Wimala M, Tamin RZ (2014) Pengembangan Konsep Perencanaan Biaya Pemeliharaan
Rutin Gedung Pendidikan di Institut Teknologi Bandung. Jurnal Itenas Rekayasa, May
2014, 11. Retrieved from https://www.researchgate.net/profile/Mia_Wimala/publication/316
282959_Pengembangan_Konsep_Perencanaan_Biaya_Pemeliharaan_Rutin_Gedung_Pen
didikan_di_Institut_Teknologi_Bandung/links/5939f83e0f7e9b32b749bae3/Pengembangan-
Konsep-Perencanaan-Biaya-Pemeliharaa
14. Yong CY, Sulieman MZ (2015) Assessment of building maintenance management practice and
occupant satisfaction of School Buildings in PERAK, Malaysia. Jurnal Teknologi (Sciences &
Engineering) 5:57–61
15. Njuangang S, Lasanthi Liyanage C, Akintoye A (2016) Performance measurement tool (PMT)
to control maintenance-associated infections. Facilities 34(13–14):766–787. https://doi.org/10.
1108/F-12-2014-0107
16. Pemerintah Indonesia (2005) Peraturan Pemerintah No. 36 Tahun 2005 Tentang Peraturan
Pelaksanaan Undang-Undang No. 28 Tahun 2002 Tentang Bangunan Gedung, September 10.
Lembaran Negara Republik Indonesia Tahun 2005, No. 83. Jakarta, DKI Jakarta, Indonesia:
Menteri Hukum dan Hak Asasi Manusia
Pedoman Teknis Pemeriksaan Berkala Bangunan Gedung, Desember 31. Berita Negara
Republik Indonesia Tahun 2010, No. 701. Jakarta, DKI Jakarta, Indonesia: Kementerian Hukum
dan Hak Asasi Manusia
18. Pemerintah Indonesia (2011) Peraturan Presiden No. 73 Tahun 2011 Tentang Pembangunan
Bangunan Gedung Negara. Jakarta, DKI Jakarta, Oktober 11. Sekretariat Kabinet, Indonesia
19. Beichelt F, Fischer K (1980) General failure model applied to preventive maintenance policies.
IEEE Trans Reliabil R-29(1):39–41. https://doi.org/10.1109/TR.1980.5220704
20. Murthy DNP, Atrens A, Eccleston JA (2002) Strategic maintenance management. J Qual
Mainten Eng 8(4):287–305. https://doi.org/10.1108/13552510210448504
21. Ahmad R, Kamaruddin S, Azid I, Almanar I (2011) Maintenance management decision model
for preventive maintenance strategy on production equipment. J Ind Eng Int 7(13):22–34
22. Lavy S, Garcia JA (2012) KPIs for facility’ s performance assessment Part II: identification
of variables and deriving expressions for core indicators. Facilities 32(5):275–294. https://doi.
org/10.1108/F-09-2012-0067
23. Chua SJL, Zubbir NB, Ali AS, Au-Yong CP (2018) Maintenance of highrise residential
buildings. Int J Build Pathol Adapt 36(2):137–151. https://doi.org/10.1108/IJBPA-09-2017-
0038
24. Vilarinho S, Lopes I, Oliveira JA (2017) Preventive maintenance decisions through maintenance
optimization models: a case study. Procedia Manuf 11(June):1170–1177. https://doi.org/10.
1016/j.promfg.2017.07.241
25. Peng R, Liu B, Zhai Q, Wang W (2019) Optimal maintenance strategy for systems with two
failure modes. Reliabil Eng Syst Saf 188:624–632, July 2017. https://doi.org/10.1016/j.ress.
2017.07.014
26. Project Management Institute (2006) Practice standard for work breakdown structures, 2nd
edn. Project Management Institute, Newton Square
27. Hidayah DN, Latief Y, Riantini LS (2018) Development of work breakdown structure standard
based on risk for safety planning on dam construction work. IOP Conf Ser Mater Sci Eng
420(1). https://doi.org/10.1088/1757-899X/420/1/012003
28. Burghate N (2018) Work breakdown structure: simplifying project management. Int J Commer
Manag Stud (IJCAMS) 3(2):1–5. Retrieved from www.ijcams.com
29. Starova-Sulejman L, Mojsoski N, Lega NH (2015) Guidelines manual for maintenance of
school buildings. Macedonian Civic Education Center (MCEC), Skopje, North Macedonia
Pedoman Persyaratan Teknis Bangunan Gedung. Jakarta, DKI Jakarta, Indonesia: Kementerian
Pekerjaan Umum
31. Saraswati R (2009) Perkembangan Pengaturan Sumber Hukum Dan Tata Urutan Peraturan
Perundang-Undangan Di Indonesia. Media Hukum IX(2):48–59
32. BPBD DKI Jakarta (2018) Data Rekapitulasi Kejadian Kebakaran di Provinsi DKI Jakarta
Tahun 2018. Dipetik Agustus 22, 2019, dari JAKARTA open data, Berbagi data untuk
transparansi. http://data.jakarta.go.id/dataset/0cb0984f-03a8-43b4-b481-7c55b24e947e/res
ource/7d595666-39bd-4803-aecc-801ac60014a9/download/REKAPITULASI-KEJADIAN-
KEBAKARAN-TAHUN-2018.csv
Evaluation of the Implementation of Fire
Safety Management Based on Work
Breakdown Structure Affecting
the Insurance Premium Costs
of High-Rise Lecture Buildings
E. P. Mahardika, F. Muslim, Y. Latief, and P. S. Nugroho
Abstract High-rise buildings, especially lecture buildings, have the potential for
fire hazards which can cause large losses. Fire risk is one of the factors that must be
minimized in a building operation where insurance is a way of transferring the risk.
However, the amount of fire insurance premiums in various insurance institutions
varies greatly due to differences in the determinants of insurance premium costs. The
purpose of this study is to evaluate the application of fire safety management in high-
rise lecture buildings in Jakarta and determine the determinants of premium costs that
are influenced by the implementation of fire safety management. This determining
factor is a tool to determine the amount of the insurance company’s fire insurance
premiums. This research refers to literature studies, surveys, interviews, and data
obtained using a questionnaire instrument. This study uses the work breakdown
structure as a tool to detail indicators that meet the fire safety management criteria
and the data are analyzed using the SmartPLS 3.0 software. Based on the survey
results, the data shows that the lecture buildings in Jakarta haven’t fully implemented
fire safety management. In addition, there is no clear relation regarding the role of
insurance in fire protection financing in lecture buildings. To address this, we propose
to consider the extent to which high-rise lecture buildings implement fire safety
management in determining insurance premium rates to improve the application of
fire safety management in high-rise buildings, by processing the data using SmartPLS
3.0 and average analysis in order to obtain priority factors in fire safety management
which can be used as a determinant of realistic premium costs. As the result, the
most priority factor in fire safety management which can be used as a determinant of
realistic premium costs is fire prevention in lecture buildings, and the least priority
factor is safety of people in the event of fire.
Keywords Fire safety management · Work breakdown structure · Insurance

premium costs · Lecture buildings
E. P. Mahardika (B) · F. Muslim · Y. Latief · P. S. Nugroho

F. Muslim
370 E. P. Mahardika et al.
1 Introduction
Building fires can occur in any type of designation of buildings, both residential
settlements, roles, and educational facilities. In Indonesia, several campus (educa-
tional) buildings occurred in the last ten years. The fire that occurred at the Campus
of the Islamic University of Malang [1] was a campus fire incident in June 2016
which was suspected to have been caused by a short circuit (electric short circuit).
The incident burned 2 of the 3 floors of the existing building and consumed several
laboratory rooms and their contents.
Campus fires not only occur in Indonesia, but these incidents also often occur
in several universities in America. It is recorded that from 2000 to 2015 there have
been 85 campus fire incidents, or an average of 7 campus fires per year, totaling
118 lives. A total of 58% of fire incidents that claimed lives originated from the
dormitory building, a student gathering place whose smoke detection equipment
was not working (damaged), while 85% of fatalities occurred in buildings that were
not equipped with sprinklers. And the cause of most events was power disruption
(Fig. 1).
The causes of fires can vary but can be categorized into two aspects, namely human
carelessness and failure of the building engineering system. These two aspects work
both during planning, construction implementation, and building operations (post-
occupancy). In Indonesia, the failure of engineering systems is higher, especially the
failure of the electrical system in buildings, such as short-circuit electricity [2]. A
good fire safety management implementation is the right effort in fire management,
starting from the planning stage to the follow-up. This action needs to be done in
Indonesia considering that Indonesia is one of the countries that have frequent fires
[3].
12
10
0
Electric short circuit Welding spark Other Unknown
Fig. 1 Graph of the cases of fires in Jakarta. Source republika.com (2018), cnnindonesia.com
(2019), and detik.com (2020)
Evaluation of the Implementation of Fire Safety … 371
To anticipate and mitigate the risk of fire, the owners of tall buildings cooperate
with an insurance company as a form of risk transfer. The form of cooperation
between individuals/legal entities and insurance companies is risk direction, either
partially or completely [4].
Therefore, each building must have fire management that plays a role in risk
prevention and countermeasures to achieve safety in fire events. The Provincial
Government of DKI Jakarta issued a Regional Regulation of the Special Region
of the Capital City of Jakarta Number 8 2008 on Fire Prevention and Prevention
to overcome the fire problem in Jakarta. The government through the Ministry of
Public Works and Public Housing also issued a provision referring to the Ministry
of Public Works and Public Housing 20/PRT/M/2009 [5] on technical guidelines
for fire protection management in urban areas, which among others state that every
public building including apartments, which are inhabited by at least 500 people, or
that have an area of at least 5000 m2 , or have a building height of more than 8 floors,
is required to apply fire safety management.
The implementation of fire safety management in buildings can be divided into
four points [6], namely:
1. Fire prevention in buildings
2. Safety of people in the event of a fire
3. Monitoring, auditing, and reviewing fire safety systems
4. Reactive monitoring (reporting, recording, and investigation)
Given the importance of fire safety, careful planning is required before the imple-
mentation of the project. According to the 6th edition of PMBOK [7], there are
several stages in planning the scope of the project, such as collecting requirements,
determining the scope of work, and creating a work breakdown structure.
The determination of fire premium rates in Jakarta is regulated in the Attachment to
the Circular letter of the Otoritas Jasa Keuangan (OJK) Number 6/SEOJK05/2017 [8]
Attachment I. The attachment determines the amount of premium tariff by the lower
limit and upper limit, which is only divided based on the occupancy of the building and
its Construction Class. The impact, according to the Indonesian Insurance Bureau,
Ministry of Finance Decree, and Otoritas Jasa Keuangan said that in the last 21 years
the growth rate of insurance premiums averaged 22% per year, while in the last
4 years (until 2017) growth according to up to only 8% per year.
This research was conducted in the province of DKI Jakarta and the object of
research is a high-rise lecture building. The focus of this research is on mechanical
electrical work in which there is fire safety management work.
2 Literature Review
2.1 High-Rise Lecture Buildings
According to Republic of Indonesia Government Regulations Number 36 of 2005 [9]

and Regional Regulation of DKI Jakarta Province Number 7 of 2010 [10], a building
is a physical form of construction work that integrates with its domicile, partly or
wholly above and/or above. On land and/or in water, which functions as a place
for humans to carry out their activities, whether for shelter or residence, religious
activities, business activities, social, cultural, or special activities.
The development of science and technology must be accompanied by high formal
education and adequate infrastructure. Therefore, a university was established which
could provide knowledge widely and globally. Meanwhile, there is an increase in the
need for lecture buildings that must be served, for that it is necessary to increase the
usability of the building either in the form of adding new buildings or constructing
new buildings in more appropriate and strategic places.
According to National Fire Protection Association 101 [11] or National Fire
Protection Association 101A [12], one type of building based on the level of occu-
pancy load is an educational building where this building has the objective of activities
related to education through classrooms with 6 (six) occupants or more. for 4 h per
day or more than 12 h per week.
2.2 Fire Safety Management Based on Work Breakdown

Structure
Given the importance of fire safety, careful planning is required before the imple-
mentation of the project. According to the 6th edition of PMBOK [7] there are
several stages in planning the scope of the project, such as collecting requirements,
determining the scope of work, and creating a work breakdown structure.
Work breakdown structure is a multi-level framework for organizing and graph-
ically representing the elements of work displayed to be completed in logical rela-
tionships [13]. Work breakdown structure is one of the tools available for a project
manager to define and structure projects [14]. A well-designed work breakdown
structure is critical to the success of the proposed cost estimation methodology [15].
WBS can also be used as a repetitive process in the form of templates useful for
future projects [16].
2.3 Insurance Premium Costs
Insurance is a tool used in reducing the risk inherent in the economy where it combines
a large enough number of units affected by the same risk to predict the potential that
losses will occur and if these losses do occur, it will be shared by all parties that are
joined together proportional [17].
The determination of fire premium rates in Jakarta is regulated in the Attachment
to the Circular letter of the Otoritas Jasa Keuangan (OJK) Number 6/SEOJK05/2017
[8] attachment I. The attachment determines the amount of premium tariff by the
lower limit and upper limit, which is divided based on the occupancy of the building
and its Construction Class. The types of coverage according to Article 247 of Kitab
Undang-Undang Hukum Dagang (KUHD):
1. Fire hazard
2. Dangers that threaten agricultural products that have not been harvested
3. The soul of one or more people
4. The danger of the sea and the danger of slavery
5. Transportation hazards on land, in rivers and inland waters
In Undang-Undang Number 40 of 2014 [18] on Insurance, paying premiums is the
obligation of an insurance policy holder. Article 1 point 29 states that the premium is
the amount of money determined by the insurance company or reinsurance company
and approved by the policyholder to be paid under the insurance agreement or rein-
surance agreement, or the amount of money stipulated under the provisions of the
legislation underlying the insurance program is obliged to obtain liability benefits.
The amount of premium for participation in the insurance to be paid has been deter-
mined by the insurance company about concerning to the circumstances of the insured
[19].
Research method is a scientific way to obtain data for specific purposes and uses
[20]. The scientific way means that research activities are based on scientific char-
acteristics, namely rational, empirical, and systematic. To answer the formulation of
these research question, a literature review was conducted and validation was carried
out using the Delphi method. The data was collected through a survey of respondents
using a questionnaire instrument and a case study method with a simulation of the
cost of fire insurance premiums for lecture buildings based on determined variables.
The questionnaire used contains indicators of fire safety management from literature
reviews and the results of leverage, clarification, and validation by experts. The ques-
tionnaire was filled out by the management of the high-rise lecture building. Each
response item in the Likert sample scale has five responses, namely very influential,
influential, moderately influential, not influential, and very unaffected. The results
of this data are analyzed descriptively and processed using SmartPLS software, and
the average analysis is to obtain priority factors in fire safety management which can
be used as a realistic determinant of premium costs.
In this research, data were obtained as many as 31 respondents building manage-
ment of the lecture building and 31 respondents are fire insurance underwriters. The
questionnaire was filled in by the building management of high-rise lecture buildings
and fire insurance agencies in Jakarta. Criteria for building management respondents
include bachelor or diploma education and 5 years of building management expe-
rience, especially in the fire prevention department of lecture buildings. And the
criteria for fire insurance underwriter include a minimum education of bachelor or
diploma, have an underwriter position in an insurance company and at least 2 years
of experience.
4.1 Mechanical Electrical Work Breakdown Structure

on High-Rise Lecture Buildings to Plan Fire Safety
Management
At this stage, interviews and discussions with experts are conducted to identify
and validate the work breakdown structure related to the components of fire safety
management for mechanical and electrical work in the high-rise lecture building.
The work breakdown structure of the literature review was then carried out by
the experts to get the initial validation, clarification, and validation related to the fire
safety management components of mechanical and electrical work in the high-rise
lecture building. Here are the results (Table 1):
4.2 The Implementation of Fire Safety Management

on High-Rise Lecture Buildings in Jakarta Based
on Work Breakdown Structure
The work breakdown structure of mechanical electrical work in the study [21] was
developed into a work breakdown structure for mechanical electrical work that meets
fire safety criteria. The points contained in the work breakdown structure of mechan-
ical electrical work are then carried out preliminary validation to several experts
to obtain verification, clarification, and validation related to the components of fire
safety management for mechanical and electrical in high-rise lecture buildings.
A book [6] entitled introduction to fire safety management explains and mentions
indicators of fire safety management. The fire safety management indicator in the
Table 1 Work breakdown WBS level 3 WBS level 4

structure fire safety
Sub work section Work package
Fire system Fire pump
Panel control
Cable power
Pipe riser
Diesel tank
Fire pipe
Valve and accessories
Hydrant box
Sprinkler head
Fire extinguisher
Pillar hydrant
Siamese connection
Compartment
Fire lift
monitoring center
Emergency lighting
Fire system testing
Fire alarm Main control panel
Announciator
grounding system
Cable feeder
Detector
Manual break glass push button
Alarm bell
Firemen telephone
Terminal box
Cable tray
Fire alarm testing
Sound system Paging
Gas –
Telephone –
Data –
MATV –
CCTV –
Security system –
Radio communication –
Audio videophone –
Building autosystem Monitoring center
Electrical Emergency lighting
Air conditioning works Pressurized fan
Transport equipment Fire lift
books [6] are added by several additional indicators from the results of the develop-
ment of work breakdown structures for mechanical electrical work that meet the fire
safety criteria. Furthermore, these indicators are used to conduct this research.
The results of the implementation of fire safety management in the high-rise
lecture building will be obtained using descriptive analysis of the data from the
questionnaire. Descriptive analysis is done by describing or explaining the collected
data. Based on data from samples of high-rise lecture buildings in Jakarta, data on
the fire safety management indicator applied to the building were obtained. The
following is a recapitulation of percentage data by high-rise lecture buildings which
are presented in tables and diagrams:
Based on these four dimensions, fire safety management hasn’t been fully imple-
mented 100% in every indicator. The lowest average implementation is the dimen-
sion of safety people in the event of a fire which is 78%, while fire protection in a
building is 83%, reactive monitoring (reporting, recording and investigation) 82%,
and monitoring auditing and reviewing fire safety systems 80%. This percentage
shows that high-rise lecture buildings in Jakarta have not fully implemented fire
safety management, which is also in line with previous research [22].
4.3 Indicators of Fire Safety Management Implementation

that Affecting Insurance Premiums Cost
Based on data testing through SmartPLS 3.0 software obtained by T-Statistic which
shows that the four dimensions of fire safety management affect the cost of premium
(Y) significantly Table 2. This is also in line with the results of the analysis that uses
the average value, where the priority of dimension influence has the same order as
the highest position, namely Fire Prevention in Building Buildings and the lowest is
People Safety in Fire Incidents Fig. 2.
Based on Table 2 it is known that the order of variable implementation of fire safety
management that affects premium costs are fire prevention, reactive monitoring,
active monitoring and people safety. This is in accordance with the journal [22] which
states that the implementation of fire safety management affects the calculation of
premium costs.
In addition, it is also obtained through SmartPLS 3.0 software with path coefficient
test as shown in Table 3, that the path coefficient of the calculation result is close to
Table 2 Significance
T-statistic
indicators of fire safety
management implementation X1 → Y 2.227 Significant
X2 → Y 1.951 Less significant
X3 → Y 2.009 Significant
X4 → Y 2.137 Significant
100%
95%
90%
85% 83% 82%
80%
80% 78%
75%
70%
65%
60%
Fire Prevention Safety of People Monitoring Reaction
Auditing Monitoring
Fig. 2 Graph of the implementation of fire safety management in Jakarta
Table 3 Path coefficient

X Y
X1 −0.291
X2 −0.069
X3 −0.194
X4 −0.280
−1. This indicates a strong negative relationship, and proves that the more applied
fire safety management, the smaller the cost of insurance premium.
5 Conclusions
Based on the discussion of the research results that have been described in the previous
chapter, it can be concluded:
1 Obtained work breakdown structure of mechanical electrical as shown in Table
1. in high-rise lecture buildings.
2 Safety management based on the work breakdown structure in the high-rise
lecture buildings in Jakarta hasn‘t fully implemented. With an average value of
81%, it means that it is in the good category
3 Factors/indicators determining premium costs based on the quality of work
breakdown structure based fire safety management in high-rise lecture build-
ings are derived from the four dimensions of fire safety management with the
order of the most priority is fire prevention in building buildings, reactive moni-
toring (reporting, recording, and investigation), monitoring, audit, and review
safety system, then people safety in fire incidents.
Acknowledgements The Authors would like to thank the financial support provided by Ministry
of Research and Technology/National Research and Innovation Agency through PDUPT Grant
2021 with contract number: NKB-220/UN2.RST/HKP.05.00/2021 managed by the Directorate for
Research and Community Engagement (DRPM) Ministry of Research and Technology/National
Research and Innovation Agency (BRIN).
References
1. LNCS Homepage. https://news.detik.com/berita-jawa-timur/d-3239846/kampus-unisma-keb

akaran-diduga-konsleting-listrik. Last accessed 2016/06/22
2. Sufianto H, Green AR (2012) Urban fire situation in Indonesia. Fire Technol 48(2):367–387
3. Nugraha R (2018) Penerapan Sistem Manajemen Kebakaran Di PT. Application of fire manage-
ment system in PT. Adiluhung. Indonesian J Occupat Saf Health 7(3):378–386. https://doi.org/
10.20473/ijoshv7i3.(2018).378
4. Kusumoardi AN, Rinitami Njatrijani S (2016) Klausula Baku dalam Polis Asuransi Kebakaran
Ditinjau dari Undang-undang Nomor 8 Tahun 1999 Tentang Perlindungan Konsumen (Studi
pada PT. Asuransi Jasa Indonesia). Diponegoro Law J 5(3):1–13
5. Peraturan Menteri PUPR No. 20/PRT/M/2009. Pedoman Teknik Manajemen Proteksi
Kebakaran di Perkotaan (2009)
6. Furness A, Muckett M (2007) Introduction to fire safety management. Routledge
7. PMBOK (2017) PMBOK Guide, 6th edn. Project Management Institute, pp 104–104. https://
doi.org/10.1002/pmj.20125
8. Jasa Keuangan O (2017) Surat Edaran Otoritas Jasa Keuangan Nomor. 6/SEOJK.05/2017
Tentang Penetapan Tarif Premi Atau Kontribusi Pada Lini Usaha Asuransi Harta Benda dan
Asuransi Kendaraan Bermotor Tahun 2017
9. Peraturan Pemerintah Nomor. 36 Tahun 2005. Peraturan Pemerintah Republik Indonesia Nomor
35 Tahun 2005 Tentang Peraturan Peaksanaan Undang-Undang Nomor 28 Tahun 2002 Tentang
Bangunan Gedung (1), pp 1–5 (2005)
10. Peraturan Daerah Provinsi DKI Jakarta Nomor 7 Tahun 2010. Peraturan Daerah Provinsi DKI
Jakarta Nomor 7 Tahun 2010 Tentang Bangunan Gedung (2010)
11. National Fire Protection Association. NFPA 101-Life safety code (2000)
12. National Fire Protection Association (2000) NFPA 101-Guide on Alternative Approaches to
Life safety code
13. Burghate M (2018) Work breakdown structure: simplifying project management. Int J
Commerce Manag 3
14. Sutrisna M, Ramanayaka CD, Goulding JS (2018) Developing work breakdown structure
matrix for managing offsite construction projects. Architect Eng Des Manag 14:381–397
15. Sequeira S, Lopes E (2015) Simple Method Proposal for cost estimation from work breakdown
structure. Procedia Comput Sci 64:537–544
16. Supriadi R, Sagita L, Latief Y, Susilo B, Rajasa M (2017) Development of risk-based standard-
ized WBS (Work Breakdown Structure) for cost estimation of apartment’s project. Int J Civil
Eng Technol (IJCIET):822–833
17. Sunyoto D, Putri WH (2017) Manajemen Risiko dan Asuransi. CAPS, Yogyakarta
18. KUHD (1936) Kitab Undang-Undang Hukum Dagang 49(1):1–9. https://doi.org/10.1016/j.sol
ener.2019. 02.027
19. Agustiranda W, Yuliani Y, Bakar SW (2018) Pengaruh Pendapatan Premi, Pembayaran Klaim,
Dan Risk Based Capital Terhadap Pertumbuhan Laba Pada Perusahaan Asuransi Yang Terdaftar
Di Bursa Efek Indonesia. Sriwijaya University
20. Darmadi H (2013) Metode Penelitian Pendidikan dan Sosial. Alfabeta, Bandung
21. Ajizah N (2018) Perencanaan Sumber Daya Pada Pekerjaan Mekanikal dan Elektrikal
Bangunan Gedung Apartemen Berbasis Work Breakdown Structure. Universitas Indonesia,
Fakultas Teknik
22. Winarningsih YR (2002) Pemodelan Pengaruh Sistem Proteksi Kebakaran terhadap Premi
Asuransi Kebakaran pada Bangunan Tinggi di Jakarta. Universitas Indonesia, Fakultas Teknik
A Critical Review of Bridge Management
System in Indonesia
Surya Dewi Puspitasari , Sabrina Harahap , and Pinta Astuti
Abstract The bridge is a crucial infrastructure in public mobilization. Recently, the

utilization of old bridges has intensely come into an option during limited funds in
many developing countries. Hence, the bridge management system plays an impor-
tant role in maintaining and reducing bridge failure. In Indonesia, the bridge manage-
ment system is called Interurban Bridge Management System (IBMS). However, the
IBMS implementation in Indonesia is still poor. This system has been implemented
throughout all provinces in Indonesia since 1992, yet many bridges were still found
damaged, even collapsed. At this moment, Indonesia is still focusing on construction,
not maintenance work. Some developed countries like the United States of America
and Japan have expanded a complex system for a bridge management system, Pontis,
and J-BMS. Thailand as a developing country is the same as Indonesia has also real-
ized the importance of having a bridge management system. Thailand has established
a Bridge Maintenance and Management System (BMMS) to guide the periodic bridge
inspection and maintenance program. In this study, the authors intend to grow the
awareness of the IBMS application in Indonesia which consist of the development of
bridge management system, the causes of poor IBMS implementation, the improve-
ment needs to enhance the IBMS implementation, and recommended aspects may
extend to current IBMS implementation in Indonesia by seeking for and integrating
the advantageous information the existing bridge management system (BMS) that
has been established and implemented in the U.S., Japan, and Thailand.
Keywords Bridge · IBMS · Maintenance · Management system
S. D. Puspitasari (B) · S. Harahap

Department of Civil Engineering, Faculty of Military Engineering, Universitas Pertahanan RI,
Bogor 16810, Indonesia
S. Harahap
P. Astuti
Yogyakarta, Yogyakarta 55138, Indonesia
382 S. D. Puspitasari et al.
1 Introduction
Indonesia has been becoming the world’s 10th largest economy and the most enor-
mous one in Southeast Asia [1], encouraging Indonesia’s infrastructure to grow
rapidly. Indonesia has a total of 18.648 existing bridges in 2019 which were reported
around 14% in unhealthy condition [2]. The increasing traffic volume and the weight
of vehicle, that pass through the bridge indicates there has been overloading which
led to being suspected as the main cause of bridges deteriorated [3]. Bridges have
an indispensable role in the transportation system to support people and goods in
their mobility and accessibility which affect social and economic development. It is
becoming substantial to avoid severe damages. The Ministry of Public Works and
Public Housing of the Republic of Indonesia is the executive of its Interurban Bridge
Management System (IBMS) through its agency Directorate General of Bina Marga.
Unfortunately, the implementation of IBMS in Indonesia is still poor. Consequently,
it has caused some bridges to be damaged and collapse during their service life, such
as the collapse of the Kutai Kartanegara bridge in 2011. An assessment of bridge
maintenance practice in Indonesia is needed to enhance the practices of the IBMS
whether the procedure, standard, bridge engineer, bridge maintenance engineer, or
bridge inspector.
On the other hand, the U.S. and Japan as developed countries have expanded their
bridge management system through the Federal Highway Administration (FHWA)
and Road Bureau of the Ministry of Land, Infrastructure, Transport and Tourism
(MLIT), respectively. Besides, our neighbour country, Thailand the same as a devel-
oping country as Indonesia, has come to realize the importance of having a bridge
management system. This country has established a bridge maintenance management
system (BMMS) to undergo a periodic bridge inspection and maintenance program.
Therefore, to enhance the practices of the IBMS in Indonesia, the authors wrote this
paper to increase the awareness-raising of the IBMS implementation in Indonesia
which consist of the development of bridge management system, the causes of poor
IBMS implementation, the improvement needs to enhance the IBMS implementation
and recommended aspects may extend to current IBMS implementation in Indonesia
by seeking for and integrating the advantageous information from the existing bridge
management system (BMS) that has been established and implemented in the U. S.,
Japan, and Thailand.
2 Development of Bridge Management System
The bridge management system is a series of actions to assure the structural perfor-
mance of a bridge structure in good condition during its service life without heavy
maintenance [4]. The BMS encompasses all activities related to bridges’ manage-
ment, for example, inspections, planning, programming, design, construction, and
maintenance [5–8]. Aiming to boost the efficacy and quality of maintaining the bridge
A Critical Review of Bridge Management System in Indonesia 383
structure, the Directorate General of Bina Marga constructed the IBMS associated
with SMEC-KinHill in 1991. Since then, all the National and Provincial Roads in
Indonesia, including bridge infrastructures, has been using the IBMS for maintainace
guidance [5, 9].
The U.S. through the FHWA has started creating the National Bridge Inspec-
tion Standard (NBIS) and made a program to train their bridge inspectors in 1971,
soon after the Silver Bridge collapsed in 1967. The NBIS was responsible for the
national policy concerning bridge inspection frequentness, inspector eligibility, advi-
sory form, and rating method. The revised NBIS was in 1988, and a few years later,
developing PONTIS for the BMS to help the authority in decision-making and plan-
ning regarding bridge monitoring, maintenance, and rehabilitation. The updated and
modified NBIS provide more comprehensive and useful information [10].
Japan evolves its J-BMS by using advanced technologies and techniques in infor-
mation processing and health monitoring. A combined life-cycle management system
was built as Doctor for Infrastructure Systems, called InfraDoctor. This system has
three main features, i.e., basic features for GIS and 3D point group data; manage-
ment and search features for GIS data; and further sophisticated management and
advanced features [11]. From this system, we could possess real-time monitoring,
structural performance, prediction of behaviour, and the type of maintenance that
could be done [12].
In 1989, Thailand’s Department of Highway (TDOH) collaborates with the
Danish Road Directorate to establish the bridge management and maintenance
system (BMMS) for more than 10,000 concrete highway bridges in Thailand [6,
13]. However, due to several constraints and limitations in utilizing them, BMMS is
not thoroughly applied in the present system. Thus, the defined works on the inspec-
tion and evaluation of bridge performance conducted by TDOH ally with TESCO
Ltd. in 2008 [6]. In 2013, Thailand, JICA and Japanese’s company set a new manual
for bridge maintenance in rural area [14].
3 Factors May Cause the IBMS Implementation

in Indonesia is Still Poor
The IBMS has been employed in Indonesia since 1993 [5], yet the implementation
of this system is still poor. There were at least 17 collapsed bridges were recorded
have taken place in Indonesia from 2007 to 2011 [15]. Recently, a suspension bridge
in West Java was also reported almost collapsed in January 2021. It is only five years
after the bridge has been completely constructed. It was presumed that there is a
shifting part on the bridge due to flooding [16]. The authors outlined three factors
that may cause the poor implementation of IBMS in Indonesia as follows:
3.1 Scarcity of Funds
Insufficient financial budgeting in bridge management has been a problem for all
countries, including Indonesia. In the period from 2013 to 2018, the Indonesian
government entrusts about 42 billion USD1 to The Ministry of Public Works and
Public Housing of the Republic of Indonesia to enhance infrastructure development.
This amount is not only for the bridge but also for water resources, highways, housing,
etc. [17]. In 2018, maintaining the 46,812 km road and the 495,698 m length of the
bridge requires 1.65 billion USD [18]. Dananjoyo has calculated one bridge main-
tenance cost for replacement works that sum up to 3,5 million USD with the IBMS
method in 2020 [19]. In addition, about 2 thousand bridges with 71,549.6 km length
were recorded in poor condition in 2019 [20]. Comparing the maintenance cost of
one bridge and the repair work needs for all bridges to the available funding is inad-
equate. The high bridge maintenance cost cannot be covered all by the government,
thus it is reasonable if the IBMS implementation in Indonesia is still unsatisfactory.
3.2 Insufficient Human Resources
The development of transportation networks in Indonesia does not equal human

resources number with bridge inspection speciality [21]. Inspection activities are
fundamental works in BMS to determine the effective treatment for the bridge at the
right time [6]. Roughly 7,398 people from different backgrounds and level of educa-
tions work for the Directorate General of Bina Marga that liable for the 32,790,18 km
provincial road, the 211,644,08 km district road, the 35,241.40 km city street, the
2100 km length of the existing toll road, and 18,648 bridges with 510,336 km length
[2]. Looking at the numbers of officials compared to the length of highways and
bridges, it does make sense if there are insufficiently qualified inspectors and analyt-
ical resources, particularly in a rural area. The bridge inspector or maintenance
officer may not certify nor an expert in bridge management. Furthermore, currently,
Indonesia still focuses on construction, not maintenance works [22].
3.3 Ignoring Standard, Procedure, Manual
The collapse of the Kutai Kartanegara Bridge on November 26, 2011, shocked people
generally and the engineering community particularly. This bridge is supposed to last
at least 40–100 years of service life [15]. However, in its 10th year of service life,
this bridge collapsed unpredictably and rapidly after the maintenance works. This is
evident that ignoring standard, procedure and manual are still becoming the biggest
issue for Indonesia in terms of IBMS application. Kawai in 2014, mentioned that an
1 1 IDR = 0.000070 USD.

accumulation of design errors, misconduct in maintenance, and reconstruction work

had driven the bridge to collapse [23]. Maintenance works in Indonesia is still weak
and carried out poorly [24]. The bridge failure is due to the alteration of deflection by
retightening the hangers clamp should have been gone in advance by the main cables
of retightening. Analyzing joint conditions must come first before retightening was
carried out [24, 25]. Ignoring standard might also happen due to human error, such
as lack of knowledge and competence, the officer did not mean to neglect it, but it
was pure owing to their lack of knowledge.
4 How to Improve the IBMS Implementation in Indonesia
Coping with some issues regarding the poor implementation of IBMS in Indonesia,
the authors propose a ‘bottom-up’ approach as a fundamental philosophy. The stan-
dard helps in proposing a strategy. The proposed strategy’s cost is compared to the
budget to accommodate the standard and customized strategy [26]. Following the
proposed concept, in the authors’ opinion, the IBMS requires the optimization model
of the bridge maintenance strategy. Some considerations are necessary to make an
optimum bridge maintenance strategy, like the structural performance view (i.e.,
deteriorating structures) and economic view (i.e., the life-cycle costing and the net
present value). The combination policy optimizes the structural performance, maxi-
mizes the net present value (NPV), and minimizes the life-cycle cost (LCC) that
has been successfully investigated the application to the wharf [27–29], so does the
bridge [19, 20, 30].
NPV describes more about the benefit, this extent maximizing NPV is the same
as maximizing the benefit value. LCC got influence from the life-cycle period of
the bridge structures, which usually between 50 and 100 years. Budget estimation
and project priority poll are for 10 or 15 years budgeting period. Hence, the most
presentable mix in planning and time scheduling on maintenance and improvement
work under budget constraints could be achieved [31]. Following this approach,
insufficient financial and ignoring standard issues could be tackled. According to the
IBMS method, the bridge maintenance cost results are lower than using the LCC
method, because the inspection and additional life-cycle cost did not exist on the
IBMS method, which are important factors [19]. The success of the optimization
model of bridge maintenance implementation depends on the data (structural condi-
tion and cost) that we have. The more available data, the more accurate analysis is,
the better the optimization model we will obtain. Thus, collecting data activities are
essential.
Insufficient human resources handling by recruiting more people, then train them
to become qualified bridge engineers, bridge inspectors, or bridge maintenance engi-
neers. Besides, changing the concern of the Indonesian government from constructing
only to construction and maintenance is also needed.
Table 1 Proposed aspects to be added for IBMS from BMS in Thailand, U. S., and Japan
Aspects Thailand US Japan References
√ √ √
Safety [7, 8, 32, 33]
√ √ √
Risk [7, 8, 32, 33]
√ √ √
Degradation model [7, 8, 33]
√ √ √
Effects of maintenance work [7, 8, 33]
√ √
Life-cycle costing × [7, 8, 33]
√ √
Optimal maintenance strategies × [7, 8, 33]
5 Recommended Aspects from BMS Implementation

in the U.S., Japan, and Thailand
Bridge management systems (BMS) employed in the United States, Japan, and Thai-
land for bridge maintenance practices, which could be applied to current IBMS proce-
dures in Indonesia are provided in Table 1. Those aspects are such as safety, risk,
degradation model, effects of maintenance work, and optimal maintenance strategy.
Safety is the assessment of structure level to know the collision possibility (prob-
ability of failure). The bridge could be assessed based on the design code. Risk,
same as safety is also an assessment type to recognize the collapse’s effect and
possibility. Deterioration of a major member, deterioration of a minor member, pier
failure because of scouring, road accident owing to sub-standard guardrails, and loss
of life due to the earthquake are five risk factors to be considered. Both aspects are
obtained from inspection activities and collected as data [33].
To forecast the structural condition (predict the degradation) of the bridge at any
given time point in the forthcoming, thus the degradation should be modelled [31].
Effects of maintenance work are the improvement works after the bridge has already
gotten the repair actions (i.e., change in physical condition and performance barome-
ters). Generally, the degradation model and the effects of maintenance work following
probabilistic model, for example, Markov Chain model [33]. The degradation model
along with the cost model (e.g., LCC) are supposed to define the optimization model,
at least for maintenance cost, repairing works, and rehabilitation works to obtain the
optimal maintenance strategies [31]. By adding those mentioned aspects, the IBMS
is expected to be able to define the most appropriate and optimum policies for the
bridge management system in Indonesia.
6 Conclusions
In this paper, some factors that may cause the poor implementation of IBMS in
Indonesia and some recommendations adopted from BMS implementation in the
U. S., Japan, and Thailand have been described. Finally, the following conclusions
could be derived:
• Improvement of IBMS practice in Indonesia is highly required by making opti-
mization of the bridge maintenance model with the structural performance and
economic consideration. However, it is going to be a more perfect model if we
also could consider the environmental issues in the optimization model for IBMS.
• Improving and updating the procedure and the standard quality as well as special
training for bridge inspectors and maintenance officers. So that, the right treatment
can be performed at the right time and on the right bridge.
• Having more detailed and comprehensive information for some aspects in IBMS,
for example, safety, risk, degradation model, effects of maintenance work, life-
cycle costing and optimal maintenance strategies are expected to be able to define
and apply for optimum maintenance works within budget.
References
1. Worldbank Homepage. https://www.worldbank.org/en/country/indonesia/overview. Last

accessed 2021/02/22
2. Data and Information Technology, General Secretary, The Ministry of Public Work and Public
Housing of the Republic of Indonesia (2020) Infrastructure statistical information PUPR 2020.
PUPR, Jakarta (In Bahasa Indonesia)
3. Marwoto S (2019) Damage evaluation application for concrete bridges based on expert systems
with android technology. Jurnal Riset Rekayasa Sipil Universitas Sebelas Maret 3:34–41
4. Sriskandan K (1990) Bridge management—an overview. In: Harding JE, Parke GAR, Ryall
MJ (1990) Bridge management—inspection, maintenance, assessment and repair. Springer
Science+Business Media Dordrecht, London, pp 17–29
5. Directorate General of Bina Marga, Ministry of the Public Works Republic of Indonesia (1993)
Bridge management system (BMS): bridge inspection manual. Directorate General of Bina
Marga, Jakarta (In Bahasa Indonesia)
6. Suksuwan N (2010) Condition rating system for Thailand’s concrete bridges. J Construct Dev
Countries 15(1):1–27
7. TDOH (2012) BMMS: bridge management and maintenance system. Department of Highway,
Bangkok
8. Suksuwan N, Hadikusumo BH (2009) Development of computer applications for bridge
management system. In: IABSE symposium: sustainable infrastructure—environment friendly,
safe and resource efficient. Sustainable infrastructure—environment friendly, safe and resource
efficient, Bangkok, pp 130–139
9. Ginting DT (2001) Study on the efficiency of bridge maintenance at BMS Bina Marga
and efforts to determine the pattern of maintenance cost allocation [Master Thesis]. Institut
Teknologi Bandung, Bandung (In Bahasa Indonesia)
10. Lee S, Kalos N, Shin DH (2014) Non-destructive testing methods in the U.S. for bridge
inspection and maintenance. KSCE J Civil Eng 1–10
11. MLIT Hompage. https://www.mlit.go.jp/road/road_e/s3_maintenance.html, Last accessed
2021/03/27
12. Miyamoto A (2012) Life-cycle performance of deteriorating structures. In: Proceedings of the
3rd IABMAS workshop on life-cycle cost analysis and design of civil infrastructures systems.
American Society of Civil Engineers, Lausanne
13. Anders B, Sorensen FB (1990) Implementation of bridge management and maintenance

systems (BMMS) in Europe and the Far East. In: Harding JE, Parke GAR, Ryall MJ (eds) Bridge
management—inspection, maintenance, assessment and repair. Springer Science+Business
Media Dordrecht, London, pp 28–38
14. Japan International Cooperation Agency (JICA) CHODAI Co., Ltd and Metropolitan
Expressway Company Limited.: Kingdom of Thailand the Project for Bridge Master Plan and
Bridge Maintenance Ability in Rural Area Report. Japan International Cooperation Agency,
CHODAI Co., Ltd and Metropolitan Expressway Company Limited, Tokyo
15. Putra GA, Rio K, Hatmojo D (2012) Study on the importance of bridge management system
(BMS) applications and standard operating procedures (POS) in maintaining the quality value
of bridge infrastructure, case study bridge infrastructure—bridges in Indonesia. In: SCAN#3
National Proceeding Conference. Universitas Atma Jaya Yogyakarta, Yogyakarta (2012)(In
Bahasa Indonesia)
16. Banten suara Homepage. https://banten.suara.com/read. Last accessed 2021/02/26 (In Bahasa
Indonesia)
17. Data and Information Technology, General Secretary (2018) The ministry of public work and
public housing of the Republic of Indonesia.: infrastructure statistical information PUPR 2018
(In Bahasa Indonesia). PUPR, Jakarta
18. Bisnis Homepage. https://ekonomi.bisnis.com/read/20180426/45/788964/dana-pemeliharaan-
jalan-jembatan-rp2370-triliun. Last accessed 2021/05/27 (In Bahasa Indonesia)
19. Dananjoyo RA, Aminullah A, Nugroho ASB (2020) Application of life-cycle cost method in
calculation of economic evaluation of bridges to determine the priority of bridge handling.
Jurnal Teknosains 9(2):91–180 (in Bahasa Indonesia)
20. Liu Min AM, Frangopol DM (2005) Multiobjective maintenance planning optimization for dete-
riorating bridges considering condition, safety, and life-cycle cost. J Struct Eng 131(5):833–842
21. The Ministry of Public Work and Public Housing of the Republic of Indonesia.: Module 2
Bridge Management System. PUPR Internal Module, Bandung (2017) (In Bahasa Indonesia)
22. Sumargo, Hakiki R, Raafidiani R (2020) Condition inspection and handling of bridges in Nias
Island, North Sumatra province by using the bridge management system method. Jurnal Sipil
Politeknik 22(2):156–168 (in Bahasa Indonesia)
23. Kawai Y, Dionysius S, Yozo F (2014) Failure analysis of the hanger clamps of the Kutai-
Kartanegara Bridge from the fracture mechanics viewpoint. J Jpn Soc Civil Eng 2:1–6
24. 逊汪 Homepage. https://www.academia.edu/5832394/A_STUDY_ON_THE_CAUSE_OF_
KUKAR_BRIDGE_COLLAPSE. Last accessed 2021/02/26
25. Luknanto Homepage. https://luk.staff.ugm.ac.id/kukar. Last accessed 2021/02/26 (in Bahasa
Indonesia)
26. Small E, Philbin T, Fraher M, Romack G (1999) The current status of bridge management
system implementation in the United States
27. Puspitasari SD, Kawabata Y, Yokota H (2019) Optimization of life-cycle management on port
mooring facilities. In: Proceedings of the 21st International Summer Symposium in junction
with the 74th JSCE Annual Conference. Japan Society of Civil Engineers, Takamatsu
28. Kawabata Y, Kato E, Tanaka Y (2018) A maintenance management plan for port mooring facil-
ities based on cost-benefit analysis—a case study. In: Proceedings of life-cycle of engineering
systems: towards an integrated vision. international association on life-cycle civil engineering,
Ghent
29. Yokota H, Hashimoto K, Tani T (2016) Sensitivity analysis on repair prioritization for mooring
facilities. In: Life-cycle of engineering systems: emphasis on sustainable civil infrastructure.
International Association on Life-Cycle Civil Engineering, Delft
30. Xie HB, Wu WJ, Wang YF (2018) Life-time reliability based optimization of bridge
maintenance strategy considering LCA and LCC. J Clean Prod 176:36–45
31. Jeong Y, Kim W, Lee I, Lee J (2018) Bridge inspection practices and bridge management
programs in China, Japan, Korea, and US. J Struct Integr Maint 3(2):126–135
32. Ministry of Land, Infrastructure, Transportation, and Tourism (2014) Manual for bridge
periodic inspection. ministry of land, infrastructure, transportation, and tourism, Tokyo
33. Mirzaei Z, Adey BT, Klatter L, Kong JS (2014) Overview of existing bridge management
systems. The IABMAS Bridge Management Committee
Crack and Corrosion Inspections
for Coastal and Marine Concrete
Infrastructure: A Review
Sabrina Harahap , Surya Dewi Puspitasari ,
and Ahmad Aki Muhaimin
Abstract As one of the biggest archipelago countries, Indonesia has an extensive

coastline where concrete was used as a construction material. Concrete can suffer
cracks that may stimulate the ingress of chloride to the rebar then induced corrosion.
Coastal and Marine Infrastructure (CMI) that are vulnerable to corrosion due to chlo-
ride attack should be considered for continual inspection or monitoring system. The
inspection must use the optimal techniques to achieve the objectives of the inves-
tigation. However, some developing countries, such as Indonesia, have no standard
related to corrosion investigation resulting in the lack of understanding and aware-
ness of concrete deterioration due to corrosion. This paper compares some inspection
techniques consist of visual observation of cracking and corrosion investigation. This
paper provides the utilization condition and the limitation of the inspection technique.
The paper also contributes to knowledge related to cracking and corrosion inspection
in coastal and marine concrete structures by providing a readily available reference
for researchers, practitioners, and society.
Keywords Coastal infrastructure · Concrete inspection · Crack · Corrosion
1 Introduction
Coastal and Marine Infrastructures (CMI) are crucial and strategic assets for the
development, improvement, and defense of any country [1, 2]. Around half of the
world’s population live around the coastline which has increased many economic
S. Harahap (B) · S. D. Puspitasari

Indonesia Defense University, Bogor 16810, Indonesia
S. D. Puspitasari
A. A. Muhaimin
The University of Tokyo, Tokyo 113-8654, Japan
392 S. Harahap et al.
benefits [3]. Indonesia as the largest archipelago country has an extensive coastal
area where concrete is used as a construction material [4]. Concrete is one of the
popular materials for coastal and marine infrastructures as well as masonry and steel
[5, 6]. In many developing countries, like Indonesia, the utilization of old coastal and
marine infrastructures in port and coastal defense has intensely come into an option
due to limited funds. Inevitably, some coastal and marine infrastructures suffered
damage over time, one of the main damages is cracking.
Coastal and marine infrastructures should be a concern because they are suscep-
tible to deterioration and corrosion due to chloride attacks [7]. Corrosion will be
induced even by tiny cracks on the concrete surface. The presence of crack is the
main factor of corrosion initiation. Cracks influence the durability of reinforced
concrete structures in aggressive environments by accelerating the ingress of corro-
sive agents to the embedded steel [8–10]. In this case, the coastal area has a role as
an aggressive environment while chloride ion has a role as corrosive agent.
Deterioration due to corrosion is an issue for concrete structures related to dura-
bility, safety, and sustainability [11, 12]. However, there is no standard nor manual
of corrosion inspection for concrete in Indonesia. This situation leads to the lack of
monitoring and the lack of inspection systems for coastal and marine infrastructure
in Indonesia. This paper attempts to present the importance of corrosion monitoring
in CMI that consists of visual monitoring and non-destructive testing.
2 Crack Inspection
Concrete structures suffer from cracks by several causes such as material, construc-
tion aging, environment, and external force that potentially decrease the durability
of the concrete [13]. Hence, crack monitoring is a crucial maintenance process for
civil infrastructure in general [14, 15] as well as coastal and marine infrastructure.
Manual inspection, as the most common technique of visual observation, is normally
inspected by the inspector by observing and monitoring the concrete surface. Manual
inspection is also conducted for underwater structure. Diver is necessary for annual
inspection to underwater structure such as breakwater and pier. However, divers have
limited time to remain the certain depth of water [16]. This method has several disad-
vantages such as time consuming, costly, and some parts of the concrete structure
might be dangerous to be observed manually [17].
Some researchers have tried to develop alternative crack inspection. Based on the
literature, there are two classifications of crack inspection methods: the image-based
method and the sensor-based method [18]. In image-based method, unmanned aerial
vehicle (UAV) able to take the image of structure surface. UAV has been utilized in
many civil engineering projects for mapping [19, 20] and monitoring [21, 22]. UAV
utilization should be considered to use in the CMI monitoring system. The UAV
was able to take the image of a crack on the concrete surface. Nonetheless, UAV-
based application has a limitation on quantitative assessment of crack width and
crack length [23]. Other type of the image-based method is a head-mounted display
Crack and Corrosion Inspections for Coastal … 393
(HMD). HMD could have 3D data of concrete structures that is more convenient
to be checked in real time and onsite. Moreover, this type could show the cracks in
which difficult to be seen by human eyes [24, 25]
In the sensor-based method, optical fiber sensor or capacitive dense sensor array
installed in the infrastructure aim to detect the real time building condition or struc-
tural health monitoring by analyze the crack detection and propagation [26–28].
However, sensor-based methods are quite expensive, inconvenient, and need physical
contact with structure in the application [18].
3 Corrosion Inspection
Corrosion is the consequence of the crack on concrete surface. Basically, corrosion

inspection divides into electrochemical measurement method and non-destructive
method [29]. Destructive method is avoided and become the last option in corro-
sion inspection. Half-cell potential (HCP) measurement is the simplest way to check
the probability of corrosion on site. This method obtains the corrosion potential
with respect to a standard reference electrode such as copper/copper sulfate elec-
trode (CSE) [29]. This measurement based on ASTM standard C876-15 [30], which
provides general guidelines for HCP evaluating corrosion in concrete structures as
outlined in Table 1. Although, some studies have reported that HCP measurement
may be unreliable [31, 32] because only provides the information and the estimation
for corrosion probability.
Polarization resistance (Rp) is another way to estimate corrosion of reinforcement
concrete. Rp is the ratio between applied voltage and the step of current [33]. The
judgement of polarization resistance based on CEB [34] is shown in Table 2. Rp
measurement able to conduct on cracked concrete, however, location with large
crack (>1 mm) is avoided. Aside from that, wet concrete has to be avoided as well
Table 1 Corrosion
Half-cell potential (mV, CSE) Corrosion estimation
probability based on ASTM
standard C876-15 −200 < E 90% no corrosion
−350 < E < −200 uncertainty
E < −350 90% corrosion
Table 2 The criteria of the

Corrosion current density, Icorr Rate of corrosion estimation
polarization resistance
(µA/cm2 )
Icorr < 0.2 Low
0.2 Icorr < 0.5 Low to medium
0.5 Icorr < 1 Medium to high
1 < Icorr High
Table 3 Grade of passivity (by anodic polarization curve)

Grade Measurement result Existence of passivity film
Grade 0 0.2 < E < 0.6 (V), Icorr > 100 µA/cm2 at least No passivity exist
one time
Grade 1 0.2 < E < 0.6 (V), 10 < Icorr < 100 (µA/cm2 ) Certain degree of passivity exist
Grade 2 0.2 < E < 0.6 (V), Icorr > 10 µA/cm2 at least
once but not to qualified to Grade 1 and 3
Grade 3 0.2 < E < 0.6 (V), 1 < Icorr < 10 (µA/cm2 )
Grade 4 0.2 < E < 0.6 (V), Icorr > 1 µA/cm2 at least
once but not qualified to Grade 1, 2 and 3
Grade 5 0.2 < E < 0.6 (V), Icorr < 1 µA/cm2 Excellent passivity exist
because it may deviate the signal and resulting erroneous reading [35]. Hence, in
CMI inspection, polarization resistance is considered to conduct on dry concrete
surface.
Anodic polarization curve is related to quality of passivity film. The larger current
density, the worse grade of passivity film is. Recently, anodic polarization measure-
ment takes about 40 min [36]. There are two methods to obtain anodic polarization
curve: immersion method and contact method. Immersion method is carried out by
immersing the concrete specimen in a solution and usually conducted in labora-
tory. While contact method is carried out by using double layer counter electrode
contacted on the concrete surface, so that, it is suitable for CMI inspection on site.
The judgement of anodic polarization curve based on the finding of N. Otsuki [37] is
shown in Table 3. Grade 0 represents no passivity film existence, Grade 1–4 represent
certain degree of passivity film existence, and Grade 5 represents excellent passivity
film existence. Table 4 illustrates the area of inspections and limitations for crack
and corrosion inspection methods discussed above, respectively.
4 Proposed Framework of CMI Inspection
Indonesia, the developing country that mainly utilizes concrete as the main mate-
rial of CMI, has develop several standards for inspection, design, implementation,
maintenance and investigation in several public infrastructures such as bridge and
road by Ministry of Public Work and Housing [38] even there are many standards in
the Indonesian National Standard for concrete and steel application [39]. However,
Indonesia has no standard code on corrosion monitoring and inspection. Corrosion
is one of the main reasons for concrete structure failure. An early corrosion detection
can reduce repairmen and rehabilitation costs [40]. Hence, it is necessary to develop
the steps, rules, and standards to minimize corrosion damage. Figure 1 shows the
proposed framework of CMI Inspection that can be conducted if the standard is
Table 4 Crack and corrosion inspection method

Method Area of inspection Limitations Reference
Manual inspection Reachable for inspectors Time consuming, costly, [17]
(road, bridge, etc.) limited inspection
Image-based method Unreachable area (offshore Limitation on [23]
platform, breakwater, pier, quantitative assessment
port, etc.) of crack width and crack
length
Sensor-based method Unreachable area or Expensive, inconvenient, [18]
specified area to investigate directly contact with the
real-time (dam, airport structure
near coastal, tunnel, etc.)
Half-cell potential Reachable for inspectors, Some studies reported [31, 32]
susceptible of corrosion, that HCP measurement
(Pier, Port, etc. especially may be unreliable
around tidal zone)
Polarization resistance Reachable for inspectors, Void (>1 mm) and wet [35]
susceptible of corrosion concrete may deviate the
(Pier, Port, etc. especially signal and resulting
around tidal zone) erroneous reading
Anodic polarization curve Reachable for inspectors, Time consuming: it takes [36]
susceptible of corrosion 40 min to conduct one
(Pier, Port, etc. especially point
around tidal zone)
Crack Inspection
- Manual inspection
- Image-based method
- Sensor-based method
No crack Crack exist
Annual inspection Corrosion Inspection

- Half-cell potential test
- Polarization resistance measurement
- Anodic polarization curve measurement
Low corrosion probability result High corrosion probability result

- Grouting the cracks - Further investigation
- Applying sensor on vulnerable - Implementation of corrosion
area for monitoring protection
Fig. 1 Proposed framework of CMI inspection

unavailable in certain countries like Indonesia. The proposed framework empha-

sizes crack inspection and corrosion inspection. This proposed framework is not an
official framework but it can illustrate the inspection plot in CMI.
At the first step, crack inspection can use or combine several methods such as
manual inspection, image-based method and sensor-based method by UAV. If cracks
are not found on the concrete surface, the inspection will be continued as an annual
inspection. If the cracks exist on the concrete surface, it is highly recommended
to following to corrosion inspection. The second step, corrosion inspection can be
conducted by electrochemical measurements such as HCP, polarization resistance
measurement, and anodic polarization curve measurement. These measurements
estimate the corrosion probability so that it is strongly suggested to do at least two
kinds of measurements to mutually validate the estimation results. Using one kind
of measurement in corrosion inspection is not recommended.
Low corrosion probability result leads to various decisions such as grouting the
cracks and applying sensor on vulnerable areas. These actions are taken to prevent
corrosion initiation and to observe the vulnerable areas, respectively. The addition of
corrosion inhibitor substances for example calcium nitrate in grouting paste can be
an option to delay corrosion initiation in concrete inexpensively [41]. High corrosion
probability result leads to conducting further investigation like destructive testing to
confirm corrosion stage. Further investigation can be conducted by taking core drill
samples of CMI structure then testing by advanced measurement in the laboratory.
Furthermore, the implementation of corrosion protection like cathodic protection
(CP) is needed to repair the concrete structure due to corrosion [42]. In general, CP
installation has a high initial cost but it is considered as an economical long-term
cost [41]. The probability corrosion results can be the fundamental data to decide
the prevention or protection of CMI due to corrosion.
5 Conclusion
Costal and marine infrastructures, as the first line in the defense of the country, are
valuable assets that need proper monitoring and inspection method. For that purpose,
this paper reviewed some literature related to crack and corrosion inspection that
enable to conduct in coastline area, but found that each method has limitations. By
the review above, following conclusion could be derived.
• Each inspection has limitations that need to be developed to obtain the proper way
of inspection. The inspectors are necessary to select the best inspection method
based on location, cost, and condition.
• Even though crack and corrosion inspection standards are not available in certain
countries like Indonesia, these inspections are highly considered to be conducted
in CMI for monitoring the corrosion stage. An early crack and corrosion detections
can reduce repair and rehabilitation costs.
• The corrosion inspections such as half-cell potential test, polarization resis-

tance test, and anodic polarization resistance measurement, are estimation and
probability only. It is highly recommended to conduct minimum two kinds of
measurements to mutually validate the estimation results.
• Corrosion inspection results are precious data that help to decide the proper
treatment in concrete structures, especially around coastline. These results can
be a deliberation to implementation of corrosion prevention (low corrosion
probability) or corrosion protection (high corrosion probability).
Future researches are necessary to develop the knowledge about proper way to
inspect crack and corrosion on concrete structure in coastline area. This paper can be
used as a tool to guide engineers and asset managers in decision-making about crack
and corrosion inspection for concrete especially in coastal and marine infrastructure.
References
1. Bastidas-Arteaga E, Schoefs F (2015) Sustainable maintenance and repair of RC coastal struc-

tures. In: Proceedings of the institution of civil engineers-maritime engineering, vol 168, No
4, pp 162–173. Thomas Telford Ltd
2. Lucio D, Tomás A, Lara JL, Camus P, Losada IJ (2020) Stochastic modeling of long-term wave
climate based on weather patterns for coastal structures applications. Coastal Eng 161
3. Creel L (2003) Ripple effects: population and coastal regions. Population Reference Bureau,
Washington, DC
4. Wahyu S, Pramono GH, Purnawan B (2009) Establishment of marine and coastal spatial data
infrastructure in Indonesia. In: Coastal and marine geospatial technologies, pp 97–103
5. Boéro J, Schoefs F, Capra B, Rouxel N (2019) Technical management of French Harbour
structures, Part 1: description of built assets. Editions Paralia CFL, vol 6, pp 1–11. Revue
Paralia, France, pp 1–7
6. McManus RS, Archibald N, Comber S, Knights AM, Thompson RC, Firth LB (2018) Partial
replacement of cement for waste aggregates in concrete coastal and marine infrastructure: a
foundation for ecological enhancement. Ecol Eng 120:655–667
7. Cai R, Yu M, Hu Y, Yang L, Ma H (2021) Influence of data acquisition and processing on
surface chloride concentration of marine concrete. Construct Build Mater 273
8. Otieno MB, Alexander MG, Beushausen H-D (2010) Corrosion in cracked and uncracked
concrete—influence of crack width, concrete quality and crack reopening. Mag Concr Res
62(6):393–404
9. Cao C, Cheung MM, Chan BY (2013) Modelling of interaction between corrosion-induced
concrete cover crack and steel corrosion rate. Corros Sci 69:97–109
10. Shaikh FUA (2018) Effect of cracking on corrosion of steel in concrete. Int J Concr Struct
Mater 12(1):1–12
11. James A et al (2019) Rebar corrosion detection, protection, and rehabilitation of reinforced
concrete structures in coastal environments: a review. Constr Build Mater 224:1026–1039
12. Chaves IA, Melchers RE, Peng L, Stewart MG (2016) Probabilistic remaining life estimation
for deteriorating steel marine infrastructure under global warming and nutrient pollution. Ocean
Eng 126:129–137
13. Jung S, Lee S, Yu J (2021) Ontological approach for automatic inference of concrete crack
cause. Appl Sci 11(1):252
14. Das Khan S, Topdar P, Datta AK (2020) Applicability of fuzzy-based visual inspection approach
for condition assessment of bridges in developing countries: a state-of-the-art review. J Inst
Eng India Ser A 101:835–846. https://doi.org/10.1007/s40030-020-00465-1
15. Fujita Y, Hamamoto Y (2011) A robust automatic crack detection method from noisy concrete
surfaces. Mach Vis Appl 22:245–254. https://doi.org/10.1007/s00138-009-0244-5
16. Chen HH, Chuang WN, Wang CC (2015) Vision-based line detection for underwater inspection
of breakwater construction using an ROV. Ocean Eng 109:20–33
17. Kim H, Sim SH, Cho S (2015) Unmanned aerial vehicle (UAV)-powered concrete crack detec-
tion based on digital image processing. In: 6th international conference on advances in exper-
imental structural engineering, 11th international workshop on advanced smart materials and
smart structures technology
18. Zhao X (2020) Automated rotational-and scaling-invariant image-based crack width moni-
toring with sub-millimeter accuracy and self-numbering label. Asian J Civil Eng 21(4):741–
749
19. Remondino F, Barazzetti L, Nex F, Scaioni M, Sarazzi D (2011) UAV photogrammetry for
mapping and 3d modeling–current status and future perspectives. Int Arch Photogramm Rem
Sens Spat Inf Sci 38(1):25–31
20. Nex F, Remondino F (2014) UAV for 3D mapping applications: a review. Appl Geomat 6:1–15.
https://doi.org/10.1007/s12518-013-0120-x
21. Ham Y, Han KK, Lin JJ et al (2016) Visual monitoring of civil infrastructure systems via
camera-equipped Unmanned Aerial Vehicles (UAVs): a review of related works. Vis Eng 4:1.
https://doi.org/10.1186/s40327-015-0029-z
22. Zhang C, Elaksher A (2012) An unmanned aerial vehicle-based imaging system for 3D
measurement of unpaved road surface distresses 1. Comput Aided Civil Infrastruct Eng
27(2):118–129
23. Kim H, Lee J, Ahn E, Cho S, Shin M, Sim SH (2017) Concrete crack identification using a
UAV incorporating hybrid image processing. Sensors 17(9):2052
24. Yamaguchi T, Shibuya T, Kanda M, Yasojima A (2019) Crack inspection support system
for concrete structures using head mounted display in mixed reality space. In: 58th annual
conference of the society of instrument and control engineers of Japan (SICE)
25. Yamaguchi T, Park T, Shibuya T, Yasojima A (2019) Development of crack inspection support
system on concrete wall using MRHMD. In: The Proceedings of JSME annual conference on
robotics and mechatronics (Robomec)
26. Antunes P, Rodrigues H, Travanca R, Ferreira L, Varum H, André P (2012) Structural health
monitoring of different geometry structures with optical fiber sensors. Photon Sens 2(4):357–
365
27. Yan J, Downey A, Cancelli A, Laflamme S, Chen A, Li J, Ubertini F (2019) Concrete crack
detection and monitoring using a capacitive dense sensor array. Sensors 19(8):1843
28. Park S, Ahmad S, Yun CB, Roh Y (2006) Multiple crack detection of concrete structures using
impedance-based structural health monitoring techniques. Exp Mech 46:609–618
29. Song HW, Saraswathy V (2007) Corrosion monitoring of reinforced concrete structures—A.
Int J Electrochem Sci 2(1):1–28
30. ASTM C 876–15 (2015) Standard test method for half-cell potential of uncoated reinforcing
steel in concrete
31. Leelalerkiet V, Kyung JW, Ohtsu M, Yokota M (2004) Analysis of half-cell potential
measurement for corrosion of reinforced concrete. Constr Build Mater 18(3):155–162
32. Frølund T, Klinghoffer O, Sørensen HE, Denmark DD (2003) Pro’s and con’s of half-cell
potentials and corrosion rate measurements. In: International Conference on Structural Faults
33. Stern M, Geary AL (1957) Electrochemical polarization: I. A theoretical analysis of the shape
of polarization curves. J Electrochem Soc 104(1):56
34. CEB, Working Party V/4.1 (1997) Strategies for testing and assessment of concrete structures
affected by reinforcement corrosion (Draft. 4), BRI-CSTC-WTBC, CEB Bulletin No. 243
35. Andrade C, Martínez I (2010) Techniques for measuring the corrosion rate (polarization
resistance) and the corrosion potential of reinforced concrete structures. In: Non-destructive
evaluation of reinforced concrete structures, pp. 284–316. Woodhead Publishing
36. Hamada H, Sagawa Y, Ikeda T, Morikawa R (2010) Several factors affecting the anodic polar-
ization curve of steel bars embedded in mortar. In: 6th international conference on concrete
under severe conditions-environment and loading, CONSEC’10, pp 201–208
37. Otsuki N (1985) A study of effectiveness of chloride on corrosion of steel bar in concrete.
Report of Port and Harbor Research Institute, pp 127–134
38. Ministry of Public Work and Housing (PUPR). https://binamarga.pu.go.id/index.php/konten/
standar-pedoman (in Bahasa). Last accessed 2021/04/25
39. National Standardization Agency of Indonesia (BSN). http://sispk.bsn.go.id/SNI/DaftarList (in
Bahasa). Last accessed 2021/04/25
40. Sadowski L (2013) Non-destructive investigation of corrosion current density in steel reinforced
concrete by artificial neural networks. Arch Civil Mech Eng 13:104–111
41. Goyal A, Pouya HS, Ganjian E et al (2018) A Review of corrosion and protection of steel in
concrete. Arab J Sci Eng 43:5035–5055. https://doi.org/10.1007/s13369-018-3303-2
42. Byrne A, Holmes N, Norton B (2016) State-of-the-art review of cathodic protection for rein-
forced concrete structures. Mag Concr Res 68(13):664–677. https://doi.org/10.1680/jmacr.15.
00083
e-Peralatan System as an Equipment
Management for Disaster Mitigation
on Indonesia National Roads
Adityo Budi Utomo , P. Gitaning, and S. Tosan Kunto
Abstract The e-Peralatan is an information system developed to optimize the cycle

time of heavy equipment owned by National Road Construction Agencies of Central
Java and Special Region of Yogyakarta. Heavy equipment is used to support the
process of construction, maintenance, and disaster mitigation that impact the national
road in the work area. The type of heavy equipment contained in the e-Peralatan is
known as DRU (Disaster Reacting Unit) that consists of excavators, motor graders,
wheel loaders, vibrator rollers, dump trucks, and dozers. This paper aims to explore
how to create the e-Peralatan system as equipment management for disaster mitiga-
tion. Preliminary data were gathered by inventorying equipment, identifying existing
business processes, and collecting documents related to equipment utilization. Then,
the preparation of utilization framework involving National Road Construction Agen-
cies of Central Java and The Special Region of Yogyakarta, local government, and
publics. We found that there is no digital equipment management system for disaster
mitigation. At the end of the dissemination, all parties have the hope of digital
equipment management that can be utilized together. Our findings indicate that
e-Peralatan provides many changes in the business process compared to conven-
tional methods. The changes are including less standby time, faster equipment rental
ordering time, cost certainty, time certainty, document administration certainty, and
process certainty.
Keywords Disaster · Heavy equipment · Digital system · e-Peralatan
A. B. Utomo (B) · P. Gitaning

Public Works Polytechnic, Semarang, Indonesia
P. Gitaning
S. Tosan Kunto
National Road Construction Agency of Central Java and The Special Region of Yogyakarta,
Semarang, Indonesia
402 A. B. Utomo et al.
1 Introduction
DRU is a set of heavy equipment used for natural disaster emergency response
activities that affect Indonesia’s national roads and bridges. DRU consists of one set
of work units consist of an excavator, motor grader, wheel loader, vibrator roller,
bulldozer, and dump truck. DRU is alerted to handle natural disaster conditions such
as landslides, floods, volcanic eruptions, and tsunamis. Floods and landslides are the
dominant or common disaster occur on national roads and bridges.
Requirements that must be met in the management of the DRU include:
a. The equipment must always be in good condition or lightly damaged
b. The equipment must always be available whenever needed by any party
c. The equipment can reach the disaster location in less than 6 h so that the heavy
equipment workshop must be placed in a strategic location prone to disaster.
To improve its productivity, it is essential to improve the performance of the
construction systems. The desired production output is achieved through high equip-
ment availability, which is influenced by equipment reliability and maintainability
[1]. Idle time, downtime, poor equipment maintenance practices, improper determi-
nation of economic life and timing of replacement, poor training of equipment opera-
tors, equipment breakdown, over maintenance of equipment, huge capital investment
during acquisition, the balance of interdependent equipment, misunderstanding the
scope of work carried out, and incompatibility between the unit cost of production
and equipment suitability for job condition were found to be the major problems that
affect construction equipment planning and management [2].
The intended users are the most critical factor in the success of information systems
and technologies; this also applies to disaster management. Moreover, as users’
interactions with an information system play a significant role in determining whether
or not it is useful, this suggests that understanding users’ attitudes toward information
systems is critical to their development [3]. The system designed includes being able
to help command post commanders in making decisions based on the data displayed
on the dashboard contained in the system [4].
National Road Construction Agencies of Central Java and The Special Region
of Yogyakarta require equipment to support the construction and maintenance
of national roads and bridges in Central Java and Special Region of Yogyakarta
Provinces. Because the intensity of disaster occurrence in the national roads in
Central Java and Special Region of Yogyakarta Provinces per year is not too much,
so if The National Road Construction Agency of Central Java and Special Region of
Yogyakarta invest funds in the form of heavy equipment, there will be a lot of idle
time. Other options for the utilization of the DRU are rent to the public project on
terms that they are always in good condition or slightly damaged also if a disaster
occurs, the DRU can be mobilized at any time.
Problems arising from the management of DRU by National Road Construction
Agencies of Central Java and The Special Region of Yogyakarta are:
a. DRU equipment rent services
e-Peralatan System as an Equipment Management … 403
1) Such as utilization of equipment by third parties through lending schemes

on which between administration and realization are out of sync, no
certainty of the rental fare, no certainty of the time of lending, the rental
application process takes a long time, also disorderly administration.
2) Utilization of equipment by the internal unit of National Road Construction
Agencies of Central Java and The Special Region of Yogyakarta is not
accompanied by good administrative documents
3) The procedure for borrowing equipment has not been well-publicized.
b. Monitoring of DRU equipment assets
1) Many equipment assets in the Agencies are not monitored for their
utilization
2) There is no periodical time of reporting system related to DRU equipment.
It is necessary to increase cooperation with local governments, disaster-affected
communities for the continued dissemination of disaster mitigation information and
to add disaster information and knowledge content from community experiences [5].
An information system is required in the management of DRU that accommodates
all parties. Characteristics of information system which will be developed are:
a. The rent system, The large number of DRU in the National Road Construction
Agencies of Central Java and The Special Region of Yogyakarta has a large
country income potential but a system must be built with characteristic are:
1) Effective and Efficient. Certain rental fare, easy to monitor the process
2) Faster. There is certainty about the time of rental, the process for submitting
a rent application must be faster
3) Better. A simple process that does not violate applicable Indonesian rules
b. Monitoring System, the system can be accessed easily by interested parties
anywhere, anytime, and in real-time.
2 Methodology
The development of e-Peralatan consists of several stages as shown in Fig. 1. The first
step is to conduct a regulatory study related to equipment operations and equipment
utilization documents. From that documents, we made a framework of e-Peralatan
involving National Road Construction Agencies of Central Java and The Special
Region of Yogyakarta, local government, and IT expert in dissemination. Based on
the e-Peralatan’s framework can be determined about the data dan developing system
that will be made. The method used in designing this management information system
is the waterfall method, which consists of four stages, including the stages of analysis,
design, coding, and testing. The stages of designing this information system include
system requirements analysis consisting of two processes [6]:
Start
Literature review Field observation Architecture software design
Trial and test

E-peralatan framework
Final report
Database creation
finish
User Interface Creation
Fig. 1 Research methodology
a. Object-Oriented Analysis (OOA). Information system design is carried out in the

Object-Oriented Analysis (OOA) stage, which is designing business processes
from existing systems, then determining the parties involved in the system
(users), designing use case diagrams¸ and designing use case descriptions.
b. Object-Oriented Design (OOD). OOD is designed by making a diagram that
shows the design of the information system created. Diagrams used at this
stage include activity diagrams, sequence diagrams, class diagrams, and Entity
Relationship Diagrams (ERD). The system design described using the Unified
Modelling Language.
Furthermore, the coding is done and followed by testing by the verification and
validation process.
3 Results and Discussions
3.1 User Interface
The main page (home) is the user interface that can be accessed at www.palan.map
geo.id (Fig. 2). This page displays simple information sections that can be accessed
by the public, which contains header section, menu section, and banner section as
well as search engine.
The new procedure for e-Peralatan by external users is designed more modern,
simple, good lock, and easy to understand by anyone, which also can synchronizing
with their email account as notifications. Users submits the letters to the equipment
manager accompanied by a description of the type of equipment, location, period
time, and purpose of use [7].
Fig. 2 Main page (home). Source www.palan
New procedure for e-Peralatan by external users can show in Fig. 3:

1. Open the website address www.palan.mapgeo.id [8] (Fig. 2)
2. Type the name of the equipment you want to rent for example “excavator” or
“loader”
3. Enter the period time of the rent
4. Click the “search” icon
Input registration number on reserva-

Registration number tion menu and upload the payment
proof
Email Notification
Fill in order Email contains Make payment and Equipment rental

form registration num- complete the terms agreement
ber
Readiness equip-
ment checking
Equipment
rental agreement
Data veri- Inputing data Processing Document Change equipment

fication based on queque to rental verification statue from availa- Delivery
process ble to unavailable order
Email Notification Rental Approval
Fig. 3 Business process of new E-peralatan. Source www.palan.mapgeo.id

Fig. 4 Availability of DRU. Source www.palan.mapgeo.id
5. The website will guide you to the equipment availability page and location.
Click on the details, to know the equipment identity (Fig. 4)
6. Click “booking” to proceed. Fill in data identity along with the residence card
or passport and application letter. Application letter format can be found on
the download menu.
7. When the file is complete the notification will be sent via email along with the
attachments to the required administrative files
8. Make payment after email notification has been received and upload the proof
of payment in the website
9. Complete the equipment retrieval file and the equipment is ready to operate
10. Perform equipment returns within the agreement period. If the time limit is
exceeded, an additional fee will be charged according to the provisions
3.2 Disaster Mitigation Management System
The thing that often happens when disasters happened is that access roads to sites
will be blocked due to landslides or floods. DRU management is required to open
access for logistic road assistance and evacuation of victims. DRU support in disaster
mitigation management that has been done by National Road Construction Agencies
of Central Java and The Special Region of Yogyakarta are:
1. Conduct a study of potential risk disasters (floods, landslides, tsunamis, and
eruptions) that impact national roads so that a map of potential risk disaster is
obtained in the red zone, yellow zone, or green zone.
2. Relocations of all DRU workshops in a strategic red zone location so that it is
easy and fastly to mobilize within a maximum of 6 h (Fig. 5)
3. Socialization about e-Peralatan’s support system to Sub Project Manager, task
force, Local Disaster Board, local government, and the community about
disaster mitigation management.
DRU SERVICE ZONING MAP
Karangjati Workshop
Sleman Workshop
Buntu Workshop
Pekalongan Workshop
Fig. 5 DRU workshop location based on disaster risk management. Source National Road
Construction Agencies of Central Java and Special Region of Yogyakarta
4. Delivery DRU to the disaster site in a clear, measurable, and effective way.
An inappropriate choice of DRU will result in low productivity and will incur
additional cost and time to open the access road [9].
5. Provides optimal DRU operation services until the disaster emergency response
period is finished.
The development of the e-Peralatan gives many advantages compared to the
conventional system. The comparative performance between e-Peralatan and the
conventional system can be seen in Table 1.
The thing that is complained about this process is that there is no certainty how
long it takes to get the permit and how much the total cost must be paid. Sometimes
the equipment manager is so busy that the approval of the permit will take a long time.
E-Peralatan is designed with a digital letter, digital document, and digital signature
so that it can simplify the permit approval. we suggest that e-Peralatan management
needs human resources for optimal services. Further studies are needed regarding
the most suitable public service model for DRU operation in disaster mitigation
management.
Table 1 Comparative performance between e-Peralatan and the conventional system

No. Parameter Conventional system E-Peralatan Description
1 Permit cost There is a fee to There is no fee for With e-Peralatan
create, duplicate, and document creation, all registration
send documents copying, and shipping and document
submissions are
done digitally
2 Permit time There is no certainty Permit time for the All processes and
of completion time whole process will document can be
from 2 to 14 days take 2 h monitored in
real-time
3 Equipment rental There is no certainty Rates are stated in the All price are listed
fee of equipment fee application in the application
4 Mobilization for It is difficult to It is easy to monitor The equipment’s
disaster mitigation monitor existing existing conditions location and
equipment conditions and location conditions data
and location recorded in the
application in
realtime
Coordination takes Mobilization Relocation in the
up to 1 day for the equipment to disaster workshop makes
equipment to arrive area less than 6 h every disaster area
at the disaster site more accessible
5 Equipment Maintenance data is Maintenance is carried Equipment
maintenance done manually so that out periodically and maintenance data
spare parts planning digitally so that spare recorded in the
is less effective parts planning is more application
effective
6 Equipment usage Manually and untidy Digitally and tidy All documents are
documentation documentation documentation done digitally in
the application
4 Conclusion
Based on the analysis, it was found that by using e-Peralatan, DRU operations
have become more optimal. DRU standby time is utilized through equipment rental
schemes for disaster mitigations or other needs. The combination of e-Peralatan and
setting up the workshop location makes every disaster site more accessible so that
it can be reachable in less than 6 h. Comparing to conventional systems, equipment
data owned by National Road Construction Agencies of Central Java and Special
Region of Yogyakarta can be better organized with e-Peralatan system information
and the public can access it anywhere and anytime. E- Peralatan can make it easier
to find out information about equipment and the condition of existing equipment.
DRU management includes purchasing (investment), procurement of spare parts,
fuel-consuming, operating hours, maintenance, and removal when it is severely
damaged in a DRU life cycle time. The life cycle time must be managed disaster
mitigation that has an impact on the infrastructure. In this paper, an attempt is made
to highlight the role of e-Peralatan in the DRU life cycle time management digitally
to reduce cost and time for contingency.
References
1. Desai P, Darade PM, Khare PP, Engineering C, Pune DYPS (2017) Overall equipment
effectiveness in construction equipments. Int Res J Eng Technol (IRJET) 4(1):1111–1115
2. Manikandan M, Adhiyaman M, Pazhani KC (2018) A Study and analysis of construction equip-
ment management used in construction projects for improving. Int Res J Eng Technol (IRJET)
5(3):1297–1303
3. Sari KP, Kanegae H, Roychani M (2021) Local risk managers ’ and disaster volunteers ’ aware-
ness of , attitudes toward , and intention to use a local government-developed disaster information
system: a case study of SIKK Magelang. In: Universitas Andalas (ed.) 2nd International confer-
ence on disaster and management. IOP conference series: earth and environmental science, vol
708 (012066) (2021). https://doi.org/10.1088/1755-1315/708/1/012056
4. Patrisina R, Meilani D, Habib M (2021) An information management system for effective
disaster relief operations. In: Andalas U (ed.) 2nd International Conference on disaster and
management. IOP conference series: earth and environmental science, vol 708(012066), pp
1–11 (2021). https://doi.org/10.1088/1755-1315/708/1/012066
5. Siswanto T, Shofiati R, Hartini H (2018) Acceptance and utilization of technology (UTAUT )
as a method of technology acceptance model of mitigation disaster website acceptance. In: The
4th international seminar on sustainable urban development, pp 1–6
6. Rajagopal D, Thilakavalli K (2017) A study UML for OOA and OOD. Int J Knowl Content Dev
Technol 7(2):5–20
7. Kirmanto D (2014) Peraturan Menteri Pekerjaan Umum Republik Indonesia Nomor
09/PRT/M/2014 tentang Jenis dan Tatacara Penggunaan Peralatan Konstruksi di Kementerian
Pekerjaan Umum (Issue 10 September 2014), pp 1–27
8. www.palan.mapgeo.id
9. Nugraha AS, Putranto L (2019) The effect of heavy equipment management on the performance
of the construction project and the construction company. In: 1st International conference of
construction, infrastructure, and materials, pp 7–14 (2019). https://doi.org/10.1088/1757-899X/
650/1/012019
The Implementation of Functional Road
Assessment on Pramuka Road Section
in Klaten District
Amirotul Musthofiah H. Mahmudah, Syahrul Anggara Wuryatmaja,

and Ary Setyawan
Abstract Roads as one of the important facilities in transportation have a role as a

distributor of goods and services, as an economic driver, so it is necessary to improve
the quality of the roads so as to improve driving safety. Based on the Minister of
Public Works Regulation No. 11/PRT/M/2010 each regent/major must conduct an
evaluation of functional road assessment on district/city roads. The district road that
was reviewed for the implementation of the research was Pramuka Street with the
function of being secondary local. The purpose of this study is to determine the
functional assessment eligibility category of roads and to provide recommendations
so that Pramuka Street can achieve eligible function. The difference with previous
studies is a review of the location, types of roads based on their function, and an
assessment of the administrative component. The data required is in the form of
technical requirements and administrative requirements. Component assessment is
carried out by means of direct measurement and assessment in the field and requires
secondary data obtained from an agency. The analysis is done by calculating the devi-
ation results for each component inspection. The results of this study concluded that
the Pramuka Street was categorized as conditional function. The recommendations
given still need improvements in geometric road techniques, pavement structure tech-
niques, complementary building techniques, traffic engineering management tech-
niques, and road equipment techniques. Road administration documents for those
that do not yet exist are also immediately completed and the document procurement
process also needs to be accelerated immediately. Recommendations are given so
that roads can achieve an eligible function condition.
Keywords Functional road assessment · Secondary local · Functional assessment

eligibility category · Recommendations
A. M. H. Mahmudah (B) · S. A. Wuryatmaja · A. Setyawan (B)

Department of Civil Engineering, Sebelas Maret University, Jl. Ir. Sutami 36 A, Surakarta,
Indonesia
412 A. M. H. Mahmudah et al.
1 Preliminary
The road is an important facility in transportation that plays a big role in people’s lives.
The road is also a connector between regions and a facility for distributing goods and
services. Roads can also play a role in increasing the economic growth of a region.
The number of residents in each region has increased every year, thus encouraging
an increase in the number of private vehicles [1]. Based on these conditions, if the
number of private vehicles and public vehicles has exceeded the road capacity, it
will cause congestion. Based on data from the Central Statistics Agency of Klaten
Regency, each year the population increases. In 2017 there was a population increase
of 0.36%. The increasing number of vehicles also encourages an increase in traffic
accidents. The cause is from humans themselves who are less focused when driving,
road infrastructure namely roads, facilities or vehicles, and the environment [2]. We
often encounter problems with road quality around us, causing a lack of safety and
comfort in driving.
Based on data from the police, the number of traffic accidents in Klaten Regency
in 2018–2020 was 2799 accident cases. Traffic accidents in the district during that
year have claimed as many as 4957 victims, more specifically 389 victims died, 14
were seriously injured, and 4554 were slightly injured. These data indicate that the
accident rate is still high in Klaten Regency, so there is a need for efforts to carry out
road evaluations. From these various problems, it is necessary to hold a study aimed
at realizing a safety road. One way to do this is the road function assessment. Road
function assessment on national roads, provincial roads, and district roads must be
carried out by the respective regional governments [3].
A study regarding the Road Function Assessment with the same objective was
carried out by [4, 5], namely analyzing the level of functional assessment eligibility
category. The research method refers to regulation of the Minister of Public Works
Number: 11/PRT/M /2010. The results of the two studies show that the road that is
evaluated is categorized as conditional function (CF).
Research on road infrastructure safety evaluation on Trans Kalimantan roads. The
background of this research is suspected to have deficiencies in technical require-
ments. Data collection was carried out for two days in November 2016, calculating
geometric parameters and observing road surface conditions and harmonizing signs,
markings and other road attachments by referring to applicable regulations. The iden-
tification results in the conclusion that there has been a deficiency in the applicable
geometric planning regulations and technical requirements [6].
Research to analyze road safety at bends based on the radius and transverse slope
of the bend. The research was conducted on the Mataram-Senggigi-Pemenang road
section. Sharp bends are very dangerous for riders because they can reduce the
rider’s visibility and centrifugal force. The magnitude of the radius is obtained from
the image of the road situation as a result of the topography and the cross-slope of
the road is obtained from the difference in road side elevation. The results of the
study showed that the bend on the Mataram-Senggigi-Pemenang road, there were
71.43% of the radius of the bend which was included in the very dangerous category
The Implementation of Functional Road Assessment … 413
and only the Amarsvati bend (STA 21 + 500) was categorized as quite dangerous
and the others were safe. The solution to this problem is the installation of bend
mirrors, speed limiting signs, sharp bend signs, lighting lamps, guardrail, enlarge
bend section [7].
In the above review, the types of roads selected for the inspection were national
and provincial roads. From previous study that has not been carried out on district
roads, it is necessary to conduct research on district roads. The difference in this
study from previous studies is that the location review of this study is based on the
function of roads, namely local and roads based on their network system, namely
secondary roads. The research does not only examine the technical requirements
but also the completeness of the administrative requirements owned by the roads.
The section of road being examined was Pramuka Street in Klaten Regency because
accidents were still found on the road that caused the victim to be injured and mate-
rially damaged. This research is primarily aims to determine functional assessment
eligibility category and to determine recommendations for each assessment of road
components so that the eligibility functional conditions on the roads can be realized.
2 Theoretical Basis
2.1 Road Function Assessment
Road function assessment according to the Regulation of the Minister of Public

Works Number: 11/PRT/M/2010 is a condition of a road that meets the technical
requirements of eligibility to provide safety for its users and administrative require-
ments that provide legal certainty for road operators and road users so that it can be
operated for the public.
2.2 Technical Requirements for Road Function Assessment
The technical components of the road function assessment consist of road geometric
techniques, pavement structure techniques, complementary building structure tech-
niques, road section utilization techniques, traffic engineering management tech-
niques, and road equipment engineering. It is important to pay attention to road
geometric planning because it is very important to get uniformity in planning the
geometrical roads and to produce smoothness and comfort for road users [8]. The
assessment components that are included in the geometric road technicalities are
horizontal alignment, vertical alignment, cross section of the road body, horizontal
alignment coordination and vertical alignment. On road pavement structures, the
results of the pavement structure testing can determine the function of the structure
and the strength of the road construction which includes the suitability of the pave-
ment structure with the class of road function, road surface conditions, and surface
drainage. On the road complementary buildings the results of this study identified
the structural functioning of each complementary building component and adjusted
it according to the functional assessment eligibility category. Technical complemen-
tary buildings consist of bridges, pontoon, culverts, parking lots, retaining walls, and
waterways. Assessment of the utilization of part of the road, which includes testing
of the road benefit space, the road property space, and the road supervision. The
components included in the Traffic Management and Engineering assessment are
road equipment that supports traffic regulation which includes road marking, signs,
separators, roundabout, sidewalks, traffic light and road crossings.
The road equipment technical assessment is almost the same as the traffic manage-
ment and traffic engineering technical assessment of its components, but in this test
it is only carried out on the road equipment specifications. Focus of the assessment
is on the conditions and dimensions of the road equipment on a road segment being
tested. The road equipment technical assessment component is road equipment that
is directly related to road users and technical equipment that is not directly related
to road users.
2.3 Administrative Requirements for Road Function

Assessment
The administrative requirements for the road function acceptance test consist of docu-
ments determining instructions for orders and prohibitions, road status documents,
road class documents, land ownership documents, road information documents, and
road environmental documents. Each document must be checked for completeness
so that it meets the road functional assessment eligibility category.
2.4 Functional Assessment Eligibility Category
The category of road function eligibility is the result of the final assessment of all
components of the road segment and entire road segment tested both technically and
administratively. Based on the Regulation of the Minister of Public Works Number. 11
PRT/M/2010, the determination of the status of road functional assessment eligibility
is divided into three categories, namely Eligible Function (EF), Conditional Function
(CF), and Ineligible Function (IF). If the result of an assessment of components or the
entire road section is CF or IF, a recommendation must be given in order to achieve
a eligible function condition (EF).
Fig. 1 Research location
3 Research Methods
3.1 Research Location
The location of the road function assessment research was carried out on Pramuka
Street, Klaten Regency with a road length of 750 m. The results of the field identifi-
cation of Pramuka Street are local roads, district roads, secondary roads, and small
roads. The research location is located in the district center where there are impor-
tant places such as police, schools, hospitals and stations. For a plan drawing of the
research location, it can be seen in Fig. 1:
3.2 Data Collection Technique
The stages of data collection consisted of collecting primary data and collecting
secondary data. The following will explain the primary data collection process:
Primary Data
The primary data inspection component consists of road geometric techniques,
consisting of rider’s visibility, pathways, median, and side gutter. Pavement struc-
ture techniques, namely pavement conditions using pavement condition index (PCI)
method, road complementary building techniques, traffic engineering management
techniques, and road equipment technical. Primary data is data that is obtained
directly in the field by means of a survey according to approved procedures. The
inspection was carried out in a comprehensive manner on Pramuka Street without
dividing the segments. In the road function assessment when carrying out a survey,
things that need to be prepared in advance are a survey form that is used to record the
results of field research and tools such as roll meters, meters, pylox, cameras, wood,
objects as high as 15 cm, stationery, and the road function assessment form.
Secondary Data
The secondary data inspection component consists of the road geometric tech-
nical inspection through road information documents, complementary building data,
dimensions of the technical utilization of road sections through the road information
document, LHRT data and all of the road administration requirements. Secondary
data is data obtained from the results of inspections/surveys of certain people and
agencies.
3.3 Data Analysis
Data analysis begins with the collection of primary data and secondary data. Devi-
ation of each inspection result will be calculated and compared with the technical
standard. To calculate each component deviation using the formula:
T−M
P= × 100% (1)
T
where P = Percentage Deviation, T = Technical Standard, M = Measurement Result.
Determination of the category/status of assessment component function eligibility
criteria based on the deviation results. Determination of the functional eligibility
category for the assessment component can be seen in Table 1.
Specifically for granting categories/status of road function eligibility to comple-
mentary building techniques using the analysis of the condition values in the Minister
of Public Works Regulation Number: 11 of 2010 with a scoring number 1–5. The
eligibility category for each weighted number will be explained in Table 2:
Assessment on the road administration documents, if the required documents
are complete or exist, the document assessment is declared eligible functional, but
if one of the documents does not exist, it will be declared conditional function,
Table 1 Determination of
Functional category Deviation Information
the functional eligibility
category of the audit Eligible function (EF) 0% Can provide safety for
component riders
Conditional function >0% Can still provide safety
(CF) for riders
Ineligible function (IF) >0% Cannot provide safety
for the rider
Source [3]
Table 2 Determination of
Condition value Functional category
eligibility category for
complementary building 0–1 Eligible function (EF)
functions 2–4 Conditional function (CF)
5 Ineligible functional (IF)
Source [9]
Table 3 Matrix of how to determine functional assessment eligibility category/criteria technical,

administrative, and a road section
Possibility EF CF IF Status/Category eligible functions for technical, administrative
aspects, and a road
1 ✓ ✓ ✓ IF
2 – ✓ ✓ IF
3 ✓ – ✓ IF
4 – – ✓ IF
5 ✓ ✓ – CF
6 – ✓ – CF
7 ✓ – – EF
Source [10]
but if the road status document is not exist, the road segment is declared ineli-
gible functional. Providing recommendations only on components that are catego-
rized as conditional functional (CF) and ineligible functional (IF). Determination of
functional status/category is given for each technical component and administrative
component as well as for all roads that are carried out by road functional assessment.
Table 3 will explain the matrix for determining the functional assessment eligibility
category for technical, administrative, and road segments:
After obtaining the functional eligibility category for each technical and adminis-
trative component, the percentage of each component functional eligibility category
value can be determined using the formula:
TE
P= × 100 (2)
TO
where P = Percentage of Each Category, TE = Total of Each Category, TO = Total
Assessment of Overall Components.
4 Research Result
After analyzing the data on all assessment components, the following will be shown
regarding the results of the categories/status of each assessment component:
4.1 Technical Geometric Roads
The geometric technical inspection of the road at the cross section of the road consists
of road lanes, road shoulders, medians, safety thresholds, and traffic safety devices.
Road shoulders and traffic safety devices are not inspected. The geometric technical
inspection of roads on horizontal alignments consists of straight sections, bends,
level intersections and parcel access. There is no inspection for access to parcels.
A geometric technical inspection of vertical alignment consists of straight sections,
climbing lanes and vertical arcs. For climbing lanes and vertical curves, no checks are
made. The following will explain the results of inspection each technical geometric
component in Tables 4, 5 and 6.
The overall result of the geometric technical inspection of the road is categorized as
conditional functional, so there is still a need for recommendations for improvement.
Recommendations given to geometric road technicalities on Pramuka Street are as
follows:
1 The slope of the pavement should be further lowered to 3% according to technical
standards so that water in the road body can quickly flow into the roadside drains.
2 It is necessary to add a rumble strip to each before entering the intersection.
3 Giving stop signs on minor roads.
Table 4 Cross section assessment results

Assessment Pathways Roadside Road Threshold of Side Traffic
Components Markings Understanding Gutter Safety
Tool
Functional CF – EF EF EF –
eligibility
category
Table 5 Horizontal alignment assessment results

Assessment Straight section Bend section Level intersection Parcel access
components
Functional eligibility EF EF CF –
category
Table 6 Vertical alignment

Assessment Straight section Ascent lane Vertical curve
assessment results
components
Functional EF – –
eligibility
category
Table 7 Results of
Assessment Types of Road Road
Assessment Pavement
components pavement pavement construction
Structures Technical
conditions strength
Functional EF CF –
eligibility
category
4.2 Pavement Structures Technical
In inspection the pavement structure, the strength of the road construction was not
carried out. The results of the technical inspection of road pavement structures can
be seen in Table 7.
The results of all technical assessment for the pavement structure of the
Pramuka Street sections are categorized as conditional functional. The following
recommendations are given so that the pavement structure can achieve eligible
function:
1. Patching any damage.
2. Surface crack filling.
4.3 Road Complementary Building Technical
In the technical inspection of complementary buildings, what was not assessed were
the pontoon, culverts, and retaining walls. The results of technical inspection of
complementary buildings can be seen in Table 8.
The results of all technical inspections of complementary buildings for Pramuka
Street are categorized as conditional functional. The following recommendations are
given so that complementary buildings can achieve eligible function:
1. Marking parking spaces on the edge of Jalan Pramuka.
2. Addition of Parking Lot Signs on Jalan Pramuka.
3. Perform cleaning upstream of the roadside canal so as not to cause disruption
of water flow to the downstream of the canal.
Table 8 Result of assessment complementary buildings technical

Assessment Bridge Pontoon Water tunnel Parking lot Retaining Curbside
Components Walls Channels
Functional EF – – CF – CF
eligibility
category
Table 9 Results of
Assessment Road Benefit Road Property Road
assessment utilization of road
Components Space Space Supervision
parts technical
Room
Functional EF EF EF
Eligibility
Category
4.4 Utilization of Road Parts Technical
The technical assessment of the utilization of road parts as a whole component can
be assessed which includes the road benefit space, the road property space, and the
road supervision room. The following below will explain the results of the technical
inspection of road sections utilization which can be seen in Table 9.
The results of all technical inspections on the use of parts of the road for the
Pramuka Street section are categorized as eligible functional, so there is no need for
recommendations.
4.5 Traffic Engineering Management Technical
The overall technical inspection of traffic engineering management, separators, traffic

signal, zebra cross is not carried out. From the results of field inspection and analysis,
the technical final results of traffic engineering management can be seen in Table 10.
The results of all technical assessment of road traffic management are categorized
as conditional functional. The following recommendations are given in order to
achieve eligible function:
1. Removal of the crossing warning signs by 50 m before the intersection.
2. Placement of signs of prohibited from turning right at the junction of Pramuka
Street and Pemuda Street so that it is moved to the roundabout crossing so that
it can be seen by drivers.
3. Signs that do not meet the standards for placing signs against the roadside must
be adjusted to the standard.
4. Control of traders who are on the sidewalk.
5. Control of those who are still parked on the road sidewalks.
Table 10 Results of assessment traffic engineering management technical

Assessment Road Sign Separator Kerb/Roundabout Side-walk APILL Zebra Clutter
components markings cross tape
Functional EF CF – EF CF – – EF
eligibility
category
4.6 Road Equipment Technical
The road equipment technical inspection consists of 2 components, namely the road
equipment technicalities related to road users and the road equipment technicalities
that are not related to the road users. Inspection of components equipment road
technical related to road users is carried out on road markings, signs, kerb/roundabout,
sidewalks, rumble strip, lighting lamps, traffic light, and separators. The inspection
of traffic light, separators and road equipment technical that is not related to road
users in this research was not carried out. The results of inspection road equipment
technical can be seen in Table 11.
The results of all technical inspections of road equipment technical are categorized
as conditional functional. The following recommendations are given in order to
achieve eligible function:
1. Markings that are damaged/faded so that they can be repainted so that they can
be seen by drivers.
2. The condition of the signs is obstructed by the leaves, so the flags must be
handled immediately so that the signs can be seen clearly by the drivers.
3. The condition of the signs that have been filled with graffiti should be replaced
with new ones.
4. The color of the signs is faded to be repaired/replaced immediately.
5. The dimensions of additional board signs that are not yet suitable, please be
adjusted to technical standards.
6. Marking the edge of the roundabout at the intersection.
4.7 Road Administration
The assessment of the completeness of road administration for the scout roads is
carried out at the Public Works and Spatial Planning Office of Klaten Regency by
verifying whether each document is complete. The following in Table 12 below will
be displayed regarding the results of road administration inspection.
The results of the examination of the administrative data for the Pramuka Street
section at the Public Works and Public Housing Office of Klaten Regency were only
the road status documents and road information documents. Road class documents,
instructions and prohibitions documents, and environmental documents unavail-
able. Land ownership documents are still in the process of land certification. With
Table 11 Result of assessment road equipment technical

Examination Road Sign Kerb/Round-about Sidewalk Clutter Lighting APILL Separator
components markings tape
Functional CF EF CF EF CF CF – –
Eligibility
Category
Table 12 Road administration assessment results

Completeness of road administrative documents and assessment results
Road status Establishment of Street class Land ownership Environment Leger Jalan
instructions and
don’ts
Available Unavailable Unavailable, Certification still Unavailable Available
only using Processing,
reference to Law Unavailable
no. 22 of 2009
Article 19
paragraph 2C
EF CF CF CF CF EF
this result, the road administration can be categorized as conditional function.

Recommendation given in order to achieve eligible function is:
1. Completing missing road administration documents and completing the docu-
ment procurement process immediately in order to ensure legal certainty for
road operators and road users.
4.8 Percentage of Functional Assessment Eligibility Category
Based on the results of the recapitulation above, it was found that there were 18
components that were eligible functional, 15 components had conditional functional
values and there were no components that were ineligible functional, both techni-
cally and administratively. Based on Formula (2), it is obtained that the percentage
of eligible function categories is 54.55%, conditional function is 45.45% and for
ineligible categories is 0%. Based on previous research, the results obtained in
the road functional assessment showed the same results, namely conditional func-
tion and some road components needed to be repaired in order to realize safe and
proper road conditions for the driver. With these results based on Minister of Public
Works Regulation No. 11.PRT/M/2010 it can be concluded that the Jalan Pramuka
section has a status or has a conditional function category. The following will be
displayed regarding the percentage of functional assessment eligibility categories of
the Pramuka Street section in Fig. 2:
5 Conclusion
1. The status/category of functional assessment eligibility category for the

Pramuka Street section is Conditional Function (CF) so that it still requires
recommendations in order to achieve the Eligible Function Status
Fig. 2 Diagram of
percentage functional Percentage of Functional Assessment
assessment eligibility Eligibility Category Pramuka Road
category Section
45.45 %
54.55 %
EF CF IF
2. Recommendations for improvements in order to achieve an eligible functionally

status on the Pramuka Street are:
• Improvements to technical components which include road geometric tech-
niques, pavement structure techniques, road complement building tech-
niques, traffic engineering management techniques, and road equipment
engineering.
• Completing road administration documents consisting of documents stipu-
lating instructions for orders and prohibitions, road class documents, land
ownership documents, and environmental documents.
– For further research, it is expected to carry out a road safety audit on the
Pramuka Roads if the level of traffic accidents increases.
– The role of related agencies is needed to complete supporting data to
facilitate research on road functional assessment such as IRI data, as built
drawings, DED documents, etc.
References
1. Amin CM (2017) Faktor-Faktor yang Mempengaruhi Pertumbuhan Kendaraan Bermotor Roda

Dua Di Kota Pekanbaru. JOM Fekon 4:1106–1120
2. Priyanto S (1998) Accident analysis on Surabaya—Gempol toll road. Eng Media 20(4):50–59
3. Anonim (2010) Peraturan Menteri Pekerjaan Umum Nomor 11/PRT/M/2010 Tentang Tata
Cara dan Persyaratan Laik Fungsi Jalan. Kementerian Pekerjaan Umum, Jakarta
4. Adwang J (2020) Analysis of road function assessment (ULFJ) on tumora national road (Bts.
Kab. Parimo)—Tambrana section number 31 km. 157 + 800 to km. 168 + 000 Central Sulawesi
Province. TEKNO, vol. 17, pp 122–138
5. Zachawerus J (2016) Uji Laik Fungsi Jalan Dalam Mewujudkan Jalan Yang Berkeselamatan
(Studi Kasus Jalan Utama Di Pusat Kota Ternate). Prosiding Seminar Nasional Teknik Sipil 6,
Surakarta, 25 October 2016, pp 169–176
6. Mayuni S, Widodo S, Sulandari E (2017) Evaluasi Keselamatan Infrastruktur Jalan (Studi
Kasus Jalan Trans Kalimantan). Prosiding Konferensi Nasional Teknik Sipil dan Perencanaan
(KN-TSP), Pekanbaru, 9 Feb 2017, pp 187–195
7. Widianty D, Rohani, Employees, IA (2019) Analisis Keselamatan Jalan Pada Tikungan

Berdasarkan Jari-Jari Dan Kemiringan Melintang Tikungan. JRS-UNAND 15:103–114
8. Anonim (1997) Tata Cara Perencanaan Geometrik Jalan Antar Kota. Departemen Pekerjaan
Umum Direktorat Jenderal Bina Marga, Jakarta
9. Anonymous (2014) Petunjuk Pelaksanaan Kelaikan Fungsi Jalan No. 09/P/BM/2014. Direk-
torat Jenderal Bina Marga Kementerian Pekerjaan Umum, Jakarta
10. Mulyono AT (2012) Formulir Laik Fungsi Jalan diwilayah BBPJN-VII. Yogyakarta
The Evaluation of Irrigation
Maintenance in Pacal Irrigation Area
at Bojonegoro Regency, East Java
Mahdika Putra Nanda, Rintis Hadiani, and Antonius Suryono
Abstract Pacal Irrigation Area, located at Bojonegoro regency and sourced from
Pacal dam, was supporting 16.624 Ha agricultural field in 5 districts including Kapas,
Balen, Sumberrejo, Kepoh Baru, and Baureno. Due to its vital function as agricul-
tural support service, Pacal irrigation network should be able to function optimally
to assist high crop yields. The maintenance of irrigation then became an important
concern. However, the previous survey showed poor maintenance toward irrigation
infrastructure, leading to non-optimal function to crop yields. This research aimed
to identify the physical condition of the irrigation network and its infrastructure in
Pacal irrigation area. The evaluation is based on the Minister of Public Works and
Public Housing Regulation No. 12/2005 about exploitation and irrigation mainte-
nance system. This particular regulation determined the appropriate maintenance
for efficient irrigation operation and performance, leading to optimal agricultural
productivity. The result showed that the primary canal was damaged in 12.55%,
indicating a minor level of damage. Meanwhile, for the network and infrastruc-
ture surrounding the area, periodic maintenance was needed to ensure the irrigation
performance running efficiently for supplying water.
Keywords Irrigation · Maintenance · Evaluation · Performance
M. P. Nanda · R. Hadiani (B)

Department of Civil Engineering, Faculty of Engineering, Sebelas Maret University, Jl. Ir. Sutami
36 A, Surakarta, Central Java 57126, Indonesia
M. P. Nanda
R. Hadiani · A. Suryono
Procurement Officials of Water Resource Operation and Maintainance IV, Banaran, Pabelan,
Kartasura District, Sukoharjo Regency, Central Java 57169, Indonesia
426 M. P. Nanda et al.
1 Introduction
Water is a vital natural resource that supports a lot of sectors in life. This element is
highly needed especially in the agricultural sector where the water supply supports
most of the production. This sector then requires good water management to fulfill
all the water need. That management is usually in the form of a dam or irrigation to
accelerate and facilitate water distribution. In Indonesia, water sources are usually
taken from a river, reservoir, groundwater, and tidal system. If the water availability
is not adequate, another solution is made to still fulfill the water need. Water scarcity
is a challenge due to increasing demand from various sources ranging from agricul-
ture, mining, industry, hydroelectric power, and the increasing human population,
especially in rural areas, the need for irrigation water is very important for agricul-
tural land [1]. This stresses how important to calculate water need is because it is
part of irrigation planning and maintenance. Irrigation management must consider
the method of water use, the time of giving water, and how much water will be used
[2]. Irrigation network performance is greatly affected by the physical condition of
the building, function channels, and other factors related to the management of the
irrigation network. Irrigation system performance is part of irrigation management
and assessments are usually carried out once a year to monitor the function and
performance of all aspects irrigation system. This assessment is carried out by Irri-
gation Area Manager by the location of the irrigation area [3]. Irrigation strategies
and water demand have been hampered by a lack of understanding of the relationship
between water scarcity, food production, food security, and sustainability [4]. The
irrigation area is the basic territorial unit for the management of water resources,
especially in rural areas. Irrigated areas are not only for the rural population and the
amount of agricultural production on a large scale but also consume a lot of water
resources [5]. This can be a factor that indirectly promotes the high productivity of
rice crops. Irrigated agriculture appears to be a prime candidate for interventions
aimed at reducing water demand and encouraging water savings [6].
Irrigation itself is the key to increase agricultural productivity [7]. In its prin-
ciple, irrigation is an effort to distribute water through canals to agricultural fields
and remove the excess water through wastegate [8]. The irrigation system includes
infrastructures, water, operation management, and human resources [9]. Irrigation
management is supported by five pillars including water availability, infrastructure,
operation management, irrigation institution, and human resources [10].
By having this irrigation system, it is hoped that the need for water can be supplied
to accelerate agricultural activities. It also can be a solution for water shortages in the
dry season due to low rainfall. Maintenance of irrigation infrastructure then occurs
to be important to promote agricultural activities. Besides that, the demand for irri-
gation’s good performance guarantees people’s food security, especially farmers.
This is based on the fact that 23% of irrigation networks are damaged, only 11%
of irrigation systems drain water regularly effective, and irrigation management is
not efficient [11]. This problem causes water management to be ineffective and inef-
ficient, leading to non-optimal agricultural production. The irrigation network will
The Evaluation of Irrigation Maintenance in Pacal Irrigation Area … 427
be customized with the water availability if the case shows that the water supply is
lower than the water need. In an irrigation network, regular maintenance is important
to support effective operation and performance. HIPPA, as the organization of agri-
cultural irrigation management, can also participate in supervising and controlling
the irrigation system.
Maintenance of damaged irrigation has become a priority. instead of constructing
a new irrigation network. rehabilitation is the solution. Due to the large number
of costs required for the rehabilitation of a large irrigation network, it is necessary
to find an optimal rehabilitation scenario that can improve water management and
hydraulic performance at the minimum possible cost [12]. The provincial government
is responsible to restore the primary and secondary irrigation canals based on the
irrigation maintenance level priority. Unfortunately, this priority data can be mistaken
due to the limited knowledge and skill of farmers and government, leading to bad
distribution of funding allocation. Therefore, technical training and data management
are considered to be important in rehabilitating irrigation.
The irrigation area which covers an area of more than 3000 Ha is the government
authority. Pacal irrigation area is under BBWS Bengawan Solo’s supervision, a public
river basin management organization. This area is located in Bojonegoro Regency
with an agricultural field of 16.624 Ha. This irrigation supplies water for 5 districts,
including Kapas, Balen, Sumberrejo, Kepoh Baru, and Baureno. However, an inter-
view from Procurement Officials for Water Resource’s Operation dan Maintainance
IV stated that a water reservoir that conveys only 24 Million m3 can not supply all the
needs for these districts. Besides that, the canal and other irrigation facilities show
some damage such as a crack on the canal network. This lead to a decrease in Pacal
irrigation performance. This research then aims to investigate the physical condition
of Pacal irrigation to function optimally and the kind of maintenance that can help
increase the irrigation performance.
2 Research Method
This research used a descriptive method to explain the variables. In principle, descrip-
tive research does not aim to form hypotheses or theory development. The character-
istic of descriptive research is objectivity or neutrality. Descriptive research is about
describes how reality is. In this case, descriptive research is different from prescrip-
tive research which is primarily concerned with the question of what reality should
be. Descriptive research is making an inventory prescriptive research is normative
[13]. The data were taken from primary and secondary data. Primary data included
traceability data and irrigation documentation. On the other hand, secondary data
include irrigation network schematic data and irrigation map area. To know the
physical condition of Pacal irrigation, evaluation is performed by investigating the
existing condition, size, traits, the scope of work, and other parameters that have
been set based on the standard and guidelines. In this research, primary data was the
basis that determines the physical condition of Pacal irrigation. Then, after knowing
Start
Data
Collection
Secondary Data
1. irrigation network Primary Data
Schematic data Irrigation network
2. Irrigation area map traceability data
Data Analysis
Analysis of
exixting Data
Determination of
the appropriate
types of
maintenence
Finish
Fig. 1 Research flowchart
the data under existing condition, the type of maintenance and development for the
irrigation can be determined to increase its function as the agricultural production
support. The research data is presented in Fig. 1, as seen.
This research was conducted in Pacal Irrigation Area, Bojonegoro Regency, East Java.
Existing condition evaluation was done to know Pacal Irrigation Area’s situation in
the part of physical infrastructure, the rice planting area, and growth, and also the
fulfillment of irrigation need. From the research in 2017, the area of rice planting in
the wet season used 15.096 Ha of agricultural field. It decreased in the first period of
the dry season to become 14.528 Ha fields. On the other hand, in the second period
of the dry season, it lowers to become 9.019 Ha. For the rice planting growth, Pacal
irrigation area reached 90.46% in the wet season, 87.75% in the first dry season, and
90.59% in the second of the dry season.
Survey and documentation of Pacal irrigation were taken throughout the canals
of 12.48 km of Klepek to Mengkuris weir. It also was taken from 8 km canals of
Mengkuris to Kerjo weir. Pacal irrigation area consists of 16.624 Ha of agricultural
field. Throughout the canals, maintenance was indeed needed such as clearing the
weed on the canal wall and fixing the cracked wall that can disturb the streamline of
water. It is seen in Fig. 2, as such.
The investigation through GPS and Google Earth is used to facilitate the damage
tracking and inventory process. Figure 3 showed the map to the track location in
Pacal irrigation, as seen.
After the investigation, the result showed the damage of Pacal irrigation infras-
tructure. The evaluation is based on Minister of Public Works and Public Housing
Regulation No. 12 of 2015 where the physical condition index is presented in Table
1, as seen.
Based on the result above, Pacal’s irrigation infrastructure condition can be
determined. It is presented in Table 2 as irrigation damage index, as seen.
Fig. 2 The damage and weed on irrigation canals
Fig. 3 Map of track on

irrigation network
Table 1 Irrigation condition

Description Condition index
index
Maximum (%)
Infrastructure
1. Main structure 13
2. Canals 10
3. Canal structure 9
Source Minister of Public Works and Public Housing Regulation
No.12 of 2015
According to the table above, it can be seen that the damage of Pacal irrigation main
infrastructure is 4.40%. It came from a lot of components such as a reservoir, sluice,
sand trap, and wastegate. On the other hand, the damage of 7.40% happened in canals
and 0.75% in canal structures. The total damage in Pacal irrigation infrastructure is
12.55%, showing a minor level of damage. For the damage, the major level is >40%,
the moderate level is 21–40%, and the minor level is 10–20%. Compared to that,
<10% level means the irrigation is in good condition. Pacal irrigation infrastructure
damage is presented in a bar chart, as shown (Fig. 4).
Maintenance for irrigation infrastructure is important for agricultural sustain-
ability [14]. Maintenance of the irrigation scheme for the whole Every year Indonesia
requires a large amount of money, an allocation provided by the central government
as well as local governments have not been able to meet these costs. requirements for
the maintenance of the irrigation scheme of each The Irrigation Area must consider
the performance of the Irrigation Area in the aspects of the physical infrastructure
condition of the network and other supporting aspects [15]. Irrigation asset manage-
ment, which includes planning, maintaining, and funding the network, is important
to be done efficiently to ensure the optimal benefit for irrigation users. Operation
and maintenance of irrigation network should be a responsibility of all, including
water user farmers association as organization of agricultural irrigation management,
provincial government, body corporate, social organization, people, and individuals
that gain benefit from this irrigation [16]. The maintenance of an irrigation system is
something that cannot be defined in technical, economic, or even institutional terms.
Rather, it must be defined as a process that can be carried out within a technical,
economic, and institutional framework and framework [17]. All should maintain
irrigation to run optimally based on its function.
4 Conclusion
From the analysis above, it can be concluded that physical damage of Pacal’s primary
canal is 12.55%, indicating a minor level of damage. Meanwhile, for the network
and infrastructure surrounding the area, regular maintenance was needed to ensure
the irrigation performance run efficiently for watering the agricultural field. That
Table 2 Index percentage of irrigation damage

Details Ratio Value Condition index
(%) (%) Inspection (%) Maximum (%)
I. Infrastructure
With the detail description as seen
1. Main structure 4.40 13.00
1.1 Weir 1.62 100.00 4.00
a. Weir crest 0.36 20.00 20.00 0.80
b. Wing wall 0.00 15.00 0.00 0.60
c. Weir floor 0.00 20.00 0.00 0.80
d. Weir shutter 0.00 20.00 0.00 0.80
e. Bridgefoot 0.22 5.00 50.00 0.20
f. Information board 0.62 10.00 70.00 0.40
g. Head gauges 0.04 5.00 10.00 0.20
h. Safety bridge fence 0.38 5.00 85.00 0.20
1.2 Weir gate’s gear that can be operated 2.78 100.00 7.00
a. Intake gate 0.68 55.00 8.00 3.85
b. Waste weir 2.10 45.00 30.00 3.15
1.3 Sand trap dan waste weir 0.00 100.00 2.00
a. Good condition of sand trap 0.00 50.00 0.00 1.00
b. Waste weir and gear of sand trap that can 0.00 50.00 0.00 0.50
be operated
2. Canal network 7.40 10.00
a. Canal capacity 2.89 50.00 26.00 5.00
b. Levee canals 0.70 10.00 31.67 1.00
c. Leaked levee canals 0.75 10.00 33.64 1.00
d. Slided/Broken/Collapsed retaining wall 3.06 30.00 46.00 3.00
of levee
3. Structure in canal network 0.75 9.00
3.1 Regulation structure 0.75 100.00 3.50
(Conveyance/Distribution/Field application
structure)
a. Conveyance/Distribution structure 0.75 100.00 9.60 3.50
3.2 Current meter 0.00 100.00 3.00
a. Current meter 0.00 30.00 0.00 0.90
3.3 Weir’s ancillary works 0.00 100.00 2.50
a. Flumes, Syphon, Spillway, Drop 0.00 60.00 0.00 1.50
Structure, Settling Basin, Inlet Drain
b. Bridgefoot, culverts 0.00 40.00 0.00 1.00
Total damage 12.55
Irrigation Network Damage

15
percentage
10
5
0
Canal
Main Structure Canals
Structure
Condi on Index 13 10 9
calcula on results 4.4 7.4 0.75
Fig. 4 Diagram for irrigation network damage
maintenance could be in the form of cleaning the weed that grew uncontrollably
around the canals, plastering the cracks or the broken canal walls, lubricating the
hinges of the sluice, coating the sluice gate with anti-corrosion paint, and change the
lost or broken bolt.
References
1. Winters P, Salazar L, Maffioli A (2010) Designing impact evaluations for agricultural projects.
Impact-Evaluation Guidelines, Technical Notes No. IDBTN-198 Strategy Development
Division. Inter-American Development Bank
2. Allen LN, Adam JWM (2020) Irrigation and water management. The Science of Grassland.
Wiley Online Library, Forages
3. Amrizal UL, Tohari B (2016) Performance evaluation of irrigation system in Cikeusik Irrigation
Area, Cirebon Regency West Java Province. In: The 2en international conference on civil
engineering research (ICCER). pp 12–18
4. Carruthers I, Rosegrant MW, Secler D (1997) Irrigation and food security in the 21st century.
Irrig Drainage Syst 11:83–101
5. Pan H, Xu Q (2010) Quantitative analysis on the influence factors of the sustainable water
resource management performance in irrigation areas: an empirical research from China. Econ
Bus Aspects Sustain 10(1):264
6. Bekchanov MJ (2016) Water res and economic 13:30–45
7. Molden D, Orr S, Pittock J, Richter B, Riddell PJ, Scott CA, Philippe VJ, Vos J, Zwarteveen M
(2020) A scale-based framework to understand the promises, pitfalls and paradoxes of irrigation
efficiency to meet major water challenges. Glob Environ Change 65
8. Peraturan Menteri Pekerjaan Umum Dan Perumahan Rakyat (2015) Eksploitasi Dan Pemeli-
haraan Jaringan Irigasi
9. Kementerian Pekerjaan Umum dan Perumahan Rakyat (2019) Modul Pengenalan Sistem Irigasi
10. Direktorat Jenderal Sumber Daya Air (2019) Modul Pemeliharaan Dan Operasi Jaringan Irigasi
11. Waskito NT, Arif SS, Maksum M, Sutanto S (2010) Study on amortiztion in irrigation system
management using knowledge management appoach. Paper on Asia Regional ICID
12. Elgamal T, Elfetyany M, Elkassar G (2019) Effect of irrigation network rehabilitation on water
management-case study: Tanta navigation canal—Egypt. Alexandria Eng J 58(4)
13. Lans W, van der Voordt T (2002) Descriptive research. Delft University of Technology. Ways
to study and research urban, architectural and technical design
14. Sharaunga S, Mudhara M (2018) Determinants of farmers’ participation in collective mainte-

nance of irrigation infrastructure. In: Kwazulu-Natal economic factors time ; Dependency on
resource household characteristics. Physics and Chemistry of the Earth 105
15. Edhisono S, Hadihardaja IK, Suripin S (2016) Optimation budgeting distribution model for
maintaining irrigation scheme. Int J Sci Eng 10(1):35–44
16. Peraturan Menteri Pekerjaan Umum (2007) Pedoman Pengembangan Pengelolaan Sistem
Irigasi Partisipatif
17. Svendsen M, Huppert W (2003) Optimal maintenance in irrigation. Irrig Drainage Syst 17:109–
128
Hazard Mitigation
Landslide Analysis Subject to Geological
Uncertainty Using Monte Carlo
Simulation (A Study Case in Taiwan)
Joni Fitra, Wen-Chao Huang, and Yusep Muslih Purwana
Abstract Landslide is the primary driver of the denudational process and sediment
source dominantly onsite. Landslides are one of the most disastrous effects in Taiwan;
groundwater or flood erosion is highly attributed to the landslide. Water induced to
the slope increases driving force and decrease resisting force causing a slope land-
slide. This condition is generally affecting slope stability. In this study, we attempt to
consider the uncertainty of the dip angle in slope stability analysis. In this research, the
Monte Carlo simulation was used to quantify the effect of the geological uncertainty.
Various sources of dip angles (with mean and standard deviation) were employed
to generate 100,000 dip angle samples. All of the dip angles employed in this study
were based on Highway no. 3 sliding events in Taiwan. Four different measure-
ment sources, i.e., Central Geological Survey (CGS, Taiwan), Compass measurement
before the sliding event, Surface measurement after the event, and LiDAR-derived
data, were employed in this study. Further, the measured dip angles were converted
to the projected dip angle based on the plane’s strike. Simulation results show LiDAR
Measurement Source provides the lowest failure probability of 16.9%, and Central
Geological Survey (CGS, Taiwan) Measurement provides the highest failure prob-
ability of 78%. Therefore, based on the engineering design concept, if the design
performed using the CGS data, the engineering design must be very conservative
compared to the design using the LiDAR data.
Keywords Landslide · Geological uncertainty · Monte Carlo simulation
J. Fitra (B) · Y. M. Purwana

Department of Civil Engineering, Sebelas Maret University, Surakarta, Central Java, Indonesia
Y. M. Purwana
J. Fitra · W.-C. Huang
Department of Civil Engineering, National Central University, Taoyuan, Taiwan (R.O.C.)
438 J. Fitra et al.
1 Introduction
Taiwan is an island formed by the active geological process on a complex of the

convergent boundary. Landslide is the primary driver of the denudational process
and sediment source dominantly onsite [1]. Landslides are one of the most geolog-
ical disasters affecting Taiwan; groundwater or flood erosion is highly attributed
to the landslide. Water induced to the slope increases driving force and decrease
resisting force causing a slope landslide. This condition is generally affecting slope
stability. The landslide has occurred widely globally, especially in the ring of fire’
zone countries, i.e., Indonesia, Philippines, Taiwan, Japan, South Korea. A landslide
occurred on Highway No. 3 near Xizhi, Taiwan (ROC) on 2010 April 25. A large
amount of sediment buried both directions of the freeway. Four people were killed,
and four cars were buried under debris [2]. For the above landslide, no rainfall or
earthquakes have been recorded before the event. The landslide may be due to the
local infiltration of groundwater between the weak planes or the weakening of the
slope’s stabilization facilities (Fig. 1).
The stability of rock slopes can be assumed as Mohr–Coulomb material expressed
in terms of the cohesion c, friction angle f and the factor of safety of the slope as
follows [3–6]:
c A + [W cos(β) − Wp − H sin(β) + T cos(θ )]tan

FS = (1)
W sin(β) + V cos(β) − T sin(θ )
c: the cohesion, F: the friction angle (°); A: the slip area plane; θ: the angle of the
rock anchor and normal vector (°); β: the dip angle (°), W: the weight of slipped,
Wp : water pressure, H: horizontal force, T: rock anchors strength affecting to design
force.
In this study, some assumption was applied for the equation. These are (1). No
Cohesion was applied, (2). No water pressure effect, (3). No anchors were applied
to stabilize the plane. If the sliding surface contains no water infiltration and zero
Fig. 1 Photograph of
freeway No. 3 landslides in
Taiwan [3]
Landslide Analysis Subject to Geological Uncertainty … 439
cohesion, Eq. (1) reduces to
tan
FS = (2)
tan (β)
While a geologist does site investigation, one of the field’s parameter measurements is
strike and dip of the possible sliding planes. Site investigation and ground description
are considered as one of geologic model uncertainty [7–9]. It is crucial to include
a dip angle in slope stability analysis that best describes the potential sliding plane.
However, it may not be realistic due to the intrinsic and extrinsic uncertainties.
Therefore, in this study, we attempt to consider the dip angle’s uncertainty in slope
stability analysis. Various sets of dip angles that correspond to different measurement
accuracy were applied in the plane failure analysis type of a simplified slope failure,
hoping to shed some light on the influences of dip angle uncertainties on the slope’s
overall stability.
2 Monte Carlo Simulation Method
The Monte Carlo simulation was used to quantify the effect of the geological uncer-
tainty on the safety factor and failure probability. Eight sources of dip angles (with
mean and standard deviation) were employed to generate 100,000 dip angle samples.
All of the dip angles employed in this study were based on the Highway no. 3 sliding
events in Taiwan, i.e., the dip angles were measured or derived close to the sliding
site. The sources include (a). Measured Central Geological Survey (M-CGS), (b).
Projected Central Geological Survey (P-CGS), (c). Measured Compass (M-Com),
(d). Projected Compass (P-Com), (e). Measured LiDAR (M-LiD), (f). Projected
LiDAR (P-LiD), (g). Measured Surface (M-Sur) and (h). Projected Surface (P-Sur)
as shown in Table 1.
Monte Carlo analysis randomly samples those parameters on the given probability
distribution and its statistical parameters. Afterward, the generated random param-
eters were employed with the equation to calculate a corresponding safety factor
several times (100,000 iterations in this study). We can obtain a series of safety
factors. We can achieve a probability distribution of safety factors accounting for
the dip angle’s uncertainty through this process. We assumed that the distribution
of all data sources of dip dingle and safety factors in this research was log-normal
distribution.
2.1 Central Geological Survey (CGS) Taiwan
The mean value and standard deviations of the sliding plane dip angle are 22.5 and
9.57, while the friction angle is considered a constant of 15°. We addressed σ 2ln x
440 J. Fitra et al.
Table 1 Variation source of

Dip angle source Mean Standard
dip angle
value (degree) deviation (degree)
Measured CGS 22.5 9.57
(M-CGS)
Projected CGS 18.75 8.08
(P-CGS)
Measured Compass 17 4.37
(M-Com)
Projected Compass 16.01 3.89
(P-Com)
Measured LiDAR 14.2 1.81
(M-LiD)
Projected LiDAR 13.66 1.41
(P-LiD)
Measured surface 14.91 2.55
(M-Sur)
Projected surface 14.7 2.47
(P-Sur)
Fig. 2 Log-normal
distribution data of central
geological survey (CGS) dip
angle
and μln x are 0.170 and 2.846. The log-normal distribution of the Central Geological
Survey dip angle is shown in Fig. 2.
Figure 2 illustrated the probability density function of the measured Central
Geological Survey (M-CGS) and projected Central Geological Survey (P-CGS) dip
angle for the non-symmetric log-normal distribution. The measured Central Geolog-
ical Survey (M-CGS) distribution data separate widely within a long-range between
4 and 60°. The distribution data of the projected Central Geological Survey (P-CGS)
separate widely within range 4–50°. Based on this graphic, the measured Central
Fig. 3 Log-normal
distribution data of central
geological survey (CGS)
safety factor
Geological Survey (M-CGS) dip angle has a broader range than the projected Central
Geological Survey (P-CGS) dip angle. Furthermore, distribution data of central
geological survey projected dip angle has a higher peak. The graph also shows a
reasonable trend, as the projected dip angle indicates the effect of transferring from
an apparent dip into the true dip.
Consequently, we calculate the factor of safety using the generated random vari-
ables. After 100,000 repeated calculations, the probability distribution curve for the
factor of safety was obtained. The number of cases with a safety factor less than
one and dividing it by the total cases were used to evaluated failure probability. The
probability density function (PDF) of the safety factor is also shown in Fig. 3. Failure
probability (Pf ) of the measured Central Geological Survey (M-CGS) data is 78%,
and failure probability (Pf ) of the projected Central Geological Survey data is 63%.
It can be concluded that the employment of the true dip angle in the analysis can
yield a smaller failure probability. However, the Pf in the two analyzed cases showed
an extremely high value leading to the slope’s failure.
2.2 Compass Measurement in the Field
The second data source that was used in this research is the compass measurement
dip angle. The probability density function of the compass measurement dip angle is
shown in Fig. 4, in which one is the measured dip angle, and the other is the projected
dip angle.
Figure 4 illustrates the probability density function of the compass measurement
dip angle for the non-symmetric log-normal distribution. The distribution data of
measured compass (M-Com) separate widely within range 6–32°, and the distribution
data of projected compass (P-Com) separate widely within range 6–30°.
442 J. Fitra et al.
Fig. 4 Log-normal
distribution data of compass
measurement dip angle
Fig. 5 Log-normal
distribution data of compass
measurement safety factor
Based on this graphic, the measured compass dip angle has a broader range than the
projected compass dip angle. Therefore, distribution data of the projected compass
dip angle has a higher peak. The probability density function (PDF) of the safety
factor is shown in Fig. 5. Failure probability (Pf ) of the measured compass data is
64.5%, and failure probability (Pf ) of the projected compass data is 56.1% factor of
safety <1 and evaluated as unstable.
Fig. 6 Log-normal
distribution data of surface
2.3 Surface Measurement of the Sliding Plane After

the Event
The third data source that was used in this research is the surface measurement
dip angle. The surface measurement was carried out after the sliding event. The
probability density function of the surface measured dip angle is shown in Fig. 6.
Figure 6 illustrates the probability density function of the surface measurement dip
angle for the non-symmetric log-normal distribution. The distribution data separate
widely within the range of 8–24°. Based on this graph, the measured surface (M-Sur)
dip angle and projected surface (P-Sur) dip angle have the same dip angle range.
Furthermore, distribution data of the projected surface dip angle has a higher peak.
Finally, we can also observe that the measured and projected surface dip angles are
close to each other, indicating that the field’s measurement was indeed close to the
true dip of the sliding plane. The probability density function (PDF) of the safety
factor is shown in Fig. 7. Failure probability (Pf ) of the measured surface data is
45.1%, and failure probability (Pf ) of the projected surface data is 42.1%, given that
factor of safety <1 is evaluated as unstable.
2.4 LiDAR Measurement
Traditional geological mapping full potential can be achieved using 3D rendering

techniques on airborne Light Detection and Ranging (LiDAR) due to the exact three-
dimensional (3D) Digital Terrain Model (DTM) information [10, 11]. The fourth
444 J. Fitra et al.
Fig. 7 Log-normal
distribution data of surface
Fig. 8 Log-normal
distribution data of LiDAR
data source that was used in this research is the LiDAR measurement dip angle. The
probability density function of the LiDAR measured dip angle is shown in Fig. 8.
Figure 8 illustrates the probability density function of the LiDAR measured dip
angle for the non-symmetric log-normal distribution. The distribution data sepa-
rate widely within the range of 8–20°. Based on this graphic, the measured LiDAR
(M-LiD) dip angle has the same range as the projected LiDAR (P-LiD) dip angle.
Therefore, distribution data of the projected LiDAR dip angle has a higher peak. We
can also see that due to this source’s high accuracy, the distribution of dip angles is
more concentrated than other sources of dip angles.
The probability density function (PDF) of the safety factor is shown in Fig. 9.
Fig. 9 Log-normal
distribution data of LiDAR
Table 2 Summary table of

Source Failure probability (Pf )
failure probability of the eight
sources Measured Central Geological Survey 78%
(M-CGS)
Projected Central Geological Survey 63%
(P-CGS)
Measured Compass (M-Com) 64.5%
Projected Compass (P-Com) 56.1%
Measured Surface (M-Sur) 45.1
Projected Surface (P-Sur) 42.1%
Measured LiDAR (M-LiD) 31%
Projected LiDAR (P-LiD) 16.9%
Failure probability (Pf ) of the measured LiDAR data is 31%, and failure probability
(Pf ) of the projected LiDAR data is 16.9%, given that safety factor <1 is evaluated
as unstable.
3 Discussions and Conclusions
Table 2 summarizes the failure probabilities of the eight sources of dip angles. Based
on Table 2 results, we realize that the failure probability values are greater than 10%,
indicating a high potential for slope failure if no stability measures were taken. In
the current analysis, we have assumed no anchor applied in the design, while in the
446 J. Fitra et al.
actual design, anchor forces were applied in the slope so that the factor of safety is
acceptable [12]. Even the safety factor is acceptable when the anchor forces were
applied, the failure probability is still not acceptable [13]. If the design is carried
out only using the CGS data with a failure probability of around 60–80%, engineers
would like to bring the Pf down to 0.01% (an acceptable failure probability for
engineering design). For example, we can expect that the corresponding cost can be
high. For example, the number of anchors can be more than the project’s budget, or
more cutting and filling may need to be performed to lower the slope geometry.
On the other hand, if the design is performed with the LiDAR data, it is evident
that the required cost to bring down the original failure probability from 16% to
0.01% must be much lower than the case designed with CGS data. Hence, it is
crucial to adopt more realistic dip angles of the analyzed slope to avoid unnecessary
engineering designs. Although adopting the CGS data for design may cost less in
the initial design phase, doing so could cost more in the actual design phase.
Acknowledgements The authors are grateful for the financial support provided by the Ministry of
Science and Technology (MOST) Shackleton Program through Grant No. MOST108-2638-E-008-
001-MY2 (Principal Investigator: Dr. Hsein Juang).
References
1. Broeckx J, Rossi M, Lijnen K, Campforts B, Poesen J, Vanmaercke M (2020) Landslide

mobilization rates: a global analysis and model. Earth-Sci Rev 201:102972. https://doi.org/10.
1016/j.earscirev.2019.102972
2. Chen L-K, Chen S-C, Ke M-C (2015) Investigation of the Freeway No. 3 landslide in Taiwan.
In: Lollino G, Giordan D, Crosta GB, Corominas J, Azzam R, Wasowski J, Sciarra N (eds)
Engineering geology for society and territory, vol 2. Springer International Publishing, pp
2093–2096. https://doi.org/10.1007/978-3-319-09057-3_374
3. Wang L, Hwang JH, Luo Z, Juang CH, Xiao J (2013) Probabilistic back analysis of slope
failure—a case study in Taiwan. Comput Geotech 51:12–23
4. Hoek E, Bray J (1981) Rock slope engineering, 3rd edn. Institution of Mining and Metallurgy,
London
5. Turner AK, Schuster RL (1996) Landslides—investigation, and mitigation: national research
council. Transportation research board special report 247. National Academy Press, Wash-
ington (DC)
6. Wyllie DC, Mah CW (2004) Rock slope engineering, 4th edn. Spon Press, London
7. Juang CH, Zhang J, Shen M, Hu J (2019) Probabilistic methods for unified treatment of
geotechnical and geological uncertainties in a geotechnical analysis. Eng Geol 249:148–161.
https://doi.org/10.1016/j.enggeo.2018.12.010
8. Keaton JR (2013) Engineering geology: fundamental input or random variable? In: Proceeding
of the geo-congress, 3–7 Mar 2013, San Diego, pp 232–253
9. Sullivan TD (2010) The geological model. In: Williams AL, Pinches GM, Chin CY, McMorran
TJ, Massey CI (eds) Geologically active: proceeding of the 11th congress of the international
association for engineering geology and the environment. CRC Press, Auckland, New Zealand,
London, Sep 5–10, pp 155–170
10. Yeh C-H, Chan Y-C, Chang K-J, Lin M-L, Hsieh Y-C (2014) Derivation of strike and dip in
sedimentary terrain using 3d image interpretation based on airborne LiDAR data. Terr Atmos
Oceanic Sci 25(6):775. https://doi.org/10.3319/TAO.2014.07.02.01(TT)
11. Yeh C-H, Lin M-L, Chan Y-C, Chang K-J, Hsieh Y-C (2017) Dip-slope mapping of sedimentary
terrain using polygon auto-tracing and airborne LiDAR topographic data. Eng Geol 222:236–
249. https://doi.org/10.1016/j.enggeo.2017.04.009
12. TGS (2011) The investigation report of the landslide at Taiwan highway No.3, Ministry of
Transportation and Communications investigation report. Taiwan Geotechnical Society, Taipei.
(in Chinese)
13. Yeh C-H, Dong J-J, Khonevisan S, Juang CH, Huang W-C (2021) The role of the geological
uncertainty in a geotechnical design—a retrospective view of Freeway No. 3 Landslide in
Northern Taiwan (under review)
The Performance of Horizontal Drain
as a Landslide Mitigation Strategy
Putu Tantri K. Sari
Abstract Landslides often occur during or shortly after heavy rains in several loca-
tions. One of the preventive efforts proposed is the removal of groundwater levels
from the subsurface. A method proven to be effective in achieving this is the appli-
cation of horizontal drain to increase the safety factor on a slope against landslides,
especially during the rainy season. This research was, therefore, conducted to summa-
rize the results of previous studies both on a field scale and numerical analysis to
determine the effectiveness of horizontal drain on landslides. The results showed the
horizontal drain has the ability to effectively lower the groundwater level at different
degrees depending on soil parameters, rain conditions, the topography of the area,
and the dimensions used. Moreover, the horizontal drain was also able to increase
the slope safety factor by 1–4% for 1-m groundwater level decrease and al-most 20%
for 9-m groundwater level decrease.
Keywords Landslide · Horizontal drain · Landslide due to rainfall
1 Introduction
Landslide, which is one of the impacts of the global climate change phenomenon, is
increasing annually in Indonesia as reported by the National Disaster Management
Agency (BNPB). It is also one of the top three disasters experienced between 2012
and 2014. Moreover, most of the landslide disasters recorded in Indonesia, especially
in West Java, were reported by [1] to be caused by factors such as a steep inclination
of the slope for more than 400, change of land usage from forest to farms and paddy
fields, thick weathered rock soil from volcanic breccia, and high rainfall.
Several methods of slope reinforcement have been developed to mitigate the
dangers of landslides and they include geotextile, soil retaining walls, sheet piles,
ground anchors, soil nailing, and others [2, 3]. However, the use of reinforcement
P. T. K. Sari (B)
Civil Engineering Department, Institute Technology of Sepuluh Nopember,
Surabaya 60111, Indonesia
450 P. T. K. Sari
is very uneconomical due to the occurrence of slope instability which triggers land-
slides during the peak-rainy season which lasts for only 1–3 months while stability
is usually experienced in the remaining 9–11 months. Therefore, the control of rain-
water infiltration into the soil is increasingly preferable and one mechanism for this
is the use of a horizontal drain which is considered to be one of the most effective
methods.
According to Tang et.al. [4], a horizontal drain is very effective in lowering the
groundwater table. It was also reported by [5] to have the ability of increasing slope
stability during rainfall, especially for dangerous slopes and embankments. More-
over, [6–8] showed it is the most economical method available in addition to its
effectiveness and the same was also reported by other studies such as [9–16]. A
horizontal drain is a hole made by drilling a manmade embankment or natural slope
covered by a perforated metal or plastic pipe [17]. It is also defined by Barret [18] as
a 3.7 cm or 1.5-inch diameter polyvinyl chloride (PVC) pipe mounted horizontally
or inclined through drilling to form a permanently installed angle. Meanwhile, there
are possible variation in drain dimensions such as the diameter and length of the
installation depending on the topography of the area, climatic and rainy conditions,
and subgrade parameters.
Several studies have been conducted through numerical modeling and field tests
to obtain an effective horizontal drain design formulation. The results from these
studies vary based on several variables and field conditions but almost all of them
have positive findings. The differences in the previous studies make it important to
have a reasonable summary and this is the focus of this research.
The advantages of using horizontal drain for landslide mitigation need to be
summarized especially in Indonesia for construction designers. This is necessary
considering their lack of information on the function, design, and advantages of this
method. Moreover, the horizontal drain is generally installed for special occasions
mostly based on previous experience without understanding the philosophy of its
design. This research, therefore, intends to provide an explanation of the design and
use of the horizontal drain.
The current state of horizontal drain practice through field applications and numer-
ical analysis was summarized using results obtained from several locations with
different conditions to determine the effective dimensions on landslides. Moreover,
the factors of safety with and without horizontal drain installation were also compared
based on the decrease in groundwater table elevation using the results from previous
research. The impact of this method on slope stability based on different soil param-
eters, rainfall, and topographic conditions was also evaluated using information from
previous studies.
2 Horizontal Drain Design Procedure
Martin et al. [13] showed that horizontal drain needs to satisfy certain basic require-
ments to perform successfully. These include the availability of sufficient size to
The Performance of Horizontal Drain as a Landslide … 451
carry the maximum water flow without any disturbance to the surrounding soil or the
development of excessive outflow pressures and absence of significant loss of flow
through infiltration along the drain. It also needs to be strong and sturdy enough to be
easy to install according to the dimensions designed and able to support the borehole
without collapsing. Moreover, the slotted or perforated length of each liner needs to
be formed to prevent soil ingress or equipped with a suitable filter and the drainage
is required to function properly without clogging and with minimum maintenance
in the long run.
Some factors usually considered for effective design of horizontal drain include the
shape, dimensions of the hole, length of the penetration into the ground as well as the
installation spacing and configuration. In addition, the monitoring process, cleaning
and inspection as well as design of the numerical modeling are also important.
Barret [18] and Santi [6] showed the formation process for horizontal drains.
According to Barret, they are flat or low angle holes drilled with temporary or
permanent case and are limited to small-diameter boreholes with a 3.7-cm or 1.5-in
diameter slotted polyvinyl chloride (PVC) casing which is installed for permanent
drains. Moreover, Santi also showed the hole of a horizontal drain is typically cased
with a small diameter perforated polyvinyl chloride pipe. There is also a horizontal
wick-drain model which is considered to be advantageous in resisting clogging,
more affordable, has the ability to deform without rupture, and easy to install even
by unskilled laborers with a minimal investment in equipment [6].
Another important aspect of horizontal drain design is the perforation size. These
perforations or slots allow water in the soil or slope to enter the horizontal drain in
order to be conveyed out of the slope. According to [19], the holes are usually at
every 120° with the width varying between 0.25, 0.5, and 1.3 mm. Wide ones are
generally applied on coarse sand or gravel soils and this allows fine sand and silt or
smaller diameter materials to enter the horizontal drain. It is important to determine
the hole to be used based on the analysis of soil in the installation area. Meanwhile,
additional filtering can be conducted by surrounding the pipe with a fibrous filter
material [20].
The horizontal penetration length and direction of the installation also need to be
analyzed to ensure effectiveness. Several studies have been conducted to obtain the
effective penetration length and the results vary depending on the topographic slope
shape and subgrade parameters. For example, [19] reported that horizontal drains
are most effective when placed in a location that holds water before reaching a zone
of instability. It was also discovered that drains need to extend far enough into the
slope to achieve the desired drawdown throughout the slope. This means they are to
be installed beyond the area experiencing landslide due to the need to remove water
from the landslide zone. Moreover, the determination of the penetration length is
also influenced by the source and direction of groundwater movement in the area,
the character of the strata or carrier layer, presence of open breccias or discontinuities,
interconnected gaps in clay, and any perched water layers [21].
Pohll et al. [15] found the water table elevations to be dropping as the drain length
increased. However, the horizontal drain becomes ineffective as it penetrates much
longer than the predicted landslide zone. It was also discovered that the maximum
452 P. T. K. Sari
reduction in subsurface water was not affected by the changes in drain length beyond
a critical length [22]. Moreover, Royster [17] recommended that drains should not
extend more than 3–5 m past the slip surface, and installations that significantly
exceed this value is considered uneconomical. Lau and Kenney [9] also showed
there is no additional benefit to installing a horizontal drain beyond the limit where
the critical slip surface intersects the top of the slope. Furthermore, Cai et al. [5]
indicated a smaller increase in safety factors as the drains extend beyond a critical
length.
Several researchers have attempted to provide quantitative methods to estimate
drain spacing. For example, Hunt [23] showed it depends on the type of material being
drained with fine-grained soils reported to require 3–8 m while more permeable
materials need up to 8–15 m. Huculak and Brawner [24] also recommended the
installation of horizontal drains on an exploratory basis at a spacing of 9–12 m when
the locations of high-permeability zones are unknown. Some researchers showed the
need to place at least one additional drain between drains of the initial exploratory
set in some cases based on site conditions and to tap the zones producing substantial
amounts of water. Moreover, Cornforth [25] indicated parallel drains are usually
1–3 m apart and drilled normally into the slope in the direction of slide movement.
These findings showed the effective spacing for horizontal drain installation varies
depending on different conditions encountered in the field. The summary of all
these studies is presented by [7] that the initial drains for parallel drains in high-
permeability soils should be spaced at 8–15-m intervals while those in fine-grained
soils should be at 1–8-m intervals.
There are 2 horizontal drain installation configuration models which are fan-
shaped or array and parallel. There is, however, no significant difference in the
maximum amount of water reduced between the two methods and they both have their
respective advantages [22]. According to Mekechuk [26], the fan-shaped method has
a faster installation process because it does not require a lot of time to set up the equip-
ment. In addition, it also has some benefits related to cost and site access through an
easier collection of water from several drains at once for conveyance off the slope
but it requires the installation of sufficient drains to provide a mean distance equiv-
alent to the guidelines provided for parallel horizontal drains. Meanwhile, parallel
configurations provide more confidence in measuring the spacing of drains and in
representing the slope with simplified models [7].
The changes in water conditions in the soil with a horizontal drain can be moni-
tored using a piezometer which is a small-diameter well or tube used to measure the
hydraulic head of groundwater. This instrument is usually applied to estimate the
initial water heights in the slope for stability modeling purposes and also to evaluate
the average drawdown over time to observe the effectiveness of the drainage system
[21]. However, the use of piezometers to observe and determine the effectiveness
of a drainage system requires understanding the possibility of a rapid change in
water levels over short distances. This is a problem in areas with sparse piezometer
coverage [6, 9].
The horizontal drain also requires some maintenance to keep it effective. This
is necessary considering the possibility of clogging by root growth, sediments, or
mineral deposits due to the time factor which needs to be cleaned. Several methods
have been developed to clean these drains. For example, Cornforth [25] showed that
high-pressure water systems are used to clean the slots and scour away sediments or
mineral deposits. Moreover, the use of chemical treatments is sometimes necessary
to remove stubborn mineral deposits. According to Santi et al. [6], horizontal wick
drains tend to resist clogging but once clogged, they cannot be easily cleaned, thereby,
requiring a new installation. The frequency of cleaning depends on the climate, local
geology and vegetation, and several other factors [27]. Cornforth [25] recommended
cleaning and inspection every 3 months in the first year of installation followed by
once in a year and then after 4 years. An exception was, however, made when there
is calcium carbonate blockage, in which case it is necessary to clean and inspect the
drain every 2 years.
Another important factor to assess the effectiveness of horizontal drain is the
method of conducting numerical modeling before the horizontal drain is installed
in the field. Several studies have used different models as observed in the use of
a hydrogeological software known as FEFLOW by [28] to analyze the transient
hydrogeological and coupled hydromechanical modeling. Tacher (2005) also used
the same software to calculate the groundwater pressure field in each landslide area.
Meanwhile, Lin et al. [12] used SEEP/W and SLOPE/W software to determine
the slope stability under rainfall seepage conditions and obtain safety factors with
and without subsurface drainages. In another study, Lin et al. [29] used a 3-D finite
element method program known as Plaxis 3D modeling to analyze groundwater flow
and slope stability.
3 The Increase of Safety Factor of Slope After Horizontal

Drain Installation
Horizontal drain has usually been applied to lower groundwater table and to ensure
slope stability for a long time. It was initially developed for road construction in Cali-
fornia in anticipation of landslides when the California Highway Division introduced
the Hydrauger method in 1939 [17, 30]). A study was subsequently conducted on its
use and the results were published by ([27, 31]). The application of horizontal drain
to ensure slope stability then expanded to several countries such as Japan, England,
Hong Kong, Germany, Czechoslovakia, Yugoslavia, Canada, France, and others as
noted in previous studies published [31, 32].
The use of the horizontal drain for slope stability was observed to be continu-
ously growing with a wide variety of conditions which were researched to determine
its effectiveness. Those studies were generally divided into 3 major groups which
include the analysis of the effectiveness of horizontal drain by observing water level
fluctuations in the field before and after its installation. It is important to note that
piezometer assay was commonly used in this research group. The second group
includes the use of laboratory tests such as the centrifuge device and the results were
454 P. T. K. Sari
later compared with those from the analysis conducted using a computer program or
verification of the landslides in the field. The third group involved the use of math-
ematical modeling with the focus on the rate of movement of water in the porous
media.
These studies were all conducted at several conditions including the presence and
absence of the influence of rain intensity, soil parameters as well as the dimensions
and elevations of the horizontal drain installed. The results of these studies showed
that (i) the presence of horizontal drain was able to reduce the groundwater table
to 3–7 m and, even in some cases, up to 15 m ([9, 11, 14, 27]), (ii) the presence of
horizontal drain was able to increase slope stability by improving the slope safety
factor ([4, 16, 33, 34]).
Nonveiller [33] also showed that horizontal drain increased the safety factor of
slope value by 5–20% while [4] show the increment was up to 21.5%. The same trend
was also observed in [16] and [34] even though the percentage increase was not stated.
Moreover, the safety factor based on the groundwater table drop from previous studies
was determined using 3 soil slope parameters which include the cohesive dominance,
non-cohesive dominance, and cohesive and non-cohesive mixture. It is, however,
important to note that the slope was considered to be a man-made embankment
standing on a soft subgrade. The results presented in Fig. 1 showed the safety factor
increases with a decrease in groundwater level in the three types of soil due to the
installation of the horizontal drain. Meanwhile, the increase in safety factor varies
from 1 to 4% for every 1-m drop in groundwater level while the value was observed
to be up to approximately 20% for 9 m.
Fig. 1 The increase of

safety factor after horizontal
drain installation on slope
4 Conclusion
The summary of the findings from previous studies showed horizontal drain is an
effective method to mitigate landslides, especially during the rainy season. It has
been proven to have the ability of reducing groundwater level in order to increase
the safety factor on a slope. This, however, depends on several factors such as the
topography of the location, soil parameters, and rain conditions and this is the reason
for the difference in the groundwater level recorded in the locations. Moreover, the
horizontal drain was also able to increase the slope safety factor for every 1-m drop
in groundwater level while the value was observed to be up to approximately 20%
for 9 m.
Some other factors affecting the effectiveness of this method include dimensions,
the shape of the horizontal drain material as well as the installation space and length.
These factors influence the ability of the drain to hold water and drain rainwater
out of the slope. Moreover, the cases summarized were observed to have different
dimensions, shapes, and results after the installation but both the field scale and
numerical analysis proved the effectiveness of this method. This means the horizontal
drain is one of the best alternatives to mitigate landslides in Indonesia. It is, however,
important to have very complete data on soil parameters especially those related to
water infiltration rate and topography of the location for the preliminary design of
the drain. In addition, the data on rain conditions at the location under review at a
certain time is also very necessary.
References
1. Qodriyatun SN (2019) Landslide and disaster mitigation, in INFO singkat

2. Sari PTK, Lastiasih Y (2018) A General formulation to describe the empirical prediction of
the critical area of a landslide. JESTEC 13(8):2379–2394
3. Sari PTK, Lastiasih Y (2020) Combined effects of vertical drain and pre-loading to reduce the
number of geotextile reinforcement for road embankment design combined effects of vertical
drain and pre-loading to reduce the number of geotextile reinforcement for road embankment
design. J Phys Conf Ser 1569:1–13
4. Tang L, Tang X, Liu W (2011) A case study of effect of horizontal drains on rainfall-induced
landslide. Adv Mater Res 253:1834–1837
5. Cai F, Ugai K, Wakai A, Li Q (1998) By three-dimensional finite element analysis effects of
horizontal drains on slope stability under rainfall by three-dimensional ® nite element analysis.
Comput Geotech 23(December):1998
6. Santi PM, Elifrits CD, Liljegren JA (2003) Design and installation of horizontal wick drains
for landslide stabilization. 2(01)
7. Cook DI, Santi PM, Higgins JD (2008) Horizontal landslide drain design: state of the art and
suggested improvements. Geol Soiety Am XIV(4):241–250
8. Rahardjo H (2012) Performance of horizontal drains in residual soil slopes. SOILS Found
51(3):437–447
9. Lau KC, Kenney TC (1983) Horizontal drains to stabilize clay slopes. Can Geotech J 21:241–
249
10. Kopecký M, Ondrášik M, Antolová D (2013) Horizontal drains as effective measure for
horizontal drains as effective measure. Stud Geotehnia Meh XXXV(March):129–141
456 P. T. K. Sari
11. Ahmed J, Ghazali MA, Mukhlisin M, Mukhlisin M (2012) Effectiveness of horizontal drains
in improving slope stability : a case study of landslide event in Putra Jaya Precinct 9, Malaysia
effectiveness of horizontal drains in improving slope stability : A case study of landslide event
in drains Putra in Jaya, no. January, 2012
12. Lin D, Hung S, Ku C, Chan H (2016) Evaluating the efficiency of subsurface drainages for
Li-Shan landslide in Taiwan. Nat Hazards Earth Syst Sci Discuss 1(January):1–22
13. Martin R (2013) Use of horizontal drains for slope stabilisation in Hong Kong use of horizontal
drains for slope stabilisation in Hong Kong. HKIE Trans 3(April):37–41
14. Pfeiffer TJ (2014) Practical considerations for the design and construction of landslide
mitigation using horizontal drains. In: 65th highway geology symposium
15. Pohll GM (2013) Design guidelines for horizontal drains used for slope stabilization, no. March
2013
16. Amatya SC, Mori M (2018) Horizontal drilling drainage as a preventive measure for water
induced landslide risk reduction : a case study from Sindhuli Road, Section I, Nepal. J Nepal
Geol Soc 55:109–122
17. Royster DL (1972) Horizontal drains and horizontal drilling: an overview. Transp Res Rec
16–20
18. Barrett RK (1980) Use of horizontal drains : Case histories from the colorado division of
highways. Transp Res Rec 783(5)
19. Diana C, Paul S, Higgins J (2012) Prediction of piezometric surfaces and drain spacing for
horizontal drain design. Landslides, no. June 2011
20. Choi (1983) Seepage around horizontal drains in hill slopes. J Hydraul Eng ASCE
109(10):1363–1368
21. Hoek E, Bray JW (1981) Rock slope engineering. In: Institution of Mining and Metallurgy,
3rd edn, p 23
22. Nakamura (1988) Landslide control works by horizontal drainage works. In: Landslides:
proceedings of the 5th symposium, vol 2. Lausanne, Switzerland, pp 893–896
23. Hunt (2005) Geotechnical engineering investigation handbook, 2nd edn. Taylor and Francis
Group
24. Huculak NA, Brawner CO (1961) The use of horizontal drains in landslide stabilization. In:
Proceedings of the 42nd annual Canadian good roads conference, pp 383–400
25. Cornforth (2005) Landslides in practice: investigation, analysis, and remedial/Preventative
options in soils
26. Mekechuk J (1992) Experience with horizontal drains as a means to improve slope stability.
In: Proceedings of the 45th Canadian geotechnical conference. Toronto, Ontario, pp 1–7
27. Smith TW, Stafford GV (1957) Horizontal drains on California highway. Soil Mech Found
Eng 83(SM-3):1300–1 to 1300–2
28. Matti B, Tacher L, Commend S (2012) Modelling the efficiency of a drainage gallery
work for a large landslide with respect to hydrogeological heterogeneity. Can Geotech J
49(November):968–985
29. Lin D, Chang K, Ku C (2020) Applied sciences three-dimensional numerical investigation on
the efficiency of subsurface drainage for large-scale landslides numerical models. Appl Sci
10:3346
30. Stanton TE (1948) California experience in stabilizing earth slopes through the installation of
horizontal drains by the Hydrauger method. In: 2nd international conference on soil mechanics
and foundation engineering. Rotterdam, the Netherlands, pp 256–260
31. Root (1958) Prevention of landslides. In Eckel EB (ed) special report 29: landslides and
engineering practice
32. Hutchinson JN (1977) Assessment of the effectiveness of corrective measures in relation to
Geolo_Gical conditions and types of slope Movem
33. Nonveiller (1981) International society for soil mechanics and 1981
34. Salmasi F, Dalir AH, Sarkarabad RN (2019) Investigation of the performance of horizontal
drains in increasing slope stability in intense rainfall conditions by numerical simulation.
Amirkabir J Civ Eng Amirkabir 51(3):2018–2020
The Analysis of Impact and Mitigation
of Landslides Using Analytical Hierarchy
Process (AHP) Method
A. Andriani, B. M. Adji, and S. Ramadhani
Abstract Landslides disaster is the third biggest disaster in the last ten years
in Indonesia. The impact of landslides has resulted in many materials and non-
material losses, such as casualties, physical impacts, environmental health, and socio-
economy. Mitigation efforts are needed to reduce and even eliminate losses due to
land-slides. In this study, the Analytical Hierarchy Process (AHP) method is used
to assess the impact and mitigation of landslides so that the right decisions can be
taken to deal with this problem. Questionnaires were conducted on 20 experts in the
fields of disaster, geotechnical, geology and environment through direct distribution
of questionnaires and google forms. The Likert scale is used as an additional method
to obtain weights from the indicators contained in the last level of landslide mitiga-
tion. The results show that using the AHP method can identify the direct impacts
felt by the community and the most dominant impacts occur due to landslides so
that the most suitable mitigation efforts can be chosen based on the environmental
conditions experiencing the disaster. Mitigation efforts before the disaster is one of
the steps that must be taken to reduce the impact caused by landslides, structural and
non-structural mitigation is carried out to mitigate the landslide hazard.
Keywords Landslide · Mitigation · AHP
1 Introduction
Indonesia was ranked first as the country with the most disaster victims in the world,
and was ranked second after Haiti as the country with the most victims for the
last 19 years [1]. This fact cannot be denied due to the geographical and geological
A. Andriani (B) · B. M. Adji · S. Ramadhani

Civil Engineering Departement, Andalas University, Padang 25163, Indonesia
B. M. Adji
458 A. Andriani et al.
conditions of Indonesia which are at the junction of the three major plates of Eurasia,
Indoaustralia and the Pacific.
Indonesia’s natural conditions such as topography, geology, and climatology are
the dominant factors causing disasters, the movement of land masses. Land mass
movement or landslide is a disaster that often occurs in hilly areas, valleys and
volcanoes [2–5]. Landslides are the movement of slope-forming material in the form
of rocks, scrap material or mixed material that moves down or out of the slope. The
movement of the slope-forming material that moves out of the slope is due to the
absorption of water into the soil until it penetrates the impermeable layer which acts
as a sliding area, causing an increase in soil weight. Landslides will have a direct or
indirect impact on the ecological condition of an area [6–8].
In the last ten years, landslides have occupied the third largest disaster with a total
of 5022 events after a whirlwind disaster with a total of 6467 events and floods with
a total of 7283 events [9]. According to calculations by the disaster management
agency in Indonesia in 2019 from January to August, landslides are the biggest
disasters after a tornado. The number of landslide disasters was 548 and the most
areas that experienced landslides were Central Java with 242 incidents. Landslides
are disasters that often occur in Indonesia. Especially in hilly areas and rivers that
have steep cliff relief. One of the causes of frequent landslides in Indonesia is the
change in land use. Land use change is caused by limited area. West Sumatra is one
of the provinces that often experiences landslides disasters. The results of monitoring
carried out by the Geological Agency, the Center for Volcanology and Geological
Disaster Mitigation, in 2011 recorded 21 times the disasters of land movements in
West Sumatra that are spread in several districts.
Landslide disasters have negative impacts, physically, socio-economically and
on environmental health. In 2019, landslides in Indonesia caused 145 people to die
and disappear, 125 injured victims, 5293 affected/displaced people, 2,298 damaged
houses and 76 damaged facilities and infrastructure and displaced victims during
landslides. in 2019, was ranked first with a score of 9473. Based on the impact
of landslides, disaster mitigation efforts are needed to reduce these impacts. UU
no. 24 (2007) concerning disaster management, mitigation is an effort or action
that can reduce risks due to disasters either through physical efforts or awareness
and enhancement of community capacity when facing disasters that occur. The steps
taken to mitigate landslides can be grouped into three parts, namely mitigation before
a landslide occurs, mitigation when a landslide occurs and mitigation after a landslide
occurs [10–13].
To determine the priority of this mitigation action requires a lot of considerations,
one of which is the most frequent impacts caused by the landslide disaster. The
scientific approach can be used as a way to get the dominant impact of landslides
and the right mitigation in reducing the impact of landslides. One method that can
be used is the Analytic Hierarchy Process (AHP) method. The AHP method is a
decision support model that uses a hierarchical arrangement and assessments derived
from human perception. Opinions or perceptions used in this method are opinions
that come from experts or people who understand the problem. AHP method is
well known and widely used for decision making. This is because the AHP method
The Analysis of Impact and Mitigation of Landslides … 459
has advantages, namely: make broad and unstructured problems into a flexible and
easy-to-understand model through deductive integration. To get optimal results, this
method requires people who are experts in their fields so that the assessment is
subjective, so sometimes there are consistent answers or responses from respondents
that are not always logical in the sense that they are in accordance with the existing
problems. On the problem of landslides, many researchers use the AHP method. The
use of the AHP method for mapping landslide-prone potential has been carried out
by several researchers [14–19]. Besides that, disaster impact analysis can also be
carried out using the AHP method [20–24].
Based on the results of previous research, it shows that AHP can be used as an
alternative that can be used to determine appropriate mitigation efforts so that the
impact of disasters can be reduced or eliminated.
AHP model uses human perception which is considered by experts as its main input.
The criteria for “expert” here refers to people who understand the problem posed,
feel the consequences of a problem or have an interest in the problem.
The main step taken in this study is a literature review, then a questionnaire is
made to the impact and mitigation of landslides. The questionnaires were distributed
to experts all over Indonesia who have experience in the fields of landslides, civil
engineering and the environment. The experts used in the study were 20 people.
From the results of the AHP analysis, only 14 experts met the criteria.
The following are the steps for the assessment using the AHP method:
2.1 Creating a Hierarchical Structure Begins with the Main

Objective
The hierarchical structure begins with a general objective and then continues with
the criteria criteria and alternative alternatives that will be sorted later. The network
is made on the basis of a comprehensive structure related to the problems being
developed.
2.2 Calculation Using the AHP Method
Assessment Matrix
Qualitative decisions are made by providing a perception of comparison that is scaled
in pairs (pairwise comparison) and becomes a quantitative assessment by assessing
Table 1 Pairwise comparison matrix

A1 A2 .. An
A1 a11 a12 .. a1n
A2 a21 a22 .. a2n
.. .. .. .. .
Am am1 am2 .. amn
Table 2 Random index value (RI)

No. 1 2 3 4 5 6 7 8 9 10
RI 0.00 0.00 0.58 0.90 1.12 1.24 1.32 1.41 1.45 1.48
the importance of an element compared to other elements. After the assessment is

carried out using the appropriate scale according to Saaty, an assessment matrix or
pairwise comparison matrix is made. Table 1 shows the paired matrix.
The values a11 , a22 ,….. amn are the comparison values for the elements of row A1
to column A1 which state the relationship of how far the importance of row A1 and
column A1 is, or how far row A1 dominates over column A1.
Normalizing Data
Which normalizes this data is done by dividing the value of each element in the
paired matrix by the total value of each column.
Calculating Priority Vector
In determining the priority value, a comparison of the total rows in the matrix is
needed, where the total rows are obtained from the sum of the total row values of the
calculated matrix that have been compared with the total value of the matrix in each
column.
Testing of Hierarchy Consistency
The difference between the AHP method and other decision making methods is
the requirement of absolute consistency. This consistency is also determined by
the maximum eigen value. If the CI value is 0, the matrix is consistent (pairwise
comparison). The inconsistency limit has been set by Saaty using the consistency
ratio (CR) and the random index value (RI), following the formula below.
CR = CI/RI (1)
where as
CI: Deviation ratio (deviation) consistency = (λmax – n)/(n−1)
λmax : The largest eigenvalue of the n order matrix
n: Matrix order
RI: Random index (Table 2)

If the pairwise comparison matrix has a CR greater than 0.100, it can be said that
the expert makes inconsistent decisions so that the judgment cannot be used.
Combining Opinions from Experts with Geomean
The geometric mean is the average obtained by multiplying all consistent data within a
sample group, then rooted to the power of that sample size. To calculate the geometric
mean the formula below is used.

GM = n (X 1)(X 2) . . . . . . (X n) (2)
where as
GM: Geometric mean
X1 : Expert-1
Xn : Expert-n.
This AHP method certainly has various disadvantages and advantages. The
advantages of the AHP method are:
a. Unity. The model created by the AHP method is an understandable and flexible
model for broad and unstructured problems.
b. Complexity. AHP can solve or solve complex or complex problems.
c. Hierarchy Structuring. AHP performs grouping by including each element or
action into different levels. Each level contains elements that are similar.
d. Measurement. The assessment on the AHP method uses a scale to set priorities
e. Consistency, the assessment used in the AHP method must be consistent so that
the level of consistency is tested for each assessment.
Meanwhile, the weakness of the AHP method is that this model is highly depen-
dent on its main input, which is human perception, so that if the judgment is made
wrong, the assessment will be meaningless.
2.3 Likert Scale
The Likert scale is a psychometric scale that has generally been widely used in
questionnaires. In this study, the Likert scale was carried out for validation and
reliability testing.
Validity Test
Validity is a measure that indicates that the measured variable is really the variable
the researcher intends to study. Validation testing is carried out using the formula
below.
n X Y − ( X )(Y )
r= √ (3)
(n(X )2 − X )2 (n(Y )2 − Y )2
Reliability Test
In this study using a single test reliability with a split technique (odd–even). For the
reliability test formula used the formula below.
rcalculted = 2r/(1 + r) (4)
Calculation of Critical Score and Rating Scale

The criterion score is the score used to calculate the rating scale. The formula used
to calculate the criterion score is as follows:
Criterium Scale = Scale value × Number of Experts (5)
In this study, the rating scale used is divided into four classes, namely strongly
disagree (1), disagree (2), agree (3) and strongly agree (4).
Combining the Weights of All Experts
Combining the scores given by the expert on each question in the questionnaire using
the formula below:

P(%) = f/z × 100 (6)
where as,
f: The sum of the scores of all experts on each question
P: Percentage of answers
Z: Total ideal score/highest criterion scale.
After obtaining the combined values of all the experts, the weight of each question
is obtained using the formula below:
Weight(%) = (P(%))/(Sum of P(%)) × 100 (7)

3.1 Assessment of Landslides Disaster Impact
The AHP method divides the level of impact into 3 parts where at level 2 the impact
of landslides is classified into four parts, namely the impact of casualties, physical
losses, environmental health and socio-economic. Furthermore, at the third level,
each indicator is compared at level two. Table 3 shows the results of the expert
assessment using AHP.
Based on the results of data processing using the AHP method, it was found that
the most dominant impact occurred was the impact of casualties with a weight of
0.358, then the impact on environmental health was 0.272, the physical impact was
0.213 and the lowest weight was the socio-economic impact of 0.157. The weighting
value obtained is in accordance with BNPB data from 2015–2020, which shows that
the casualty impact is the most caused by landslides. In the data released by BNPB,
the highest impact on casualties was in 2017, in which 163 people died, 185 were
injured and 59,640 affected victims.
Table 3 Landslide impact assessment with AHP method

No. Impact of Landslides Weight Score Total
1 Causalities impact 0.358
Affected victims/relocation 42.84 15.35
Injured victims 25.22 9.04
The victim of died 31.93 11.44
Total 100.00
2 Physical impact 0.213
The transportation track was cut off 32.98 7.01
Lost of the settlement 34.15 7.26
The infrastructure damage 32.88 6.99
Total 100.00
3 Environmental health impact 0.272
Deteriorating environmental sanitation 31.96 8.71
Disruption of ecosystem balance 32.23 8.78
Loss of vegetation 35.81 9.76
4 Socio-economic impact 0.157
Incidence of crime 43.13 6.75
Disruption of economic stability 56.87 8.90
Total 1.00 100.00 100.00
3.2 Assessment of Landslides Disaster Mitigation
The magnitude of the impact caused by landslides needs to be anticipated, with miti-
gation efforts, the impact that occurs due to landslides can be reduced or eliminated.
Based on the time of occurrence, mitigation efforts can be divided into mitigation
before landslides, during landslides and after landslides. The method used can be
done structurally or structurally. Based on the results of the expert assessment, the
mitigation efforts that need to be done are shown in Table 4.
From Table 4, it can be seen that disaster mitigation measures can be carried
out before having the greatest weight value compared to mitigation efforts during
and after a disaster. It is hoped that with mitigation efforts before a landslide disaster
occurs, the resulting impact can be reduced and can minimize casualties, so preventive
Table 4 Landslide mitigation assessment with AHP method

No. Mitigations Weight 1 Weight 2 Weight 3 Total
1 Mitigation before landslides 0.57
Structural mitigation 0.57
Making maps of areas prone to landslides and 34.53 11.29
evacuation routes
Establish an early warning system and install 34.53 11.29
ground motion monitoring devices
Construction of landslide retaining structures 20.94 10.11
Non structural mitigation 0.43
The outreach to the community about the 20.47 5.07
impact and mitigation of landslides
The holding of landslide simulations to the 17.67 4.38
affected communities
Making posters about landslide disaster 19.07 4.73
information
Vegetation arrangement on slopes 20.93 5.19
The Arrangement of RTRW (Regional Spatial 21.86 5.42
Planning) in areas with potential for landslides
Total 1.00
2 Mitigation during Landslides 0.22
Making a special distribution track for aid 50.00 5.35
Construction of a proper shelter 50.00 5.35
Coordination between the parties involved 48.00 5.55
A responsive team of evaluators 52.00 6.02
Total 1.00
(continued)
Table 4 (continued)
No. Mitigations Weight 1 Weight 2 Weight 3 Total
3 Mitigation after landslides 0.21
Improvement of slope physical 33.33 3.88
Replanting vegetation 34.78 4.05
Improvement of residential areas and 31.88 3.71
infrastructure buildings
Rehabilitation of communities affected by 34.59 2.97
landslides
Making of settlements in accordance with 32.33 2.78
applicable standards
Enforce a RTRW (regional spatial plan) policy 33.08 2.84
that is suitable for landslide-prone areas
Total 1.00 1.00 100.00
measures must be prepared as early as possible, preventive efforts are made to avoid
a major impact on areas where disaster is indicated. Natural disasters can be local
(micro) but can also be widespread, causing macro impacts. One way that can be
done to mitigate disasters is through spatial planning, so as to reduce physical and
economic losses due to disasters. Mitigation activities include risk assessment and
determination of potential disaster.
4 Conclusions
Based on the results of the assessment using AHP, the impact of casual-
ties/displacement is the impact with the highest weight of 15.35. For this reason,
pre-disaster mitigation efforts need to be carried out to reduce the impact, such as
by making maps of landslide disaster locations and easy-to-reach evacuation routes
and secondly, creating a landslide early warning system and installing a rainfall
monitoring device.
Acknowledgements Publication of this article is supported by Engineering Faculty of Andalas

University, fiscal year 2022 and we would like to thank all those who have helped this research,
especially all the experts who were involved in filling out the questionnaire.
References
1. BNPB (2018) Buku IRBI 2018 Diterbitkan Oleh : Badan Nasional Penanggulangan Bencana
2. Suriadi AB, Arsjad M, Hartini S (2014) Analisis Potensi Risiko Tanah Longsor di Kabupaten
Ciamis dan Kota Banjar, Jawa Barat. Maj Ilm Globe 16:165–172
3. Nathania B, Muira F (2017) Remote sensing and GIS approach for landslide susceptibility
mapping: a case study in Hofu City, Yamaguchi, Japan. Int J Environ Geosci 1(1):28–35
4. Wieczorek G (1996) Landslides: investigation and mitigation. Chapter 4—Landslide triggering
mechanisms. Transp Res Board Spec Rep 247
5. Young R, Norby L, Wieczorek GF, Snyder JB (2019) Monitoring slope movements. Geol Monit
11:245–271
6. Polemio OP (2000) Landslides—Rainfall as a landslide triggering factor : an over view of
recent international research. Landslides Res Theor Pract Thomas Telford, London 3
7. Berglar K, Gaedicke C, Lutz R, Franke D, Djajadihardja YS (2008) Neogene subsidence and
stratigraphy of the Simeulue forearc basin, Northwest Sumatra. Marine Geol 253:1–13
8. Moradi S, Rezaei M (2014) A GIS-based comparative study of the analytic hierarchy process,
bivariate statistics and frequency ratio methods for landslide susceptibility mapping in part of
the Tehran metropolis, Iran. Geopersia 4(1):45–61
9. Yanuarto T, Pinuji S, Utomo AC (2019) Buku saku Tanggap Tangkas Tanguh Menghadapi
Bencana 48
10. Tokmechi Z (2016) Landslide mitigation and its risk controlling landslide mitigation and its
risk controlling. May 2016
11. Warburton J (2015) Peat landslides. Elsevier Inc.
12. Sultana N, Tan S (2021) Landslide mitigation strategies in southeast Bangladesh: lessons
learned from the institutional responses. Int J Disaster Risk Reduct 62:1–18
13. Iinran BAM (2012) Malino- Manipi landslide, south Sulawesi Indonesia, vol. 2012
14. Oktorie O (2017) A study of landslide areas mitigation and adaptation in Palupuah Subdistrict,
agam regency, West Sumatra Province, Indonesia. Geogr Geogr Educ 1(1):43–49
15. El Bchari F, Theilen-willige B, Malek HA (2019) Landslide hazard zonation assessment using
GIS analysis at the coastal area of Safi (Morocco), no. July, pp 15–20
16. Roy J, Saha S (2019) Landslide susceptibility mapping using knowledge driven statistical
models in Darjeeling district, West Bengal, India. Geoenvironmental Disasters 6(11):1–18
17. Moradi M, Bazyar MH, Mohammadi Z (2016) GIS-based landslide susceptibility mapping by
AHP method , a case study, Dena GIS-based landslide susceptibility mapping by AHP method,
a case study, Dena City, Iran. no. Dec 2016
18. El Jazouli A, Barakat A (2019) GIS-multicriteria evaluation using AHP for landslide
susceptibility mapping in Oum Er Rbia high basin (Morocco). 7
19. Mondal S, Maiti R (2013) Integrating the analytical hierarchy process (AHP) and the frequency
ratio (FR) model in landslide susceptibility mapping of Shiv-khola watershed, Darjeeling
Himalaya. 4(4):200–212
20. Eddy AI, Dwi D, Putranto A, Affandi AK (2018) Evaluation of land subsidence impact at
lowland areas using analitycal hierarchy process (AHP) Method. 7:1–7
21. Munthafa AE, Mubarok H, Teknik J, Universitas I (2017) Penerapan metode analytical hier-
archy process dalam sistem kata Kunci : analytical hierarchy process, consistency index, Maha-
siswa Berprestasi. keywords : analytical hierarchy process, consistency index, achievement
student b. Kelebihan dan Kelemahan Metode AHP. 3(2):192–201
22. Whitaker R, Foundation CD (2017) The analytic hierarchy process—what it is and how it is
used. 0255(October)
23. Szabó S, Gácsi Z, Balázs B (2016) Specific features of NDVI, NDWI and MNDWI as reflected
in land cover categories. Landsc Environ 10(3–4):194–202
24. Studi P, Informatika T, Tinggi S, Ilmu M, Nusa K, Jakarta M (2015) Multi criteria analysis
Menentukan point weight XI(1):11–19
The Stability of a Slope on Soft Soil Using
the Hardening Soil Model
Yerry Kahaditu Firmansyah and Maharani Putri Dewanty
Abstract The development of reservoir construction does not always stand on with
good carrying capacity soil, on the part of the existing soil layer with a soft to very
soft constancy. About 10% or more of the total area in various regions in Indonesia
or more precisely about 20 million hectares are soft-land which is composed of soft
soil and moss soil. If we look at the problem of soft clay soil, which is very influential
in the construction of the construction and its success, soil repair must be done first;
therefore, the construction is not dented already the predetermined and intentional
age. There is a soil repair solution that is minimal in cost but very good and effective
for maintaining the stability of the landfill, namely by using a soil repair system with a
retaining wall. The Finite Element Model is a method of research to solve a problem in
a study or field of mathematics and physics. This method presents several parameter
estimates for an amount of distinct scheme nodal constructions not known. The
Hardening Soil Model Standard, or what can be abbreviated as (HS) is a continuation
of the advanced soil behavior model. But the value or number of soil stiffness can
be described, analyzed, and explained in more detail and accurately using stiffness.
There is three dissimilar effort arduousness, called triaxial load stiffness (E50 ), triaxial
loading stiffness (Eur ), and one-way load stiffness (Eoed ). In a convincing case, the
Hardening Soil Model or (HS) method is proven to be more truthful and closer to
field conditions. It can be seen and has been proven from numerous previous studies
that show the Hardening Soil Model (HS) method is very distinctive and accurate
by the outcomes of field testing. The charge of the safety factor or (SF) obtained
after the investigation of slope constancy on soft soil using hardening soil modeling
analysis is 1.59.
Keywords Reservoir construction · Soil improvement · Hardening soil model
Y. K. Firmansyah (B) · M. P. Dewanty (B)

Universitas Pembangunan “Veteran” Jawa Timur, Raya Rungkut Madya, Surabaya,
East Java, Indonesia
468 Y. K. Firmansyah and M. P. Dewanty
1 Introduction
Temporary water or shelter is often referred to with the reservoir construction should
be built as sturdy as possible so that it can function optimally [1]. most of the soft
soils found in Indonesia are probably composed of Holocene clay deposits. This
Holocene material is obtained from volcanic eruptions in which there is a potentially
very significant volcanic ash content [2]. The nature of soft soil is fundamentally
different from other jobs soils found on coastal plains or in inland areas, due to the
leaching of freshwater [3]. And it can be identified that the existence of derivative
volcanic soil has a very significant effect on the properties of the soil. In some cases,
the deposits seen from a geological perspective are very young [4], where there is
a relationship and a relationship between the rate of deposition and the degree of
consolidation, which can be seen in Fig. 1 in the following. In terms of degrees of
consolidation, the average for deltaic clays in the Southeast Asia region can range
between and begin 20 to 100% reliant on where the location is located. Furthermore,
for sediments in the Indonesian region, for example, the grade of the alliance is
prophesied (on Fig.1) around 70 to 80%. It indicates that the soft clay in the area
around the coast or the sea is experiencing very little consolidation.
Fig. 1 The relationship between sedimentation rate and the consolidation degree [4]
The Stability of a Slope on Soft Soil Using the Hardening Soil Model 469
Among a lot of obstacles, there is one of the obstacles in an embankment activity

on soft soil is the number of landslides. Several other problems are frequently encoun-
tered when constructing roads in soft soils located around swamps, especially if there
is no reinforcing material available for the road [5]. An embankment material sinks
frequently into the softer soil because it is not sturdy enough to hold the ridge soil
[6]. In general and basically, the general purpose of soil improvement is to surge the
comportment volume and shave strength of soil, upsurge the modulus of soil, reduce
the compressibility of soil, control the stability of the soil (reduction and inflamma-
tion), improve the quality of construction resources. Material and helps to diminish
the effects of damage to work processes in a zone and its surroundings. A sheet pile
is a structural tool where it is designed to be able to withstand horizontal forces when
embedded in the ground. The sheet pile can usually also be used as a holding system.
The pile of sheets is made of different materials, which play an important role in
their use and application [7].
The above calculations that have been analyzed will then be carried out by an
approach called the Hardening Soil or (HS) model method. The admiration of this
model is based on very unpretentious restriction documentation of a predefined test
and empirical formula. The HS Model is realized in most profitable FE ciphers
which purpose has been designed to analyze geotechnical problems [8]. Developed
by Schanz, Vermeer, and Bonnier, this rudimentary form of the Hardening Soil (HS)
elastoplastic model is realized into the Plaxis FE cipher [10]. The key basis of the
HS model can be further described and enlightened by illustrating the outlines of
the resulting surface on the Roscoe p-q stress invariant plane, which can be seen in
Fig. 2. The criteria of Mohr–Coulomb shear strength will later limit the level of shear
stress. On the other hand, the hardening of the volumetric deviator and the pre-failure
is controlled by the two yield surfaces. The HS model is entirely isotropic in terms of
elastic range and terms of elastoplastic. Therefore, inherent anisotropy is absent or
absent, and the available stress contradicts experimental evidence mainly or focused,
e.g. in fine-grained deposits [11, 12].
Fig. 2 Basic HS model in

terms of surface yield and
elastic range [8]
Fig. 3 Sepat reservoir location, Surabaya, East Java
In the previous HS version, the relationship of stress to hyperelastic strain was

functional and applied in the elastic range over which Young’s modulus was depen-
dent on the stress. Eur (Erefur , σ) through a law called Poisson’s law of power and
constant ratio Vur . In an HSS version, increasing the small strain stiffness is based on
and depends on the stress and the accumulated shear strain, whereas in the hypoelastic
relationship, the actual reference shear modulus Greft changes between one value and
another Gref
ur and G0 .
ref
2 Project Information in General
2.1 Description on a Experimental Test Site
The site that the experiment held is located at Sepat reservoir, Surabaya, East Java.
The Sepat reservoir has a water holding capacity of 13,000 m3 . Furthermore, the
alternative reinforcement in the talud area of the reservoir is used corrugated concrete
sheet pile (CCSP) type W400A. Figure 3 shows the Sepat reservoir location.
2.2 Method of Construction
The CCSP installation method uses the jack-in method. This method is considered
environmentally friendly, meaning that it does not cause damage to surrounding
buildings. This material is in installed on the inside of the reservoir and surrounds
the area with type of material used is W400A. Figure 4 shows the CCSP material
properties.
Fig. 4 CCSP material properties
2.3 Soil Field Investigation
Results of soil investigations that have been carried out show that the condition of
the soil layer has a very soft to soft consistency at a depth of 0 m to a depth of 5 m.
The description on layer 1 is Inorganic clay with high plasticity, brown color, and
layer 2 is Inorganic clay with high plasticity, grey color. Figure 5 shows the soil
stratification.
Fig. 5 Soil stratification

Fig. 6 An example of overshooting which is an error in modeling hysterical behavior during cyclic
cutting [15]
3 Approaches for Analysis
3.1 Review of the HS Base Model
The HS construction of the extension of the small strain stiffness was first proposed by
Benz and his colleagues [13, 14], whereupon the so-called original model was called
the Hardening Soil Small or (HSS). This new construction presents and provides
an unusual but specific algorithm to govern the present hypoelastic rigorousness in
the flexible assortment of the basic HS model. There are further restrictions that are
gotten on or after a test that is carried out by standard then extends or empirical
formulas. This is depicted in the systematic schematic which can be seen in Fig. 6.
Related to the problem of overshooting in the HSS model, it will be reviewed in a
critical and detailed way [15]. Subsequent minor load reversals are accomplished by
reloading or rebuilding, which can cause unexpected irregularities.
3.1.1 Mechanism of Deviatoric Plastic
The formula of the yield surface controlling the deviatoric hardening mechanism [9]:
qa σ1− σ3 σ1− σ3
f1 = −2 − γ P = 0 for q < qf , (1)
E 50 qa − σ1− σ3 E ur
3.1.2 Mechanism of Volumetric Plastic
The equation for the surface of the smooth cover:
q2
f2 = + p 2 − p20 = 0, (2)
M 2 r 2 (θ )
3.1.3 Barotropy of Rigidity
In the HS model basic [9], loading and unloading Eur and secant E50 The modulus
of stiffness and stress depend on the following force rules and the corresponding
empirical barotropy function fσ :
E ur (σ ) = E ur
ref
f σ (σ ), (3)
ref
E 50 (σ ) = E 50 f σ (σ ), (4)
m
σ3 + c cot ∅
f σ (σ ), = . (5)
Pr e f + c cot ∅
4 Results and Disscussions
The safety factor (SF) is generally used to measure stability on a slope in geotechnical
engineering. In this study, finite element analysis was used to obtain deformation,
bending moment at CCSP, and safety factors that occurred on slopes. Slope geometry
data will be modeled according to field data. Figure 7 shows the input geometry
model.
Fig. 7 Input geometry model

Fig. 8 Distribution of stress that occurs in the slope area
The use of CCSP material as deep as 14 m looks safe because the distribution
of soil stress does not exceed the depth of the material (This can be seen in Fig. 8).
The SF value obtained depends on the outcomes of analysis modeling using the
finite element is 1.59. This means that the use of CCSP as an amplifier for the Sepat
reservoir is safe. Based on the material strength control, the type of material W400A
has obtained a bending moment value of 10.4 t.m while the permitted capacity based
on the material brochure for type W400A is 20.1 t.m, meaning that the material is
still safe to use. Furthermore, the permit deflection is 0.5% x the elevation difference
is 25 mm, while the horizontal displacement that occurs is 13.06 mm, meaning that
the deflection is still a safe limit. Figure 9 shows the force on CCSP material.
Fig. 9 Force distribution for a. Bending moment occured is 10.4 t.m and b. Deformation occured
is 13.06 mm
5 Inference
This case focuses more on the study of an analytical approach using the Hardening
Soil (HS) model. Furthermore, the conclusions obtained are:
a. From the talud stability modeling results in the Sepat reservoir, it is known that
the critical condition when the reservoir is dry or without water filling.
b. Talud in the Sepat reservoir has the potential for landslides to occur if there
is no reinforcement in the talud area. In order to avoid landslides, corrugated
concrete sheet pile (CCSP) reinforcement type W400A with a total length of
14 m is used and installed on the inside of the reservoir.
c. The safety factor (SF) value of the use of corrugated concrete sheet pile (CCSP)
type W400A with a total length of 14 m is 1.59. This is still within safe limits
and in accordance with the recommendation of SNI 8460: 2017 “Geotechnical
Design Requirements”
d. The used of the CCSP W400A material as reinforcement for the Sepat reservoir
has taken into account several aspects, including: When the reservoir is full
of water, the reservoir is empty, the load that may occur in the reservoir area,
digging into the ground, material for permit capacity, and deflection limits for
the use of these materials.
Acknowledgements We would like to thank UPN “Veteran” Jatim for backing and investigation
funding. Next, the data for this paper is provided by Citraland’s Developer Properti and Mr. Bangun.
References
1. Irmawan M, Firmansyah YK, Refani AN (2016) Modification of reinforced concrete retaining

wall at Pilangbango reservoir construction project madiun, East Java. ICCER
2. Barry AJ, Rachlan A (2001) Embankment on soft soils in North Java. Embankments on
soft soils. In: Proceedings of the preconference volume, international conference on in situ
measurement of soil properties and case histories, Bali, Indonesia
3. Younger JS (1998) Ground improvement options for highways in Indonesia. In: Proceedings
of the short course and seminar on ground improvement, application to Indonesia soft soils.
Universitas Kristen Indonesia and Asian Institute of Technology in Cooperation with Himpunan
Ahli Tanah Indonesia, Jakarta, pp S3/i–S3/32, 23–26 Aug 1988
4. Cox JB (1970) The distribution and formation of recent sediments in South East Asia. In:
Proceedings of the second South East Asia conference on soil engineering, 1970, pp 30–47
5. Muhammad A, Indrasurya BM, Noor EM, Yulian FA (2020) Road embankment full- scale
investigation on soft soil with geotextile stabilization. Int J GEOMATE 19(71):145–152
6. Hardiyatmo HC (2013) Geosynthetic for highway engineering-planning and application
7. Eskandari L, Kalantari B (2016) Basic types of sheet pile walls and their application in the
construction industry—a Review. EJGE 16
8. Marcin C, Andrzej T (2020) Refinement of the hardening soil model within the small strain
range. Acta Geotech. https://doi.org/10.1007/s11440-020-00945-5
9. Schanz T, Vermeer PA, Bonnier PG (1999) The hardening soil model: formulation and verifica-
tion. In: Beyond 2000 in computational geotechnics—10 years of plaxis. Rotterdam, Balkema,
pp 1–16
10. Brinkgreve RBJ, Kumarswamy, S, Swolfs WM, Zampich L, Ragi Manoj N (2019) Plaxis
finite element code for soil and rock analyses. Plaxis BV , Bentley Systems, Incorporated,
Philadelphia
11. Lings ML, Pennington DS, Nash DFT (2000) Anisotropic stiffness parameters and their
measurement in a stiff natural clay. Géotechnique 50(2):109–125, 27
12. Ması̌n D, Rott J (2014)Small strain stiffness anisotropy of natural sedimentary clays: review
and a modelActa Geotech9(2):299–312
13. Benz T (2007) Small-strain stiffness of soils and its numerical consequences. Ph.D. thesis,
University of Stuttgart
14. Benz T, Vermeer P, Schwab R (2009) A small-strain overlay model. Int J Numer Anal Methods
Geomech 33:25–44
15. Niemunis A, Cudny M (2018) Discussion on “Dynamic soilstructure interaction: a three-
dimensional numerical approach and its application to the Lotung case study”. Poor perfor-
mance of the HSS model. Comput Geotech 98:243–245
Software Performance of Risk-Targeted
Maximum Considered Earthquake
(MCER ) Calculation
Windu Partono
Abstract Website software (WS) for Risk-targeted Maximum Considered Earth-

quake (MCER ) acceleration calculation was announced in 2019. Single user non
website software (stand-alone software/SAS) was also released in June 2020. The
MCER acceleration is divided into three different values MCEG (peak ground accel-
eration), MCER -SS (short period) and MCER -S1 (1-s period) spectral acceleration.
The purpose of the SAS software development is to help civil engineers living in
areas having no internet connection or internet connection problems. Both software
packages should yield the same MCER result when the same building location or
position or coordinate is entered as an input. This paper describes the evaluation
of SAS in calculating the MCER compared to the result calculated using the WS.
The evaluation was performed at three different islands in Indonesia; Sumatra, Java
and Sulawesi islands and at 154, 115 and 81 coordinate positions on these islands,
respectively. More than 10% difference in the data was observed when the MCER
acceleration was calculated at the area having (MCEG and MCER -S1 ) greater than
0.6 g (g is gravitational acceleration). The minimum 10% difference in data was also
observed when the MCER -SS was greater than 1.5 g. The largest difference was also
observed when the coordinate positions were located close to seismic sources (fault
source traces).
Keywords MCER · Website software · Stand-alone software
1 Introduction
The Indonesian seismic code for buildings and other structures design was released
in 2019 [1]. The code was implemented as part of the development of Indonesian
seismic hazard maps 2017 [2–4]. Some of the information described in this code was
developed following the American Standard Code for Building Design ASCE/SEI
W. Partono (B)
Civil Engineering Department, Faculty of Engineering, Diponegoro University,
Semarang 50275, Indonesia
480 W. Partono
7-16 [5, 6]. Compared to the previous 2012 Indonesian seismic code, two different
software were announced as a complementary part of the 2019 code. However, only
the website software was developed for the 2012 seismic code. All the software
created during the development of the 2012 and 2019 seismic codes is related to
the calculation of the Risk-targeted Maximum Considered Earthquake (MCER ) and
seismic Response Spectrum Acceleration (RSA) for seismic forces in the design of
buildings and other structures.
The additional 2019 seismic code software was created to meet the requirement
civil engineers or users living in areas having internet connection problems. The
website software (WS) was first released in 2019 and after the improvement of
the software performance, the final version of the website software was released in
February 2021. A new and additional software, also known as stand-alone software
(SAS) or binary software, was also created in 2020. The final stand-alone software
was also released in February 2021. Both WS and SAS were developed using different
programming languages. Although WS and SAS were developed using the same
algorithm and procedures, the restrictions and differences in syntax and semantics
that exist in these two programming languages sometimes lead to different results
[7–10]. The main objective of the software application used by [7–10] related with
the calculation of probabilistic seismic spectral acceleration [11, 12].
Fig. 1 Three different location of WS and SAS evaluation, Sumatra, Java and Sulawesi islands
Software Performance of Risk-Targeted … 481
This paper describes the evaluation of the MCER acceleration values calculated
using WS and SAS software. The evaluation was conducted at three different loca-
tions: Sumatra, Java and Sulawesi islands and performed at 154 coordinate (longitude
and latitude) positions on Sumatra Island and performed at 114 coordinate positions
on Java Island and 81 coordinate positions on Sulawesi Island (Fig. 1). The evalu-
ation was performed in terms of the mapped MCEG peak ground acceleration and
MCER at short period (MCER -SS ) and at 1-s period (MCER -S1 ).
2 Methodology
The MCER acceleration (MCEG , MCER -SS and MCER -S1 ) at a specific position or
coordinate P can be calculated using linear interpolation of the MCER values calcu-
lated at four different points close to P. Figure 2 shows the basic concept of spectral
acceleration calculation at point P. S1, S2, S3 and S4 represent four MCER accel-
erations (MCEG or MCER -SS or MCER -S1 ). SP represents the MCER acceleration
(MCEG or MCER -SS or MCER -S1 ) calculated at point P. The distance “d” between
two points is 0.1° or approximately 11.13 km. If the point P is located between two
other points, the MCER acceleration can also be calculated using linear interpolation
of the two points. Table 1 shows an example of spectral acceleration MCEG calcu-
lation at point P. Table 2 shows an example of MCER -SS calculation at point Q. The
difference of the MCEG or MCER -SS results calculated using WS and SAS can also
be seen in Tables 1 and 2.
The differences of MCEG , MCER -SS or MCER -S1 calculated using the two
different type of software WS and SAS were then evaluated and compared to deter-
mine the performance of the WS and SAS software. The evaluation was performed
Fig. 2 Linear interpolation

of MCER acceleration S1
calculation at specific S2
coordinate point P
S1 S2
1 S1 S2 2
SP
P
d
3 S3 S4 4
S3 S4
S4
S3
482 W. Partono
Table 1 Sample spectral acceleration MCEG calculation at P (110.57, −6.53)

Point no Longitude (degree) Latitude (degree) MCEG (g) MCEG at P (g)
1 110.5 −6.5 0.198 0.2032 (WS)
2 110.5 −6.6 0.211 0.2031 (SAS)
3 110.6 −6.5 0.2
4 110.6 −6.6 0.212
Table 2 Sample spectral acceleration MCER -SS at Q (99.73764, 1.05033)

Point no Longitude (degree) Latitude (degree) MCER -SS (g) MCER -SS at Q (g)
1 99.7 1.1 1.5 1.4762 (WS)
2 99.7 1.0 1.5 1.5 (SAS)
3 99.8 1.1 1.176
4 99.8 1.0 1.5
based on the total data (percentage of the total points in one study area) having MCEG
or MCER -SS or MCER -S1 differences, as calculated using WS and SAS software,
greater than 0.01 g.
3 Result and Discussions
Calculation of response spectrum acceleration design at specific building locations

is an important step in seismic design for buildings. The purpose of this analysis is to
obtain the MCER , SDS (short period) and SD1 (1-s period) spectral acceleration design.
According to [1], another parameter used for response spectrum acceleration design
is TL (the long transition period). Only the MCER difference performance of the two
software packages was evaluated in this study. Figure 3 shows the analysis results of
the WS and SAS difference performance calculated at 154 points on Sumatra island.
Comparative analysis of the MCER calculation between the WS and SAS software
was also conducted on Java and Sulawesi islands. Figure 4 shows the analysis results
of the WS and SAS difference performance conducted on Java island. Figure 5 shows
the analysis results for Sulawesi island.
Based on the calculation results conducted using WS and SAS at Sulawesi island,
approximately 10–11% of the total 81 MCEG , MCER -SS and MCER -S1 data have a
minimum 0.01 g difference. The largest MCEG , MCER -SS and MCER -S1 differences
between the WS and SAS calculation were observed on Sumatra island. At least 21–
28% of the total 154 investigation data had a minimum 0.01 g difference. The largest
MCEG , MCER -SS and MCER -S1 differences for these two islands were observed at
all points located close to seismic source traces. Figures 3 and 5 show the significant
differences of MCEG , MCER -SS and MCER -S1 calculated using the WS and SAS
Fig. 3 MCEG, MCER -SS and MCER -S1 difference performance calculated for Sumatra island
(a) and the percentage of the total investigation data having at least 0.01 g differences (b)
Fig. 4 MCEG, MCER -SS and MCER -S1 difference performance calculated for Java island (a) and
the percentage of the total investigation data having at least 0.01 g differences (b)
Fig. 5 MCEG, MCER -SS and MCER -S1 difference performance calculated for Sulawesi island
(a) and the percentage of the total investigation data having at least 0.01 g differences (b)
484 W. Partono
software for Sumatra and Sulawesi islands. Figure 4 shows the performance of the
WS and SAS software used for Java island. Based on this figure, a minimum of 1%
up to a maximum of 9% of the total 115 data were observed to have at least 0.01 g
difference. The largest difference (9%) among the total of 115 data was observed for
MCER -SS calculation. However, only 1% of the total 115 MCER -S1 results have at
least 0.01 g difference.
To evaluate the performance of these two software packages especially at the
position coordinates close to seismic sources, the analysis was performed at the
area where the MCEG and MCER -S1 reached ±0.6 g and the MCER -SS reached
±1.5 g. These two 0.6 g and 1.5 g values are two constant MCER acceleration values
calculated based on the combination of deterministic, probabilistic and RTGM (Risk-
targeted Ground Motion) seismic hazard analysis and following the same procedure
proposed by [13]. Figures 6a, 7a and 8a show the predicted area where the MCEG or
MCER -SS or MCER -S1 reached constant acceleration values. The distribution area
of the three constant acceleration values MCEG , MCER -SS and MCER -S1 at one
island is not always equal or in the same area. The performance of WS and SAS in
calculating MCEG , MCER -SS and MCER -S1 was also investigated at three different
cross sections. The purpose of the analysis was to find the predicted position where
significant MCEG , MCER -SS and MCER -S1 differences were observed. Figure 6a,
b, 7a, b and 8a, b show three different positions (cross-sections A, B and C) for
MCEG , MCER -SS and MCER -S1 evaluation. For each cross-section, the analysis
was conducted at 21 coordinate position points: from point A1 (98.84, 1.31) to A21
(100.49, 2.21) at cross-section A (Sumatra island); from point B1 (106.75, −7.41)
to B21 (107.01, −6.09) at cross-section B (Java island) and from point C1 (119.31,
−1.48) to C21 (120.55, −1.06) at cross-section C (Sulawesi island).
Fig. 6 Constant MCEG , MCER -SS and MCER -S1 area on Sumatra island (a), detailed position
of constant area and cross-section A (b), MCEG , MCER -SS and MCER -S1 distribution calculated
using WS and SAS at cross-section A (c)
Fig. 7 Constant MCEG , MCER -SS and MCER -S1 area on Java island (a), detailed position of
constant area and cross-section B (b), MCEG , MCER -SS and MCER -S1 distribution calculated
using WS and SAS at cross-section B (c)
Fig. 8 Constant MCEG , MCER -SS and MCER -S1 on Sulawesi island (a), detailed position of
constant area and cross-section C (b), MCEG , MCER -SS and MCER -S1 distribution calculated
using WS and SAS at cross-section C (c)
The largest MCEG , MCER -SS and MCER -S1 difference results calculated using
the WS and SAS software were observed at the coordinate points close to seismic
fault traces, shown in Fig. 6c and 8c. When the MCEG and MCER -S1 acceleration
at one position coordinate was less than 0.6 g, the MCEG and MCER -S1 calculated
using WS and SMS software were nearly equal. The same performance was also
486 W. Partono
observed when the MCER -SS calculated at one position coordinate was less than
1.5 g, when the MCER -SS calculated using the WS and SAS software were almost
equal. The MCEG , MCER -SS and MCER -S1 calculated at all 21 position coordinates
at Java island, (Fig. 7c), were almost equal due to the condition that there are no
coordinate positions having MCEG and MCER -S1 greater than 0.6 g or MCER -SS
greater than 1.5 g.
The MCEG , MCER -SS and MCER -S1 constant area at Sumatra island was observed
in the western part, while the MCEG , MCER -SS and MCER -S1 constant area was
predicted at the small south-western area of Java island. However, there are two
MCEG , MCER -SS and MCER -S1 constant areas at Sulawesi island, at the middle and
northern parts of the island. The seismic source traces were predicted close to the
MCEG , MCER -SS and MCER -S1 constant area [13].
4 Conclusions
Significant differences of MCEG , MCER -SS and MCER -S1 calculated using WS and
SAS software were observed at Sumatra and Sulawesi islands at coordinate positions
close to the seismic sources area. Based on the calculation results performed at
Sumatra island (154 points), the difference minimum of 0.01 g was observed at 21%
(32 points) to 28% (43 points) of the data. However, for three spectral accelerations
calculated at Sulawesi island (81 points), the difference minimum of 0.01 g was
observed at 10–11% (approximately 9 points) of the data.
The MCEG , MCER -SS and MCER -S1 acceleration values calculated at 151 coor-
dinate positions on Java island using WS and SAS software were almost equal. Based
on the calculation results conducted at 115 coordinate points, a maximum of 5% (5
points) of data with the minimum difference of 0.01 g were observed for the MCEG
and MCER -S1 spectral acceleration. For MCER -SS calculation, 9% (10 points) of the
data with the minimum 0.01 g were observed on the Java island.
Acknowledgements Thank you to the Dean of the Faculty of Engineering, Diponegoro University,
Semarang, Indonesia, for the approval of the research funding through the 2021 Strategic Research
Scheme.
References
1. SNI 1726 (2019) Seismic resistance design codes for building and other structures. National
Standardisation Agency of Indonesia
2. National Center for Earthquake Studies: Indonesian Seismic Sources and Seismic Hazard Maps
(2017) Center for research and development of housing and resettlement. Ministry of Public
Works and Human Settlements
3. Partono W, Irsyam M, Wayan SI, Asrurifak M, Kistiani F, Sari UC (2017) Development of

design response spectrum acceleration for building based on new Indonesian Seismic Hazard
maps 2017. In: AICCE conference, 21–22 August 2019, Penang Malaysia (2019), Lecture
Notes in Civil Engineering, vol 53. Springer Nature Switzerland AG
4. Sengara IW, Irsyam M, Sidi ID, Mulia A, Asrurifak M, Hutabarat D, Partono W (2019)
New 2019 risk-targeted ground motions for spectral design criteria in Indonesian seismic
building code. In: 4th International conference on earthquake engineering & disaster mitiga-
tion (ICEEDM 2019), vol 156. Padang Sumatra Barat, E3S Web of Conference, 26–27 Sept
2019
5. ASCE/SEI 7–16 (2017) Minimum design loads for building and other structures. American
Society of Engineers
6. Stewart JA, Seyhan E (2013) Semi-empirical nonlinear site amplification and its application in
NEHRP site factors. Pacific Engineering Research Center (PEER) Report 2013/13, University
of California, Berkeley, USA
7. Irwandi I, Umar M, Khaizal J, Asrurifak M, Usama F, Ridwan M (2020) The Neo-deterministic
seismic hazard map (NDSHA) of Sumatra compared with official 2010 and 2017 derived
from PSHA method. In: NCWE & ISSCE 2019, IOP conferences series materials science and
engineering, vol 712. IOP Publishing
8. Danciu L, Monelli D, Pagani M, Wiemer S (2010) GEM1 hazard: review of PSHA software.
GEM Technical Report 2010–2, GEM Foundation, Pavia, Italy
9. Syahbana AJ, Aditramulyadi DD, Irsyam M, Asrurifak M, Hendriyawan (2020) Application of
modified PSHA USGS software in Java Island Bedrock Peak ground acceleration and hazard
curve with 2475 years return period. Int J Adv Sci Technol 29(7):3138–3148
10. Pailoplee S, Palasri C (2014) CU-PSHA a Matlab software for probabilistic seismic hazard
analysis. J Earthq Tsunami 08(04):1450008
11. Mulargia F, Stark PB, Geller R (2017) Why is probabilistic seismic hazard analysis (PSHA)
still used. Physics of the Earth and Planetary Interiors, Elsevier, pp 63–75
12. Sianko I, Ozdemir Z, Khoshkholghi S, Garcia R, Hajirasouliha I, Yazgan U, Pilakoutas K
(2020) A practical probabilistic earthquake hazard analysis tool: case study Marmara region.
Bull Earthq Eng 18:2523–2555
13. Leyendecker EV, Hunt EJ, Frankel AD, Ruktales KS (2000) Development of maximum
considered earthquake ground motion maps. Earthq Spectra 16(1):21–40
Seismic Microzonation of Yogyakarta
Province Based on 2019 Risk-Targeted
Maximum Considered Earthquake
Windu Partono , Ramli Nazir, Frida Kistiani, and Undayani Cita Sari
Abstract Yogyakarta Province is located close to two active seismic sources, Opak
Fault which crosses the province area and the South Java subduction source which
is located at the south of Java island. The province is located at the southern part of
Java island. Based on the Indonesian earthquake database, from 1984 to 2015 at least
seven earthquakes struck this province. The 2006 earthquake with 6.2 Mw magnitude
was the largest earthquake to hit this area, causing approximately 88,249 buildings
to be totally destroyed and 98,343 buildings to collapse. Most of the destroyed and
collapsed buildings were constructed based on the old version of the seismic design
code. Improvements in earthquake research have already been conducted in this
area and the latest research was conducted at 2016. According to the New Indone-
sian seismic code 2019, improvements in seismic acceleration for building design
need to be adjusted in this area. This paper described the seismic microzonation of
Yogyakarta Province based on the New Indonesian seismic maps and 2019 seismic
code in terms of the Risk-targeted Maximum Considered Earthquake. The analysis
was performed by conducting a combination of three seismic hazard analysis, namely
probabilistic, deterministic and Risk-targeted Ground Motion. Based on the Risk-
targeted Maximum Considered Earthquake data calculated and its distribution in this
area, an area with a maximum 10 km radius from the Opak Fault trace was detected
as the largest acceleration area. This area can be used as an indicator of a dangerous
area of the province when subject to earthquake ground motion.
Keywords Deterministic · Probabilistic · Risk-targeted maximum considered

earthquake · Risk-targeted ground motion
W. Partono (B) · F. Kistiani · U. C. Sari

Civil Engineering Department, Faculty of Engineering, Diponegoro University,
R. Nazir
Center of Tropical Geoengineering, School of Civil Engineering, Faculty of Engineering,
Universiti Teknologi Malaysia, 81300 Skudai, Johor, Malaysia
490 W. Partono et al.
1 Introduction
Yogyakarta Province is located on the Java island and close to Central Java Province.
Yogyakarta Province is divided into four districts, Bantul, Gunung Kidul, Kulon
Progo, Sleman, and one city, Yogyakarta the province’s capital city. According to [1],
the province is located close to two active seismic sources, the South Java Subduction
Megathrust and Opak Fault trace. According to Sutiono et al. [2], the first predicted
earthquake to strike Yogyakarta Province occurred in 1840 and by 2015 there were
seven different earthquakes caused by two different seismic sources (the Subduc-
tion Megathrust and Opak Fault). The 2006 earthquake that occurred in this area
was caused by the Opak Fault source and was declared as the largest earthquake
to strike this area and the southern part of Central Java Province. Elnashai et al.
[3] stated that 4659 people from Yogyakarta Province area died and 19,401 people
were injured. However, this 2006 earthquake was also struck part of Central Java
Province. According to [3], 1,057 people from Central Java Province were killed
and another 18,526 people were injured (BAPPENAS/The Indonesian Ministry of
National Development Planning). Elnashai et al. [3] also stated that 186,592 houses
in Yogyakarta Province and another 172,104 in Central Java Province were totally
destroyed or damaged by the 2006 Opak Fault earthquake. According to [3], there
were 8 precisely known 2006 earthquake epicentre positions. Figure 1 shows these
earthquake epicentre positions declared by 8 different institutions together with
information on the buildings that collapsed or were destroyed by 2006 earthquake.
Based on the 2002 Indonesian seismic code [4], Yogyakarta Province is located
in region 3 and has a peak horizontal ground acceleration of 0.3 g (g = gravitational
Fig. 1 Epicentre position and total buildings destroyed due to 2006 earthquake.
Seismic Microzonation of Yogyakarta Province ... 491
acceleration). During the 2006 earthquake, the peak horizontal ground acceleration
in the study area was predicted at a maximum 0.5 g [3]. The seismograph equipment
installed in the study area detected that the predicted peak horizontal ground accel-
eration maximum was 0.34 g and the maximum peak vertical ground acceleration
was 0.3 g [3]. According to this seismic data analysis, the study area was not strong
enough to resist the earthquake events having a minimum peak ground acceleration of
0.5 g. The existing buildings constructed before 2006 should be evaluated against
earthquakes having a minimum peak ground acceleration of 0.5 g, especially in the
areas close to the Opak fault trace (Bantul, Sleman districts and Yogyakarta City).
This paper describes the seismic microzonation analysis conducted for the whole
of Yogyakarta Province. The analysis was performed by using the 2017 seismic
hazard maps and following the 2019 Indonesian Seismic Code [5] for determining the
Risk-targeted Maximum Considered Earthquake. Three basic seismic hazard anal-
yses (probabilistic, deterministic and Risk-targeted Ground Motion) were combined
together to obtain the Risk-targeted Maximum Considered Earthquake (MCER ) [5–
7]. The MCER acceleration values obtained at this study area were used as a parameter
to evaluate the risk performance of the study area.
2 Methodology
The seismic hazard analysis used for calculating probabilistic spectral acceleration
was conducted based on a 2% probability of occurrence in 50 years, or a 2500-
year return period earthquake. Three different seismic sources (shallow crustal fault,
subduction interface and deep intraplate) and eight attenuation functions [8–10] for
a shallow crustal fault, [11–13] for a subduction interface and [12, 14] for the deep
intraplate model were applied to calculate the probabilistic spectral acceleration.
Uniform hazard curves obtained from the probabilistic analysis were then used to
calculate the spectral acceleration (Uniform Hazard Ground Motion/UHGM). Equa-
tion (1) shows the basic model of probabilistic seismic hazard analysis, where “v” is
the annual rate, and Pm(m) and Pr(r) are the probability of magnitude and distance.
¨
λ(a) = ν P(a > a*)|m, r| Pm (m)Pr (r)dr dm (1)
The deterministic seismic hazard analysis was performed by using the same atten-
uation function used for probabilistic analysis. The deterministic analysis for each
seismic source model was calculated using the average of three different attenuation
functions and extending the maximum magnitude used in the probabilistic analysis
and the minimum distance for each investigation point to all seismic sources. The
84th percentile of deterministic spectral acceleration was applied for developing the
MCER acceleration.
Risk-targeted Ground Motion (RTGM) analysis for each investigation point in
the study area was performed following the same approach proposed by [15]. The
purpose of this analysis is to obtain modified probabilistic analysis result by applying

a probability of building collapse of 1% in 50 years. Equations (2) and (3) show
the basic approach for calculation of the RTGM. The “β” value used in these two
equations represents the logarithmic standard deviation, while “c” and “c10%” repre-
sent the spectral acceleration and the 10th percentile capacity of building collapse,
respectively. For the 2019 seismic code, the value β = 0.65 was applied to calculate
risk-targeted ground motion acceleration calculation.
Figure 2 shows three examples of risk-targeted ground motion calculated for three
different positions A, B and C. Figure 3 shows the positions of the Risk targeted
Ground Motion calculation. The analysis was performed for a short period (0.2 s)
and a long period (1 s) of spectral acceleration. Tables 1 and 2 show the RTGM
analysis results (RTGM 0.2 and 1 s) and the corresponding risk coefficient obtained
from this analysis. The risk coefficient was calculated by comparing the RTGM value
and the UHGM spectral acceleration from the probabilistic hazard analysis.
Fig. 2 Risk-targeted ground motion calculation at three position A, B and C for spectral acceleration
0.2 s (a) and 1 s (b).
Fig. 3 Risk-targeted ground motion calculation position.
Table 1 Risk targeted ground motion calculation for 0.2 s spectral acceleration.
Point no Longitude (degree) Latitude (degree) UHGM (g) RTGM (g) Cr
A 110.3 −7.9 1.1497 1.1070 0.9302
B 110.4 −7.9 1.3449 1.2378 0.9204
C 110.5 −7.9 1.7599 1.5815 0.8986
Table 2 Risk targeted ground motion calculation for 1 s spectral acceleration.

Point no Longitude (degree) Latitude (degree) UHGM (g) RTGM (g) Cr
A 110.3 −7.9 0.5364 0.4931 0.9192
B 110.4 −7.9 0.6026 0.5379 0.8927
C 110.5 −7.9 0.7278 0.6446 0.8858

lnc−(lnc10% −1.28β )2
1 −
F(c) = e√ 2β 2
(2)
cβ 2π
∞
Pcollapse = F(c)P[S > c]dc (3)
0
The Risk-targeted Maximum Considered Earthquake (MCER ) calculation was

conducted by applying Risk-targeted Ground Motion and 84th percentile determin-
istic calculation results. To perform the calculation, direction factors 1.1 and 1.3 were
applied for the spectral acceleration of 0.2 and 1 s respectively. The basic proce-
dure for the Risk-targeted Maximum Considered Earthquake (MCER ) acceleration
calculation followed the method proposed by [16].
Risk-targeted Maximum Considered Earthquake distribution maps were developed

for the whole of the study area. Three different spectral acceleration distributions
(MCEG , MCER -SS and MCER -S1 ) were performed by applying spatial analysis.
The purpose of developing these three different seismic microzonation maps was to
evaluate the distribution of the minimum to maximum spectral acceleration in the
study area. Figure 4 shows the distribution of MCEG , MCER -SS and MCER -S1 . Based
on the 2019 seismic code, the minimum and maximum MCEG values were distributed
between 0.35 and 0.6 g, the MCER -SS values were distributed between 0.85 and
Fig. 4 MCEG (a), MCER -S1 (b) and MCER -SS (c) distribution at the study area.
1.5 g and the MCER -S1 were distributed between 0.4 and 0.67 g. All the maximum
Risk-targeted Maximum Considered Earthquake were distributed from 10 km north-
west to 20 km south-east of the Opak Fault. According to these three Risk-targeted
Maximum Considered Earthquake, the maximum ground acceleration at the study
area was predicted to occur between 10 north-west to 20 km south-east of predicted
Opak fault trace.
Evaluation of the MCER 0.2 s (MCER -SS ) and 1 s (MCER -S1 ) distribution was
also performed at a specific line (cross-section). The purpose of this evaluation was
to obtain the information of deterministic, probabilistic and Risk-targeted Ground
Motion analyses in developing a combined analysis of the Risk-targeted Maximum
Considered Earthquake (MCER ) spectral acceleration. The analysis was performed
from coordinate positions 1 (110.1, −7.9) to 13 (110.7, −7.9) (see Fig. 3). All thir-
teen coordinate positions were distributed at the same interval or distance. Figure 5
shows the distribution of the three spectral accelerations and the related Risk-targeted
Maximum Considered Earthquake (MCER ) spectral acceleration developed from
these three values. Only spectral accelerations of 0.2 and 1 s are displayed in
these two figures because the PGA distribution is almost equal to spectrum of 1 s.
Fig. 5 PSHA, DSHA, RTGM and MCER performance at cross section 1–13 for spectra of 0.2 s
(a) and 1 s (b).
Maximum values of 1.5 and 0.6 g capped for the 0.2 and 1 s spectral acceleration
were applied following the same procedure as proposed by [5, 16]. The maximum
spectral acceleration was detected at all points close to the predicted Opak Fault
trace.
4 Conclusions
Seismic microzonation of Yogyakarta Province was conducted based on the Risk-

targeted Maximum Considered Earthquake by combining three probabilistic, deter-
ministic and Risk-targeted Ground Motion analyses. The analysis was developed
based on all the earthquake data and earthquake sources used for developing the
Indonesian Earthquake Maps 2017. The maximum spectral acceleration MCEG ,
MCER -SS (spectrum 0.2 s) and MCER -S1 (spectrum 1 s) was detected within a
10 km north-western to 20 km south-east radius from Opak Fault trace.
Maximum values of 1.5 g for 0.2 s (MCER -Ss) period and 0.6 g for MCEG and
the 1 s (MCER -S1 ) period spectral accelerations were obtained in the study area. The
predicted maximum spectral acceleration and earthquake danger area associated with
the Opak Fault was detected in a radius of 10 km from the Opak Fault trace.
Acknowledgements Thank you to the Dean of the Faculty of Engineering, Diponegoro University,
Semarang, Indonesia, for the approval of the research funding through the 2021 Strategic Research
Scheme.
References
1. National Center for Earthquake Studies: Indonesian Seismic Sources and Seismic Hazard Maps
(2017) Center for research and development of housing and resettlement. Ministry of Public
Works and Human Settlements
2. Sutiyono A, Prastistho B, Prasetyadi C, Supartoyo (2018) Opak fault a comparative review. In:
IOP conference series, earth and environmental science 212, IOP Publishing
3. Elnashai AS, Sun JK, Gun JY, Sidarta D (2007) The Yogyakarta Earthquake of May 27 2006.
Mid-America Earthquake Center, MAE Center Report 07-02
4. SNI 1726 (2002) Seismic resistance design for building structures. National Standardization
Agency of Indonesia
5. SNI 1726 (2019) Seismic resistance design codes for building and other structures. National
Standardization Agency of Indonesia
6. ASCE/SEI 7-16 (2017) Minimum design loads for building and other structures. American
Society of Engineers
7. Sengara IW, Irsyam M, Sidi ID, Mulia A, Asrurifak M, Hutabarat D, Partono W (2020) New
2019 risk-targeted ground motions for spectral design criteria in Indonesian seismic building
code. In: 4th international conference on Earthquake engineering and disaster mitigation
(ICEEDM 2019) Padang Sumatera Barat 26–27 September (2019), E3S Web of Conference,
vol 156
8. Boore DM, Stewart JP, Seyhan E, Atkinson GM (2013) NGA-West 2 equations for predicting
PGA, PGV, and 5%-damped PSA for shallow crustal earthquakes. Earthq Spectra. Earthq Eng
Res Inst 30(3):1057–1085
9. Campbell KW, Bozorgnia Y (2013) NGA-West2 Campbell-Bozorgnia, ground motion model
for the horizontal components of PGA, PGV and 5%-damped elastic pseudo acceleration
response spectra for periods ranging from 0.01 to 10 s. Pacific Earthquake Engineering Research
Center, PEER Report 2013/6, pp xii+75
10. Chiou BSJ, Youngs RR (2013) Update of the Chiou and Youngs NGA ground motion model for
average horizontal component of peak ground motion and response spectra. Pacific Earthquake
Engineering Research Center
11. Abrahamson NA, Addo KO, Atkinson G, Chiou B, Gregor NJ, Silva W, Youngs RR (2014)
Ground motion characterization for BC hydro SSHAC level 3 study. In: Tenth U.S. national
conference on earthquake engineering, frontiers of earthquake engineering, IONCEE, 21–24
July 2014, Anchorage, Alaska
12. Atkinson GM, Boore DM (2003) Empirical ground motion relations for subduction zone
earthquakes and their application to Cascadia and other regions. Bull Seismol Soc Am
93(4):1703–1729
13. Zhao JX, Irikura K, Zhang J, Fukushima Y, Somerville PG, Asano A, Ohno Y, Oouchi T,
Takahashi T, Ogawa H (2006) An empirical site classification method for strong motion stations
in Japan using H/V response spectral ratio. Bull Seismol Soc Am 96(3):914–925
14. Youngs RR, Chiou SJ, Silva WJ, Humphrey JR (1997) Strong ground motion attenuation
relationships for subduction zone earthquakes. Seismol Res Lett 68(1):58–73
15. Luco N, Ellingwood BR, Hamburger RO, Hooper JD, Kimball JK, Kircher CA (2007) Risk-
targeted versus current seismic design maps for the conterminous United States. In: Structural
engineers association of California 2007 convention proceedings. California, pp 163–175
16. Leyendecker EV, Hunt EJ, Frankel AD, Ruktales KS (2000) Development of maximum
considered earthquake ground motion maps. Earthq Spectra 16(1):21–40
Liquefaction Potential of Volcanic
Deposits During Lombok Earthquake
in 2018
Muhajirah
Abstract The Lombok earthquake and the Palu earthquake in 2018 were earth-
quakes that caused major damage in the two referenced areas. The Lombok earth-
quake occurred over a longer period than the Palu earthquake, but the Palu earthquake
was able to cause enormous liquefaction. This study aimed to determine the behavior
of the volcanic deposits of Mount Samalas during the corresponding earthquake.
The research included a physical properties test, SEM photos, a shear strength test,
a shaking table test, and grading tests before and after the soil samples were given
cyclic loads. The results of the grading test after the shaking table test showed that
there was the addition of fine grains before and after the shaking table test. This
indicates that the pumice particles are scoured on the surface, resulting in the small
grains becoming locked between the larger particles. This condition caused the shear
strength to increase and the pore pressure to decrease significantly. The cavity in the
pumice particles is thought to reduce the seismic waves during an earthquake. For
the large particles, although the gradation curve has liquefaction potential, the pore
water pressure did not increase significantly. However, pumice has a diameter of less
than 0.075 mm because it does not have a cavity, therefore that it has the potential
to experience liquefaction such as sand boiling.
Keywords Lombok earthquake · Pumice sand · Liquefaction
1 Introduction
In 2018, Lombok Island was shaken by a series of destructive earthquakes. The first
earthquake with a magnitude of 6.4 M occurred on July 29, 2018, in the northern
part of the island of Lombok. The intensity of the earthquake on the MMI VIII scale
was recorded in the Loloan area. The epicenter was on land 47 km northeast of
Mataram City at a depth of 24 km. A week later, on August 5, 2018, an earthquake
again shook the island of Lombok that was 6.9 Mw . Seismologically, based on the
Muhajirah (B)
Department of Civil Engineering, Mataram University, Mataram, Indonesia
500 Muhajirah
distance between the two epicenters, magnitude, and the time of occurrence, the
two earthquakes were related or were dependent events. The 6.4 Mw earthquake
was a foreshock of the 6.9 Mw earthquake. Four days later, on August 9, 2018,
strong tremors resulted from an earthquake that was 6.3 Mw . The epicenter of the
earthquake was estimated to be to the southwest of the epicenter of the previous
earthquake, closer to Mataram city [1].
The geological phenomena that emerged after the 6.4 Mw (July 29, 2018) and
6.9 Mw (August 5, 2018) earthquakes were dominated by land surface deformation
in the form of landslides, lateral spreading, and liquefaction. The types of landslides
that occurred included rockfalls and combined sliding debris. After the 7.0 Mw earth-
quake, several phenomena arose including the fractures that released water along with
mud and sand. The fractures that occurred were in the form of longitudinal fractures
and round holes. The elongated soil cracks due to liquefaction were in the form of
longitudinal cracks with a width of 2–10 cm. From these cracks, sometime after
the earthquake, water came out together with mud and/or sand. This phenomenon
occurred in front of the village office of Selengan in North Lombok Regency [2].
The Palu earthquake occurred on Friday afternoon on September 28, 2018. It was
7.4 Mw centered 26 km north of Donggala Regency in Central Sulawesi Province.
This earthquake caused a tsunami in Palu City and widespread liquefaction in the
Petobo and Ballaroa areas. The series of large earthquakes that occurred on the island
of Lombok for two months did not cause liquefaction effects like the one in Palu
City.
The liquefaction events that occurred in Japan, New Zealand, and Indonesia indi-
cate that most of the liquefaction occurs in sedimentary soils resulting from volcanic
eruptions. Of the three countries, Japan suffered the most losses due to the impact of
the earthquake. Several earthquakes resulted in liquefaction in parts of Japan. The
liquefaction susceptibility of this area is thought to be because most of the land is
covered by volcanic deposits that are easily destroyed. These deposits are widespread
in the southern part of Kyushu known as Shirasu.
The Hokuseibu earthquake in 1997 resulted in heavy damage to the land reclama-
tion which was covered with Masado soil. Masado soil is crushable granite residual
soil. An investigation to determine the characteristics of the volcanic soil following
the earthquake in Japan was carried out by Aramaki et al. [3] and Hyodo et al. [4].
Orense et al. [5] and Liu et al. [6] also researched the soil liquefaction that occurred
in New Zealand due to the Edgecumbe earthquake in 1987 and the Canterbury earth-
quake in 2010–2011. Most of the New Zealand areas are built up by the volcanic
eruptions concentrated in the Taupo and Rotorua regions, which are known as the
Taupo volcanic zone.
On May 27, 2006, Yogyakarta and its surrounding areas experienced an earthquake
with a magnitude of 6.3 on the Richter scale. The areas that were damaged by the
earthquake included Bantul Regency and Klaten Regency. According to Soebowo
et al. [7], most of Yogyakarta is made from the volcanic deposits of Mount Merapi.
The phenomenon of liquefaction that occurred after the earthquake was in the form
of sandblasting holes, wells drying up and being covered in fine sand, lateral ground
Liquefaction Potential of Volcanic Deposits … 501
movement, land subsidence, and the emergence of water on top of the ground shortly
after the earthquake [7–9].
1.1 The Liquefaction Research in Volcanic Deposits
Kusumawardani et al. [8] conducted tests on soil samples from the UMY Yogyakarta
campus. The cyclic triaxial test was carried out on the soil samples. The sample was
prepared using the dry pluviation method with a relative density of 25%. In this test,
the sample was loaded with a frequency of 0.05 and 0.1 Hz. The results showed that
for the number of cycles equal to 100 at the same density (Dr = 25%), the soil
sample with a load frequency of 0.1 Hz reached a pore pressure ratio of 1.0 while
the soil sample with a frequency of 0.05 Hz reached pore pressure ratio of 0.8.
Mase [10] conducted a shaking table test using Opak River sand from Imogiri
District in Bantul Regency. He conducted an enlargement test on the Kali Opak sand
particles. The test results, when examined through a stereo microscope, showed that
the particles looked solid with no cavities (Fig. 1).
The sand samples had a field density of (Dr = 28.40%). The amplitudes of
acceleration (amax ) used to move the shake table were 0.30, 0.35, and 0.40 g at
frequencies of 1.4, 1.6, and 1.8 Hz. The duration of the shaking for each acceleration
was 8, 16, and 32 s. The three amplitudes of acceleration applied to the sand model
showed the same tendency, specifically an increase in the excess pore-water pressure
ratio (ru ) to reach a value of one. The increase in the excess pore pressure at 32 s was
the highest. This indicates that over a long duration of shaking time, the accumulated
increase in excess pore pressure will be greater.
Kusumawardani [8] and Mase [10] evaluated the liquefaction potential of non-
porous particles (Fig. 2). The pore water pressure increased after the soil sample was
subjected to a cyclic load and reached the point of liquefaction [10]. Even though
Fig. 1 The shape of sand

particles in the Opak River
[10]
502 Muhajirah
the cyclic triaxial test was carried out at low frequencies, the clean sand sample still
reached a liquefaction condition [8].
This research was conducted in order to determine the behavior of the volcanic
deposits of Mount Samalas following an earthquake. The research was carried out
on the soil samples taken from the city of Mataram and the fine sand samples that
came through the cracks during the Lombok earthquake.
1.2 Liquefaction Mechanism
Liquefaction is defined as the transformation of soil from a solid into a liquid. This
occurs due to the increased pore water pressure and reduced effective stress. It gener-
ally occurs in less cohesive soils that are saturated with water. When the soil is
subjected to rapid loads such as an earthquake load or other dynamic loads, the pore
water cannot be drained in the short amount of time when the loading occurs. This
loading condition is known as an undrained condition. If the soil is less dense, the
less cohesive soil that receives cyclic loads tends to experience compaction. When
the cavity between the soil grains contains water and the water cannot be drained,
the excess pore water pressure will increase. Figure 2 shows the changes that occur
in soil particles due to cyclic loads causing excess pore water pressure [11].
At a certain depth below the ground surface, Eq. (1) can explain the stress condi-
tions when liquefaction occurs. The undrained condition causes the excess pore
pressure to increase. At one point, the excess pore water pressure reaches the total
stress value and the effective stress is close to or equal to zero [12].

σ =σ −u (1)
where:

σ effective stress,
Fig. 2 Cyclic loads cause soil particles to move and excess pore pressure increase [9]
σ total stress,
u pore water pressure.
Not all soil deposits are susceptible to liquefaction. In general, liquefaction
requires the following three conditions: (i) the cohesionless soil has a relative density
of loose to medium dense, (ii) a shallow groundwater level to ensure that the soil is
saturated, and (iii) a sufficiently strong ground motion or a cyclic load.
Several factors can influence the susceptibility of the soil to liquefaction. Based
on the field investigations and laboratory test results, the liquefaction characteristics
of cohesionless soils are influenced by the following factors: grain size distribution,
soil type, relative density, earthquake loading characteristics, effective vertical stress
and over-consolidation, the age of the soil, the origin of the soil, the seismic strain
history, the degree of saturation and the thickness of the sand layer [13].
2 Materials and Method
The research was carried out in two stages: (1) a series of tests conducted on the soil
samples from the City of Mataram, and (2) gradation tests and SEM photos conducted
on the sand boil samples from Selengan village in North Lombok Regency.
Research procedures include:
(1) Pumice sand taken from Kebon Tallo in Mataram city. It was excavated to a
depth of 2 m for the shaking table test. Furthermore, pumice sand was used in
the analytical sieve test to obtain material with a potential liquefaction gradation
curve.
(2) Testing the physical properties of the pumice samples using ASTM standards
including a water content test, sieve test, relative density test, and dry weight
density test.
(3) SEM (Scanning Electron Microscope), the particles used included the material
retained from the No. 10 sieve (grain diameter 2.00 < D < 4.75 mm) and No.
200 sieve (grain diameter 0.074 < D < 0.106 mm).
(4) The shear strength test was used to determine the effect of pumice particle
size on the shear strength value. There were four variations of particle size
used in this test: type A pumice with a diameter of less than 4.75 mm, type B
pumice with a diameter of 0.075–4.75 mm, type C pumice with a diameter of
0.106–4.75 mm, and type D pumice. 0.250–4.75 mm in diameter.
(5) In the shaking table test, the dry pluviation method was used to form a pumice
sample with a relative density of 30%. For dynamic loads, a cyclic load in
the form of a sine wave was used with an acceleration amplitude of 0.38 g
for 20 s in order to study the effect of cyclic loads on the pumice particles.
Sieve analysis tests were carried out on the pumice samples before and after
the shaking test.
504 Muhajirah
The evaluation of the liquefaction potential using a shaking table was conducted
by measuring the pore water pressure increment of the sample. To determine the
development of excess pore water pressure, three Pore Pressure Transducers, abbre-
viated as PPT, were placed on the side of the sample box at a distance of 0.10 m
(Upper PPT), 0.25 m (Middle PPT), and 0.40 m (Base PPT) from the bottom of the
sample box.
For the sand boil sample, only the sieve analysis test and SEM photos were
carried out. The sieve analysis test was carried out to determine the particle size
and percentage. The purpose of the SEM photos was to determine the shape of the
particles.
3.1 Physical Properties of Mataram Pumice
The unique thing that can be seen in Table 1 is that the dry weight density of the
pumice sand is very low (0.54 gr/cm3 ). The void ratio of pumice (3.22) is greater than
that of normal sand. This indicates that the low dry weight density of the pumice is due
to the large number of cavities present in the pumice particles. The soil classification
is SP, which refers to poorly graded sand with silt and gravel.
The shape of the Mataram pumice particles is subrounded to subangular [14].
The liquefaction susceptibility is influenced by grain characteristics such as particle
Table 1 Physical
No. Properties Mataram pumice
characteristics of Mataram
pumice 1 The specific gravity, G s 2.28
2 Water content, w % 114.37
3 Grain size
Gravel mm 22.23
Sand mm 69.75
Silt and clay mm 8.02
D50 mm 1.20
4 Coefficient of uniformity, 23.53
Cu
5 Coefficient of 0.32
curvature,Cc
6 USCS classification SP
7 Bulk density,ρs gr/cm3 1.16
8 Dry density,ρs−dry gr/cm3 0.54
9 Void ratio, e 3.22
size, gradation, and particle shape. Soil particle size is related to soil composition.
Well-graded soils are generally less susceptible to liquefaction than poorly graded
soils. Soil with a rounded particle shape will compact more easily when it undergoes
shear stress than soil with an angular particle shape.
According to Youd and Poulos, when an earthquake occurs, spherical sand parti-
cles with a low relative density (loose) will more easily experience changes in the
composition of the particles due to the shear stress and/or cyclic loads. Therefore,
soils with a rounded particle shape are more susceptible to liquefaction compared to
soils with an angular grain shape.
3.2 Scanning Electron Microscope (SEM)
Figure 3a shows the presence of cavities in the pumice particles. The cavities look
like an elongated channel (arrow) and some of the cavities are not connected to other
cavities (isolated cavities). Figure 3b shows a particle shape that is the result of the
crushing of larger sand grains and the existence of small craters which are the end
of isolated cavities (dotted white circles).
3.3 Shear Strength Characteristic
In Fig. 4, it can be seen that the shorter the range of soil particle sizes, the greater the
friction angle of the soil or the increase in the soil shear strength. In this test, the size
of the grain diameter contributed to the value of the shear strength of the soil. This
can be seen in the figure where the type D particles have the largest internal friction
angle.
(a) (b)
Fig. 3 SEM photograph of Mataram pumice: a retained No. 10 sieve and magnification of 3000×,
b retained of No. 200 sieve and magnification 1000×
506 Muhajirah
3.4 Shaking Table Test
Figure 5 shows the recorded Excess Pore Pressure (EPP) of a pumice sample at an
acceleration of 0.38 g for 20 s. From this figure, the change in excess pore pressure
is not very significant. The maximum pore pressure in Upper PPT and Middle PPT
is only up to 0.10 kPa whereas the Base PPT records the maximum pore pressure at
0.20 kPa. Liquefaction condition did not occur in this sample. The maximum pore
pressure recorded for the three transducers did not reach the effective stress value at
57
D
56
Friction Angle, ϕ ( º )
55
54 C
53
B
52 A
51
0.001 0.01 0.1 1 10
Particle Size (mm)
Diam A : < 4,75 mm, Diam B : 0,075 – 4,75 mm,
Diam C : 0,106 – 4,75 mm, Diam D : 0,25 – 4,75 mm
Fig. 4 Shear strength of Mataram pumice
2
EPP/Effective Stress (kPa)
t = 20 s
0
0 5 10 15 20 25 30 35
Time (s)
EPP in Upper PPT Effective Stress in Upper PPT
EPP in Middle PPT Effective Stress in Middle PPT
EPP in Base PPT Effective Stess in Base PPT
Fig. 5 Excess pore water pressure with time at an acceleration of 0.38 g and time 20 s
each level. The effective stresses in the Upper PPT, Middle PPT, and Base PPT were
0.38, 0.95, and 1.52 kPa respectively.
3.5 Degradation of Mataram Pumice
The gradation curve of the pumice sample was in the potential liquefaction zone
(see Fig. 6). There were differences in the amount of the pumice sand before and
after the shaking table test. The sieve analysis test showed that there was a reduction
in the number of particles in the first sieve and the addition of particles in each
subsequent sieve. Based on Fig. 6, the particle size ranges were 0.425–0.85 mm and
0.106–0.25 mm, and particles smaller than 0.074 mm with an increase of 3.0, 1.5 and
4.2% respectively. The particle size ranges of 2.00–4.75 mm, 0.85–2.00 mm, 0.25–
0.425 mm and 0.074–0.106 mm decreased by 1.3, 5, 5, 1.6 and 0.4% respectively.
The increase in the percentage of the number of particles in a certain size range
indicates that there was a change in the shape of the particles from larger particles to
smaller particles during the shaking table test. This condition was also confirmed by
Wesley [15], indicating that pumice particles are easily destroyed in wet conditions.
Gravel Sand Silt & Clay

100
90
80
Percentage Finer (%)
70
60 b'
50 a'
40
30 b a
20
10
0
10 1 0.1 0.01 0.001
Particle size (mm)
a-a’ curve : most liquefiable,
Before shaking
test After shaking b-a and a’-b’ curve : potentially
liquefiable
Fig. 6 Degradation of Mataram pumice

508 Muhajirah
Fig. 7 SEM photograph of

sand boil from Selengan
village
3.6 Sand Boil of Selengan Village
Sand boil particles have a diameter of less than 2 mm and 65% passed the No. 200
sieve. The particles were made up of pumice sand, thought to be the result of the
crushing of larger particles (Fig. 7). Based on the results of the Atterberg test, the
particles were determined to be non-plastic. A similar gradation curve was obtained
from the sand boil during the Padang earthquake in 2009 [16].
4 Conclusions
Based on the tests that have been done, it is concluded that:

1. The results of the grading test after the shaking table test show that there was an
addition of fine grains before and after the shaking test. This indicates that the
pumice particles were scoured on the surface, and that the small grains become
locked between the larger particles. This condition causes the shear strength to
increase and the pore pressure to decrease significantly.
2. The cavities in the pumice particles are thought to reduce the seismic waves
during an earthquake. For large particles, although the gradation curve has
liquefaction potential, the pore water pressure does not increase significantly.
However, pumice has a diameter of less than 0.075 mm because it does not have
cavities. This means that it has the potential to experience liquefaction like in a
sand boil.
Acknowledgements The author wishes to thank Dr. es sc. tech. Ir. Ahmad Rifai, MT. and Ir.
Agus Darmawan Adi, M.Sc. Ph.D. for their contributions to various aspects of the experimental
work. The author would like to thank the Head and staff of the Soil Mechanic and the Structural
Engineering Laboratories of Gadjah Mada University for providing the laboratory facilities and for
their assistance conducting the shaking table test.
References
1. Maipark, di Lombok G, dan Donggala P, di Selat Sunda S. http://www.maipark.com/assets/upl

oads/publication/Gemuruh-di-Lombok-%20Palu-dan-Donggala-%20Senyap-di-Selat%20S
unda.pdf
2. Tim Pusat Studi Gempa Nasional (PusGen) (2018) Kajian Rangkaian Gempa Lombok
Provinsi Nusa Tenggara Barat, Indonesia. Pusat Penelitian dan Pengembangan Perumahan dan
Permukiman, Badan Penelitian dan Pengembangan, Kementerian PUPR, Cetakan Pertama,
Bandung
3. Aramaki N, Okabayashi T, Hyodo M (2007) Liquefaction characteristics of crushable volcanic
soil “Shirasu”. In: 4th International conference on earthquake geotechnical engineering, June
25–28, Paper No. 1288
4. Hyodo M, Nakata Y, Aramaki N, Hyde AF, Inoue S (2000) Liquefaction and particle crushing
of soil. In: Proceeding 12th world conference on earthquake engineering
5. Orense RP, Pender MJ, O’Sullivan AS (2012) Liquefaction characteristics of pumice sand,
EQC Project 10/589
6. Liu L, Orense RP, Pender MJ (2015) Crushing-induced liquefaction characteristics of pumice
sand. In: NZSEE Conference
7. Sebowo E, Tohari A, Sarah D (2007) Studi Potensi Likuifaksi di Daerah Zona Patahan Opak
Patalan-Bantul Jogjakarta. Prosiding Seminar Geoteknologi Kontribusi Ilmu Kebumian Dalam
Pembangunan Berkelanjutan, Bandung
8. Kusumawardani R, Suryolelono KB, Suhendro B, Rifa’i A (2014) The loading frequency
effects of Yogyakarta’s sand under cyclic Triaxial testing. Int J Civil Environ Eng 14(2)
9. Muntohar AS (2009) Evaluation of peak ground acceleration using CPT data for liquefaction
potential. In: 4th annual international workshop & expo on Sumatera tsunami disaster and
recovery
10. Mase LZ (2017) Experimental liquefaction study of Southern Yogyakarta using shaking table.
Jurnal Teknik Sipil 24(1):11–18
11. Lenart L (2008) The response of saturated soils to a dynamic load. Acta Geotechnica Slovenia
1:37–49
12. Das BM (1993) Principles of soil dynamics. PWS-KENT Publishing Company, Boston
13. Geotechnical Design Procedure (2015) GDP-9: Liquefaction Potential of Cohesionless Soils.
Revision #3, Geotechnical Engineering Bureau, Department of Transportation. State of New
York
14. Hiden et al (2017) The Isopach mapping of volcanic deposits of Mount Samalas 1257 AD
based on the values of resistivity and physical properties. Geosciences 7(67)
15. Wesley LD (2010) Geotechnical engineering in residual soils. Wiley, Hoboken
16. Hakam A, Suhelmidawati E (2013) Liquefaction due to September 30th 2009 earthquake
in Padang. In: The 2nd international conference on rehabilitation and maintenance in civil
engineering, Procedia Engineering, pp 140–146
Fault Structure Interpretation
on the Western Part of East Java Using
Second Vertical Derivative
Wien Lestari , Amien Widodo , Dwa Desa Warnana ,

Firman Syaifuddin, Rusba Saputra Rivensky, and Bagoes Idcha Mawardi
Abstract The National Earthquake Center (PusGeN) in 2017 stated that East Java
was traversed by several active faults that extended towards Bali and Flores. The
gravity research was carried out to map the fault structure in the western part of
East Java. Gravity method is applied to identify subsurface geological structures by
utilizing variations in Earth’s gravity caused by differences in density of subsur-
face rocks. The processed data is satellite gravity data (TOPEX) as many as 110
data spread in Lamongan, Gresik, and Surabaya City. Gravity Topex is a Geodesy
satellite altimeter launched by NASA to measure the altitude of the satellite above
the closest sea surface point to a very high precision. Gravity data that collected
from Topex/Poseidon satellite are already processed into free air anomaly, further,
it is performed Bouguer Correction and Terrain Correction, resulting in Bouguer
Anomaly data which is then filtered using Second Vertical Derivative (SVD). The
results of SVD 2-D modeling have a high level of accuracy because the faults are
well identified and highly correlated with the PuSGeN faults map (2017). The curve
of the slice profiling results on the Surabaya Fault shows that the maximum value of
SVD is 0.1 mGal/m2 and the minimum value is 0.2 mGal/m2 which is identified as
a reverse fault. The curve of the slice profiling results on the Waru Fault shows that
the maximum value of SVD is 0.2 mGal/m2 and the minimum value is 0.3 mGal/m2
which is identified as a reverse fault.
W. Lestari (B) · A. Widodo · D. D. Warnana · F. Syaifuddin · R. S. Rivensky · B. I. Mawardi

Geophysical Engineering Department, Institut Teknologi Sepuluh Nopember, ITS Raya Street,
Surabaya 60111, East Java, Indonesia
A. Widodo
D. D. Warnana
F. Syaifuddin
512 W. Lestari et al.
Keywords Bouguer anomaly · TOPEX gravity · Fault · Second vertical derivative

(SVD)
1 Introduction
1.1 Background
Indonesia occupied a very active tectonic zone because the world’s three large plates
and nine other smaller plates had encountered each other in the Indonesian territory
and form complex plate meeting routes [1]. The existence of interaction between
these plates located Indonesia as an area that is very prone to earthquakes. Victims
of material and human lives because of the earthquake disaster were mostly caused
by ignorance of a fault in an area, so there was no anticipatory measure against
the movement of the fault. Therefore, active fault mapping is very necessary for
the development of spatial planning in Indonesia because from now on, building
construction must consider aspects of disasters, one of which is earthquakes.
Geodetic studies using the Global Positioning System (GPS) have been carried
out on surface deformation measurements in Java [2]. Its GPS deployment identified
an active extension or extension of Flores back-arc thrust westward along 300 km on
land to East Java along the Kendeng Fault as described by Fig. 1 [2]. The results of
this study denoted a new major earthquake threat for East Java and a tsunami threat
to Bali, Lombok, Nusa Tenggara, and other shores along the Flores Sea. Informa-
tion about near-surface faults forms the basis of studies of deformation studies in
deeper areas. The western part of East Java, including Surabaya and its surround-
ings, is traversed by several geological structures including the active Lidah and
Fig. 1 Primary faults on Java Island [2]

Fault Structure Interpretation on the Western Part … 513
Kedung Waru faults which are associated with the Lidah Anticline, the Gayungan
Anticline, and the Kedung Waru Anticline, causing the western part of East Java and
its surroundings to have the potential for earthquakes [3, 4]. The National Earthquake
Center (PusGeN) published the 2017 national earthquake map with the discoveries of
various new active faults in Indonesia as well as the Kendeng Zone of East Java which
are located the Surabaya Fault, the Waru Fault, the Blumbang Fault, and the Cepu
Fault. Their slip displacement of about 0.05 mm with a magnitude < 7 [5]. Earthquake
potential sources are crucial in determining earthquake hazard, therefore geometric
information and location of potential sources are very important [6]. Geodetic studies
have been carried out, however, data or research on fault mapping on the western part
of East Java and its surroundings is still incomplete, so further research elaboration
is needed. The subduction paths that showed in Fig. 1 run through the Northeast–
Southwest (NE-SW) to a relatively East–West (EW) which has occurred since the
Oligocene to the present has took place in a very complicated Tertiary geological
setting on Java Island [7]
Gravity is a passive method that can be implemented to identify subsurface geolog-
ical structures by utilizing variations in Earth’s gravity caused by differences in
subsurface rock density [8–10]. Based on research related to the priority of potential
earthquakes on the island of Java using multi-criteria analysis of decision-making, it
is stated that the Kendeng Surabaya Fault is one of the main priorities for mapping
so that this research will identify related subsurface structures covering the area.
Surabaya, Gresik and Lamongan [11, 12]. Interpretation of the fault can be carried
out using the Second Vertical Derivative (SVD) method [13] to indicate the type of
fault by delineated the highest and lowest anomalies beneath the subsurface [14].
2 Basic Theory
2.1 Regional Geology
The Java Island physiography at the eastern part are divided into 5 zones, namely:
the Southern Mountain Zone, the Solo Zone, the Kendeng Mountain Zone, the
Randublatung Depression Zone, and the Rembang Zone [15]. Physiographically,
the research area is located in the Rembang Anticlinorium, where the rock forming
it consists of shallow marine sediments, clastic sediments, and carbonate rocks.
Figure 2 informed about three directions of the structural pattern considering the
tectonic setting of Java Island, namely the Northeast–Southwest (NE-SW) direction
which is called the Meratus pattern, the North–South (NS) direction or the Sunda
pattern and the East–West (EW) direction.
At the rest of the Cretaceous a subduction zone was built in the Karangsam-
bung area continuously to the Meratus Mountains in Kalimantan. This zone created
a geological structural framework that traverses Northeast-Southwest then during
Tertiary the subduction zone is appeared at the southern of Java Island and the
Fig. 2 Tectonic setting of Java Island [2]
structure formed is East–West trending. The Eurasian Plate and the Australian Plate
produces the main force of north–south compression. This force forms an oblique
wrench fault pattern in the Northwest-Southeast direction, which is more or less in
line with the Late Cretacious ridge pattern [7, 16]. In the Pliocene–Pleistocene period
the main stress direction is still similar, namely North–South, the tectonic activity of
this period produces an upward and folded structure patterns in an East–West direc-
tion that can be recognized in the Kendeng Zone. In general, Java Island has an active
geological structure which is dominated by shear and upward faults with downward
faults which become minor structures. Based on The National Earthquake Hazard
and Source Map of 2017 issued by the National Earthquake Center (PusGeN), it is
declared that East Java is traversed by several active faults, including the Surabaya
Fault, the Waru Fault, the Blumbang Fault, and the Cepu Fault [5].
2.2 Gravity
The Gravity Survey is a geophysical method which is based on measuring the varia-
tion of gravity values at a certain location [8, 17]. Figure 3 illustrated the Newton’s
Law as the basic formula of Gravity Method (Eq. 1) which is explained that the
magnitude of the gravitational force between two substances or materials is propor-
tional to the two masses of substances and inversely proportional to the square of the
centre distance between them [18, 19].
Fg = G(m1 · m2 )/r2 (1)

Fig. 3 The concept of the

force of attraction between
two modified bodies [18]
where Fg is the force of attraction (N) between the mass m1 and m2 , G is the Universal
Gravity Constant (6.674 × 10−11 Nm2 /kg2 ), m1 and m2 are the mass of the object
(kg), and r is the distance between the centers of the two objects (m).
2.3 Second Vertical Derivative (SVD) of Gravity
Second Vertical Derivative (SVD) was started popular in 1951 for gravity interpreta-
tion [20]. This method is performed by calculating the gradient to increase the subtle
gravity data feature which is not visible visually from the original data [21]. The
SVD method is based on the second derivative of the Bouguer Anomaly equal to
zero contact of the density contrast between the two layers of lithologies, in this case
the contact is a fracture or intrusion [22, 23]. The SVD of the vertical component
of gravity, gz, can be calculated in the spatial domain from the horizontal gradients
by using the Laplace’s Equation. The SVD can be approximated by the second hori-
zontal derivative of the gravity data in particular for an elongated anomaly along the
y-axis [24]. The existence of the fault structure is identified by the change in the value
of the second order vertical gradient which is very contrasting [17, 25]. In addition,
the type of fault can be defined by Eq. 2 and Eq. 3 below:
2 2
∂ g
min < ∂ g max (Normal Fault) (2)
∂z 2 ∂z 2
2 2
∂ g
min > ∂ g max (Thrust Fault) (3)
∂z 2 ∂z 2
Where g is gravity acceleration (m/s2 ), z is vertical component.

3 Methodology
3.1 Acquisition Data
This study used gravity data from the TOPEX satellite (https://topex.ucsd.edu/cgi-
bin/get_data.cgi). Figure 4 exhibited the 110 acquisition points of gravity data that are
scattered in Lamongan, Gresik, and Surabaya City districts. The area of data collec-
tion is 18 km × 17 km with a distance between stations of 1.8 km. The measuring
station crosses the active structures are overlaid with the revised map 2017 of earth-
quake hazard and source published by The National Earthquake Center [5]. The
gravity data obtained from TOPEX is in the form of Free Air Anomaly data, then
it will performed Bouguer Correction (Bouguer Correction) and Terrain Correction,
in consequence, Bouguer Anomaly data is generated and displayed through Kriging
method.
Figure 5 displayed the complete processing where the Gravity Bouguer Anomaly
were devided into regional and residual anomaly. Utilization the Second Vertical
Derivative (SVD) method is to generate shallow sources of anomalies hence, the
anomalies resulting from this process are residual anomalies. Second Vertical Deriva-
tive (SVD) is able to determine the boundaries of the existing structures in the research
area.
Fig. 4 Distribution of gravity data points

Fig. 5 Research flowchart
4 Result and Analysis
The gravity method is carried out by implementing variations in the earth’s gravity
caused by differences in the density of subsurface rocks. The gravity data extracted
from the satellite then corrected and calculated to produce Bouguer Anomaly. The
results of the calculation of gravity data carried out by the interpolation process
(gridding) to produce a Bouguer Anomaly map with dimensions of 42 × 46 grid
point. Figure 6 described the Bouguer Anomaly map which has a range of values
from −20.8 to 10.3 mGal.
Fig. 6 Bouguer anomaly map

The distribution of anomalies are grouped into three parts, namely high anomalies
with values ranging from 10.3 to −2.5 mGal marked in red to yellow. This high
anomaly is distributed at the northern area of the study, high anomaly indicates the
presence of rocks that have high density. Medium anomaly in the range −2.5 to −
12.6 mGal, characterized by yellow to green color, this anomaly is deployed in the
middle area of the study. Low anomalies are marked by light to dark blue with a
value of −12.6 to −20.8 mGal, these low anomalies are at the western area, low
anomalies predicted the existence of rocks with low density.
Second Vertical Derivative (SVD) analysis is used to clarify the location and
identify the type of fault. Second Vertical Derivative is carried out to bring the
shallow effect of its regional influence because the geological structure will have a
maximum absolute value and a minimum absolute value which is limited to a value
of zero (0) or close to zero as the limit of geological characteristics. The Bouguer
anomaly map will be the input for calculating SVD (2-D) data.
The interpretation stage is carried out by integrating the results of SVD modeling
with active fault structure mapping data published by the National Earthquake Center.
The results of the SVD 2-D map are interpreted with five trends described by Fig. 7
(dotted black lines), the interpretation is based on the similarity of trends by consid-
ering the maximum and minimum values of global SVD. The results of this inter-
pretation described that the interpretation of the existence of a fault using SVD has a
high level of accuracy. This is shown in Fig. 8 where illustrated the highly correlated
results between SVD-2D modeling and active fault structure mapping published by
Fig. 7 Second Vertical Derivative (SVD) map with fault trend lines interpretation
Fig. 8 Second vertical derivative (SVD) map with trend interpretation with fault correlation of the
National Earthquake Study Center (PuSGen) and the slicing lines
PuSGeN (2017) [5] namely traversed by the Baribis–Kendeng Segment of Surabaya

fault (top) and the Waru segment (bottom). It is detected that the five faults inter-
preted on the Second Vertical Derivative (SVD) anomaly map have a similar trend
that is relatively close to the fault from the National Earthquake Study Center. Subse-
quently, slicing the SVD anomaly map was carried out to qualitatively determine the
direction of the fault, which was done by cutting perpendicularly from the fault.
In the first slicing (Fig. 9), resulting in a profile of SVD values, the maximum
value of SVD is 0.1435 mGal/m2 , this value is smaller than the minimum value which
is 0.1678 mGal/m2 , the negative sign on the minimum value curve is not predicted as
the value, but to distinguish the minimum and maximum values on the SVD curve.
Based on this analysis and the equation to determine the direction of the fault, it is
possible to identify the form of reverse fault structures.
In the second slicing that represented by Fig. 10, the maximum value of SVD
is 0.2076 mGal/m2 smaller than the minimum value which is 0.2668 mGal/m2 .
According to this analysis hence the form of reverse fault structure can be identified.
Fig. 9 Profile curve of the Second Vertical Derivative - 1
Fig. 10 Profile curve of the Second Vertical Derivative - 2
5 Conclusion
Based on the research that has been done, the following conclusions can be implied
as:
1. The results of gravity data processing using SVD 2-D modeling have a high
level of accuracy to interpret the fault, indicated by the highly correlated results
between SVD-2D modeling and active fault structure mapping published by
The National Earthquake Centre - PuSGeN (2017).
2. Slicing curve of the Surabaya Fault SVD profiling exhibited the SVD’s
maximum value is 0.1 mGal/m2 , this value is smaller than the minimum value
of 0.2 mGal/m2 which is identified as a reverse fault.
3. Slicing curve of the Waru Fault SVD profiling showed the maximum value of
SVD is 0.2 mGal/m2 smaller than the minimum value of 0.3 mGal/m2 which is
predicted as a reverse fault.
References
1. Bird P (2003) An updated digital model of plate boundaries. Geochem. Geophys Geosyst 4(3)
2. Koulali A, Susilo S, McClusky S, Meilano I, Cummins P, Tregoning P, Lister G, Efendi J,
Syafii MA (2016) Crustal strain partitioning and the associated earthquake hazard in the eastern
Sunda-Banda Arc. Geophys Res Lett 43:1943–1949
3. Gafoer S, Ratman N (1999) Peta Geologi Lembar Jawa Bagian Timur Skala 1:500.000
4. Supandjono JB, Hasan K, Panggabean H, Satria D, Sukardi (1992) Peta Geologi Lembar
Surabaya & Sapulu, Jawa
5. National Earthquake Study Centre (Pusat Studi Gempa Bumi Nasional), Indonesian Earthquake
Hazard and Source Map of 2017, Bandung
6. Widodo A, Syaifuddin F, Lestari W, Warnana DD (2020) Earthquake potential source identifi-
cation using magnetotelluric data of Kendeng thrust Surabaya area. In: E3S Web Conference,
vol 156, p 01002. https://doi.org/10.1051/e3sconf/202015601002
7. Sribudiyani IP, Muchsin N, Sapiie B, Ryacudu R, Asikin S, Kunto T, Harsolumakso AH,
Astono P, Yulianto I (2003) The collision of the East Java microplate and its implication for
hydrocarbon occurrences in the East Java Basin Proc. Indon Petrol. Assoc., 29th Ann. Conv.
Indonesian Petroleum Association (IPA). https://doi.org/10.29118/IPA.1530.03.G.085
8. Kearey P, Brooks M, Hill I (2002) An introduction to geophysical exploration, p 281 (2002)
9. Firdaus S, Setianto A (2018) Interpretasi Struktur Geologi berdasarkan Citra Landsat 8, SRTM
dan Anomali Medan Gravitasi Satelit di Cekungan Jawa Timur Utara, ISSN, p 12 (2018)
10. Sota I (2011) Estimation of fault structure using gravity method (Pendugaan Struktur Patahan
Dengan Metode Gaya Berat). Positron 1(1), (2011). https://doi.org/10.26418/positron.v1i1.
1565
11. Lestari W, Widodo A, Syaifuddin F, Warnana DD (2020) Research priority of the potential
earthquake on the java island using decision making analysis. In: E3S Web conference, vol
156, p 03003. https://doi.org/10.1051/e3sconf/202015603003
12. Rochman JPGN, Warnana DD, Widodo A, Syaifuddin F, Lestari W, Mahsa A (2018)
Application of gravity method for local geological structures identification in Surabaya. In:
EAGE-HAGI 1st Asia pacific meeting on near surface geoscience and engineering pp 1–5
13. Febriyansyah D, Haerudin N, Setiadi I, Study of hydrocarbon sub-basin patterns using spectral
decomposition analysis (Studi Pola Sub-Cekungan Hidrokarbon menggunakan Analisis Spec-
tral Decomposition, 2D Modeling and 3D Modeling based on Gravity Data for the Longiram
Region, East Kalimantan (Pemodelan 2D dan Pemodelan 3D berdasarkan Data Gayaberat
Daerah Longiram, Kalimantan Timur), p 16
14. Fitriastuti A, Aristo, Putri FF (2019) Identification of subsurface structures using gravity
method analysis of first horizontal derivative (FHD) and second vertical derivative (SVD),
for Earthquake disaster mitigation efforts in Wonosobo Regency, Central Java Province (Iden-
tifikasi Struktur Bawah Permukaan Menggunakan Metode Gaya Berat Analisis First Hori-
zontal Derivative (FHD) dan Second Vertical Derivative (SVD), Guna Upaya Mitigasi Bencana
Gempa Bumi di Kabupaten Wonosobo, Provinsi Jawa Tengah), The 12th Earth Science National
Proceeding (SeminarProsiding Seminar Nasional Kebumian ke-12), p 11
15. Vanbemmelen RW (1949) The geology of Indonesia, Government Printing Office, The Hague,
vol 1A
16. Hamilton W (1973) Tectonics of the Indonesian Region, pp 3–10
17. Parera AFT, Yusuf M (2015) Three dimensional modeling of gravity anomalies and identi-
fication of local faults in determining fault types in the Sidoarjo Region (Pemodelan Tiga
Dimensi Anomali Gravitasi dan Identifikasi Sesar Lokal dalam Penentuan Jenis Sesar di Daerah
Sidoarjo), p 10
18. Telford WM, Geldart LP, Sheriff RE (1990) Applied geophysics, 2nd edn. Cambridge
University Press
19. Hinze WJ, Von Frese R, Saad AH (2013) Gravity and magnetic exploration: principles,
practices, and applications. Cambridge University Press, New York
20. E. T. A. (1951) The second derivative method of gravity interpretation. Geophysics 23
21. Sumintadireja P, Dahrin D, Grandis H (2018) A note on the use of the second vertical derivative
(SVD) of gravity data with reference to Indonesian cases. J Eng Technol Sci 50(1):127–139
22. Setiadi I (2018) Pattern of subsurface geological structure and configuration of the North West
Java Basin based on gravity analysis (Pola Struktur dan Konfigurasi Geologi Bawah Permukaan
Cekungan Jawa Barat Utara Berdasarkan Analisis Gayaberat), p 14
23. Sarkowi M (2014) Gravity exploration (Eksplorasi Gayaberat), Graha Ilmu
24. Sari L, Sehah, Hartono (2018) Application of second vertical derivative (SVD) on gravity
data to identify the presence of faults along the South Serayu Mountains, Banyumas Regency
(Penerapan Second Vertical Derivative (SVD) Pada Data GravitasiUntuk Mengidentifikasi
Keberadaan Patahan di Sepanjang Pegunungan Serayu Selatan Kabupaten Banyumas). J Ter
Fis 1(2):8. https://doi.org/10.20884/1.jtf.2018.1.2.1127
25. Zain MA, Rozi MF, Septikasari AN, Nuruddianto M, Supriyanto, Zarkasyi A (2020) Study of
application of derivative analysis methods on gravity potential data (Studi Penerapan Metode
Analisis Derivatif pada Data Potensial Gravitasi). Open Sci Framework. https://doi.org/10.
31219/osf.io/m9dg7
Lesson Learned from Weathering Clay
Shale Residual Interface Shear Strength
Testing Method
Fathiyah Hakim Sagitaningrum, Samira Albati Kamaruddin, Ramli Nazir,

Budi Susilo Soepandji, and Idrus M. Alatas
Abstract Clay shale has been a problematic soil due to its weathering characteristic.
Weathering is the degradation of shear strength due to exposure to water and air. As
the depth of soil increases, the weathering effect would decrease, leading to different
weathering grades. On a slope, the different weathering grades would lead to an
interface slope failure. Until now, laboratory methods on finding the interface shear
strength between the weathering grades seldom been investigated. This research will
propose a method and give an evaluation of the results. An interface direct shear
test has been conducted until its residual state between an undisturbed clay shale and
weathering clay shale. The weathering condition of the bottom clay shale is achieved
by soaking and drying the bottom of the sample using a two-days wetting-drying
cycle. Results showed scattered values with different behavior for the soaked and
unsoaked conditions for increasing weathering days. The soaked condition resulted
in an increasing cohesion, increasing average stress ratio, and decreasing friction
angle. The unsoaked condition resulted in decreasing cohesion, increasing friction
angle, and increasing average stress ratio. Previous research shows that, the residual
shear stress should have decreased as the weathering days increases. The cause for
the contrary results is due to soil swelling and the absence of a separator when
F. H. Sagitaningrum (B) · S. A. Kamaruddin

Razak Faculty of Technology and Informatics, University of Technology Malaysia,
54100 Kuala Lumpur, Malaysia
S. A. Kamaruddin
R. Nazir
Centre of Tropical Geoengineering, University of Technology Malaysia,
81310 Johor Bahru, Malaysia
B. S. Soepandji
Civil Engineering Department, Faculty of Engineering, University of Indonesia,
Depok 16424, Indonesia
I. M. Alatas
National Science and Technology Institute, Jakarta 12630, Indonesia
524 F. H. Sagitaningrum et al.
weathering the bottom clay shale. In conclusion, the test conducted is not quite
suitable to understand the interface shearing behavior of weathering clay shale. Thus,
a modification is suggested for future research according to the identified causes.
Keywords Clay shale · Interface shear strength · Direct shear test · Residual shear
stress · Weathering
1 Introduction
Landslide is known to be a common problem in civil engineering. One of the causes of

landslide is the existence of problematic soil, such as clay shale within the soil profile.
In Java island, clay shale had been a challenging case for engineers due to its rapid
decreasing of shear strength. Some of the clay shale landslides recorded were the
Cipularang tollway landslide [1], Hambalang landslide [2], and the Semarang-Bawen
landslide [3].
Clay shale is known for its unique property of swelling and weathering, influenced
by its minerals [3, 4]. Not only due to its minerals, but weathering might also happen
due to exposure of clay shale to water and air. Due to weathering, the shear strength
of clay shale would decrease. However, the effect of weathering decreased as the
clay shale layer deepened. The different characteristics are defined as weathering
grades. When the clay shale layer got closer to the surface, the weathering grade
would increase, and the shear strength would decrease.
As the Semarang-Bawen slope, the different strengths of the soil layer would
initiate a landslide at the interface of the two soil [3]. However, the behavior of the
interface shear strength between the two soils are rarely discussed, especially for the
different weathering grades.
Thus, this research tried to propose the interface weathering method. The method
would describe the interface shear strength behavior of an undisturbed clay shale
with a weathering clay shale. It would also evaluate and offer suggestions for further
research.
2 Methodology
The proposed method involved an interface direct shear test on different weathering
grades of clay shale soil. The samples of clay shale were taken in an undisturbed
state. The samples were from the Semarang-Bawen toll road KM 483 + 200. It was
classified as the Kerek and Kaligetas formation clay shale, dominated by claystone
and shale classification [5].
Lesson Learned from Weathering Clay Shale … 525
From the undisturbed state, the samples were cut to fit the shear box. The samples
were divided into two sections of the shear box where the lower part of the shear box
would act as the weathering clay shale and the upper part would act as the undisturbed
clay shale. The schematic of the sample can be seen in Fig. 1.
The interface shear test was conducted as a single-stage direct shear test until
its residual state with three normal stresses: 35, 70, and 104 kN/m2 . Due to the
limitation of the device, the effect of fissures and pore water pressure reading was
ignored. Overall, the flowchart of the research can be seen in Fig. 2.
The consequence of using a single-stage direct shear test meant that the samples
could only be used once for each test. Two conditions determined the number of
samples: water at the interface and weathering days. The effect of water at the inter-
face differentiated the samples into two states: testing the sample five minutes after the
eight-days wetting–drying cycle and after one day of the eight-days wetting–drying
cycle. For each condition, two samples were determined. The other condition repre-
sents several cycles of the eight-days wetting–drying cycle. From previous research
in a similar location, the clay shale would completely slake after 48 days [6]. The
48 days were divided into an eight-day wetting–drying cycle with a two-time soaking
period. Thus, a total of six sample series were needed. In total, twenty-four samples
were needed.
Initially, four samples were tested beginning interface direct shear test until its
residual state. Later, the remaining samples were air-dried for two days. The bottom
part of the samples was then soaked for five minutes, followed by another drying for
four days. After soaking the bottom part of the samples for five minutes, different
treatments were used in the testing. Sample 1 and 2 were tested after air-drying the
samples for five minutes, whereas samples 3 and 4 were tested one day after the
soaking. Both tests were conducted until its residual state. The testing of the samples
is terminated after the 48 days of slaking was reached.
Fig. 1 Sample schematic for

the interface direct shear test
From the test, several results were extracted. Three parameters from the weathering
interface shear strengths: interface cohesion, interface friction angle, and interface
stress in the residual state. To understand the behavior of the parameters, a linear
trendline was used to observe the trend of the values.
Additionally, the average stress ratio is also considered. The average stress ratio
is the average ratio of shear stress and normal stress. The term was used to simplify
the moving interface stress between one weathering day to another.
3.1 Cohesion, Friction Angle, and Average Stress Ratio

at Residual
The interface cohesion results were shown in Fig. 3. From the graph, the values were
Fig. 2 Research schematic

of the interface weathering
direct shear test
generally scattered. The scattered values can be seen for both the soaked (Sample 1
and 2) and dried (Sample 3 and 4) conditions. The trend for the soaked conditions
was not similar but had a positive value. However, the dried conditions had two
different behavior. Sample 3 had a positive trend, whereas sample 4 had a negative
trend. Overall, both at soaked and dried conditions, the trend gave a positive value.
It means that as the weathering days increases, the interface cohesion increases.
The interface friction angle results are shown in Fig. 4. From the graph, the
values were generally scattered. The scattered values can be seen for both the soaked
(Sample 1 and 2) and dried (Sample 3 and 4) conditions. Contrary to the interface
cohesion, the interface friction angle had a distinct behavior for the soaked and
drying conditions. As seen in the graph, both the soaked samples exhibit a positive
trend, whereas the dried samples exhibit a negative trend. It means that right after a
wetting-drying cycle, the friction angle at the interface is higher than after the soil
had some time to dry itself. The trend was consistent with the increasing weathering
days, meaning that the testing condition of the sample is also crucial when portraying
the field condition.
Direct shear used a linear equation to determine the cohesion and friction angle.
Thus, the stress behavior might not be fully representing the shear strength condition.
The average stress ratio was used to understand the stress behavior throughout the
test comprehensively. The results for the average stress ratio were shown in Fig. 4.
From the graph, the trend shows similar trend for both the soaked and dried
conditions. The values were generally scattered for all of the samples. The trend for
all of the samples showed a positive value. It means that the average stress ratio of
both the soaked and dried conditions when testing would increase as the weathering
days increases.
Fig. 3 Interface cohesion for increasing weathering days

3.2 Discussion
From the test results, the values were seen to be consistently scattered. Thus, evaluation
for regression, R2 values, were needed. The results were summarized in Table 1.
From the R2 values, most of the samples were indeed very scattered. The precision
of the data was not good. Identification of the reasons from the scattered values was
done.
First of all, it was assumed that the samples were prepared similarly. Thus, the
samples would have given similar results in return. The samples were indeed from
the same location, same method, same undisturbed condition, and went through the
same cutting method. However, even when the samples were similarly treated, the
results did not behave similarly. A simple linear trendline was established, but the
trend shows differently between one another.
Fig. 4 Interface friction angle for increasing weathering days
Table 1 R2 values for the

Samples Interface Interface Interface average
tests
cohesion friction angle stress ratio
Sample 1 0.0320 0.0006 0.0452
Sample 2 0.4684 0.1326 0.4089
Sample 3 0.0020 0.3375 0.1007
Sample 4 0.1010 0.3255 0.4686
Second, the results indicated that using a single-stage method might not be the
correct method. Both the values and trends were random. Theoretically, past research
showed that with the increase of weathering days, the residual shear stress of clay
shale should have decreased [3]. But, as shown in Fig. 5, the residual stress increases
instead. There were several reasons for the behavior, but the significant factor might
be due to the nature of clay shale as a swelling soil. As the soil swells when immersed
in the water, the tested interface when shearing might have changed. Thus, the shear
stress observed might have been the fresh clay shale, the weathered clay shale, or
the accurate interface. Clay would naturally bind the particles, which effects seen at
the opened interface after testing in Fig. 6.
Mesri pointed out that there should be a thin boundary that separates the two
samples of direct shear before testing [7]. The partition would keep the interface as
the shearing plane and act as a stopping point of the water effect. As seen in Fig. 6,
the top part of the undisturbed clay shale was also affected by the water from the
bottom weathered clay shale. Thus, the interface shear strength might not be precise.
Fig. 5 Interface average stress ratio for increasing weathering days

Fig. 6 Visual of interface

for different weathering
degree after testing in soaked
condition
Sample 1 Sample 2
4 Conclusion
For determining the interface shear strength of different weathering grades of clay
shale, a method was proposed. The results of interface cohesion, interface friction
angle, and the average stress ratio from the interface direct shear test was evaluated.
Several conclusions were determined from the analysis:
1. Values of the interface cohesion, interface friction angle, and average stress ratio
was scattered. The trend of both the soaked and dried conditions was random.
The R2 values indicated that the results of the test had low precision.
2. As most of the trendline indicated a positive trend, the results were not in
agreement with previous research [3] where the cohesion and friction angle
should have decreased.
3. The adverse trends of the results might have been due to the swelling tendency
of clay shale. The swelling would disrupt the shearing plane at the interface.
4. No partition was used in the interface direct shear test. Thus, water from the
bottom part of the shear box was able to seep through the interface to the upper
undisturbed clay shale. Water would reduce the interface shear strength and
cause the undisturbed clay shale to deteriorate.
5. Finally, the single-stage shearing method is not be suitable for the weathering
process.
It can be concluded that the interface direct shear test for the weathering clay
shale was not suitable to obtain the objective. Future research should consider the
lessons learned to enhance the proposed method. Future improvement should at least
consider the use of the multistage direct shear method, include a pore-water pressure
measurement on the test, use of thin partition to separate the upper and bottom part
of the sample and use a separate weathering phase from the interface direct shear
test.
The future research suggestions would give no water intrusion to the upper undis-
turbed part of the clay shale and confirm that the shearing would be at the interface
of the different weathering grades.
Acknowledgements First Author is a PhD student funded with International Doctoral Fund (IDF)
from University of Technology Malaysia.
References
1. Irsyam M, Susila E, Himawan A (2007) Slope failure of an embankment on clay shale at

KM 97+500 of the Cipularang toll road and the selected solution. In: International symposium
on geotechnical engineering, ground improvement, and geosynthetics for human security and
environmental preservation, Thailand, pp 531–540
2. Widjaja B (2008) Engineering characteristics of Bukit Sentul clayshale based on laboratory and
in situ tests. In: Huang, Mayne (eds) Geotechnical and geophysical site characterization, pp
1231–1237. Taylor & Francis Group, London
3. Alatas IM, Kamaruddin SA, Nazir R, Irsyam M (2016) Effect of weathering on disintegration
and shear strength reduction of clay shale. Jurnal Teknologi 78(7–3):93–99
4. Sadisun IA, Shimada H, Ichinose M, Matsui K (2005) Study on the physical disintegration
characteristics of Subang claystone subjected to a modified slaking index test. Geotech Geol
Eng 23:199–218
5. Thaden RE, Sumadirdja H, Richards PW (1975) Geologic map of the Magelang and Semarang
quadrangles, Java. In: Geological Survey of Indonesia, Ministry of Mines, Jakarta
6. Alatas IM (2017) Kesan luluhawa terhadap kekuatan ricih syal lempung dalam penentuan
parameter kestabilan cerun. Universiti Teknologi Malaysia, Malaysia
7. Mesri G, Huvaj-Sharihan N (2012) Residual shear strength measured by laboratory test and
mobilized in landslides. J Geotech Geoenviron Eng 138(5):585–593
Shear Strain Evaluation on Analysis
of Additional Clay Liner Layer Modeling
in Ngipik Landfill, Gresik
Siti Nurlita Fitri
Abstract The discovery of new faults in the East Java region has led to different
perspectives of analyzing dynamic infrastructures including the landfills and one of
the major factors in understanding the local site effect is the shear strain. The Ngipik
landfill in Gresik, East Java has been using an open dumping method which involves
direct disposal of wastes into the surface pit since 2002. There is, therefore, the need to
implement strategies to avoid groundwater contamination and one of the appropriate
methods is the addition of clay liner. Therefore, this research applied different forms
of this method in four simulated modelings and four profiles to assess its efficiency in
terms of the surface area dynamic properties using finite element analysis. The result
showed the shear strain reduced by approximately 40% from the existing condition.
This, therefore, shows the importance of adding a layer such as the clay liner in open
dumping landfill.
Keywords Clay liner · Shear strain · Dynamic
1 Introduction
Several factors are included in the calculations associated with civil structures, espe-
cially the dynamic calculations. For example, the development of seismicity in the
eastern region of Java significantly influences infrastructural analysis as indicated
in the new faults recently identified by Kuncoro et al. [1]. This is in the form of
the new active Kendeng fault located along Central-East Java with a strain rate of
0.99–1.93 mm/year. This further strengthens the importance of the local site effect
in evaluating an infrastructure, including a landfill.
The local site effect is the reaction of soil to seismic motion on the ground surface.
It has been reported that the most critical step in determining the possible character-
istics of earthquakes is to assess the ground motion attributes on the surface when
accounting for the conditions of local geological and geotechnical sites [2].
S. N. Fitri (B)
Civil Engineering Department, Universitas Sebelas Maret, Surakarta, Indonesia
534 S. N. Fitri
Ngipik landfill is located in Gresik, East Java, and the waste treatment system
applied is an open dumping system without a protective layer. Meanwhile, one of the
layers recommended by the Environmental Protection Agency (EPA) is the clay liner.
This is a layer of cohesive soil with certain parameters which is alternatively used to
prevent leachate from contaminating the groundwater [3]. Moreover, Soemitro et al.
[4] found the geological formation of the Lidah Structure with Quarterage where the
landfill is located to be consisting of clay which is black and blackish, hard when
warm, small-fossil content, and sandstone clay. It also has a fairly thick clay content
on the layer and this makes the area vulnerable to earthquakes.
Dynamic analysis including simulation and evaluation has been conducted on
the landfill in several countries and the results showed the landfill has a significant
contribution to groundwater pollution due to the large strain on the protection layer
of the solid waste. Moreover, the maximum shear strain produced in the landfill for
low-frequency input movements is significantly higher due to the fact that the seismic
inertial forces at all depths are in phase at lower frequencies. It is also possible to
reduce the stiffness of the saturated soils through repetitive shear strain based on
the consideration that soil stiffness modifications have the ability to drive long-term
results [5, 6].
The preliminary data of the Ngipik landfill dynamic properties have been deter-
mined using microtremor analysis in previous studies. Fitri et al. [7] showed the
shear strain is approximately 3.5 × 10−3 to 12.5 × 10−3 with high vulnerability to
dynamic failure observed in the east-southwest area. The soil vulnerability index was
also estimated to be 2.04–7.16 with the same prone area [8].
This research was, therefore, conducted to determine the effect of adding clay
liner to the new model of municipal solid wastes in Ngipik Landfill through the use
of dynamic analysis. Several models of clay liner were analyzed to determine the
shear strain value on the surface zone despite the existence of values from previous
research in order to determine the effectiveness of clay liner construction.
2 Methodology
2.1 Ngipik Landfill Description
The geotechnical properties were evaluated using the previous 3 borehole investi-
gations [9]. The SPT results showed the soil mainly consists of 51% clay and 44%
silt and a small amount of sand at 4%, high plasticity (mostly more than 50% with
PI), and an average porosity estimated at 50%. The location of the borehole test and
profile of soil are presented in Fig. 1.
Shear Strain Evaluation on Analysis … 535
BH3
BH1
BH2
(a)
BH-1 BH-2 BH-3
0.00 m SOFT
-5.00 m MEDIUM
-10.00 m VERY STIFF
-15.00 m HARD
-20.00 m
-25.00 m
-30.00 m
45.00 m 55.00 m
(b)
Fig. 1 a Location of SPT test, b Soil stratigraphy in Ngipik Landfill
Figure 1a shows the location of the SPT test while Fig. 1b describes the stratig-
raphy of the research site which was determined through statistical analysis of the 3
bored holes. The SPT investigation of BH1, BH2, and BH3 show the soil has clay
with different consistencies and this involved the soft and medium at the top surface
while those at 15 m depth are dominated by stiff clays. Meanwhile, the soil stratig-
raphy was used to determine the properties needed for the landfill simulation model
as indicated in Table 1.
Table 1 The soil properties of the models

Soil properties k (m/day) C ϕ γ (kN/m3 ) μ E
Clay liner 0.00864 30 0 13.1 0.35 4500
Solid waste 3.58 × 10−2 10 15 3 0.25 1200
536 S. N. Fitri
Ngipik Landfill is located in Gresik and has been functional since 2002. It has
a total landmass of 6 ha but only 4 ha are being used for disposal with the landfill
estimated to be 10–12 m high. Moreover, the municipal solid waste mound is placed in
the clay soil deposited before the waste is dumped without treatment or compaction.
Several models of the mound were, however, simulated according to the existing site
as shown in Fig. 2.
Figure 2 shows the 5 models with different shapes and approximately 115–160 m
widths to be analyzed. Moreover, the addition of clay liner divides the cross-section
of the landfill into 5 parts as shown in Fig. 3.
Figure 3 shows the solid waste mound models with some variations of the clay
liner. The first is the existing model without a clay liner layer, the second is the clay
liner design with 1 m thickness around the solid waste mound while the third and last
models were simulated with 2 m thickness of clay liner as recommended by EPA.
Finite Element Model was used to analyze the behavior of the site due to dynamic
modeling while the dynamic properties were determined using Probabilistic Seismic
Hazard Analysis. The results showed the Peak Ground Acceleration (PGA) in the
landfill bedrock with a 2% chance of being exceeded in 50 years or a 2475-year
earthquake return duration is 0.2 g [10]. The initial ground motion was used for the
bedrock while the Nonlinear Earthquake site Response Analysis (NERA) software
was applied to determine the ground motion to model the dynamic properties on the
surface.
FEM model was simulated using Plaxis 2D at a constant-coefficient of α = 5.5
and β = 5 while the average natural frequency on the surface ground was found to be
1.9 Hz based on a previous study [8]. The mound of landfills with dynamic properties
which were determined based on the graph of acceleration vs time in NERA software
was applied as an input model in Plaxis and the result produced is the shear strain
value in the geometry models. Meanwhile, the final result was obtained from the
section of shear strain in the surface layout using all the landfill models and profiles.
This research was conducted on the dynamic analysis of a finite element method with
the profile variations of municipal solid mound waste in different clay liner models
analyzed. The results are, therefore, presented in the following Figs. 4, 5, 6, 7 and 8.
(a)
12m 7m 12m
5m
70m 30m 40m 30m 20m 30m 35m
140m 115m
(b) (c)
5m 2m
7m 10m
10m
3m
3m
30m 30m 30m 30m 30m 40m 20m 25m 40m
160m 125m
(d) (e)
10m
80m 40m
120m
(f)
Fig. 2 Solid waste mound in Ngipik Landfill. a The sections of the landfill, b Profile 1, c Profile
2, d Profile 3, e Profile 4, f Profile 5
538 S. N. Fitri
Waste
Exsisting Soil
(a)
Clayliner 1m
Waste
1m
Exsisting Soil
(b)
Clayliner 2m
Waste
2m
Exsisting Soil
(c)
Clayliner 2m
Waste
2m
Exsisting Soil
(d)
Fig. 3 Addition of clay liner, a Model 1, b Model 2, c Model 3, d Model 4
Figure 4 shows the shear strain parameter in the surface around profile 1 and
the results of the dynamic parameters in several conditions were discovered to have
different shear strain values as indicated in Fig. 1. The maximum shear strain was
estimated to be from 6 × 10−3 up to 10−2 and, according to Nakamura et al. [11],
10−2 indicates the site is collapse while 10−3 means there is crack and settlement. A
slight reduction in the maximum shear strain was, however, observed in Models 1 to
4 of profile 1 and this means the dynamic conditions do not have an influence on the
site.
The summary of all the results is presented in Fig. 9 with the graph observed
to illustrate the relationship between shear strain and clay liner modeling in several
profiles of the solid waste mound. All the profiles have the same trend of falling
slightly from Models 1–4. Moreover, the maximum change in shear strain was found
to be 40% of the initial condition due to the addition of 2 m clay liner as indicated in
Profile 4 while the minimum was 2% as observed in Profile 1. This was attributed to
the difference in the shapes of mound and soil properties. Furthermore, the maximum
result was discovered in Profile 2 due to its location as the highest area of amplification
Fig. 4 Shear strain analysis of Profile 1
[8]. The increase in the shear strain indicates a better condition based on dynamic
behavior. It is, however, important to note that a value ranging between 10−6 and 10−5
indicates an elastic characteristic of the soil and this usually leads to the vibration
on the surface. Meanwhile, a higher shear strain value of 10−2 –10−1 indicates a
collapsed structure due to dynamic conditions. This, therefore, means the site has
540 S. N. Fitri
a significant effect in identifying seismic ground movements which significantly

intensify or deamplify at the last moment just before impacting the surface of the
ground or the basement of man-made structures [12].
Fig. 9 Shear strain in all the models and profiles
4 Conclusion
Clay liner addition in Ngipik open dumping landfill was modeled in this study and the
results showed similarity in all variations. The maximum difference was observed to
be up to 40% of the existing condition and this was discovered in the 2 m clay liner
model. This means it is important to consider the use of clay liner as a replacement
material. It is, however, recommended that further study needs to analyze basic local
sites in different geological areas.
542 S. N. Fitri
References
1. Kuncoro H, Kartini GAJ, Meilano I, Susilo S (2019) Identifikasi Mekanisme Sesar Di

Bagian Timur Pulau Jawa Dengan Menggunakan Data Gnss Kontinyu 2010–2016. Semin
Nas Geomatika. https://doi.org/10.24895/sng.2018.3-0.1069
2. Parvez IA, Rosset P (2014) The role of microzonation in estimating earthquake risk. In:
Earthquake hazard, risk and disasters
3. Agency EP (2000) Landfill manuals: landfill site design. Environmental Protection Agency,
Ireland
4. Soemitro NSRAA, Warnana DD, Characterization of leachate distributions at Ngipik municipal
solid waste disposal site
5. Pain A, Annapareddy VSR, Sarkar S (2019) Seismic stability analysis of municipal solid waste
landfills using strain dependent dynamic properties. Indian Geotech J. https://doi.org/10.1007/
s40098-018-0314-6
6. Zania V, Tsompanakis Y, Psarropoulos PN (2008) Seismic distress and slope instability of
municipal solid waste landfills. J Earthq Eng. https://doi.org/10.1080/13632460701574605
7. Fitri SN, Soemitro RAA, Warnana DD (2020) Mitigasi Bencana Berdasarkan Nilai Shear Strain
Tanah Akibat Gempa Bumi Pada Tempat Pembuangan Akhir (Tpa) Ngipik, Gresik. Rekayasa
Sipil 14(3):227–231
8. Nurlita Fitri S, Asih Aryani Soemitro R, Dewa Warnana D, Sutra N (2018) Application of
microtremor HVSR method for preliminary assessment of seismic site effect in Ngipik landfill,
Gresik. In: MATEC web of conferences. https://doi.org/10.1051/matecconf/201819503017
9. Sutra N et al (2020) The impact of open dumping method in Ngipik landfill investigated with
electrical resistivity tomography (ERT) and very low-frequency electromagnetic (VLF-EM).
Int J GEOMATE. https://doi.org/10.21660/2020.73.52066
10. Irsyam M et al (2020) The new national seismic hazard maps for design of buildings and infras-
tructures in Indonesia. In: 16th Asian regional conference on soil mechanics and geotechnical
engineering, ARC 2019
11. Nakamura Y, Sato T, Nishinaga M (2000) Local site effect of Kobe based on microtremor. In:
Proceedings of sixth international conference Seism. Zo. EERI, 12–15 November 2000, Palm
Springs, California
12. Chen Y, Wu Z (2004) International handbook of earthquake and engineering seismology. Acta
Seismol Sin. https://doi.org/10.1007/bf03191403
Tsunami Hazard in Cilacap City Due
to the Megathrust of West-Central
Java Segment
Wahyu Widiyanto and Sanidhya Nika Purnomo
Abstract The source of the earthquake and tsunami in the Indian Ocean, particularly
in South Java, is needed for mitigation purposes that are beneficial to life on the
expanding south coast of Java. Cilacap City is one of the important cities in driving
development in south coast of Java. This study aims to perform a tsunami simulation
with the study area of Cilacap City. The simulations are carried out with software
which is based on the shallow water equations. The epicenter varies in the form of
coordinate points in the megathrust segment of west-central Java. The simulation
results show that Cilacap City is prone to tsunami disasters because the tsunami
amplitude on the coast can reach 10.7 m and the arrival time of the tsunami is
between 10 and 20 min. These results may complement previous data assessing the
potential for tsunamis along the Sumatran Java Ring of Fire.
Keywords Tsunami hazard · Megathrust · Java · Cilacap
1 Introduction
The megathrust that stretched from the northern part of Sumatra Island to the islands
of Nusa Tenggara is the source of the earthquakes and tsunamis that are very intense.
Earthquakes seem to occur regularly that is often felt by people on the west coast of
Sumatra, the south coast of Java and Bali and Nusa Tenggara. Likewise, although
with a lower frequency of occurrence, tsunamis have become the recognized type of
disaster in the region. The greatest tsunami in history that occurred in modern era
has certainly not disappeared from our memory, namely the 2004 Aceh Tsunami [1].
W. Widiyanto (B) · S. N. Purnomo

Department of Civil Engineering, Universitas Jenderal Soedirman, Purwokerto, Indonesia
S. N. Purnomo
Department of Hydraulic and Ocean Engineering, National Cheng Kung University,
Tainan, Taiwan
544 W. Widiyanto and S. N. Purnomo
Before and afterwards, tsunamis also occurred in the megathrust zone of the same ring
of fire as the Aceh tsunami, i.e. the 1992 Maumere tsunami [2], the 1994 Banyuwangi
tsunami [3, 4], the 2006 Pangandaran tsunami [5] and the 2010 Mentawai tsunami
[6]. While studies and mitigation measures have been carried out, more extensive
studies and follow-up are needed to raise the parties up prepare for possible disasters.
Therefore, this study aims to provide information on the potential for a tsunami
disaster in the city of Cilacap.
Cilacap is a densely populated city with a variety of vital economic activities on the
south coast of Java. This city is administratively located in Central Java Province. The
population of Cilacap region in 2020 which is living in the coastal area are 655,000.
The Cilacap coastal area (Fig. 1) is the center of economic activity, containing a cargo
port, a fishing port, two large power plants, a cement factory and an oil refinery. In
addition, there is a Pendem Fort, which has cultural heritage status and is a valuable
asset that cannot be easily valued in currency as well as the Indonesian naval base.
With this in mind, it is important to evaluate the city of Cilacap against the risk of a
tsunami.
The assessment of the potential tsunami height is carried out at critical facilities
in Cilacap City. These facilities are described below. The first is Nusakambangan
Prison. Nusakambangan Prison is the best super tight security prison in Indonesia.
This prison is located on the island of Nusakambangan, one of the outer islands of
Indonesia. The Nusakambangan Island in the north used to be a water area called
Segara Anakan. Meanwhile, the southern part is directly facing the Indonesian Ocean.
When it was first established, Nusakambangan Prison was a prison for political land.
Currently, however, Nusakambangan Prison is used for big criminals and terrorist
7
6
5
2
3 4
Fig. 1 Study area covers Cilacap City and its surroundings. Important points include (1) prison at
Nusakambangan Prison, (2) Port of P.T. Pelindo III, (3) PT. Pertamina, (4) Fort Pendem, (5) Cilacap
Ocean Fishing Port (PPSC), (6) Karangkandri PLTU, and (7) Adipala PLTU
Tsunami Hazard in Cilacap City Due … 545
prisoners. The oldest prison in Nusakambangan is Batu Prison, built in 1925. In

the 2006 Pangandaran tsunami, one of the prisons near the south coast, Permisan
Prison, recorded the highest increase of any other tsunami-hit area. Futhermore, the
second assessed point is P.T. Pelindo III, a state-owned company (BUMN) operating
in the transportation sector, having a role in the management and control of 43 ports
in 7 Indonesian provinces. Established on December 1, 1992, P.T. Pelindo III has
businesses and services in the form of increasing connectivity and competitiveness
between regions across Indonesia, marine pilot and towing services, marine logistics,
canal and basin dredging. The company also serves oil and natural gas raw materials,
supplies terminals and energy (oil and gas), power plants, shore connections, fuel
supply and waste management.
The evaluation has also been conducted on the third point, namely P.T. Pertamina
(Persero) Refinery Unit IV Cilacap. This company is one of seven series of processing
units in Indonesia with the largest production capacity of 348,000 barrel per day and
the most comprehensive facilities. This refinery is a strategic value as it supplies 34%
of the national fuel requirement or 60% of the fuel requirement of Java Island. The
refinery, built in 1974, also produces asphalt and base oil for Indonesia’s infrastructure
needs. The fourth point is the fortress of Pendem. This object can be found in Teluk
Penyu, Cilacap. This fort was built by the Dutch in 1861. At the time, Pendem Fort,
built in stages over an area of 6.5 ha over 18 years, served as defense against attacks
from the southern side of Cilacap. Currently, Pendem Fort is a cultural heritage as
well as a historical and cultural tourism object.
The fifth point is the Cilacap Ocean Fishing Port (PPSC). This port complex is
located in the village of Tegalkamulyan in the South Cilacap district. This port is
the only large fishing port on the south coast of Java and directly faces the Indian
Ocean. This port is expected to become a center for the growth and development
of an integrated fishing industry. The sixth site is PLTU Karangkandri, also called
PLTU 2 Central Java. This is a steam power plant that has been built since December
29, 2003 with an investment value of USD 510 million, with a capacity of up to
6,000 megawatts. The last point evaluated is PLTU Adipala, which is a power plant
managed by P.T. Indonesia Power with a capacity of 1 × 660 MW. In contrast to
other PPDE generators, this plant uses a Supercritical Boiler with a steam pressure
of up to 25.4 MPa.
Many tsunami studies for the regions of Sumatra and Java were conducted from
different research perspectives. The tsunami documentation in the form of field
survey results was conducted by [1, 3, 7, 8] for the Banyuwangi tsunami. While [2,
5, 9, 10] provide field survey data for the Pangandaran tsunami. The recent tsunami
triggered by flank collapse of anak Krakatau volcano occurred in Sunda Strait in
2018 was reported by [11].
This study is a brief analysis to assess whether Cilacap City is susceptible to
tsunamis and how high a potential tsunami could be. The possible causes of the
earthquake and tsunami are limited to the West-Central Java segment (Fig. 2).
10°0’0’’N
Mgt. Aceh-Andaman
Mgt. Nias- Mgt. North Su- Mgt. Philippines
Simelue lawesi
Mgt. Batu
Mgt. Papua
Mgt. Mentawai-
0°0’0’’
Siberut
Mgt.
Mentawai-
Pagai
Legend
10°0’0’’S
Mgt. Enggano Subduction

Mgt. West- Fault
Mgt. Selat Sunda Central Java Mgt. East Java Mgt. Sumba Segment
95°0’0’’E 110°0’0’’E 125°0’0’’E 140°0’0’’E
Fig. 2 Earthquake sources in Indonesia (Mgt. = Megathrust). This study uses megathrust of West-
Central Java segment (see green box) as tsunamigenic to assess tsunami hazard in Cilacap city (
Source of the figure: PuSGeN 2017)
Table 1 shows the coordinates of each assessed location. These coordinates are
near the beach, but not directly on the coastline. The coordinate for location 1 is right
on the edge of the Indonesian Ocean. Location 2 and 3 have a point at the mouth of
the Donan River. Point 4 is directly across from Indonesian Ocean, but has north to
south direction so that the coastline is in the same direction as the dominant wave
propagation direction. Whereas the locations 5, 6 and 7 are harbors, the evaluation
point is thus at the port mouth between the two breakwaters.
Table 1 Assessed location

No. Site name Long. (°) Lat. (°)
1 Lapas Permisan Nusa Kambangan 108.882833 −7.745443
2 Pelabuhan P.T. Pelindo—TNI AL 108.987000 −7.735000
3 Dermaga Kilang Minyak P.T. Pertamina 109.004000 −7.748000
4 Benteng Pendem 109.022000 −7.750000
5 Pelabuhan Perikanan Samudera Cilacap 109.025000 −7.732000
6 PLTU Karangkandri 109.082642 −7.695936
7 PLTU Adipala 109.129000 −7.696000
2 Method
In this study, tsunami waves propagation were calculated by the long wave equation
or shallow water equation. Considering that the simulation area is large enough,
a spherical coordinate system was applied. Whereas the numerical solution uses a
leap-frog scheme which can be expressed:
n+1/2 n−1/2
ηi, j − ηi, j 1
n
Pi+1/2, j − Pi−1/2, j
n
1
+ +
t R cos ϕ i, j ψ R cos ϕ i, j
n n n+1/2 n−1/2
cos ϕi, j+1/2 Q i, j+1/2 − cos ϕi, j−1/2 Q i, j−1/2 h i, j − h i, j
=− (1)
ϕ t
n+1/2 n+1/2
j − Pi+1/2, j ηi+1, j − ηi, j
n+1 n
Pi+1/2, gh
+ − f Q i+1/2,
n
j =0 (2)
t R cos ϕ i+1/2, j ψ
n+1/2 n+1/2
j+1/2 − Q i, j+1/2
Q i,n+1 ηi, j+1 − ηi, j
n
gh
+ + f Q i,n j+1/2 = 0 (3)
t R i, j+1/2 ϕ
where η is the water surface elevation; (P, Q) denote the volume fluxes in X (West–
East) direction and Y (South-North) direction, respectively; (ϕ, ψ) denote the latitude
and longitude of the Earth; R is the radius of the Earth; g is the gravitational accel-
eration and h is the water depth. And the term –∂h/∂t reflects the effect of transient
seafloor motion, can be used to model landslide-generated tsunamis. F represents
the Coriolis force coefficient due to the rotation of the Earth.
Equations (1)–(3) are then solved by the COMCOT (Cornell Multi-grid Coupled
Tsunami model) [13]. This model is one of the most widely used tsunami models
and has successfully modeled many tsunami cases. Tsunami generation originating
from seabed deformation was modeled with elastic finite fault plane theory. This
approach requires input in the form of epicenter, focal depth, fault length and width,
strike angle, dip angle, slip angle, rake angle, and slip.
The tsunami propagation model from source to points assessed requires bathy-
metric data. This study implements Batnas (6 arcsec resolution) as bathymetry data.
These data are sourced from the Geospatial Information Agency (Indonesian: BIG).
Bathymetry data were used to form the simulation layer. In this study, three layers
of simulation are used which form a nested grid system. COMCOT uses explicit
staggered leap-frog finite difference schemes to solve numerical Eqs. (1)–(3). From
this simulation a tsunami wave height will be obtained in the assessed area, in this
study is Cilacap City.
3 Result
The results of the running model are processed to show the area affected by the
tsunami. Figure 3 shows the beginning of the tsunami generation in the form of
sea level changes due to the deformation of the seabed as a result of the Mw 8.7
earthquake. The red color represents the rise of the water level and the blue color
represents the drop of the water level, so it is a wave that has a peak and a valley
that represents the start of the tsunami wave. Furthermore, this wave propagates at
high speed, so it slows down when it reaches the coastal area. Wave energy turns into
rising waves and creeps upward in the form of attempts. Figure 4 shows the tsunami
height of about 2–6 m approaching the south coast of Ciamis, Tasikmalaya, Garut
and Cianjur districts as it spread about 10 min after the earthquake. The tsunami
will slow down and reach the coast of Cilacap in about 36 min after it is generated
(Fig. 5).
The detection of the water level provides the information that the wave height
varies at the examined points. As can be seen in Fig. 6, the Permisan Nusa Kambangan
Prison experienced a tsunami that was not too high, 2.1 m. At sites 2 and 3 in the
river, no rise in the water level was detected. This perhaps because the numerical
measurement point is not placed at the correct elevation, for example considered on
Fig. 3 The beginning of the Time = 122.53 s

simulation, a change in sea
level due to the deformation
of the sea floor by the
Latitude (S)
earthquake Mw 8.7. The red
meter
color stands for the rise in
the water level and the blue
color for the fall of the water
level from the MSL
Longitude (E)
Fig. 4 Tsunami propagation Time = 1250.83 s

process after 660 s from the
earthquake. Dots circle
indicates tsunami is reaching
Latitude (S)
the shoreline
meter
Longitude (E)
Fig. 5 The tsunami reached Time = 2203.85 s

the city of Cilacap after
traveling from the source of
the earthquake in the
Indonesian Ocean
Latitude (S)
Longitude (E)
Fig. 6 The simulation 15 Permisan Nusakambangan

results take the form of a rise Benteng Pendem
Water surface elevation (m)
Pelabuhan Ikan PPSC
in the water level due to the 10 PLTU Karangkandri
tsunami at several observed PLTU Adipala
points in the city of Cilacap 5
0
0 2000 4000 6000
-5
-10
-15 Time (s)
land by the Batnas data. The propagation of the tsunami into the river actually causes
a significant increase in river water level over a considerable distance as in the case
of the 2011 Tohoku Japan Tsunami.
The next location, Benteng Pendem, based on simulations is expected to expe-
rience a tsunami with an amplitude of 1.8 m. This is not too high, but needs to be
paid attention, considering that Pendem Fort is the location of an old building with a
structure without reinforced concrete. A moderate run-up can not only collapse but
also scatter a building that is predominantly masonry and brick. It needs to remark,
an investigation by [12] of the 2018 Palu tsunami found that tsunamis 3.5–4 m high
can displace 56 tons of reinforced concrete bridges. The low tsunami in front of
Pendem Fort was likely due to the wave refraction, which caused the wave direction
to be deflected from the original direction by almost 90°, reducing energy and wave
height. For the last 3 locations, namely locations 5, 6 and 7, the height of the tsunami
that occurred with a strength of more than 3 m was included in the category of the
major tsunami, more precisely 3.2, 6.7 and 10.7 m, respectively. These number is
consistent with result by [13]. The risk of a tsunami attack could damage the break-
water in the three ports. The simulation results at the seven locations show that the
Adipala PLTU has the highest potential of tsunami wave, which in this computation
produces ≈11 m. However, it seems that this has been eagerly awaited since the
PLTU was designed. For example, the breakwater chosen is quite heavy, 25 tons
each of the protected layer stone [14]. However, this breakwater structure has yet to
be prepared to be tested by a large wave, in this case a tsunami.
For damage control purposes, the arrival time of a tsunami is also a parameter
that needs to be known. The results of this study show that a tsunami will hit this
location within 45–60 min. Especially for the LP Permisan location, the arrival time
will be half an hour earlier. Early warnings can be announced by the authorities and
processed by the community for security reasons.
4 Conclusion and Recommendation
The results of the tsunami simulation show that the coast of Cilacap has significant
potential for a tsunami. Medium height tsunamis can occur around Fort Pendem and
the Cilacap Ocean Fishing Port (PPSC). Even at moderate height, the tsunami will
hit a densely populated area with casualties. Higher tsunamis that can occur at the
PLTU Karangkandri and the PLTU Adipala are 6.7 and 10.7 m high, respectively.
The strong force could destroy the breakwater and pier, although the power house
facilities on land are likely to remain safe. To find out more about the effects of the
tsunami on land, a tsunami run-up must be simulated. High-resolution topographic
data is required for this calculation. Tsunami arrival times between 45 and 60 min
provide an opportunity for early warning measures that can save live.
This study is limited to the use of tsunami sources in the megathrust segment of
west-central Java. In order to assess it more accurately, it is necessary to consider
the source of the earthquake and the tsunami for other nearby megathrust segments,
such as the Enggano segment in the western part and the East Java segment, and
consider different scenarios. A more detailed analysis, covering many possibilities,
can be done by making a probabilistic tsunami hazard assessment.
References
1. Borrero JC (2005) Field survey of northern Sumatra and Banda Aceh, Indonesia after the
Tsunami and earthquake of 26 December 2004. Seismol Res Lett 76:312–320
2. Hidayat D, Barker JS, Satake K (1995) Modeling the seismic source and tsunami generation of
the December 12, 1992 Flores Island, Indonesia, earthquake. Pure Appl Geophys PAGEOPH
144:537–554
3. Tsuji Y et al (1995) Field survey of the East Java earthquake and tsunami of June 3, 1994. Pure
Appl Geophys PAGEOPH 144:839–854
4. Maramai A, Tinti ST (1994) The 3 June 1994 Java Tsunami: a post-event survey of the coastal
effects. Nat Hazards 1997. https://doi.org/10.1023/A:1007957224367
5. Lavigne F et al (2007) Field observations of the 17 July 2006 Tsunami in Java. Nat Hazards
Earth Syst Sci 7:177–183
6. Lay T et al (2011) The 25 October 2010 Mentawai tsunami earthquake (Mw 7.8) and the
tsunami hazard presented by shallow megathrust ruptures. Geophys Res Lett 38:2–6
7. Borrero JC et al (2009) The tsunami of 2007 September 12, Bengkulu province, Sumatra,
Indonesia: post-tsunami field survey and numerical modelling. Geophys J Int 178:180–194
8. Borrero JC et al (2011) Field survey of the March 28, 2005 Nias-Simeulue earthquake and
Tsunami. Pure Appl Geophys 168:1075–1088
9. Fritz HM et al (2007) Extreme runup from the 17 July 2006 Java tsunami. Geophys Res Lett
34:1–5
10. Imamura F, Gica E, Takahashi T, Shuto N (1995) Numerical simulation of the 1992 Flores
tsunami: Interpretation of tsunami phenomena in northeastern Flores Island and damage at
Babi Island. Pure Appl Geophys PAGEOPH 144:555–568
11. Widiyanto W et al (2020) Run-up, inundation, and sediment characteristics of the 22 December
2018 Sunda Strait tsunami, Indonesia. Nat Hazards Earth Syst Sci 20:933–946
12. Widiyanto W, Santoso PB, Hsiao SC, Imananta RT (2019) Post-event field survey of 28
September 2018 Sulawesi earthquake and tsunami. Nat Earth Syst Sci 19:2781–2794
13. Widiyantoro S et al (2020) Implications for megathrust earthquakes and tsunamis from seismic
gaps south of Java Indonesia. Sci Rep 10:1–11
14. Widiyanto W (2014) Assessment of armor layer stone weight of breakwater in ports of
logending, PLTU Adipala and Cikidang. Civ Eng J. https://doi.org/10.24002/jts.v12i4.637
15. Pusat Studi Gempa Nasional (PuSGeN) (2017) Peta Sumber dan Bahaya Gempa Indonesia
Tahun 2017 (Map of Indonesia Earthquake Sources and Hazards in 2017). Pusat Peneli-
tian dan Pengembangan Perumahan dan Permukiman, Badan Penelitian dan Pengembangan,
Kementerian Pekerjaan Umum dan Perumahan Rakyat, Bandung. (in Bahasa Indonesia)
Assessment of the Conditioning Factor
for Flash Flood Susceptibility Potential
Based on Bivariate Statistical Approach
in the Wonoboyo Watershed in East Java,
Indonesia
Entin Hidayah , Gusfan Halik , and Wiwik Yunarni Widiarti
Abstract Flash floods that occur suddenly, which cause damage to the weirs or
embankments, immediately threaten human life. Identifying the causes of a flash
flood is very important to reduce its negative impact. This paper examines changes
in flash flood disasters in the Wonoboyo watershed based on estimates of flash flood
hazard, land-use changes, and rainfall depth distribution patterns. The method of
predicting susceptibility to flash flood hazards is based on various environmental
factors that are integrated with GIS. Three bivariate statistics consisting of the Statis-
tical Index (SI), Frequency Ratio (FR), and Predictor Rate (FP-PR) model are applied
to select the best Flash Susceptibility Index (FFHSI) model. Changes in land use are
then explored based on the conditioning factor for a flash flood. In the final stage, the
estimation of areal rainfall uses Inverse Distance Weighting (IDW) to describe the
position of rain and flash flood events. The best statistical bivariate statistical approach
for the FFHSI is FR. Assessment of environmental factors using the FFHSI shows
that 21% of the catchment area has moderate to high until very high vulnerability
levels. Changes in land cover significantly affect flash floods, especially changes
from forest to agricultural land or settlements. The distribution pattern and intensity
of rainfall are closely related to the location of the flash flood. This study results can
guide future flood mitigation measures.
Keywords Flash flood · FFHS · Land-use change · Rainfall distribution ·

Wonoboyo watershed
E. Hidayah (B) · G. Halik · W. Y. Widiarti

University of Jember, Jember 68121, Indonesia
G. Halik
W. Y. Widiarti
554 E. Hidayah et al.
1 Introduction
Flash floods are local floods that occur suddenly with a high, difficult to predict peak
flow [1, 2]. The occurrence of high rainfall intensity triggers flash floods. This kind
of flood occurred in the rivers in the Wonoboyo watershed in East Java province.
Flash floods in a short time can burst the embankments, damage the infrastructure,
and generate runoff that flows over settlements and rice fields in the watershed area
[3–7].
In addition to the rainfall, the natural terrain condition is a driving factor of the
susceptibility to flash floods. A steep topography drives the water down to the estuary
more quickly, endangering the river bed stability [8]. Especially during extreme
rain-fall, flooding in this area is unavoidable. Other driving variables of the floods
are often meteorological conditions in general, including low atmospheric pressure
and land breezes that can cause storms and the entrance of air containing water
vapor to the catchment area that produces heavy rains [9]. Furthermore, according to
Pradhan [10], both the height and the slope of the hills significantly affect the floods
occurrence, namely that steeper slopes resulting in faster flows. Accurate estimation
of flood hazard susceptibility is essential to measure the relationship among the
various spatial and temporal driving factors.
Furthermore, several studies suggest that changes in land use, distribution, and
rain intensity are also the causes of floods. The land-use change in urban areas’
growth contributes to increased flash floods [11]. The land-use change induces a
hydrological response in the watershed in the form of a higher peak flow rate and a
shorter concentration time, increasing the difficulty to avoid flash floods.
Several researchers have used various approaches and environmental factors that
affect flash floods to identify the flash flood potentials. The commonly used factors
are slope, convergence, Topographic Wetness Index (TWI), profile curvature, river
networks, density, and land use. These factors have shown significant results to obtain
potential susceptibility [12–15]. On the other hand, some researchers also used the
NDVI factor.
In addition to determining the relevant factors, using the proper method is also
very important in selecting a capable flash flood hazard model. Previous research
has applied several methods including: Analytic Hierarchy Process (AHP) [16–18],
decision tree analysis (DT) [15], weight of evidence [19, 20], Statistical Index [21],
Entropy Shannon [21], and FR [20, 22]. The decision tree method provides better
performance than AHP [12]; however, the decision tree method does not consider
the temporal changes of various factors such as SPI, land use, and dynamic NDVI
[15]. Compared to other methods, the relatively simple FR method applied in several
flash flood studies has shown excellent training and validation results [9, 12, 20].
In addition, the FR method has provided good results for mapping the landslide
hazard in the Wonoboyo watershed [23]. SI model gave the highest flood prediction
result of 98.72% compared to weighting factors and the Shannon entropy model with
predictive rates of 97.6 and 92.42% [21].
Assessment of the Conditioning Factor for Flash Flood Susceptibility … 555
Based on the area and recording of flash flood events in the Wonoboyo watershed,
this paper aims to develop a flash flood susceptibility map in the Wonoboyo water-
shed. The best model is obtained by comparing three methods: SI, FR, and FR-PR
integration to find a better and more accurate model. Next, the impact of the causes
of a flash flood is assessed with changes in land use and the areal rainfall events in
the study area.
2 Study Area, Floods, and Flood Inventory Mapping
2.1 Study Area
Wonoboyo watershed, East Java Province, between 113°50 E and 113°56 E, and
between 7°49 N and 7°46 N (Fig. 1) is the study area. This area is topographically
hilly, flanked by two mountains, namely Mt. Malang and Mt. Rajekwesi, of which
the heights are 1177 and 762 m above sea level, respectively. The watershed covers
an area of 3968 ha.
Fig. 1 Study area and flood inventory mapping at Wonoboyo Watershed

The climate has a distinct tropical monsoon with two seasons, namely rainy and
dry seasons. The rainy season starts in October and ends in March. The dry season
begins in April and ends in September. Based on the Agency of Civil Work of East
Java Province, the mean annual rainfall in the study area varies from 460 to 2319 mm.
2.2 Historical Records of Floods and Flood Inventory

Mapping
The historical records of the flood event were obtained from field surveys and various
mass media reports. Base on the information gathered, 9 flash floods occurred in two
decades. Typically, flash floods used to occur during heavy rain that lasted three days.
The events and the impacts of the flash floods are described in Table 1.
The sampling of flash flood events using field observation data. The results in
Fig. 1 show that 8859 pixels with a surface area of 44 ha are flood-affected areas
with 78% for training and 22% for validation. The pixel size is 30 × 30 m.
3 Methodology
The methodology implemented in this study comprised of the following main steps:
(i) determination of flood conditioning factors for flooding; (ii) estimating, evalu-
ating, and mapping the best flash flood susceptibility index by comparing methods:
SI, FR, and integrating FR-PR; (iii) statistical analysis of the changes in land covers
and analyzing the distribution of rainfall on 6 flood events.
3.1 Flood Condition Factor
Flash flood parameterization is complicated because a flash flood has many causes
generally interrelated. For an integrated analysis of the factors that influence a flash
flood, it is crucial to determine data collection in essential variables to build a base
map representing the conditioning factors for the flash flood in the area under study.
Based on previous studies, this study uses 10 factors associated with the flash flood,
which are shown in Table 2. [13–15, 19]. The 10 factors are slope, Topography
Position Index (TPI), TWI, profile curvature, plan curvature, aspect, convergence
index, the density of river networks, land cover, and NDVI. Of the ten factors, 7
geomorphological factors are derived from DEM: slope, TPI, TWI, profile curvature,
plan curvature, aspect, and convergence index. Geomorphological data were obtained
from the National DEM with a resolution of 8 m × 8 m. 2 other factors, namely land
use and NDVI, are derived from Landsat 8 in 2018 and Landsat 7 in 2000, a resolution
Table 1 The historical record on the rainfall depths, the affected villages, and the flood hazards on several flash floods
Flood day evens (rain gauge) 18-Jan-08 31-Jan-16 29-Jan-17 27-Jan-18 25-Jan-19 31-Oct-20
Historical Bluncong D-2 75 0 23 7 0 0
rainfall depth D-1 7 44 0 67 22 0
(mm)
D 27 15 96 190 74 55
Total 109 59 119 264 96 55
Wonoboyo D-2 75 0 0 9 18 0
D-1 7 50 0 18 17 0
D 27 150 100 120 50 80
Total 109 200 100 147 85 80
Exposure village Wonoboyo and Wonoboyo Wonoboyo and Pandak, Leprak, Wonoboyo Leprak, and
Rajekwesi, Leprak and Wonoboyo Wonoboyo
Impact of flash flood The breakdown of The breakdown The breakdown The breakdown Small Overflow into settlement, road, rice
the leave, and the of the leave, and of the leave and of the weir, and landslide field damage
overflow of flood overflow into damage to rice overflow into
Assessment of the Conditioning Factor for Flash Flood Susceptibility …
into roads and settlements fields the bridge,

settlements mosque, and
settlement
557
Table 2 Identification driving factors of the flash flood susceptibility

Factor The preparation procedure of preparation of each
factor in Fig. 2 and its relationship with flood
susceptibility
Slope (SL) According to Pradhan [10], slope significantly
affects flooding, where steeper slopes produce
heavier flow. On the other hand, flat lowlands
appear more prone to flooding. The Wonoboyo
watershed ranges in slope from 0.008° to 52.602°
Topography position index (TPI) TPI represents a morphometric index value that
describes the difference in height between focused
cells and neighboring cells [19]. The TPI value in
the Wonoboyo watershed ranges from −11.46 to
10.38
Topographic wetness index (TWI) The TWI value describes water accumulation’s
tendency to slope according to gravity, which
controls flow direction. A high TWI value
indicates that an area is increasingly vulnerable to
saturated soil surfaces and areas that have the
potential to produce surface runoff [24–26].
Equation (1) is the formula for TWI
[As]
T W I = ln tanβ (1)
Whereas represents the upstream contribution area,
and β represents the slope angle. The TWI value in
the Wonoboyo watershed ranges from 3.151 to
19.013
Profile curvature (Prof. C) Curvature of the profile is a morphometric factor in
the form of surface curvature in the vertical plan
along the slope [27]. The profile curvature value at
3.516
Plan curvature The curvature of the plan is another morphometric
(Plan C.) factor that affects runoff. This parameter is in the
form of a contour line that is formed between the
horizontal plane and the ground surface. The value
is the difference between the size of the area and
the runoff (affects the surface runoff) and those
with convergent runoff (the area’s infiltration
capacity) [9, 10, 27]. The plan curvature value at
4.737
Aspect (A) Aspect is the direction of the slope. The output
raster value is the compass direction of the aspect
factor. The aspect factor in the Wonoboyo
watershed is grouped into eight classes
(continued)
Table 2 (continued)
Factor The preparation procedure of preparation of each
factor in Fig. 2 and its relationship with flood
susceptibility
Coverage index (CI) CI is a terrain parameter that shows the relief
structure as a set of convergence areas (channels)
and divergent areas (ridges The CI value ranges
from −93.116 to 92.441
River density (RD) RD affects the level and intensity of flooding
significantly. The denser the network and the area
around the river, the more prone the area to
flooding events [28, 29]. Network density is the
division of the flow length (m) to the watershed
area (km2 ). The RD value ranges from −93.116 to
92.441
Land use (LU) Land use contributes to infiltration rates in forests
and vegetated areas that support infiltration. In
contrast, residential areas and grasslands
accumulate surface runoff due to impermeable
layers that reduce the infiltration capacity. The five
land cover classes identified for reclassification are
scrubs, followed by plantation area, forest,
settlement, and water
Normalized difference vegetation index NDVI is an index used to measure vegetation
(NDVI) characteristics in an area [30]. NDVI can
sufficiently represent the decrease in land cover
based on the image’s local brightness [31].
Equation (2) is a formula to compute NDVI [32]
N DV I = (N I R−V I S)
(N I R+V I S) (2)
VIS shows the spectral reflectance measurement
obtained in the visible wave region in red (band 5).
NIR shows the spectral reflectance measurements
obtained in the infrared wave region (band 4). The
NDVI value ranges from −1 to 1 [33]
of 30 m × 30 m. The river network map is obtained from the delineation of Indonesian

Topographical Map data at a scale of 1:25,000. Hydro-climatological data in the form
of rain data are obtained from the Bondowoso Technical Operating Unit of Water
Resource Management (UPT PSDA) and the Agency of Public Works and Water
Resources of Bondowoso Regency.
8 factors in the form of numerical variables are classified into several classes
using statistical quantile classification according to the data grid’s value [15]. 2 other
factors, namely land use and aspect, are nominal variables and are classified based
on the feature class.
Fig. 2 Maps of flood hazard

susceptible conditioning
factors, a S, b TPI, c TWI, d
Prof. C, e Plan C. f A, g CI,
h RD, i LU, j NDVI
Fig. 2 (continued)
Fig. 2 (continued)
Fig. 2 (continued)
3.2 Estimation of the Flash Floods Hazard Susceptibility

Index
The estimation of flash flood hazard is based on statistical observations between the
inventory of flood events and the flood conditioning factors. This stage comprises four
steps (1) Weighting each class’s factors using various statistics bivariate methods. The
weights for each class of geomorphological and environmental factors are computed
by using GIS, resulting in FFHSI values. (2) The validation of the model is carried
out by comparing the Area Under Curve (AUC) values between the flash flood hazard
susceptibility predicted by the model with the actual flash flood locations from the
field survey. The values of the indices were then grouped into six intensity classes
by using natural break reclassification.
Statistical Index
The SI model is a bivariate statistical approach introduced by Van Westen [34]
Initially, this model was to estimate landslide hazards. Recently, this model has been
developed to be applied to flash flood hazard areas [9, 12]. The SI method describes
the weight value for the parameter class in the form of the natural logarithm of the
flash flood density in that class divided by the flash flood density across the map.
The equation used is as in Eq. 3 [34]:

Li j
Di j LT
Wi j = ln( ) = ln (3)
D Pi j
PL
whereas Wij is the weight assigned to class i given from factor j; Dij is the flash flood
density in class i from factor j; D shows the totalflash flood density on the entire
map; L ij is the number of pixels with flash flooding in class i from factor j; LT is the
total number of flash floods on the entire map; Pij is the number of pixels in class i
of parameter j, and PL is the total number of pixels on the entire map.
The results of the calculation can be interpreted that if the value of Wij is positive,
it indicates a viable and strong relationship between the class and the flash flooding
distribution. The higher the score, the stronger the relationship. Conversely, if the
Wij value is negative, there is no correlation between the class and the occurrence of
flooding. Furthermore, the FFHSI value can be calculated by adding all the weights
for the number generated as in Eq. 4.
R=n
F F H S I SI = Wi j (4)
R=1
Frequency Ratio and Predictor Rate Method

The frequency ratio method was used in [9, 10] to estimate the probabilistic relation-
ship between dependent variables and the independent variables and to determine
the weighting coefficient values in each flood-related variable class Eq. 5 shows a
formula used to estimate the values in Frequency Ratio.
m n
F R = Fi Fi / A j Aj (5)
i−1 j=1
where Fi is the number of pixels with flash floods for each class of each factor; Aj is
the number of pixels for each class of each factor; m is the number of classes in the Fi
factor; n is the number of factors in the study area [35]. The FR value interpretation
shows that if the value is more than 1, there is a correlation between the class and
the occurrence of flooding. The higher the value, the stronger the relationship.
Based on the calculations for the FR sole conditioning factor, FFHSI can be
calculated by aggregating the FR of all factors by using Eq. 6:
n
F F H S IF R = FR (6)
j=1
The next step is to calculate the relative frequency (RF) using Eq. 7 by normalizing
the FR value (probability value range {0–1}).

F Ri j
R F = m (7)
i=1 F R i j
In RF, all factors have the same weight; FR-PR integration considers the disad-
vantages of these. Weights using the level predictor (PR) are calculated by assessing
each flood conditioning factor with the training data set in Eq. 8 [34]
(R F max − R F min )
PR = (8)
(R F max − R F min ) Min
Furthermore, the FFHSI value is the sum of the results of each FP-PR factor with
each RF class as in Eq. 9.

F F H SI PR = (P R × R F) (9)
3.3 Preparation of Datasets for Training and Validation
This stage evaluates and compares the training and prediction capability. The last
step includes validating the model based on statistical metrics such as percentage of
sensitivity, and specify, as shown in Eqs. 10 and 11.
The sensitivity indicates the proportion of incidents correctly classified as flash
floods. Accuracy means the balance of incidents correctly classified as flash floods
and correctly classified as non-flash-floods.
TP
sensitivit y(y) = (10)
T P + FN
TN
speci f it y(x) = (11)
FP + T N
where TP (True Positive) and TN (True Negative) indicate correctly classified

incidents and FP (False Positive) and FN (False Negative) indicate wrongly classified
incidents [13]. The ROC plot is a sensitivity versus specificity graph, statistically
unaffected by prevalence. It is calculated by estimating the parameters for the infinity
limit. The Area Under the Curve (AUC) is estimated from the ROC curve as Eqs. (12)
show the overall performance of the vulnerability model [36]. An AUC value that
is close to 1.0 indicates that the model is very good and can classify the flash flood
area correctly.
n+1 1
AU C = (xi − xi+1 )2 × (yi − yi+1 ) (12)
i=1 2
3.4 Estimation of Rain Spatial Distribution by Using IDW
Rainfall is the main factor that causes flash floods. The higher the rainfall intensity,
the higher the flooding potential will be. For a study on the causes of flash floods in an
area, spatial rain distribution is essential for describing the locations where the flash
floods occurred. The spatial rain distribution is estimated by using Inverse Distance
Weighting (IDW) method [36]. IDW is a spatial interpolation method suitable for the
prediction of rainfall possibility [37]. Besides, estimates of mean regional rainfall

using the IDW method are slightly better than the forecast resulting in the other
three ways (ANUDEM, Spline, and Kriging). The IDW method is also easier to
implement in GIS [38]. IDW interpolation determines cell values using a weighted
combination of a set of sample points. The weight is a function of the inverse distance.
The data recorded by these devices are used to estimate mean regional rainfall. The
study focused on determining flash flood causes. The estimate of spatial rainfall
distribution in various times of occurrence (in Table 1) of flash floods occurrences is
used to identify the effect of rainfall on the flash floods historically.
4.1 Assessment of FFHS from SI, FR, and FR-PR
In this paper, the potential of the flash floods on the Wonoboyo watershed has been
assessed by using SI, FR, and FP-PR a bivariate statistical approach integrated with
geospatial techniques. These methods are used to find a correlation between a flash
flood in its causes.
In the SI method, all factors are calculated to obtain the resultant weight for each
conditioning factor, as shown in Fig. 3. The slope factor in classes 1 and 2 has a
positive value, and the lower the class, the stronger the value. TPI in classes 2–4
shows positive values. TWI has positive scores in grades 4–7, and the higher the
class, the stronger the relationship. The curvature of the profile with the curvature
of the plan has the opposite positive value strength. Aspects have positive values
in classes 6, 7, 8, and 10. Convergent in class 1 and 2 has a positive and strong
value in class 1. River density indicates that the lower the class, the stronger the
relationship. The use of paddy fields and settlements has a strong positive value. The
Fig. 3 The computed SI value variations for each class of the conditioning factors
Fig. 4 The computed FR value variations for each class of the conditioning factors
NDVI in classes 1 and 3 had positive scores, and the lower the score, the stronger
the relationship.
The FR method applied to measure the prediction capability from selected flash
flood driving factors is shown in Fig. 4. Figure 4 shows that the dominant factors
that drive flash floods are land use, followed by aspect. The other six other factors
(river density, TWI, slope, NDVI, TPI, convergence index) have equal importance.
The land-use factor is very closely related to flash flood occurrence. It mainly occurs
in the rice fields and settlements class because it has the two highest FR values.
The map in Fig. 5 shows the estimation and identification of flash flood potentials.
The map shows that the most susceptible areas to flash floods are the ones with the
lowest slopes. The lower the slope, river density, NDVI, and convergence index, the
higher the flash flooding potential.
Figure 5 shows a graph of the FR-PR method applied to measure the FP-PR of
the selected flash flood driving factors. It shows the dominant factors that cause flash
floods are density (3.76), land use (3.73), followed slope (3.26). The other six factors
are NDVI (2.14), Convergence index (2.13), TPI (1.59), Prof Curvature (1.38), TWI
(1.26), Plan Curvature (1.23), and Aspect. (1.00).
Figure 6 and 7 show a graphical comparison of the hazard levels for each model.
The FR and FR-PR models have almost the same hazard pattern at moderate to high
Fig. 5 Flash flood

conditioning factors from
predictor rate
Fig. 6 Map of FFHSI at

Wonoboyo Watershed with
method: a SI, b FR, and c
FR-PR
Fig. 7 Comparison of level of hazard between SI, FR, and FP-PR
to very high levels. These two models show that a very low to moderate hazard level
has a high percentage value, but has decreased from a moderate to high to very high
hazard level. On the other hand, the SI model has small percentage values at very low
to moderate hazard levels but shows high percentage values at moderate to high until
very high hazard levels downstream of the watershed. In general, from the modeling
results, the highest percentage value is found in the FR method with a low hazard
level with a value of 38%, while the lowest percentage value is found in the FR
method with a very high hazard level with a value of 2%.
4.2 Validation of Flash Flood Hazard Susceptibility Map
The validation of the FFHSI model shows that the model has performed well for the
SI model and very well for the FR and FR-PR. The AUC value represents this for
success curve of 91.06%, and the AUC value for the prediction curve of SI, FR, and
FR-PR is 79.38, 90.75, and 90.40, respectively. FR model is the best model from
others. Therefore, the prediction of flash flood hazard susceptibility resulting in this
the study can be used as a reference for flood disaster management plans for disaster
management and decision makers.
Table 3 The change in land use

Land-use Area (%) % Change in land use from 2000 conditions
2000 2018
Rice field 4.6 4.7 2.84
Plantation 58.5 58.2 −0.5
Forest 17.4 17.3 −1.06
Settlement 1.6 2.0 19.84
Shrubs 17.8 17.8 0.13
4.3 Effect of Change in Land Uses to Flash Floods

Occurrences
The change in land cover is analyzed by using a statistical approach. The land-use
data of the year 2000 is used as the baseline, and the data of the year 2018 is used to
show the change in land use (Table 3).
Of the total Area at Wonoboyo watershed of 4008 ha, the land uses most consid-
erable proportion is for plantations (58%). The land use between 2000 and 2018 does
not show any significant change in the total area ratio. The most significant change in
land use was the 19.84% increase in settlement, followed by rice fields’ rise and the
decrease in plantations and forests. Based on Table 2, there is no significant evidence
that the change in land use had conditionend the flash floods. Nevertheless, the rise of
rice fields and settlements will increase the surface runoff and the infiltration capacity
of an area, hence conditionend flooding.
4.4 The Effect of Rainfall on the Flash Floods Occurrence
Figure 7 shows the distribution pattern of rainfall depth and the distribution of six
flash flood historical occurrences in 2008, 2016, 2017, 2018, 2019, and 2020. It
shows that the distribution of flash flood occurrences is related to the distribution of
the rainfall depths. Based on Table 1, the rainfall depth of 55 mm has conditioned
the flash floods (Fig. 8).
The flash flood that was triggered by consecutive rainfall for 3 days, as shown in
Table 1, would have been worse if one of the variables had become extreme. Based
on this study’s results, it is essential to consider factors that affect the occurrence of
flash floods to determine the proper preparation and mitigation to minimize the risks
of future flash floods.
Fig. 8 The rainfall distribution pattern and the flash floods occurrences in Wonoboyo Watershed
5 Conclusions
Based on the comparative analysis of three statistical bivariate methods, the frequency
ratio approach is the best method to describe the flash flood potentials, as implied
from the AUC model value of 90.75. It is found that land use, the density of river
networks, and slope are three dominant factors that conditionend the flash floods.
Among the classes in the land-use factor, rice fields and settlement have the most
significant effects. As for the factor of the density of the river network, it is found
that the denser the river network is, the more susceptible it is to flash floods. Besides,
the distribution pattern of rain and the intensity will determine the location of flash
floods.
Rain intensity of more than 50 mm per day is the threshold in the occurrence of
flash floods. The results of this study can be used as guidelines in the management
of watershed spatial planning and steps to minimize the occurrence of flash flood.
References
1. Elkhrachy I (2015) Flash flood hazard mapping using satellite images and GIS tools: a case study
of Najran City, Kingdom of Saudi Arabia (KSA). Egypt J Remote Sens Sp Sci 18(2):261–278.
https://doi.org/10.1016/j.ejrs.2015.06.007
2. Hapuarachchi HAP, Wang QJ, Pagano TC (2011) A review of advances in flash flood
forecasting. Hydrol Process 25(18):2771–2784. https://doi.org/10.1002/hyp.8040
3. Antarafoto (2008) Banjir Desa Wonoboyo, antarafoto
4. BPBD (2016) BPBD Jatim Bantu Korban Banjir dan Tanah Longsor di’, BPBD. https://web.
bpbd.jatimprov.go.id/2016/02/01/bpbd-jatim-bantu-korban-banjir-dan-tanah-longsor-di-bon
dowoso/. Accessed Feb 27 2021
5. Beritaekspress (2017) Akibat Banjir, Akses ke Desa Wonoboyo Jatim Lumpuh, beritaekspres
6. Terbitannews (2020) Akibat Hujan Deras, Banjir dan Longsor Terjadi di Desa Leprak
Bondowoso, Terbitannews
7. Faktualnews (2018) 3 Kecamatan di Bondowoso Diterjang Banjir Bandang, faktual-
news. https://faktualnews.co/2018/01/28/3-kecamatan-bondowoso-diterjang-banjir-bandang/
60730/. Accessed 27 Feb 2021
8. Gaume E, Borga M (2008) Post-flood field investigations in upland catchments after major
flash floods: proposal of a methodology and illustrations. J Flood Risk Manag 1(4):175–189.
https://doi.org/10.1111/j.1753-318x.2008.00023.x
9. Cao C, Xu P, Wang Y, Chen J, Zheng L, Niu C (2016) Flash flood hazard susceptibility mapping
using frequency ratio and statistical index methods in coalmine subsidence areas. Sustainability
8(9) (2016). https://doi.org/10.3390/su8090948
10. Pradhan (2009) Flood susceptible mapping and risk area delineation using logistic regression,
GIS and remote sensing. J Spat Hydrol 9(2):1–18
11. Jodar-Abellan A, Valdes-Abellan J, Pla C, Gomariz-Castillo F (2019) Impact of land use
changes on flash flood prediction using a sub-daily SWAT model in five Mediterranean
ungauged watersheds (SE Spain). Sci Total Environ 657:1578–1591. https://doi.org/10.1016/
j.scitotenv.2018.12.034
12. Costache R, Hong H, Pham (2020) QB (711) Comparative assessment of the flash-flood poten-
tial within small mountain catchments using bivariate statistics and their novel hybrid integra-
tion with machine learning models. Sci Total Environ 711:134514. https://doi.org/10.1016/j.
scitotenv.2019.134514
13. Bui DT, Tsangaratos P, Ngo PTT, Pham TD, Pham BT (2019) Flash flood susceptibility
modeling using an optimized fuzzy rule based feature selection technique and tree based
ensemble methods. Sci Total Environ 668:1038–1054. https://doi.org/10.1016/j.scitotenv.2019.
02.422
14. Popa MC et al (2020) Spatial assessment of flash-flood vulnerability in the Moldova river
catchment (N Romania) using the FFPI. J Flood Risk Manag 2019:1–10. https://doi.org/10.
1111/jfr3.12624
15. Khosravi K et al (2018) A comparative assessment of decision trees algorithms for flash flood
susceptibility modeling at Haraz watershed, Northern Iran. Sci Total Environ 627:744–755.
https://doi.org/10.1016/j.scitotenv.2018.01.266
16. Lappas I, Kallioras A (2019) Flood susceptibility assessment through GIS-based multi-criteria
approach and analytical hierarchy process (AHP ) in a River Basin in Central Greece, IRJET,
pp 738–751
17. Stefanidis S, Stathis D (2013) Assessment of flood hazard based on natural and anthropogenic
factors using analytic hierarchy process (AHP). Nat Hazards 68(2):569–585. https://doi.org/
10.1007/s11069-013-0639-5
18. Kazakis N, Kougias I, Patsialis T (2015) Assessment of flood hazard areas at a regional scale
using an index-based approach and analytical hierarchy process: application in Rhodope-
Evros region, Greece. Sci Total Environ 538:555–563. https://doi.org/10.1016/j.scitotenv.2015.
08.055
19. Costache R (2019) Flash-flood potential assessment in the upper and middle sector of Prahova
river catchment (Romania). A comparative approach between four hybrid models. Sci Total
Environ 659:1115–1134. https://doi.org/10.1016/j.scitotenv.2018.12.397
20. Khosravi K, Nohani E, Maroufinia E, Pourghasemi HR (2016) A GIS-based flood suscepti-
bility assessment and its mapping in Iran: a comparison between frequency ratio and weights-
of-evidence bivariate statistical models with multi-criteria decision-making technique. Nat
Hazards 83(2):947–987. https://doi.org/10.1007/s11069-016-2357-2
21. Khosravi K, Pourghasemi HR (2016) Flash flood susceptibility analysis and its mapping
using different bivariate models in Iran: a comparison between Shannon’s entropy, statistical
index, and weighting factor models. Environ Monit Assess. https://doi.org/10.1007/s10661-
016-5665-9
22. Costache R, Zaharia L (2017) Flash-flood potential assessment and mapping by integrating the
weights-of-evidence and frequency ratio statistical methods in GIS environment—case study.
Bâsca Chiojdului River Catchment (Romania). J Earth Syst Sci 126(4):1–19. https://doi.org/
10.1007/s12040-017-0828-9
23. Efendi D, Hidayah E, Hasanuddin A (2020) U KaRsT, Mapping of landslide susceptible zones
by using frequency. Ratios Bluncong 5(1):126–141. https://doi.org/10.30737/ukarst.v3i2
24. Tehrany MS, Pradhan B, Jebur MN (2015) Flood susceptibility analysis and its verification
using a novel ensemble support vector machine and frequency ratio method. Stoch Environ
Res Risk Assess 29(4):1149–1165. https://doi.org/10.1007/s00477-015-1021-9
25. Beven K (1984) Infiltration into a class of vertically non-uniform soils, Hydrol Sci J 29(4):425–
434. https://doi.org/10.1080/02626668409490960
26. Manfreda S, Di Leo M, Sole A (2011) Detection of flood-prone areas using digital elevation
models. J Hydrol Eng 16(10):781–790. https://doi.org/10.1061/(asce)he.1943-5584.0000367
27. Duman TY, Can T, Gokceoglu C, Nefeslioglu HA, Sonmez H (2006) Application of logistic
regression for landslide susceptibility zoning of Cekmece Area, Istanbul, Turkey. Environ Geol
51(2):241–256. https://doi.org/10.1007/s00254-006-0322-1
28. Fernández DS, Lutz MA (2010) Urban flood Hazard zoning In Tucumán Province, Argentina.
Using GIS and multicriteria decision analysis. Eng Geol 111(1–4):90–98. https://doi.org/10.
1016/j.enggeo.2009.12.006
29. Glenn EP, Morino K, Nagler PL, Murray RS, Pearlstein S, Hultine KR (2012) Roles of Saltcedar
(Tamarix spp.) and capillary rise in salinizing a non-flooding terrace on a flow-regulated Desert
River. J Arid Environ 79:56–65. https://doi.org/10.1016/j.jaridenv.2011.11.025
30. Falahnsia AR (2014) Vegetasi dengan metode skoring menggunakan citra satelit di, pp 400–416
31. Ullah K, Zhang J (2020) GIS-based flood hazard mapping using relative frequency ratio method:
a case study of Panjkora River Basin, Eastern Hindu Kush, Pakistan. PLoS One 15(3):1–18.
https://doi.org/10.1371/journal.pone.0229153
32. Ullah K, Zhang J, Id JZ (2020) GIS-based flood hazard mapping using relative frequency ratio
method: a case study of Panjkora River Basin, Eastern Hindu Kush, Pakistan. PLoS 15(3):1–18.
https://doi.org/10.1371/journal.pone.0229153
33. Abdulwahid WM, Pradhan B (2017) Landslide vulnerability and risk assessment for multi-
hazard scenarios using airborne laser scanning data (LiDAR). Landslides 14(3):1057–1076.
34. Van Westen CJ, Rengers N, Terlien M (1997) Prediction of the occurrence of slope instability
phenomena through GIS-based hazard zonation. https://doi.org/10.1007/s005310050149
35. Regmi AD et al (2014) Application of frequency ratio, statistical index, and weights-of-evidence
models and their comparison in landslide susceptibility mapping in Central Nepal Himalaya.
Arab J Geosci 7(2):725–742. https://doi.org/10.1007/s12517-012-0807-z
36. Bartier PM, Keller CP (1996) Multivariate interpolation to incorporate thematic surface data
using inverse distance weighting (IDW). Comput Geosci 22(7):795–799. https://doi.org/10.
1016/0098-3004(96)00021-0
37. Chen FW, Liu CW (2012) Estimation of the spatial rainfall distribution using inverse distance
weighting (IDW) in the middle of Taiwan. Paddy Water Environ 10(3):209–222. https://doi.
org/10.1007/s10333-012-0319-1
38. Yang X, Xie X, Liu DL, Ji F, Wang L (2015) Spatial interpolation of daily rainfall data for
local climate impact assessment over Greater Sydney Region. Adv Meteorol. https://doi.org/
10.1155/2015/563629
Flood Management Strategies
in Indonesia: A Lesson Learned
from Pepe River, Central Java
Rian Mantasa Salve Prastica and Amalia Wijayanti
Abstract Floods occur every year and threaten communities worldwide. Indonesia
is among the countries faced with this problem and still struggling to seek practical
solutions to date. Further, as the country’s centers of trade, business, and technology,
the island of Java cannot even perform better in dealing with floods. To further under-
stand possible damage and losses, this research investigated structural mitigation
measures as part of flood management in the Pepe River, notably river channeliza-
tion to increase hydraulic efficiency. For the analysis, it used HEC-RAS to observe
the river capacity. Existing conditions revealed that the river could not accommodate
the predicted floods based on the return periods derived from the synthetic unit hydro-
graph of Nakayasu. However, after running the mitigation scenario wherein hydraulic
modification was applied, flood risk could be reduced by lowering the river water
level. Based on the results, it can be concluded that Indonesia needs to pay more
attention to other aspects: ecology, green infrastructure, and water governance, to
achieve more effective flood mitigation measures because structural mitigation only
provides a temporary solution.
Keywords Flood risk · Disaster mitigation · Operation and maintenance
1 Introduction
Floods are natural disasters that often occur in Indonesia [1–3]. Some of its water-
sheds are degraded and, hence, experiencing critical problems [4]. In some areas
R. M. S. Prastica (B) · A. Wijayanti

Civil Engineering Department, Gadjah Mada University, Yogyakarta 55281, Indonesia
e-mail: [email protected]; [email protected]
A. Wijayanti
R. M. S. Prastica
Water Infrastructure Construction Technology Study Program, Public Works Polytechnic,
576 R. M. S. Prastica and A. Wijayanti
with low-lying and level terrains on Java Island, floods persistently become the
main problem during the rainy season [5, 6]. Because of these issues, various water
infrastructures must deal with diverse challenges, such as the decreasing perfor-
mance of rivers, dams, and other flood-defense structures. Some scholars argue that
the island’s water management has shown signs of improvements [6–8] in various
aspects. However, evidence indicates that some watersheds suffer hydrological and
hydraulic problems [4, 8].
There are many mutually sustainable components in a watershed, including infras-
tructure, biotic and abiotic ecosystems, water quality [9], and humans. Therefore,
watershed management must consider their safety, preservation, and sustainability.
Further, characteristics and problems vary across watersheds, meaning that the
decision-making process to address what appears to be the same problem needs
to consider their respective hydrological principles, local wisdom, and ecosystems.
Finally, infrastructure operations and maintenance in the watershed are crucial for
reducing susceptibility to hazards. The infrastructure in question includes bridges,
rivers, irrigation networks [10], drainage channels [11], reservoirs, and dams.
As an essential component of a watershed [12], the river is a medium that stores,
retains, and transports water and serves as a drainage network connecting springs to
estuaries within an area defined by topographic boundaries. The river is a complex
ecosystem wherein water, sediments, and abiotic components exist and interact. Due
to the complexity, river management must operate on a principle that sees the water
body as a whole from the upper to the lower reaches and proceed in a sustainable
manner. If met accordingly, these terms can help solve problems that are naturally
inseparable from the river, e.g., flooding in level to gently sloping areas.
Flood is a natural phenomenon that can occur as a result of decreased infiltra-
tion capacity [13]. Increasing frequency and dimensions of flood occurrences are
signs of watershed degradation, which entails infiltration capacity reduction and
higher surface flow coefficient. Therefore, flood and watershed capacity analyses are
necessary so as to capture trends based on historical data and predict the expected
probability and frequency of occurrences according to the return period. For this
reason, the research conducted rainfall data analysis, prediction of rain intensity at
predefined return periods, frequency test, and hydrological analysis to calculate the
river discharge according to the rainfall return period. Then, the calculated flood
occurrences based on the river discharge were inputted to a model designed to deter-
mine predicted channel capacity. In Indonesia, floods mostly strike areas with small
and flat slopes and inadequate river channel planning and management. Therefore,
this study used HEC-RAS software to predict whether or not the channel could
accommodate flood discharge at a certain rain return period. If the model shows that
the existing channel cannot accommodate the calculated flood discharge, then it is
necessary to redesign the observed channel [14].
This study was located in Pepe River, Pepe Subwatershed, Central Java, Indonesia,
as depicted in Fig. 1. According to the above issues, the research aimed to investigate
the river’s hydraulic properties, particularly channel capacity. In addition, it evaluated
existing mitigation measures to tackle floods in the Pepe River in particular and
Indonesia in general. There are three specific questions answered in the research: Is
Flood Management Strategies in Indonesia … 577
Fig. 1 The location of the

Pepe Subwatershed
structural mitigation sufficient to prevent flooding? Are there other feasible solutions
to sustaining the mitigation efforts? What causes the difficulties in flood mitigation
in the current situation?
2 Method
The rainfall data were collected from the rain gauge stations distributed in the Pepe
Subwatershed and then processed using the Thiessen polygon method [15] to obtain
maximum daily rainfall. Afterward, the frequency analysis of maximum daily rainfall
for various return periods (i.e., 5, 10, 25, 50, 100, 500, and 1000-years) was conducted
using four statistical techniques: Normal Distribution, Gumbel Distribution, Log-
Normal Distribution, and Log-Pearson Type III Distribution. The data frequency
was tested for validity using the chi-square and Kolmogorov–Smirnov goodness-of-
fit tests. Then, the Nakayasu method was employed to calculate the flood discharge
at each predetermined return period [2].
Then, to model the existing river hydraulics, the limiting conditions for flood
occurrences were set in HEC-RAS software, such as the flood events at the seven
return periods. The flood discharges were denoted by Q5 , Q10 , Q25 , Q50 , Q100 , Q500 ,
and Q1000 .
Modeling with HEC-RAS involves several steps. First, the existing condition was
analyzed and observed. If the results show an overflow, a new river channelization
design (i.e., river channel dimensions) is analyzed as the selected structural mitigation
scenario. The hydraulic design proposed in the research used the flood discharge at
the 100-year return period (Q100 ). Second, the model built in HEC-RAS was re-run
with the new dimensions derived from the latest hydraulic analysis to observe the
river flow condition at different return periods.
In HEC-RAS software [16], the flow simulation occurs in an open channel so that
users can study the flow pattern along the channel. On a laboratory scale, simulations
factor in actual conditions by flowing water into a channel in a physical model.
Virtually, hydraulic numerical simulations can also be calculated using a computer
application software with a mathematics-based model. Mathematical models mimic
the physical phenomena of river flows in an actual or prototype channel using a
mathematical equation describing the relationship between flow variables: geometric,
kinematic, and dynamic variables. When the river flows are measured and observed
from a physical model, the flow parameters can be obtained by calculating or solving
the mathematical equation in the model.
The Cengklik Dam is located in Margorejo Village (Ngemplak District, Boyolali,

Central Java). It was built in 1923–1931 during the Dutch colonization and designed
and constructed under the Dutch East Indies Government to irrigate several subdis-
tricts in Boyolali Regency. The Bengawan Solo River Basin Center currently manages
the Cengklik Dam under the Directorate General of Water Resources, the Ministry
of Public Works and Housing.
The three rain gauge stations closest to the Pepe Subwatershed were Sleman
Geophysical Station, Semarang Meteorological Station, and Nganjuk Geophysical
Station, as presented in Fig. 2. The rainfall data recorded at these stations were
obtained from dataonline.bmkg.go.id and the Bengawan Solo River Basin Center.
The Thiessen polygon was created based on the station’s geographical position,
dividing the into areas affected by the rainfall data at each station (in %). Flood
occurrences predicted from the rain return period are depicted in Fig. 3. The analysis
used the Nakayasu method to produce all predicted flood occurrences for further
analysis.
According to the flow simulation in HEC-RAS software, all stations of the river
cannot accommodate the design flood in all return period. Figure 4 describes the
water level of the river by using several data of return period in STA 15 + 700 of the
location samples.
Based on the model of the existing river conditions in HEC-RAS, it was found that
the entire river overflowed and could not accommodate the 10, 25, 50, and 100-years
flood events. Therefore, it is necessary to redesign the channel dimensions.
Structural modification is a popular flood mitigation measure in Indonesia.
Government and policy-makers often decide to perform channelization by enlarging
the river dimension to increase its capacity to accommodate floods of varying inten-
sity and magnitude. Hence, this research factors in channelization strategies as the
selected structural mitigation scenarios in the flood simulation. The new HEC-RAS
simulation, as depicted in Fig. 5, revealed that the river could accommodate the
predicted floods even without significant runoffs appearing in the simulation result.
Fig. 2 Illustration of the Thiessen method used to calculate areal mean rainfall in Pepe Subwater-
shed
Pepe Sub-watershed synthetic hydrograph unit

500.00
Design flood (m3/s)
400.00
300.00
200.00
100.00
0.00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Hour
2-year 5-year 10-year 25-year
50-year 100-year 500-year 1000-year
Fig. 3 Nakayasu’s synthetic unit hydrograph for flood occurrences predicted for each rain return
period in Pepe Subwatershed
However, some researchers argue that structural flood mitigation has many weak-
nesses, especially related to sustainability [17], ecological aspect [12], and prospects
of economic benefits for the society [18]. For instance, water-sensitive urban design,
or WSUD for short, is currently the leading topic discussed among urban manage-
ment researchers and practitioners as a flood management strategy. It is also believed
to be a potential solution to flood issues in Indonesia.
Over the last decades, water-related disasters have stolen the spotlights on various
political, social, and scientific stages due to their impacts on many sectors and are
often related to accelerating climate change dynamics and urban sprawl. As a leading
center for urban water research, Australia has successfully established a new research
field focusing on WSUD and waves a flag of new topic discovery worldwide. WSUD
brings about improved environmental development in terms of water quantity and
quality [19]. The United Kingdom, China, the Netherland, and several other countries
prone to water-related disasters have adopted this concept and can offer their residents
promising results to deal with urban stormwater management problems in the future.
However, some countries claim that the concept cannot be implemented. Despite this
negative result, the question remains: Can the WSUD concept be applied to solving
flood issues in Indonesia? Some scholars argue that Indonesia’s water system does
not meet the conditions of the novel approach and that the WSUD concept does
not encourage integrated water resource management. However, there are several
reasons for Indonesia to implement WSUD effectively to manage water quantity and
quality.
Green infrastructure, or GI, is one of the elements of WSUD. Many Indonesian
regions have started introducing and applying the GI concept, especially since the
infrastructures are widely available and easy to construct. In previous studies, it is
mentioned that urban settings have the infrastructure arrangement but not proper
installation, such as in the case of constructed wetlands [20], green roofs [21],
Fig. 4 Pepe River’s flow simulation in HEC-RAS software for a 10-year, b 25-year, c 50-year and
d 100-year floods
Fig. 4 (continued)
and rainwater harvesting [22]. Recently, society has developed awareness of the
wide availability of GI in its surroundings and installed some systems personally
without supervision from the experts. To tackle this issue, the government has legis-
lated GI installation procedure standards as community guidelines. The GI instal-
lation benefits society in many ways. First, materials are easily found and locally
sourced. Second, locally available green materials are less expensive than grey infras-
tructure. However, even though the installment procedures have been published,
some green infrastructures are still built without following the measurement stan-
dards. In many cases, people decide to install large-sized rainwater harvesting tanks
without factoring in mathematical concepts. Also, the product manuals or installation
guidelines sometimes provide complicated measurements that should otherwise be
translated to layman’s terms and encourage time efficiency.
Besides structural mitigation and green infrastructure, river restoration has been
a higher priority for most regional governments in today’s urban areas [23, 24].
As a result of urban sprawl and rapid population growth [25], urban areas face a
never-ending challenge: riverine floods. On the other hand, rivers play a pivotal role
in rural and urban populations [26]. Farmers depend on river water for irrigation [10];
an unreliable river system can disturb water irrigation flows and decrease agricultural
production. Rivers undoubtedly provide water for daily use, sustain aquatic habitats,
and maintain the food chain in the environment. Without the nutrients provided by
rivers, the population of the aquatic habitat will deteriorate [27]. This is possible
when the river water quality is declining [28], partly due to human activities that
adversely affect the water [29].
At the same time, Indonesia still lacks adequate water quantity management.
Challenges often arise from inconsistent coordination between governmental agen-
cies in conducting integrated flood management [3]. Similarly, joint efforts between
the responsible ministries and local communities have yet to enforce strict rules and
effective measures. These issues can lead to overlapping organizational functions
Fig. 5 Pepe River’s flow simulation in HEC-RAS software (after redesigning the dimensions) for
a 10-year, b 25-year, c 50-year and d 100-year floods
Fig. 5 (continued)
and programs and cause some measures to cease, receive no sufficient follow-ups,
and be implemented partially [8]. An example includes a flood prevention infrastruc-
ture in Tegal (Central Java, Indonesia), which was destroyed by the society because,
instead of protecting the affected areas, it worsened the flood impacts and mate-
rial losses. This issue indicates the lack of well-formulated coordination between
central and provincial governmental agencies. Even though poor responses to water-
related disasters are the most pressing problem, the stakeholders involved (i.e.,
local, regional, and national authorities) still cannot agree on a well-synchronized
communication during and after the disaster.
To sum up, structural mitigation strategies like channelization can entail an enor-
mous change to flood mitigation, which may raise negative responses from varying
parties. Another perspective might include providing rooms for other solutions, such
as green infrastructure, river restoration, and legal coordination, and making them
priorities in the management of the Pepe River and many rivers in Indonesia. Even
though structural mitigation can tackle floods, this solution offers a temporary result,
especially when it fails to integrate other aspects into its implementation.
4 Conclusion
Based on the flood and mitigation scenario in this research, structural mitigation
proves effective in preventing floods. Before a selected measure (i.e., channelization)
is introduced, the Pepe River’s upper, middle, and lower reaches cannot accommo-
date the floods predicted from the rain return period. Redesigning channel dimensions
leads to increased capacity that allows the river to accommodate said floods. However,
the proposed mitigation appears to have negative effects, especially when other
factors are not considered in the implementation. Therefore, other existing options
like green infrastructure, community-based river restoration, and water governance

need to be further developed to achieve sustainable flood mitigation in the Pepe River
and other rivers with similar characteristics in Indonesia.
Acknowledgements The research grant funding this study was provided by the Directorate of
Research, Universitas Gadjah Mada, Number 2458/UN1/DITLIT/DIT-LIT/PT/2021.
References
1. Prastica RMS, Apriatresnayanto R, Marthanty DR (2019) Structural and green infrastructure

mitigation alternatives prevent Ciliwung River from water-related landslide. Int J Adv Sci Eng
Inf Technol 9(6):1825–1832
2. Prastica RMS, Maitri C, Hermawan A, Nugroho PC, Sutjiningsih D, Anggraheni E (2018)
Estimating design flood and HEC-RAS modelling approach for flood analysis in Bojonegoro
city. In: IOP Conference series: materials science and engineering, vol 316(1)
3. Lin E, Shaad K, Girot C (2016) Developing river rehabilitation scenarios by integrating land-
scape and hydrodynamic modeling for the Ciliwung River in Jakarta, Indonesia. Sustain Cities
Soc 20:180–198
4. Kefi M, Mishra BK, Masago Y, Fukushi K (2020) Analysis of flood damage and influencing
factors in urban catchments: case studies in Manila, Philippines, and Jakarta, Indonesia. Nat
Hazards 104(3):2461–2487
5. Wedawatta G, Kulatunga U, Amaratunga D, Parvez A (2016) Disaster risk reduction infras-
tructure requirements for South-Western Bangladesh: perspectives of local communities. Built
Environ Proj Asset Manag 6(4):379–390
6. Kim J, Gim THT (2020) Assessment of social vulnerability to floods on Java, Indonesia. Nat
Hazards 102(1):101–114
7. Marfai MA et al (2008) Natural hazards in Central Java Province, Indonesia: an overview.
Environ Geol 56(2):335–351
8. Marfai MA, Sekaranom AB, Ward P (2015) Community responses and adaptation strategies
toward flood hazard in Jakarta, Indonesia. Nat Hazards 75(2):1127–1144
9. Prastica RMS (2018) Employing Galerkin method and field study for model calibration
to perform sediment transport modelling in Agathis Lake Universitas Indonesia. In: AIP
conference proceedings, vol 2014
10. Lindenschmidt KE, Carr MK, Sadeghian A, Morales-Marin L (2019) CE-QUAL-W2 model
of dam outflow elevation impact on temperature, dissolved oxygen and nutrients in a reservoir.
Sci Data 6(1):312
11. Siqueira RC Moura1 PM, das Graças Silva1 TF (2019) Methodology for the construction of
an urban flood hazard chart. Rev Bras Recur Hidricos 24:1–16
12. Talbot CJ et al (2018) The impact of flooding on aquatic ecosystem services. Biogeochemistry
141(3):439–461
13. Prastica RMS, Wicaksono D (2019) Integrated multimodal disaster mitigation management for
urban areas: a preliminary study for 2-d flood modeling. In: IOP conference series: materials
science and engineering, vol 650(1)
14. Prastica RMS, Maitri C, Nugroho PC, Hermawan A (2017) Analisis Banjir dan Perencanaan
Desain Transportasi Sungai di Kota Bojonegoro. Media Komun Tek Sipil 23(2):91
15. Prastica RMS, Adi DAR, Famila N (2020) Mitigasi banjir dan alternatif pemeliharaan
infrastruktur keairan pada sub-DAS code Yogyakarta. Tek J Sains dan Teknol 16(1):25–33
16. Geravand F, Hosseini, SM, Ataie-Ashtiani B (2020) Influence of river cross-section data reso-
lution on flood inundation modeling: Case study of Kashkan river basin in western Iran. J
Hydrol 584(Dec 2019):124743
17. Chan SL, Wey WM, Chang PH (2014) Establishing disaster resilience indicators for Tan-Sui
river Basin in Taiwan 115(1)
18. De Vincenzo A, Jacopo Molino A, Molino B, Scorpio V (2017) Reservoir rehabilitation: the new
methodological approach of economic environmental defence. Int J Sed Res 32(2):288–294
19. Kuller M, Bach PM, Ramirez-Lovering D, Deletic A (2017) Framing water sensitive urban
design as part of the urban form: a critical review of tools for best planning practice. Environ
Model Softw 96:265–282
20. Sheng X et al (2020) Management of rural domestic wastewater in a city of Yangtze delta
region: performance and remaining challenges. Bioresour Technol Rep 11(Jul):100507
21. Liu C, Li Y (2016) Measuring eco-roof mitigation on flash floods via GIS simulation. Built
Environ Proj Asset Manag 6(4):415–427
22. Zingiro A, Okello JJ, Guthiga PM (2014) Assessment of adoption and impact of rainwater
harvesting technologies on rural farm household income: the case of rainwater harvesting
ponds in Rwanda. Environ Dev Sustain 16(6):1281–1298
23. Setyoasri YP, Prastica RMS (2020) Rapid assessment of river watershed health and vulnerability
level for restoration strategy: a study of river systems in Indramayu West Java, Indonesia. In:
IOP conference series: earth and environmental science, vol 423(1)
24. Prastica RMS, Kurniawan RP, Apriatresnayanto R (2020) Environmental vulnerability assess-
ment and stormwater management to enhance watershed performance. J Phys Conf Ser
1625(1)
25. Machado RAS, Oliveira AG, Lois-González RC (2019) Urban ecological infrastructure: the
importance of vegetation cover in the control of floods and landslides in Salvador/Bahia, Brazil.
Land use policy 89(Oct):104180
26. Kozlowski M, Yusof YM (2016) The role of urban planning and design in responding to climate
change: the Brisbane experience. Int J Clim Chang Strateg Manag 8(1):80–95
27. Geist J, Hawkins SJ (2016) Habitat recovery and restoration in aquatic ecosystems: current
progress and future challenges. Aquat Conserv Mar Freshw Ecosyst 26(5):942–962
28. Björklund K, Bondelind M, Karlsson A, Karlsson D, Sokolova E (2018) Hydrodynamic
modelling of the influence of stormwater and combined sewer overflows on receiving water
quality: benzo(a)pyrene and copper risks to recreational water. J Environ Manage 207:32–42
29. Tedford M, Ellison JC (2018) Analysis of river rehabilitation success pipers river Tasmania.
Ecol Indic 91(Sep 2017):350–358
Comparison of Suitable Drought Indices
for Over West Nusa Tenggara
Humairo Saidah, Heri Sulistiyono, and I Dewa Gede Jaya Negara
Abstract The drought has had a significant impact on West Nusa Tenggara. Drought
experienced in some areas almost every year, causing crop failure and even causing
famine in 1954 and 1966. This study aims to find the best method for obtaining
the meteorological drought index because it is essential to understand the drought
characteristics in the study area in guiding policymakers in early anticipation of this
region’s drought hazard. This study compares two drought index methods, SPI and
PDSI, and tests their proximity to drought data in these locations. Evaluation is carried
out to determine the best approach to choose for conducting a drought assessment.
The best-fitted method was determined statistically by counting the number of dry
months produced by the model and comparing it with the drought experienced over
West Nusa Tenggara. The results obtained are the SPI method has an accuracy rate
of 58% and a correlation coefficient r about 0.06. Meanwhile, the PDSI method has
a better accuracy rate of 75% and a correlation coefficient r about 0.51. The PDSI
method is also the most superior in accurately predicting the arrival of a dry month
compared to SPI. So, it can be concluded that the PDSI method is the better method
for evaluating and detecting dry periods in West Nusa Tenggara.
Keywords SPI · PDSI · Drought index
1 Introduction
Drought is one of the phenomena that occur due to seasonal circulation under
the global climate’s influence, characterized by water availability far from water
H. Saidah (B) · H. Sulistiyono · I. D. G. J. Negara

H. Sulistiyono
I. D. G. J. Negara
588 H. Saidah et al.
needs for living, agriculture, and other activities. The impact of drought events is
widely recognized. Drought is directly related to public health problems, usually
resulted from poor sanitation to spread disease, and can widely lead to poverty and
underdevelopment of an area.
Drought is a normal phenomenon, a recurrent feature of climate variability. It
differs from aridity, which is described to low rainfall regions permanently as a
feature of the site climate condition. Drought is a major natural hazard of nature
in almost all climate zone types, although it varies significantly from one region to
another spatially and temporally [1, 2]. Drought is an abnormal period of dry weather
sufficiently prolonged caused by a lack of precipitation that causes hydrological
imbalance and moisture deficiency in water use requirements [3].
Drought analysis is needed to support agricultural activities towards food security.
Agriculture, especially rice, is the largest water consumer, which is very prone to
drought. Calculation of drought analysis usually produces an index that describes
the severity of the drought. Drought is classified into three categories, namely mete-
orological drought, agricultural drought, and hydrological drought. Meteorological
drought is related to the amount of rainfall that occurs under normal conditions during
a season. Agricultural dryness is associated with reduced water content in the soil
(soil moisture), and hydrological dryness is related to reduced water supply from the
earth.
Meteorological drought describes reduced rainfall which is the first indication
of drought conditions. There are many analytical methods for detecting this mete-
orological drought with various advantages and disadvantages. Among the widely
used methods are SPI (Standardized Precipitation Index), PDSI (Palmer Drought
Severity Index), Decile Index, PNI (Precipitation Normal Index), WSI (surface water
supply index), RDI (Reconnaissance Drought Index), Standardized Precipitation
Evapotranspiration Index (SPEI) and others [4–9].
This study chooses two methods, namely SPI and PDSI, to be applied in the study
location. The SPI model has the advantage of being simple. It only requires rainfall
input and it can determine the severity of drought by measuring the rainfall deficit at
various periods based on normal conditions [4, 5]. However, the use of SPI model in
characterizing drought level should be done with caution [10]. The PDSI model is
commonly used to evaluate droughts that have occurred, especially in semiarid areas
and dry sub-humid climates [11, 12], where West Nusa Tenggara is a semiarid area.
The PDSI model is also a standardization for all country’s local climates to show
relative drought or rainfall conditions [13]. The PDSI method can also assess and
predict drought [14].
Comparison of Suitable Drought Indices … 589
Fig. 1 West Nusa Tenggara region [15]
2.1 Study Area
The research conducted on the two largest islands in West Nusa Tenggara Province,
namely Lombok and Sumbawa islands, located at coordinates 8°10’–9°5’ latitude
and 115°46’–119°5’ Longitude, which has a total area of around 20.164.84 km2 (see
Fig. 1) [15].
2.2 Data Availability
The data needed for the SPI model is rainfall, while the input data for the PDSI model
is rainfall, evapotranspiration, soil type, soil structure, and land use. The primary
data for this study are the monthly rainfall in the last 21 years from 1997 to 2018,
which were collected from West Nusa Tenggara Regional Infrastructure Information
Center. All available data used for drought index analysis, then the model’s accuracy
evaluation using only the indexes generated from 2012 to 2017. It is because the
drought experienced data to be compared to has only been recorded since 2012.
The selection of rain stations used in this study is based on two mains of consid-
eration: length of data records and data completeness (minimum blank data). In the
calculation, SPI uses monthly rainfall data with a fairly long recording period as
described in the WMO-No.1090 WMO SPI User Guide that the length of the data
used ranges from 20 to 30 years, and it is better if it is more than 30 years [16].
2.3 Research Method
This study analyzed the drought index from SPI and PDSI then compared the results
with experienced drought data collected from Regional Board for Disaster Manage-
ment (BPBD). The month where the drought occurred will be an indexed score of −
1, and when the drought has not happened, it will be an indexed score of 0. The next
step is to provide a drought index score calculated by the SPI and PDSI methods.
In the SPI and PDSI methods’ indexing system, both have almost the same range of
index values, where the index value of less than −1 indicates a drought. It makes the
comparison of the two methods easier because they have the same indexing class.
The next step is calculating the duration of drought by accumulating drought events
every month started with a score of −1 and ended when the score is 0.
The accumulated drought duration score was obtained from SPI and PDSI methods
then compared to the drought scores from the BPBD data. For the same month that
shows the same score, for example, both show drought or no drought, it is interpreted
as suitable and given a score of 1. And if it does not match, then the score given
is 0. The suitability assessment is then carried out by summed up the number of
match scores. The higher the number of corresponding scores, the better the method
produced a drought index.
Method of completing the missing data
In some areas, the recording of rain data is sometimes incomplete. It could be due to
administrative errors or recording equipment damaged. Then estimating data tech-
nique must be done to complete it. Ideally, estimating missing data is carried out by
comparing the data from several close stations and correlated with the test station
[17].
In this study, completing the missing rain data was carried out by a normal ratio
method, a simple method to fill in the missing rainfall data based on rainfall data
from several close stations at the same time and compared it with the annual rainfall
data of each station. The formula in completing missing data at stations is [18]:

1 m N x
Px = Pi (1)
m i=1 Ni
Where: Px = estimated value of rainfall for the ungauged station (mm); Pi =

rainfall values of other rain gauges used for estimation; Nx = Normal annual rainfall
data for the ungauged station; Ni = normal annual precipitation of other surrounding
stations; m is the number of surrounding stations.
Calculating of Drought Indices
SPI computation includes of matches probability density function of the distribution
of Gamma is defined by:
x
1 −x
g(x) = α
t a−1 e β (2)
β (a) 0
Where α > 0 is the form parameter; β > 0 is a scale parameter; and x > 0 is the
total monthly rainfall. While (α) is the Gamma function defined as:
∞
(α) = ∫ y ∝−1 e−y dy (3)
0
The values of α and β are estimated for each rain station using the following
formula:

1 4A
α= 1+ 1+ (4)
4A 3

x ln(x)
β = and A = ln(x) − (5)
α n
Where n is the amount of rainfall data.

Then calculating the cumulative chance then the Gamma spread integrated with
x to give G (x):
x x
1 −x
G(x) = x g(x) d x = t a−1 e β d x (6)
0 βα(a) 0
Then substitution of I = βx , so the formula (6) becomes:

1
G(x) = t a−1 e−t dt (7)
(a)
Since the gamma function is undefined for x = 0, the value of G (x) becomes:
H (x) = q + (1 − q)G(x) (8)
Where: q = m/n; m is the number of rainfall events 0 mm in the rain data series.
SPI value calculation:

C0 + C1 t + C2 t 2
Z = SPI = − t − ; for 0 < H(x) ≤ 0.5 (9)
1 + d1 t + d2 t 2 + d3 t 3

and transformation of gamma distribution: t = ln (H (x))
1
2 ;

C0 + C1 t + C2 t 2
Z = SPI = + t − ; for 0.5 < H(x) ≤ 1.0 (10)
1 + d1 t + d2 t 2 + d3 t 3

and transformation of gamma distribution: t = ln (1−H1(x))2 ;
Calculation of PDSI value uses the formula:

Z Z Z
X= + X ; and X = − 0.103 (11)
3 j−1 3 j 3 j−1
The number of c0 = 2.515517; c1 = 0.802853; c2 = 0.010328; d1 = 1.432788;

d2 = 0.189269; d3 = 0.001308.
The index is calculated from water supply and demand computed using a complex
water-balanced system based on recorded precipitation, air temperature, and the
soil characteristics of the site. The calculation of potential evapotranspiration was
proposed by Thornthwaite. This is required only daily averaged temperatures and
latitude of the site to calculate the maximum amount of sunshine duration [19].
The SPI and PDSI classification index describe a relative moisture condition
within 9 categories, as shown in Table 1.
The hydrologic accounting procedure of the water balance model involves the
production of time series of the runoff values, recharge to the soils, moisture loss
from the soils, and transpiration, as well as estimates of their potential values. From
these time series, average monthly values of the potential and the actual values are
calculated over the calibration interval. The coefficient of evapotranspiration for a
particular month is the quotient of an average of the actual value ET and the average
of the potential value PET for that month. Similar coefficients are b, giving the ratio
of average recharge to the soils R and its potential value PR, g, giving the ratio of
average runoff RO and its potential value PRO, and giving the ratio of average loss of
moisture from the soils L to its potential value PL. Here the overbar denotes a long-
term average. Palmer’s Climatically Appropriate for Existing Conditions (CAFEC)
Table 1 The SPI and PDSI classification

SPI Class PDSI Class
≥2 Extremely wet ≥4.00 Extremely wet
1.5 to1.99 Severely wet 3.00 to 3.99 Severely wet
1.00 to 1.49 Moderately wet 2.00 to 2.99 Moderately wet
0.5 to 0.99 Mild wet 1.00 to 1.99 Slightly wet
−0.49 to 0.49 Near normal 0.50–0.99 Incipient dry spell
−0.99 to -0.5 Mild drought 0.49 to −0.49 Slightly dry
−1.49 to -1.00 Moderately drought −0.50 to −0.99 Moderately dry
−1.99 to -1.50 Severely drought −1.00 to −1.99 Severely dry
≤ −2.0 Extremely drought −2.00 to −2.99 Extremely dry
precipitation; ˆP is a function of the potential evapotranspiration, potential recharge,

potential runoff and potential loss, weighted by their appropriate coefficients.
3.1 Data Preparation
The first step of this research is quality data control by checking its completeness.
The requirements of drought index calculation, especially for the SPI model, must
be filled in completely or without any missing data during that period of use. Then
if there is blank data, it needs to be interpolated to predict it using formula 1.
This study calculated SPI and PDSI index, which are divided into several regional
groups. Lombok Island was divided into four regions and two regions for Sumbawa
Island. First, the calculation of the regional average rainfall is calculated using the
Thiessen polygon method, then it is followed by the analysis of the monthly rainfall.
Monthly rainfall is the primary input data in calculating the drought index. Regional
averaged monthly rainfall for each region is presented in Fig. 2.
Other data used in PDSI calculations are air temperature. The air temperature data
is needed to calculate the potential evapotranspiration value, which is then used to
determine the criteria for wet and dry months based on rainfall data and potential
evapotranspiration rates. Air temperature data throughout West Nusa Tenggara is
presented in Fig. 3.
350
300
250
200
150
100
50
0
West Lombok Central Lombok North Lombok
Dompu Sumbawa East Lombok
Fig. 2 Regional averaged monthly rainfall over West Nusa Tenggara

30.0
29.0
28.0
27.0
26.0
25.0
24.0
23.0
22.0
21.0
20.0
West Lombok Central Lombok North Lombok

Dompu Sumbawa East Lombok
Fig. 3 Averaged monthly air temperature over West Nusa Tenggara
0
Jul-15
Nov-15
Jan-15
May-15
Jul-16
Nov-16
Jan-16
May-16
Jul-17
Nov-17
Jan-17
May-17
Jul-18
Nov-18
Jan-18
May-18
Mar-15
Sep-15
Mar-16
Sep-16
Mar-17
Sep-17
Mar-18
Sep-18
-1
-2
-3
-4
SPI
Fig. 4 The example of SPI index for West Lombok. The negative value shows the strength and
drought duration. The orange line indicates the start of the dry periods
3.2 Calculation of Drought Indices
In the SPI and PDSI methods, drought occurs when the index value continuously
gives a negative value and reaches a drought intensity with an index value of −1 or
less. A positive index shows that the rain is greater than the median and a negative
index shows that the rain is smaller than the median.
Firstly, the Indices of SPI and PDSI for any location were calculated based on the
long-term precipitation record for the desired period. The SPI model is well known
and becomes the most popularly used drought index [20] (Fig. 4).
The data input in the PDSI method is rainfall, groundwater capacity (WHC),
and potential evapotranspiration. The advantage of this method is that in addition to
producing index values, it also creates a coefficient of climate parameters, namely
the evapotranspiration coefficient, the recharge coefficient, the runoff coefficient,
and the moisture loss coefficient. From this coefficient, it can be calculated the
23
18
13
8
3
-2
Jan-15
Mar-15
May-15
Sep-15
Jan-16
Jul-15
Mar-16
May-16
Sep-16
Jan-17
Nov-15
Jul-16
Mar-17
May-17
Sep-17
Jan-18
Nov-16
Jul-17
Mar-18
May-18
Sep-18
Nov-17
Jul-18
Nov-18
-7
-12
PDSI
Fig. 5 The example of PDSI value for West Lombok. The negative value shows the strength and
drought duration. The blue line indicates the start of the dry periods
Table 2 West Lombok Drought experienced data

BPBD Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec
2014 0 0 0 0 0 0 −1 −1 −1 −1 −1 0
2015 0 0 0 0 0 0 0 0 −1 −1 −1 −1
2016 0 0 0 0 0 0 0 0 −1 −1 −1 −1
2017 0 0 0 0 0 0 0 −1 −1 −1 −1 −1
2018 0 0 0 0 0 0 0 −1 −1 −1 −1 −1
rainfall that occurs during a specific month support evapotranspiration, runoff, and
soil moisture reserves which are considered normal conditions. Palmer’s drought
index classification is divided into 9 classes with zero indexes as a normal condition.
Besides, the Palmer method requires land-use data, soil type, and soil texture (Fig. 5).
3.3 Comparison of Drought Index Values
The accuracy performance test of the model in evaluating the drought period was
conducted by comparing the SPI and PDSI model’s indices against drought data
from Regional Board for Disaster Management (BPBD) West Nusa Tenggara. BPBD
recorded a dry period in which the community did not have any access to water,
both rainwater and surface water. The month when the drought recorded actually
figured the water shortage that occurred on site. BPBD data for water deficiency
were presented in numerical format, where the months with the drought occurrence
are given a score of −1, while months without drought are provided 0 (see Tables 2
and 3).
The calculation of the model’s suitability is started by first giving a dumb score
on the index produced from both methods. The SPI and PDSI values of < −1 are
given −1 score for that means the dry period has started, and a score of 0 for the
596
Table 3 Determination of both model’s drought period scores and their matches to BPDB
Score of Jan- 15 Feb- 15 Mar- 15 Apr- 15 May- 15 Jun- 15 Jul- 15 Aug- 15 Sep- 15 Oct- 15 Nov- 15 Sums (2015–2018)
SPI 0 0 0 0 0 0 0 0 0 −1 −1
PDSI 0 0 0 0 0 0 −1 −1 −1 −1 −1
BPBD 0 0 0 0 0 0 −1 −1 −1 −1 −1
SPI matches to BPBD 1 1 1 1 1 1 0 0 0 1 1 32
PDSI matches to BPBD 1 1 1 1 1 1 1 1 1 1 1 37
H. Saidah et al.
Table 4 Correlation
Region r SPI r PDSI
coefficients between SPI and
PDSI to BPBD data drought West Lombok 0.282 0.512
period Central Lombok 0.064 0.192
North Lombok 0.192 0.641
East Lombok 0.048 0.543
Sumbawa −0.266 0.643
Dompu 0.076 0.529
85% 81%
80%
81%
77% 77%
75% 75%
70% 67%
65% 60% 63%
60% 58% 56% 58%
55%
50% 48%
45%
40%
West Central North East Lombok Sumbawa Dompu
Lombok Lombok Lombok
% match SPI % match PDSI
Fig. 6 The matches percentage of SPI and PDSI drought period to BPBD data
month with no drought indicated by the index values of both models are > −1. Next,
check the matches of the score. Months with the same score will get a score of 1, and
months with a different score will get 0. The correlation coefficients were calculated
between the scores of the models and BPBD drought data (Table 4). The summary
of the total match scores can then be added to find how suitable the model is. The
higher the score, the higher the suitability drought index, especially in the dry month
(Table 3).
The matches score of the drought period was then summed up and calculated the
matches percentage (see Fig. 6).
Based on the calculation results presented in Fig. 6, it can be concluded that
the PDSI method provides a better estimation of the drought period in the study
area than the SPI method. The calculation of suitability produced the result that the
PDSI method has an average suitability percentage of 75% and a mean correlation
coefficient (r) of 0.510, and the SPI method has an average suitability percentage of
58% and a correlation rate (r) of 0.06.
The estimation of drought duration is done by accumulating a drought index
score. The more scores we get, the longer the period of the drought. A technique in
accumulating drought index score started when a score shows −1 that indicates a dry
period has occurred and ends at a score of 0, which means there is no more drought.
The same calculation is started again when a score of −1 is found.
The cumulative score obtained will show the length of the dry month. The higher
scores mean the longer the duration of the drought. The calculated drought period
results in a cumulative score for 2015–2018, presented in Fig. 7.
Figure 7 shows that the performance of the PDSI method appears to have a better
ability to detect prolonged droughts in 2014 and 2016 than the SPI method. The SPI
drought index often does not indicate the suitable occurred drought time. On the
other hand, PDSI can detect the drought that happens relatively close to the BPBD
data. The model accuracy is only reviewed for the presence or absence of the drought
event, regardless of its depth, magnitude and severity. So, the results of drought index
calculations that show specific criteria are ignored. Comparative data used is only
for the drought period, while the wetness periods and the rank are ignored.
Besides requiring more data input, the PDSI method has a higher level of difficulty
than the SPI method in the calculation process. So that in its application, the SPI
method is still interesting to apply.
Central Lombok District Drought

West Lombok District Drought Period Period
0 0
Dec-14 May-16 Sep-17 Dec-14 May-16 Sep-17 Feb-19
-2 -2
-4 -4
-6 -6
SPI PDSI BPBD SPI PDSI BPBD
North Lombok District Drought East Lombok District Drought Period

Period
0 0
Dec-14
-2 May-16 Sep-17 Feb-19 Dec-14
-2 May-16 Sep-17 Feb-19
-4 -4
-6 -6
-8 -8
-10 -10
SPI PDSI BPDB SPI PDSI BPDB
Sumbawa District Drought Period Dompu District Drought Period
0 0
Dec-14
-2 May-16 Sep-17 Feb-19 Dec-14 May-16 Sep-17
-2
-4
-4
-6
-8 -6
-10 -8
SPI PDSI BPDB SPI PDSI BPDB
Fig. 7 Cumulative score of the SPI, PDSI, and BPBD drought period
4 Conclusions
The suitability of the drought indices calculating techniques compared to experi-

enced drought data, the PDSI method can evaluate and determine the drought period
properly for West Nusa Tenggara. The PDSI method gave an accuracy rate of 75%
for dry months tested conditions, while SPI was only 58%. The correlation number
of indexes correlated with drought occurrence is about 0.51 for PDSI and 0.06 for
SPI.
References
1. Gao L, Zhang Y (2016) Spatio-temporal variation of hydrological drought under climate change
during the period 1960–2013 in the Hexi Corridor. China J Arid Land 8:157–171. https://doi.
org/10.1007/s40333-015-0022-3
2. Zhang J, Sun F, Lai W, Lim WH, Liu W, Wang T, Wang P (2019) Attributing changes in future
extreme droughts based on PDSI in China. J Hydrol 573:607–615. https://doi.org/10.1016/j.
jhydrol.2019.03.060
3. McMahon T (1982) Arenas A D. Methods of computation flows. Gustard 107
4. Pathak AA, Dodamani BM (2020) Comparison of meteorological drought indices for different
climatic regions of an Indian river basin. Asia-Pac J Atmos Sci 56:563–576. https://doi.org/10.
1007/s13143-019-00162-5
5. Chelton DB, Risien CM (2020) A hybrid precipitation index inspired by the SPI, PDSI, and
MCDI. Part II: application to investigate precipitation variability along the West Coast of North
America. J Hydrometeorol 21:1977–2002. https://doi.org/10.1175/JHM-D-19-0231.1
6. Zou L, Xia J, She D (2017) Drought characteristic analysis based on an improved PDSI in the
Wei River Basin of China. Water 9:178. https://doi.org/10.3390/w9030178
7. Haied N, Foufou A, Chaab S, Azlaoui M, Khadri S, Benzahia K, Benzahia I (2017) Drought
assessment and monitoring using meteorological indices in a semi-arid region. Energ Procedia
119:518–529. https://doi.org/10.1016/j.egypro.2017.07.064
8. Abbasian MS, Najafi MR, Abrishamchi A (2021) Increasing risk of meteorological drought in
the Lake Urmia basin under climate change: introducing the precipitation–temperature deciles
index. J Hydrol 592:125586. https://doi.org/10.1016/j.jhydrol.2020.125586
9. Swain S, Patel P, Nandi S (2017) Application of SPI, EDI and PNPI using MSWEP precipitation
data over Marathwada, India In: 2017 IEEE International Geoscience and Remote Sensing
Symposium (IGARSS) pp 5505–5507. https://doi.org/10.1109/IGARSS.2017.8128250
10. Tirivarombo S, Osupile D, Eliasson P (2018) Drought monitoring and analysis: Standardised
precipitation evapotranspiration index (SPEI) and standardised precipitation index (SPI). Phys
Chem Earth Parts A/B/C 106:1–10. https://doi.org/10.1016/j.pce.2018.07.001
11. Quiring SM, Papakryiakou TN (2003) An evaluation of agricultural drought indices for the
Canadian prairies. Agric For Meteorol 118:49–62. https://doi.org/10.1016/S0168-1923(03)000
72-8
12. Dehghan S, Salehnia N, Sayari N, Bakhtiari B (2020) Prediction of meteorological drought in
arid and semi-arid regions using PDSI and SDSM: a case study in Fars Province. Iran J Arid
Land 12:318–330. https://doi.org/10.1007/s40333-020-0095-5
13. Huang S, Li P, Huang Q, Leng G, Hou B, Ma L (2017) The propagation from meteorological
to hydrological drought and its potential influence factors. J Hydrol 547:184–195. https://doi.
org/10.1016/j.jhydrol.2017.01.041
14. Vasiliades L, Loukas A (2009) Hydrological response to meteorological drought using the
Palmer drought indices in Thessaly. Greece Desalination 237:3–21. https://doi.org/10.1016/j.
desal.2007.12.019
15. Badan Pusat Statistik Provinsi Nusa Tenggara Barat (2017) Luas Daerah Nusa Teng-
gara Barat 2015–2017. https://ntb.bps.go.id/indicator/153/56/1/luas-daerah-nusa-tenggara-
barat.html. Last accessed 30 Mar 2021
16. World Meteorological Organization (2012) Standardized precipitation index user guide
17. Adhyani NL, June T, Sopaheluwakan A (2017) Exposure to drought: duration, severity and
intensity (Java, Bali and Nusa Tenggara).In: IOP conference series: earth and environmental
science, vol 58, p 012040. https://doi.org/10.1088/1755-1315/58/1/012040
18. De Silva RP, Dayawansa NDK, Ratnasiri MD (2007) A comparison of methods used in
estimating missing rainfall data. J Agric Sci 3:101–108
19. Schrier van der G, Jones PD, Briffa KR (2011) The sensitivity of the PDSI to the Thornth-
waite and Penman-Monteith parameterizations for potential evapotranspiration. J Geophys Res
Atmos 116. https://doi.org/10.1029/2010JD015001
20. McKee TB, Doesken NJ, Kleist J (1993) The relationship of drought frequency and duration
to time scales. 6
Small Debris Flow Simulation Using
MORPHO2DH
Puji Harsanto, Dandy Darvin Septiandy, Berli Paripurna Kamiel,
and Nursetiawan
Abstract Several studies based on experiments or field surveys reported the effect
of de-bris flow from the eruption of volcanoes on river morphology. However, there
has been a notable absence of scientific literature dealing with the prediction of
river morphology at the middle or downstream after several weeks or months due to
the remnants of material sediment from an eruption. The remaining material sedi-
ment at upstream will produce a small debris flow during the rainy season. This
study analyzes the riverbed change due to small debris flow using the 2-dimensional
numerical simulation in iRIC-Morpho2DH. The prediction of riverbed erosion on
several segments shows that the erosion increases when the small debris occurs.
The maximum erosion occurs when the sediment discharge ratio is 2 times, then
decreases in the simulation of the sediment discharge ratio 3 times. Base on the field
survey, the river segment with maximum erosion, the bank protection was failed.
Keywords Erosion · Sedimentation · Morpho2DH · Sediment discharge ratio ·

Gajah Wong River
P. Harsanto (B) · D. D. Septiandy · Nursetiawan

Yogyakarta, Yogyakarta, Indonesia
D. D. Septiandy
Nursetiawan
B. P. Kamiel
Department of Mechanical Engineering, Faculty of Engineering, Universitas Muhammadiyah
Yogyakarta, Yogyakarta, Indonesia
602 P. Harsanto et al.
1 Introduction
Mount Merapi is one of the most active volcanoes in Indonesia. The material sediment
from its eruptions is an essential key for river morphology in rivers that originated
from that volcano. Gajah Wong River, which flows through Yogyakarta City, is one
of the rivers that originated on the slopes of Mount Merapi. A riverbank protection
is built on the river to protect the bank erosion. However, some constructions failed
because of the excessive erosion. Regarding [1], it is crucial to mitigate riverbank
erosion in the river to get the proper design of riverbank protection.
Morpho2DH is a two-dimensional flow and bed open-source modeling system that
is integrated into iRIC [2]. The purpose of this study was to analyze riverbed erosion
due to the concentration of bedload transport. The advantage of this software is that
the sediment transport concentration modeling can be set beyond the equilibrium
condition to model the flow of debris. This is relevant because which would prevent
instances such as the Gajah Wong River flood.
2 Study Area
This research was conducted in the segment of Gajah Wong River along ± 1 km
due to the markers in the Prenggan area (−7.822571°, 110.399046°) as the upstream
point to the Jagalan GW137 area (−7.828972°, 110.396000°) as the downstream
point. The research location can be seen clearly in Fig. 1.
Fig. 1 Cross section GW157-GW137 of Gajah Wong River

Small Debris Flow Simulation Using MORPHO2DH 603
Fig. 2 Sediment grain size in Gajah Wong River
2.1 Discharge and Cross Section Data
In this study, two types of data are used, namely primary and secondary data. The
primary data is in the form of sediment grain size based on laboratory test results
from ± 2 kg of sediment samples taken directly from the Gajah Wong River with a
total of 3 test samples (upstream, middle, downstream) which were tested for filter
analysis in the laboratory. The secondary data in cross-section data and design flood
discharge data is obtained from the Public Works Service of BBWS Serayu-Opak.
The design flood discharge is 82.97 m3 /s [3].
2.2 Grain Size Data
Three samples of sediment were obtained from the upstream, middle, and down-
stream of the river field survey. As much as ± 2 kg was obtained from each section,
then the sediment grain-sizes were determined through [4]. The results of the grading
test are presented in Fig. 2.
3 Two Dimensional Flow and Bed Change Equations
Morpho2DH is a computational analysis program developed by Hiroshi Takebayashi

of Kyoto University, which models the debris and sludge flow in rivers. In 2009,
the model was added to RIC-Nays, which is free software developed by RIC. The
simulation can predict the transport and deposition of debris flow and mud in the
horizontal distribution of erosion depths. The equations of debris and mudflow consist
of flowing water and sediment, which is treated as a single-phase continuous fluid
[5]. The mass conservation equation for the mixture of water and sediment is as
follows:
∂h ∂hu ∂hv E
+ + = (1)
∂t ∂x ∂y c∗
where,
∂t = time (s)
∂h = flow depth (m)
u, v = velocity in the x and y axis/direction (m/s)
c* = sediment deposition concentration in static deposition layer
E = bottom erosion velocity
When the cohesive characteristics of riverbed layers can be neglected and the
bed reaches equilibrium, where the rate of bed erosion is equal to the rate of sedi-
ment deposition, the rate of erosion of the bed can be estimated using the following
equation:
E
√ = c∗ tan(θ − θe ) (2)
u + v2
2
where,
θ = channel bed slope
u sin θx + v sin θy
sin θ = √ (3)
u2 + v2
where,
θx , θ y = bed slope in the x and y axis/direction
θe = bed slope balance refers to the mean depth of sediment concentration c
If it is assumed that a laminar flow layer is formed near the bottom and a turbulent
flow layer is formed in a laminar layer with a constant mean depth of sediment
concentration c, the following relationship is obtained.
(σ/ρ − 1)c hs
tan θe = tan ∅s (4)
(σ/ρ − 1)c + 1 h
where ∅s is angle of repose, the momentum conservation equation is as follows [8].
∂hu ∂huu ∂hvu ∂zb 1 ∂P τbx

+ + = −gh − − (5)
∂t ∂x ∂y ∂x ρm ∂x ρm
∂hv ∂huv ∂hvv ∂zb 1 ∂P τby

+ + = −gh − − (6)
∂t ∂x ∂y ∂y ρm ∂y ρm
where,
g = gravitational acceleration
zb = bed elevations
P = pressure (assumed as static pressure)
Then the debris flow density is as follows.
ρm = (σ − ρ)c + ρ (7)
where,
σ = sediment density
P = water density
τbx , τby = shear stress in the x and y axis/direction
When the turbulent flow area is dominant, the shear stress equation is as follows.
u
τbx = τy + ρfb (u2 + v2 ) √ (8)
u + v2
2
u
τby = τy + ρfb (u2 + v2 ) √ (9)
u + v2
2
If τy is the yield stress, use the following relationship for the non-cohesive
material:
15
c
τy = (σ − ρ)c ghs cos θ tan ∅s (10)
c∗
where fb = coefficient of friction, the equation is as follows.
Cmu 2
fb = a a = k/6 (11)
8

hs −2
5
4 (1 − c) 3 σ 2 13 hs
fb = kf + kd (1 − e )c + 9a 2
1 − (12)
25 2
c3 ρ dm h
where Cmu is the resistance coefficient of the sludge flow, kf = 0.16, kd = 0.0828,
and e is the particle inelastic coefficient, and dm is mean particle size of the sediment
from the debris flow. So, the basic elevation equation becomes:
∂z b E
=− (13)
∂t c∗
The simulation on iRIC: Morpho2DH displays the results in the form of a scalar of
flow velocity values. For example, a scalar display of the simulated flow velocity
under 2 times sediment discharge ratio conditions can be seen in Fig. 3. The results
indicate that under non-equilibrium sediment transport, more sediment supply, which
is a small debris flow model, can produce the increase of velocity in some conditions.
The velocity also increases the shear stress on the river bed. In these conditions, the
erosion will increase. This condition is demonstrated in cross-sections around 151
and 152. This cross-section shows that the erosion maximum occurs under two times
sediment supply, then erosion decrease consider to the increasing of sediment supply.
(b)
(a)
Crack
(c)
Fig. 3 The simulation result for velocity (a) and cross section elevation after 240 s (b), and condition
in the field (c)
Regarding to the research findings, simulation of river bank erosion should be

done before determining stream corridor reservations, especially along meandering
rivers in urban area [6].
5 Conclusion
The two-dimensional model for flow and riverbed morphology was performed using
iRIC: Morpho2DH. The model can simulate flow, bed degradation, aggradation,
which are similar to real river conditions. In general, the maximum flow velocity in
non-equilibrium conditions occurred when the sediment concentration was less than
two times the equilibrium condition. When the sediment rate was more than two
times the equilibrium condition, the aggradation will increase rapidly. The model
should be utilized when designing riverbank protections.
References
1. Ikhsan J, Fujita M, Legono D, Rahardjo AP, Harsanto P (2020) Dynamics of lahar-affected river
tributaries of the Progo river after the 2010 Mt. Merapi eruption, IOP Conf Ser Earth Environ
Sci 437:012009
2. Takebayashi H, Fujita M (2020) Numerical simulation of a debris flow on the basis of a two-
dimensional continuum body model. Geosciences 10(2):45
3. Nugroho AK (2015) Aplikasi Sig Untuk Evaluasi Sistem Jaringan Drainase Sub Das Gajahwong
Kabupaten Bantul. SEMNASIF
4. Nasional BS (2012) SNI ASTM C136: 2012: Metode uji untuk analisis saringan agregat halus
dan agregat kasar. BSN, Jakarta
5. Egashira S, Honda N, Itoh T (2001) Experimental study on the entrainment of bed material into
debris flow. Phys Chem Earth Part C 26(9):645–650
6. Murniningsih S (2018) Study of erosion and sedimentation in urban areas for river with meander
pattern. IOP Conf Ser Earth Environ Sci 191
Simulation of Debris Flow Using
“SIMLAR” in the Watershed of Gendol
River, Indonesia
Hendy Dwi Cahyo, Jazaul Ikhsan , and Ani Hairani
Abstract The eruption of Mount Merapi in 2010 released more than 140 million
cubic meters of pyroclastic material, which still settles on cliffs and riverbeds. Gendol
River is one of the rivers affected by the accumulation of sediment. When there is
heavy rain in the river’s upper reaches, rainwater will carry lava material at high
speed called debris flow, which will cause damage to the area in its path. Sediment
control buildings such as Sabo dams are built to reduce the impact of debris flows.
To predict the impact of debris flows, it could be used using a software application.
In this study, the debris flow simulation is used the SIMLAR V.2.1 application. The
simulation uses data, namely rain data, to calculate the flood hydrograph using the
Nakayasu method. The other data are grain size distribution data to determine the
sediment material and a topographic map in the form of DEM to determine the length
and area of the simulated watershed and modify the Sabo dam building’s existing
conditions in Gendol River. The simulation results show that at the beginning of
the simulation hour to 0.5 with a discharge of 0.299 m3 /s produces a debris volume
of 539.997 m3 with a velocity of 0.198 m/s, at peak discharge with a discharge of
107.729 m3 /s has a debris volume of 452,129.320 m3 with a velocity of 1.916 m/s.
The simulation at hour of 15 with a discharge of 2.909 m3 /s, produces a volume of
75,038.150 m3 with a velocity of 0.451 m/s.
Keywords SIMLAR V.2.1 · Sabo dam · Debris flow · Merapi
1 Introduction
Debris flow is a common occurrence in mountainous areas around the world, both
volcanic and non-volcanic. These floods have caused major property damage and
H. D. Cahyo · J. Ikhsan (B) · A. Hairani

Department of Civil Engineering, Universitas Muhammadiyah Yogyakarta, Yogyakarta, Indonesia
A. Hairani
610 H. D. Cahyo et al.
deaths, as well as significant morphological changes along riverbeds and mountain

slopes [1–4]. Indonesia is one of the countries with the most active volcanoes in
the world, which is more than 30% of the world’s active volcanoes in Indonesia
[5]. One of them is Mount Merapi, Mount Merapi is located on the island of Java
and administratively belongs to the Sleman Regency, Yogyakarta Special Region.
As an active volcano, of course, it has the potential to erupt at any time—it will
happen again. It is recorded that Mount Merapi experienced its last eruption in 2010.
According to [6]. The eruption in 2010 was the largest eruption of Mount Merapi
since 1872. The eruption in 2010 released more than 140 million cubic meters of lava
heading south, which resulted in the accumulation of lava deposits. It has the potential
to become a debris flow, especially in the Gendol River. A system of sediment control
was built, namely the Sabo dam, to anticipate debris flow impacts. It is hoped that the
Sabo dam will reduce the impact of the disaster due to debris flow in Gendol River.
The preparation of disaster mitigation is also one of the efforts to reduce the effects
of debris flow. One of the ways to mitigate disasters is to simulate debris flows. In this
case, the simulated lava flood was carried out using the SIMLAR application version
2.1. SIMLAR is a numerical modelling application to estimate the direction of the
distribution of the lava flood that will occur. This application also has a function
to analyze hydrology, analyze the lava flood propagation with 2D simulations, and
display the simulation results in the form of a Geographical Information System
(GIS) model to be used as an early warning system.
1.1 The Gendol River Watershed
Watershed (DAS) is defined as an area bounded by a ridge topography that receives

and collects rainwater and other nutrients and then flows it to the river network and
comes out at one point (outlet). The Gendol River Watershed is shown in Fig. 1.
1.2 Simulation Lahar (SIMLAR)
Simulation Lahar (SIMLAR) is a debris flow simulation application. The application

consists of three sub-chapters of the programs; namely, the first is a flood hydrograph
calculation program. The second is a hydrograph calculation program due to the
collapse of a natural weir. The third is a 2D debris-flow simulation program. To
estimate the debris flow direction, SIMLAR uses the debris flow modelling method
with a numerical model in which sludge water is also considered a fluid unit. The
debris flow simulation is based on a partial differential equation that regulates the
debris flow as follows:
Simulation of Debris Flow Using “SIMLAR” in the Watershed … 611
Fig. 1 Research locations in Gendol River watershed, Indonesia
1. Mass conservation equation
∂h ∂M ∂N
+ + =0 (1)
∂t ∂X ∂y
2. The momentum and force conservation equation (x direction)
∂M ∂(u M) ∂(u M) ∂(v M) ∂H τ bx

+β +β = −gh − (2)
∂t ∂x ∂x ∂y ∂x ρT
The momentum and force conservation equation (y direction)
∂M ∂(u M) ∂(v N ) ∂H τ by
+β +β = −gh − (3)
∂t ∂x ∂y ∂x ρT
where:
h = Flow height
t = Time (s)
M = Debris flow rate per unit width x direction (m2 /s)
N = Debris flow rate per unit width y direction (m2 /s)

β = Momentum correction
H = Flow depth (m)
g = Gravity
τbx = The component of the riverbed shear stress in the x direction
τby = The component of the riverbed shear stress in the y direction
ρT = Mass flow density
u = Average velocity in the x direction
v = Average velocity in the y direction
3. The equation for the conservation of sediment in the river

∂ Zb ∂q Bx ∂q By
c∗ + + =0 (4)
∂t ∂x ∂y
where:
c∗ = Concentration of riverbed sediments
q Bx = Sediment discharge in the x direction
q By = Sediment discharge in the y direction
q Bx + q By = The amount of base sediment (bedload)
4. Cliff erosion considered with the equations of Ashida, Egashira and Kamamoto
q p = q p∗ u ∗ d (5)

1 1 τ∗sc 3 h 1
q p∗ = F0∗ τ∗2 1 − 2 (6)
ε τ∗ d sin θ

u ∗ = gh I (7)
u 2∗
τ∗ = (8)
σ
gd ρ
−1
2
tan θ
τ∗SC = τ∗C cos θ1 1 − (9)
μ
where:
qp = Cliff erosion per unit width (m2 /s)
q p∗ = Cliff erosion per unit width (dimensionless)
u∗ = Basic shear speed (m/s)
I = The slope of the energy flow line
θ = The slope of the cliff

F0∗ = Coefficient = 0.01
d = Sediment diameter
5. The discharge formula in 2D simulation using Meyer—Peter and Muller

1,5
q∗b = 8 τ∗ − τ∗c (10)
where:
q∗b = Discharge volume per width (m3 /s)
τ∗c = The average value of the critical friction stress
τ∗ = Shield’s number
1.3 Sabo Dam
According to [7], the word “sabo” comes from Japanese, “sa” means sand, and
“bo” means prevention or control. The broad understanding is the erosion control
system, sediment, cold lava, and landslide countermeasures to protect human life
from the threat of disaster. The sabo system is a synthesis technology from mechan-
ical, civil engineering, and vegetative forestry construction with aspects of land and
river channel conservation in river catchments. The sabo dam building is a building
that controls the flow of debris or cold lava that occurs during heavy rains, and the
sabo dam is built across the river channel [8, 9]. The Sabo dam building’s working
principle is sediment control by holding, accommodating and flowing the material
or sand carried by the debris flow and bringing it downstream.
2 Research Method
2.1 Research Location
The research location that simulated debris flow was in Gendol River, which is admin-
istratively included in the Sleman Regency, Yogyakarta Special Region, Indonesia.
The watershed area used in this study is 55.663 km2 with a river length of 20.231 km.
2.2 Research Data
The research data used is secondary data obtained from Balai Sabo Yogyakarta. The
data acquired will be inputted to perform simulations on the SIMLAR application.
The author does not collect data directly in the research area. The secondary data
obtained are:
Rainfall Data
This study’s Rainfall data were rainfall data from Ngandong, Sorasan, Plosokerep
stations (Fig. 2). The rainfall data used are hourly rainfall data that occurred in 3
stations at the same hour with 3 h in April 2015. The rainfall selection is the most
significant rainfall intensity in the 2015–2019 period.
Sediment Material Data

The grain size analysis from Balai Sabo Yogyakarta found that d50 was coarse sand
with a grain size of 0.80 mm. The study of the characteristics of the lava sediment
in Gendol River shows that the lava sediment is coarse sand. The analysis results
are used to classify the type of debris flow, namely mudflow. Mudflow is a flow
dominated by sedimentary material with an acceptable size of 0.3–1 mm, and grain
below 0.1 mm is 10–30% (Fig. 3).
Topographic Data
The topographic data used were obtained from the website (http://tides.big.g.id/
DEMNAS/), which has a spatial resolution of 0.27 or about 8 m × 8 m. The data
is in the form of topographic information on DEM (Digital Elevation maps). The
model with the .tif format is then converted into Asc format to be inputted into the
SIMLAR software. The DEM map is then modified by increasing the river channel
bed’s elevation to be assumed as a Sabo dam building at the elevated point.
Fig. 2 Rain station location

100 100
90
86
80
Percent Passing (%)
70
60 63
50
40 43
30
27
20
17
10 9
0
0.01 0.10 1.00 10.00
Particle Diameter (mm)
Fig. 3 Grain size distribution chart [11]
3.1 Flood Hydrograph
The hydrograph used in this research is the synthetic hydrograph of the Nakayasu
method. The results of the calculation of the Nakayashu Synthesis Unit Hydrograph
on the SIMLAR application obtained the discharge value starting at 0.5 h (30 min)
to the 15th hour (900 min) as follows (Fig. 4).
120
100
Discharge (m3/s)
80
60
40
20
0
0 2 4 6 8 10 12 14 16
Time (hour)
Fig. 4 Nakayasu method flood hydrograph [10]

Fig. 5 Distribution map of simulation results
Table 1 SIMLAR simulation

Hour Discharge (m3 /s) Volume (m3 ) Velocity (m/s)
result
0.5 0.29 539.99 0.19
5 66.58 581,309.16 1.58
7 31.04 341,761.91 1.16
15 2.0 75,038.15 0.45
3.2 Debris Flow Simulation
The simulation is carried out with an initial time of 0.5–15 h according to hydrograph
calculations, and analysis is carried out every 30 min. The simulation results are then
processed using the Arc-Map application to produce a raster map of the distribution
of debris flow. The following is a map of the distribution of the simulation results in
the SIMLAR application.
The simulation results is shown in Fig. 5, it can be explained that the simulation
results on Gendol River are shown in Table 1.
4 Conclusion
The SIMLAR application program can model debris flows by producing a map of
each hydrograph distribution that has been predetermined. The length and the velocity
of debris flows are influenced by the river’s cross-section and the Sabo dam building
in the river. The simulation results show that at the beginning of the simulation hour
to 0.5 with a discharge of 0.299 m3 /s produces a debris volume of 539.997 m3 with
a velocity of 0.198 m/s, at peak discharge with a discharge of 107.729 m3 /s has a
debris volume of 452,129.320 m3 with a velocity of 1.916 m/s. The simulation at
hour of 15 with a discharge of 2.909 m3 /s, produces a volume of 75,038.150 m3 with
a velocity of 0.451 m/s. It means that the alert condition starts when a velocity of
debris flow is about 1.916 m/s.
Acknowledgements The author is grateful for University of Muhammadiyah Yogyakarta,

Indonesia, through the internal research grant 2021 and the Ministry of Education, and Culture,
Indonesia through PDUPT Research Grand 2021 which have supported author to conduct the
research and publicizing in this workshop.
References
1. Takahashi T (1991) Debris flow. IAHR monograph series, Rotterdam, Balkema

2. Hunt B (1994) Newtonian fluid mechanics treatment of debris flows and avalanches. J Hydraul
Eng ASCE 120(12):1350–1363
3. Huang X, Garcia MH (1997) A perturbation solution for Bingham-plastic mudflows. J Hydraul
Eng ASCE 123(11):986–994
4. Shrestha BB, Nakagawa H, Kawaike K, Baba Y (2008) Numerical simulation on debris-flow
deposition and erosion processes upstream of a check dam with experimental verification. In:
Annuals of Disaster Prevention Research Institute, Kyoto University, No. 51 B, pp 614–624
5. Pratomo I (2006) Klasifikasi gunung api aktif Indonesia, studi kasus dari beberapa letusan
gunung api dalam sejarah. Indonesian J Geosci 1(4):209–227
6. Kholiq MA (2017) Simulasi aliran banjir lahar pasca erupsi gunung merapi 2010 terhadap
keberadaan sabo dam di sungai gendol. J Teknisia 22(2):410–415
7. Munir MD (2019) Bangunan sabo dam, fungsi dan potensinya sebagai bagian dari geowisata
gunung api merapi. J Lingkungan dan Bencana Geologi 10(2)
8. Ikhsan J, Legono D, Rahardjo AP, Harsanto P, Fujita M (2020) Dynamics of lahar-affected river
tributaries of the Progo river after the 2010 Mt. Merapi eruption. In: Earth and Environmental
Science
9. Khairi MAW, Rozainy MR, Ikhsan J (2020) Smoothed particle hydrodynamics simulation for
debris flow: a review. IOP Conf. Ser. Mater. Sci. Eng. 864(1): https://doi.org/10.1088/1757-
899X/864/1/012045
10. Pengembangan Sistem Prakiraan dan Peringatan Dini Aliran Debris di Kali Putih Kabupaten
Jember
11. Yuniawan, R (2019) Hasil uji analisis saringan agregat halus dan kasar. Balai Sabo 1(1)
Wind-Generated Wave Simulation
on Payangan Beach Utilizing DELFT3D
Enggar Setia Baresi, Retno Utami Agung Wiyono,

and Wiwik Yunarni Widiarti
Abstract The Southern Coastal area of Jember has been damaged almost yearly by
coastal flooding and high waves, especially Payangan Beach. A study on high waves
is urgently needed for disaster risk preparedness. This study is conducted to model
the high waves and simulate the effect on the beach. Delft3D-Wave model is used
to provide a spatiotemporal characteristic of the event. The focus of the study is to
analyze the high waves on Payangan Beach generated by wind forces. The domain
model is a curvilinear grid with a grid size ranging from 130 m until 900 m. National
bathymetric data and CDS Copernicus wind data were utilized as data sources. The
simulation outcomes have shown that waves are generated far from the coastline
across the ocean inside the fetch area. The simulations show the significant wave
height increase during the high waves event according to specific meteorological
characteristics from 1.6–2.0 m on normal days and exceeding 3.2 m when the high
waves hit.
Keywords High waves · Delft3D · Wind-generated waves
1 Introduction
Shoreline is the point of intersection between land and sea. In this area, ocean energy
reacted to the land vice versa. The system that works in this reaction mostly caused
by the natural movement of the ocean which transferring energy into the system.
Coastal zones are directly affected by the ocean forces especially beach and near-
shore zone. Therefore, this particular area is the most dynamic in all coastal areas
[1].
E. S. Baresi · R. U. A. Wiyono (B) · W. Y. Widiarti

Civil Engineering Department, Universitas Jember, Jember, Indonesia
W. Y. Widiarti
620 E. S. Baresi et al.
The most common form of ocean energy is wave. The linear waves theory or small-
scale wave amplitude theory is derived from Laplace equation for irrotational flow.
The equation gives solution of periodic velocity for irrotational flow which later can
be utilized to derive equations from various waves characteristic for example water
level, velocity, particle acceleration, waves propagation speed, and others [2].
According to the previous study conducted in Southern Seas of Jember, this partic-
ular area has high amount of energy transfer about 190 MWh/m/year with significant
wave height up to 4 m [3]. In Payangan Beach, a beach located in Southern Seas of
Jember, high waves occurred annually e.g. 2018 [4]. There is no known study that
simulate spatial variation of high waves in this area. Thus, this study is crucial for
the various purpose related to high waves effect on the Payangan Beach coastal area.
This study aims to simulate high waves and analyze the significant wave height
in Payangan Beach in the last five years. Every year, this particular location hit by
high waves in late July when the atmosphere pressure shift in the Indian Ocean. But
only in 2020, the high waves occurred in May, two months prior to the usual period
of the event. This study focused to analyze the high waves on the study site on a
small-scale simulation by using national bathymetric data.
2 Methodology
2.1 Study Area
Payangan Beach exposed directly to waves coming from Indian Ocean that can
measure up roughly from 2 to 5 m in a bad weather condition. Average significant
wave height in Payangan Beach during normal days is ranging from 1.6 to 2.0 m.
This beach has natural defense against large waves and even massive tidal wave in a
form of sand dunes and highest slope compared to other two beaches nearby (Fig. 1).
113.5813 Payangan Beach
8.4364
Fig. 1 Study location in east Java (left); the zoom-in map of study domain and study site at Payangan
Beach (right)
Wind-Generated Wave Simulation on Payangan Beach Utilizing … 621
Table 1 Wind data distribution

Wind speed Wind direction (%) Days (%)
(knot) N NE E SE S SW W NW
0–3 0.00 0.00 0.00 0.00 0.14 0.00 0.00 0.00 0.14
3–6 1.74 2.90 1.88 0.39 21.11 0.24 3.77 0.43 32.46
6–9 3.38 4.11 2.71 0.24 16.91 0.24 5.89 0.97 34.44
9–12 2.42 2.46 1.21 0.14 13.91 1.01 6.76 0.97 28.89
12–15 0.14 0.05 0.10 0.00 1.59 0.29 0.24 0.05 2.46
15–18 0.00 0.00 0.10 0.00 0.77 0.29 0.14 0.00 1.30
18–21 0.00 0.00 0.00 0.00 0.10 0.05 0.00 0.00 0.14
21–24 0.00 0.00 0.00 0.00 0.05 0.05 0.00 0.00 0.10
30–33 0.00 0.00 0.00 0.00 0.05 0.00 0.00 0.00 0.05
Days 7.68 9.52 5.99 0.77 54.64 2.17 16.81 2.42 100.00
2.2 Wind Data Analysis
Wind data of this study was obtained from Class III Meteorological Station located
in Banyuwangi which was the closest wind station to the study site. The data that
was obtained from Class III Meteorological Station is the final wind data derived
from u10 and v10 wind data. In the equation below [5, 6], the ∅ means meteorology
wind direction angle in radians.
u = −|V| sin ∅ (1)
v = −|V| cos ∅ (2)

|V| = u2 + v2 (3)
In addition, spatial variation of wind data namely ERA5-Interim wind data was
obtained from Climate Data Store Copernicus that was measured by Aeolus satellite.
The data that was gathered classified within 3 speed interval, 0 < U < 3, 3 < U < 6,
6 < U < 9, up to 30 < U < 33.
Table 1 shows wind data distribution and Fig. 2 shows wind rose of the study
location. Based on the data, the main direction of wind is South (225°) which is
54.64% from total wind data.
2.3 Wave Simulation Using Delft3D
Delft3D-WAVE is able to compute various geospatial even meteorological parame-

ters, wave dispersion calculation, wave generation by wind force, non-linear inter-
action between the waves, and wave dissipation [7]. In this study, Delft3D-WAVE is
utilized based on the Simulating Waves Nearshore (SWAN) spectral model.
Fig. 2 Wind rose based on a peak wind direction; b mean wind direction
Fig. 3 a Depth generated from bathymetric data projected on ESRI visual; b simulation grid
In this study, bathymetric data that was used is the bathymetric data obtained
from Indonesian Geospatial Information Agency with 6-arcsecond resolution. Wind
input used in this simulation is taken from Class III Meteorological Station which
have temporal resolution of 24 h (daily). In addition, wind data from ERA5-Interim
provided by CDS Copernicus which have temporal resolution of 1 h (hourly) is used
to generate meteo file as input file in Delft3D simulation. The bathymetric data is
mapped in Fig. 3a. The grid used within the study domain ranges from 138.346 m
to 946.482 m shown in Fig. 3b.
3 Results
3.1 Simulation Results
Generally, results of Delft3D-WAVE simulations in 2017, 2018, 2019, and 2020

show significant wave height of 2.5 m to 4.5 m nearshore and around 5 m offshore.
The high waves in 2018 almost similar to those in 2017 with the difference margin
around 5–10%.
Figures 4 and 5 show simulation results based on SWAN modeling in Delft3D
when high waves occurred on July 25th, 2017. Figure 4a shows significant wave
height of 2.5–4.5 m at the coastal area. Those high waves may be the cause of coastal
flooding in the coastal area as occurred on July 25th, 2017. Figure 4b shows peak
wave period more than 11 s in offshore while in coastal area, peak wave period ranges
from 6 to 10 s. Figure 5 shows that waves approaching the beach at angle of 80° and
shifting to 70° then around 55° to 60° when the waves arrive at the beach.
Fig. 4 Simulation result on July 25th, 2017. a Significant wave height; b peak wave period
Fig. 5 Peak wave direction

Fig. 6 Comparison between

simulated significant wave
heights and CDS Copernicus
observation data on July
2017
When high waves occurred, maximum wave height at 20 km offshore from

Payangan Beach reached 3.9 m until 5.5 m. The simulation results coincide with
the observed data from CDS Copernicus.
Figure 6 shows comparison between simulated wave heights on July 25th, 2017
and CDS Copernicus observation data. The results show that the highest simulated
wave height was 5.8 m at 6 am, while based on CDS Copernicus observation data, the
highest wave height was 5.2 m at 7 am. The simulated results show close prediction
of CDS Copernicus data. The waves were generated from 8.55 m/s wind speed.
Simulation result in 2019 is different from those in 2017 and 2020 because the
wave height is outstanding. The peak direction (Fig. 7a) not from 80° but vary from
62° to 77.5°. The significant wave height reached 6 m to 7 m in nearshore of Payangan
Beach (Fig. 7b).
Figure 8a shows comparison between simulated wave heights on July 26th, 2019
wave height was 6.9 m at 8 am, while based on CDS Copernicus observation data,
the highest wave height was 5.9 m at 8 am. Although there are simulated results
overestimated the observed wave height, the simulation was able to reproduce the
similar time of maximum wave height at 8 am. The outstanding high waves were
also confirmed by the captured video shown in Fig. 8b.
Fig. 7 Simulation result on July 26th, 2019. a Peak wave direction; b significant wave height
Fig. 8 a Comparison between simulated significant wave heights and CDS Copernicus observation
data on July 26th, 2019; b Payangan beach high waves documentation [8]
Fig. 9 a Comparison between simulated significant wave heights and CDS Copernicus observation
data on May 2020; b Payangan beach high waves documentation on May 27th 2020 high waves [9]
Different from 2017 and 2019, in 2020 the high waves occurred not in July but in
May. Figure 9a shows comparison between simulated wave heights on May 27th 2020
wave height was 5.6 m at 6 am, while based on CDS Copernicus observation data,
the highest wave height was 5.9 m at 6 am. The waves were generated from 7.42 m/s
wind speed. The simulated results slightly underestimated the observed wave height.
However, the simulation was able to reproduce the similar time of maximum wave
height at 6 am. The high waves were also confirmed by the photo shown in Fig. 9b.
4 Structural Measures
In order to protect the beach along with the community living within the area of
study in Payangan Beach, one of the best solutions is to create safety measures based
Fig. 10 High waves a without sea wall as protective measures; b with sea wall as protective
measures
on the simulated wave height. Based on consideration and effectiveness according

to the events, the best measures considered is to build a sea wall.
Sea wall is a coastal defense structure built parallel to the coastline to mitigate
coastal flooding by blocking and or reflecting waves force to another direction so the
wave will less affect the area protected by the sea wall. Sea wall vary in shape and
type which based on the site condition. Payangan Beach condition require sea wall
because of the direct wave effect on the beach during high waves.
Figure 10 show illustration of Payangan Beach without sea wall (Fig. 10a) and
with sea wall (Fig. 10b). Without sea wall, waves will propagate to the coastal area
and inundate residential area. By constructing sea wall, waves will be reflected and
the residential area will be protected from severe inundation.
5 Conclusions
High waves occurred at Payangan Beach is significantly high because of the direct
exposure of wind and wave from the open ocean. Wind meteorological direction
is from the South with wind speed vary from 3.9 m/s to more than. When high
waves occurred, the wind speed vary from 5 m/s to 16 m/s and generate waves with
significant wave height of 3.4–5 m.
The high waves occurred in 2015, 2016, 2017, and 2018 show significant wave
height of 2.5–4.5 m nearshore and around 5 m offshore with difference margin around
5–10%. However, in 2019 high waves occurred with significant wave height reached
6–7 m in nearshore of Payangan Beach. The significant wave height creating massive
wave that visible from the distance. In 2020, high waves occurred two months prior
than the other five years which always occurred in late July.
Preventive measures must be taken into consideration to protect the coastal area
and residential area located in Payangan Beach. The most effective solution is to build
coastal protection structure to cope with high waves and inundation that caused the
coastal flooding. Sea wall has high efficiency to hold waves energy and to protect the
residential area from inundation. However, with this solution, there may be sediment
transport around the seawall. Further study is necessary to analyze coastal protection
structure and the effect to the beach.
References
1. Coastal Engineering Research Center (1984) Shore Protection Manual, Vol 1. US Army Corps
of Engineers, Washington DC
2. Triatmodjo B (1999) Teknik Pantai [Coastal Engineering]. Beta Offset, Yogyakarta
3. Warpindyasmoro HS (2018) Wave energy potency in east java coast. MATEC Web Conf 177:10–
18
4. High Waves Attacked a Shop in Payangan Beach. https://www.youtube.com/watch?v=6yZlZB
shfas. Last accessed 2020/12/11
5. Ostrenga D (2019) Derive wind speed and direction with MERRA-2 wind components. https://
disc.gsfc.nasa.gov/information/data-in-action?title=Derive%20Wind%20Speed%20and%20D
irection%20With%20MERRA-2%20Wind%20Components. Last accessed 2021/7/2
6. Yuwono N (1982) Teknik Pantai, Vol I. Biro Penerbit Keluarga Mahasiswa Teknik Sipil FT
UGM, Yogyakarta
7. Deltares (2014) Delft3-Wave User Manual. Deltares, Delft
8. High waves in Payangan on July 2019. https://www.youtube.com/watch?v=iATXCUStZUQ.
Last accessed 2021/6/1
9. High waves in Jember Payangan on May 2020. https://www.youtube.com/watch?v=rMFgTJ
DImaA. Last accessed 2021/6/1
Risk Reduction and Safety Management
Incorporating Cultural Attributes
into Disaster Risk Reduction-Based
Development Plans in Indonesia
Yusron Saadi
Abstract Frequent occurrences of natural disaster in Indonesia cause major prob-

lems in the form of financial losses and social disturbances. Disasters with tremen-
dous impacts disrupt the administration of state and hinder the development
processes, mainly because the events require expensive and long recovery process.
The issuance of Law No. 24 of 2007 on Disaster Management has triggered a
paradigm change in disaster management in Indonesia. The management of disaster
which previously tended to be responsive and spontaneous has shifted toward preven-
tive and more focused on disaster preparedness and risk reduction. This law explicitly
states the responsibility of the government in the implementation of Disaster Manage-
ment in the form of disaster risk reduction and it integration to development programs.
In addition, in the implementation of Disaster Management the government has the
authority in determining disaster management policies in line with national devel-
opment policies. But it is undeniable that there were many cases where different
approaches between government and community emerged and become sensitive
issues in dealing with disasters management particularly in disaster risk reduction.
In this paper a series of intensive community involvement in the form of focus
group discussions has been proposed and should be carried out by government in the
preparation of development plans. This includes early round of activities followed
by formal round of consultation before a draft document of development plans is
prepared. Accommodating local wisdom through this robust community engage-
ment is essential in the governmental strategy preparation process. It is a valuable
addition and an effective integration in the action strategies of integrated disaster
risk reduction so that an appropriate planning decision including the resulting miti-
gation plan and disaster-resistant infrastructure accepted by the community can be
implemented in the disaster prone areas.
Keywords Disaster management · Disaster risk reduction · Data and geospatial

information · Cultural attributes · Disaster-resistant infrastructure
Y. Saadi (B)
Center for Disaster Risk Management, Department of Civil Engineering, University of Mataram,
Mataram, Indonesia
632 Y. Saadi
1 Introduction
Disaster events in Indonesia have a high intensity since this archipelago country
is located in disaster-prone areas known as the Ring of Fire. Indonesian National
Disaster Management Agency, known as BNPB, revealed that there were 2952
disaster events in Indonesia during 2020 [1]. The number is based on data recorded
by BNPB between 1 January 2020 and 31 December 2020 at 3 pm of West Indone-
sian Time. The vulnerability to the disaster caused Indonesia to be referred to as the
Disaster Laboratory. Multiple natural hazards in various regions have contributed to
growing awareness that disasters can occur anytime. It is therefore necessary to make
adequate efforts to anticipate and to avoid the devastating impacts of the disaster.
A disaster is an event or series of events that threaten and disrupt people’s lives
and livelihoods, caused either by natural factors and or non-natural factors or human
factors resulting in the emergence of fatalities, environmental damage, property
losses, and psychological impacts [2]. Disaster management is part of national devel-
opment, i.e. a series of disaster management activities before, during and after a
disaster. In Law No. 24 of 2007 on Disaster Management [2] it is stated that “the
implementation of disaster management is a series of efforts which include the deter-
mination of development policies that risk the onset of disasters, disaster prevention
activities, emergency response and rehabilitation”. One of the responsibilities of
the government in implementing disaster management is in the form of disaster
risk reduction and integration of disaster risk reduction with development programs
while the authority of the government in the implementation of disaster management
includes the establishment of disaster management policies in line with national
development policies [2]. At present there are many government policies or legal
products that regulate the implementation of disaster management as stipulated in
the Government Working Plan in 2007 which was promulced through Presidential
Decree No. 19 of 2006 where disaster mitigation and management has become one
of the 9 national development priorities.
It is recognized that the perspective of government is often considered as contra-
dictive to the perspective of people who are affected by disaster and disrupting the
recovery process and future mitigation plan. This is where the role of disaster manage-
ment actors must have competence and a fundamental understanding of the concept
of disaster to bridge these differences including readiness in the management of data
and geospatial information which is indispensable to minimize losses and accelerate
the rehabilitation and reconstruction process in the affected areas.
The impact of disaster on the community is strongly influenced by the level
of vulnerability to the threat of such disasters. In order to understand the aspects
that make people vulnerable to disasters, factors such as economic, socio-cultural,
institutional, political and psychological factors that make up people’s lives should
be considered. Thus, the vulnerability of disasters is basically socially constructed
[3].
According to Setianto [4], when dealing with disasters people used to embrace
and adopt the belief of “fatalism” that nothing can be done against disasters, people
Incorporating Cultural Attributes into Disaster Risk Reduction … 633
must live with disasters and accept disasters as they are. Nowadays, people try to
reduce loss of life and property with measurable pre-disaster preparation in line with
disaster management. The problem is that academic studies of disasters in Indonesia
are still very limited [5] and from the social aspect of disaster, the high frequency
of disasters does not necessarily raise interest to conduct studies [6]. Furthermore,
Abdullah [5] mentioned that a disaster is always considered as a new experience
and something unprecedented. It is considered as something that has not become a
collective knowledge and experience so that it has not been integrated in life and
social policy.
2 Disaster and Disaster Management
Disaster Management is a science that studies disasters including all related aspects,
especially disaster risk and prevention of the disaster. The cooperation of all parties
involved in disaster management is important, therefore there must be equalizing
perceptions about disaster and disaster management. This is conducted through an
agreed system or rules of play, namely Disaster Management System. The parties
involved must carry out comprehensive and continuous work programs or activ-
ities in each field. Periodic and momentary responses will lead to failure in the
implementation of work programs since disasters will occur repeatedly [7].
Figure 1 shows the disaster management cycle based on the basic format put
forward by Carter [8] and has been developed by Setianto [4] according to disaster
management needs. This picture describes that if steps or activities have been taken
during the Prevention and Mitigation and Preparedness phase and disaster occurs
then the Emergency Response phase has started and continued with Post-Emergency
phase before returning to the Prevention and Mitigation phase. The Prevention and
Mitigation phase is necessary to deal with the possibility of future disasters and
should be interpreted as the Prevention and Mitigation phase in the next Disaster
Management Cycle.
Natural disaster events are irregular events disrupting normal pattern of life. The
irregularity in frequency, location and intensity cause natural disasters are compli-
cated to predict. Prevention, reduction, avoidance and self-recovery activities from
disaster impacts require a series of activities, both before, during and after a disaster,
referred as natural disaster management or disaster management.
Reflecting on various disasters in the past, disaster management actors must
include elements of the government, non-governmental institutions and the commu-
nity. Apparently, there is increasing recognition of the value of community-led initia-
tives that facilitate emergency management, risk reduction and community resilience
[9]. In order for disaster management to run effectively and efficiently in accordance
with the objectives of Law No. 24 of 2007 [2], disaster management actors must be
competent and certified [10]. Therefore, the role of Disaster Management Profes-
sional Certification Body becomes very important and needs to be empowered. In
634 Y. Saadi
Fig. 1 Disaster management cycle (adopted and translated with permission from Setianto [4])
addition, the readiness of data management and geospatial information is very neces-
sary to minimize losses and accelerate the rehabilitation and reconstruction process
in the affected areas.
In recent years various parties have prepared disaster risk maps but the lack of
standardization in the preparation resulted in the production of various types of
disaster risk maps developed by each institution. Setianto [4] stated that fundamental
understanding of the concept of disaster becomes a strong basis in mapping disaster
risk that can be applied to a Geographic Information Systems (GIS) and presented
spatially in the form of Threat Maps, Vulnerability Maps, Capacity Maps and Disaster
Risk Maps respectively. Figure 2 describes methods containing stages required to
produce a comprehensive Disaster Risk Map in accordance with disaster management
objectives.
Figure 2 indicates the importance of data collection. The availability of data is
essential and makes considerable contribution to analyzing and diagnosing broad
scale and discrete problem. Understanding the larger scale problems mean that inap-
propriate areas for development can be avoided completely. Detailed data allow the
authority to identify the potential hazards that are likely to be encountered.
In developed countries with a broad experience on disaster management, the devel-
opment plans includes requirements relating to consultation with the community.
Fig. 2 Risk mapping methods (adopted and translated from StIRRRD [11])
Development plans, often referred to as strategic plans, are prepared by local coun-
cils with input from the community [12]. As such, any type of strategic plan produced
by local council must consider the avoidance or mitigation of natural disasters.
3 Cultural Attributes in Disaster-Oriented Development

Plans
The existence of Law No. 24 of 2007 on Disaster Management [2] shows that devel-
opment policy in Indonesia that was previously less disaster-minded slowly but surely
began to change. Previously disaster management tended to be responsive and spon-
taneous. Now disaster management has started to move towards preventive and more
focused on disaster preparedness and risk reduction. Disaster risk reduction is one
of a series of disaster management activities that take place in conditions before
disasters occur (pre-disaster).
At a time when the role of the government is demanded more in disaster risk reduc-
tion, there are often differences in definitions and attitude towards disasters between
the government and the community alike. The government’s perspective in disaster
management is often perceived the opposite way by the community in the location
of disasters. Research conducted on one of the villages on the southwest slope of
Mount Merapi in Central Java Province shows that there is a difference in perception
between the government and the community on the threat of volcanic eruptions [6].
The government has been in a long persuasion to transmigrate villagers to one of
the islands in Indonesia but the element of rejection appear within the community
who decided to stay put on the slope of the volcano and adjusting themselves to
636 Y. Saadi
the dangers based on local knowledge systems in the form of direct monitoring and
story-telling of hereditary myths about volcanic activities. From economic aspect,
when government set the exclusion zone as preparedness measures for volcanic
hazard, the community often sees it as controversial. Recent study in other continent
severely affected by a prolonged volcanic eruption suggests that the governmental
decision to extent the exclusion zone because of volcano increasing activity met with
large disagreement, particularly from people who lost income during the period of
extension [13].
In the case of an earthquake in West Sumatra Province, research by Effendi [14]
showed that local people believe natural disasters are not a symptom of nature. The
indigenous Minangkabau view and impose natural hazards as well as all natural
symptoms as natural or normal as possible within the framework of their diverse
cultural cosmology. Some people view earthquakes as entirely natural events that
are scientifically understandable but others view the hazard as divine interference
with human life so that they can only be understood in religious discourse as a test
or punishment [6].
The results of the above research show that socio-cultural diversity in Indonesia
affects the perceptions toward natural disasters, the impacts or victims caused, and
post-disaster management. Understanding the cultural system will help the govern-
ment in trying to find answers of why a community acts on a particular pattern of
natural disasters as well as their impacts. By exploring this potential, it is expected
that the community will be capable to independently manage pre-disaster, emergency
response, rehabilitation and reconstruction process.
The government should not feel right on its own and ignore the values and norms
within the community in the form of indigenous knowledge that speaks of natural
events and the potential consequences. Various efforts in the form of technocracy
and partnerships with stakeholders provide opportunities for people to live better
and not become victims in the event of a disaster. Technocratic actions supported
by local wisdom will confirm that community development in the framework of
mitigation and social development is based on high cohesion [6]. In the framework of
adoption of local wisdom, mitigation that maintains harmony and local cultural-based
will be adaptive and can be transformed into a modern mitigation system through
partnerships with governments and non-governmental organizations to produce a
more effective and successful mitigation systems. This will strengthen the mitigation
strategies already attached to the community.
In planning infrastructure development, the government should accommodate
local wisdom in the community. In addition, knowledge of the history of disasters
obtained from local community will greatly help the development actors in planning
the right infrastructure. The results based on theoretical calculation may not be suit-
able for different region and produce an over-estimated or under-estimated result.
Marking flood water level, for instance, has been recognized for a long period of
time. The evidence of flood levels with the year of occurrences can be seen in a wall
or any surviving infrastructure.
The practice of engraving the flood level on tablets of stone, set in gardens or house
walls seem to have started more than three centuries ago [15]. In more simple fashion
Fig. 3 Flood level-marking based on personal information (Courtesy of West Lombok public works
and housing office, 2017)
the determination of flood elevation plans in the planning of river embankments pay
attention to traces of flooding based on hereditary experience of the surrounding
community (see Fig. 3).
In certain infrastructures such as reservoirs, operation and maintenance are based
on standard operating procedure (SOP) that has been prepared. Other anticipation
measures that Operational and Maintenance Officers must be equipped with knowl-
edge of the Emergency Preparedness Plan (EPP) including warning systems and
activities to increase public awareness of possible hazards [16]. It is important that
the SOP and EPP accommodate the environmental consideration and the involvement
of the community in deciding if the risk is acceptable for their particular area.
Countries with advanced knowledge in disaster management, prepare the develop-
ment planning based on the need for reducing risk and emergency management plans.
Experiences show that Indonesian are familiar with discrete hazardous events such
as annual flooding and forest fires as well as ongoing adversity resulting from disas-
ters. Adaptive strategies have been developed to ensure community wellbeing. These
strategies may works well for small-scale disasters but the community requires a more
effective strategy to anticipate larger-scale disasters. Effective strategies are perceived
by community as a key component in ensuring well organized emergency manage-
ment response. Government officials as disaster management actors are required
to be in the forefront in implementing effective strategies and the community are
eagerly expecting their part.
A schematic concept has been proposed to be applied in Indonesia within the
context of disaster risk reduction. Figure 4 shows that consultation with the commu-
nity is an important aspect in the process of preparing development plan. It should
begin with engaging communities to allow them to forward opinion based on expe-
rience and local knowledge. During the early round of the processes, the community
638 Y. Saadi
Fig. 4 Conceptualization of incorporating cultural attributes into development plans
has the opportunity to incorporate traditional knowledge, values and practices into
disaster related development plans. The community can persuade government that the
community perspective in disaster management and recovery initiatives are relevant
to the government plan.
Under early round of the process, a large amount of information collection occurs.
The relevant existing knowledge may have to be reviewed and the gaps with the
existing information gathered by government may need to be identified. Once the
required information has been obtained, the process continues to formal round of
consultation with the involvement of the experts. The consequences of development
planning associated with natural hazards require specialist reports. During this formal
round of consultation, constraints and limitation have been taken into account in the
preparation of development plans. The end result is the development plans which
incorporate the risks of development associated with natural disasters in a manner that
is aligned with the expectation of the community. In the event that developments are
considered in areas subject to natural disasters, the primary focus is on the likelihood
of the event with consideration of the potential consequences.
Disaster risk-based approach allow for a holistic consideration of both the conse-
quences and the likelihood of a disaster, thereby allowing the government to make
robust planning decision regarding natural disasters and future development. Devel-
opment planning accompanied by a pre-event recovery planning [17], for example in
the form of licensing and approval related to potential disasters, and also considers a
“development moratorium” in which development decisions are halted for a certain
time limit as well as zoning for evacuation infrastructure. For disaster risk reduction,
land administration planning must be made in such a way, disaster risk-based and
accommodate cultural attributes. It is an inevitable that the regional development
planning should incorporate traditional knowledge, values and practices to address
disaster-related risk.
4 Conclusions
Traditionally the planning approach for addressing natural disaster has been based on
the likelihood of the occurrence of an event and disaster risk reduction have always
been focused on the government. Recent developments require the need to increase
community involvement in activities related to disaster risk reduction.
Based on the above description, some conclusions can be drawn as follows:
1. High frequency of disasters in Indonesia has led to the establishment of Law
Number 24/2007 on Disaster Management as well as triggering a paradigm
change in disaster management from responsive and spontaneous approaches
to a preventive and more focused on preparedness and disaster risk reduction.
2. The existence of socio-cultural diversity in Indonesia which causes multi percep-
tions on natural disasters, the level of impacts or the number of casualties,
and post-disaster management requires government to understand such cultural
system in order to obtain information about community responses toward certain
pattern of natural disasters along with their impacts.
3. Disaster management actors must be competent and certified, have a funda-
mental understanding of the concept of disaster risk reduction including readi-
ness in the management of data and geospatial information that is indispensable
to minimize losses and accelerate the rehabilitation and reconstruction process
in the affected areas.
4. In the implementation of disaster risk reduction and infrastructure development
planning, the government should accommodate local wisdom in the form of
traditional attributes and values within the community including the absorption
of information about the history of disasters from indigenous people.
5. It is important that there is a robust community engagement and communication
process to enable community and the government to reach an agreement on risk
acceptability as well as mitigation option. Consultation with the community is
a key aspect in the process of preparing strategies for disaster risk reduction. A
concept of development planning related to disaster risk reduction should take
into account the risks of development associated with natural hazards.
Acknowledgements The author wishes to thank the GNS Science of New Zealand for their
support during New Zealand Comparative Study Programme 2015 through the project of Strength-
ened Indonesian Resilience: Reducing Risk from Disaster (StIRRRD). The appreciations are also
addressed to Dr. Agung Setianto for his permission to adopt and translate the picture of Disaster
Management Cycle and to the West Lombok Public Works and Housing Office for providing
photographs presented in this paper.
References
1. BNPB (2021) Indonesian disaster infographics 2020 in www.bnpb.go.id (in Bahasa Indonesia)
640 Y. Saadi
2. Indonesia State Gazette (2007) Law of the Republic of Indonesia number 24 year 2007 on
disaster management. State Gazette of the Republic of Indonesia Year 2007 Number 66 (in
Bahasa Indonesia)
3. Twigg J (2001) Sustainable livelihood and vulnerability to disaster. The Disaster Mitigation
Institute
4. Setianto A (2015) Remote sensing application and geographic information system for mapping
disaster prone areas. In: Proceeding of disaster risk reduction action plan workshop, strength-
ening Indonesia’s resilience through disaster risk reduction, pp 33–44. Mataram (in Bahasa
Indonesia)
5. Abdullah I (2008) Construction and social reproduction of natural disasters, working papers in
interdiciplinary studies No. 01. Graduate School of UGM, Yogyakarta (in Bahasa Indonesia)
6. Anantasari E (2015) Living harmoniously with disasters: integration of socio-cultural aspects
(local wisdom) in disaster risk reduction. In: Proceeding of disaster risk reduction action
plan workshop, strengthening Indonesia’s resilience through disaster risk reduction, pp 45–52.
Mataram (in Bahasa Indonesia)
7. Nurjanah D, Kuswanda D, Siswanto A (2012) Disaster management. Alfabeta, Bandung (in
Bahasa Indonesia)
8. Carter WN (2008) Disaster management: a disaster manager’s handbook. Asian Development
Bank, Manila
9. Kenney CM, Phibbs S (2015) A Maori love story: community-led disaster management in
response to the Otautahi (Christchurch) earthquakes as a framework for action. Int J Disaster
Risk Reduction
10. Saadi Y (2016) The role of universities in capacity building of disaster management actors.
National Disaster Risk Reduction Month Commemoration, StIRRRD Technical Session
(TS#3), Manado (in Bahasa Indonesia)
11. StIRRRD (2015) Handout of disaster risk reduction action plan workshop, strengthening
Indonesia’s resilience through disaster risk reduction, pp 49–57. Mataram (in Bahasa Indonesia)
12. Grace ES (2015) Incorporating natural hazards into development plans in New Zealand,
strengthening Indonesia’s resilience through disaster risk reduction. New Zealand Comparative
Study Programme
13. Monteil C, Barclay J, Hicks A (2020) Remembering, forgetting, and absencing disasters in the
post-disaster recovery process. Int J Disaster Risk Sci 11:287–299
14. Effendi N (2007) Disaster: experiences and cultural values of Minangkabau people. Indonesian
Soc 33/2 (in Bahasa Indonesia)
15. Archer D (1992) Land of singing waters, rivers and great floods of Northumbria. Spredden
Press, Stocksfield, England
16. Saadi Y (2013) Post-construction problems of Embung in Lombok Island and the operation
and maintenance works. Proc Eng 54:648–660
17. Becker J (2015) Pre-event recovery planning, strengthened Indonesian resilience: reducing risk
from disaster (StIRRRD), New Zealand Comparative Study Programme
Identification Risk Potential Hazard
of Railway Project Based on the Work
Breakdown Structure to Improve Safety
Performance
D. V. Aryanto, L. S. Riantini, R. A. Machfudiyanto, and Y. Latief
Abstract Railway construction work is carried out in developing or developed coun-

tries around the world. The railway was built as a means of mass transportation
to support connectivity between places. Railway construction work is one of the
complex construction projects. Construction accidents can hamper many work activi-
ties. It is not uncommon for construction accidents to occur in construction that affects
people, property, public safety, and the environment. For this reason, it is necessary to
identify and analyze the risk of hazards from the beginning to prevent the occurrence
of these construction accidents. This study aims to identify the potential hazards
in each work package by reviewing work methods, work activities, and WBS-based
resources for standardized railway construction using a qualitative approach. This re-
search is expected to produce safety risk sources with the highest potential hazards in
the railway construction works. The risk of construction accidents can be reduced or
even zero risks of construction accidents (zero accident) during construction project
activities.
Keywords Railway · Work breakdown structure · Risk · Safety performance
1 Introduction
In the past 5 years, Indonesia has participated in building infrastructure to become

a developed country. One of the infrastructures currently under construction now
is the Railway construction project. This is in line with the concept of the National
Railway Master Plan [1], which was initiated by the Directorate General of Railways
which aims to solve problems due to limited road service capacity so that the rail-
ways are ready for the implementation of national transportation capable of ensuring
the movement of people and goods throughout Indonesia. Advanced technology,
complex structure, quality, and high technical standards are required for long-term
D. V. Aryanto (B) · L. S. Riantini · R. A. Machfudiyanto · Y. Latief

Universitas Indonesia, Jawa Barat 16424, Indonesia
642 D. V. Aryanto et al.
use [2]. Therefore, in planning the construction of a railway track, it must be done
carefully with to obtain effective and efficient development results [3].
In general, the construction project building process is an activity that contains
a lot of dangerous elements or substances [4]. Based on the study [3], Railway
line construction has been investigated as one of the projects that causes the highest
number of construction accidents. Particularly in the construction of the Burma-Siam
Railway, the number of workers who fell victim to work accidents was 385.45 per
1000 workers [5]. The increase in cases of construction accidents that occur is due
to a combination of unsafe acts and unsafe conditions [6].
In carrying out a railway construction project, the project plan must be appro-
priately considered, so that timeliness, quality, and cost can be achieved [7]. The
planning process of the construction of the Railway Line construction needs to be
considered carefully to achieve success in the rampant construction of the Railway
Line construction. Correct and systematic planning and preparation of stages is the
key to a successful project, for that WBS is needed as a basis for project planning.
WBS is very important in project planning. Planning that begins with the formation
of the WBS needs to be explained as each construction activity as well as a struc-
tured work breakdown according to the scope of work and the standardized levels to
produce good construction products. Each level of WBS that has been defined brings
WBS to a level that is more and more complex namely activity. So that activities are
significantly affected by risks, and that risks will impact work safety [9]. Thus, in
this research the author will use the WBS method to identify the risks that occur in
each activity in a work package to minimize all potential hazards.
2 Research Objectives
Identify sources of risk that are potentially dangerous and affect Construction Safety
in Railway Line construction based on standardized WBS.
3 Literature Review
3.1 WBS of Railway
Railway Railroad tracks are a commercial service responsible for operating the
rail transportation system that attracts passengers or goods [10]. Based on [11] as
stated in Chapter 1 Article 1 paragraph 3, railway infrastructure is railway lines and
stations, including the facilities necessary for the railway facility to be operated.
Rail supporting facilities complement the operation of rail transportation, which
can provide convenience and comfort to users of rail transport services. Railway
Identification Risk Potential Hazard of Railway Project Based … 643
infrastructure in more detail can be classified as Rail lines or roads, Station buildings,
Bridges, Signals, and telecommunications.
Work Breakdown Structure Work Breakdown Structure (WBS) is a hierarchical
arrangement of project components that breaks the project deliverable into smaller
parts, making it easier for the project team to manage and implement [12]. Meanwhile,
according to PMI [13], WBS is a foundation tool to help the planning team project
overcome big project uncertainties by making it a project with a lower level of
uncertainty. The creation of the WBS produces a foundational document that provides
the basis for planning costs, resources, project schedules, and changes [14]. WBS
has benefits that determine the success of the project; some of the benefits of WBS
are to reduce the complexity of the project, create a schedule and cost estimation,
compile a budget, plan risk management, and identify activities.
3.2 Construction Safety Performance
According to Endroyo et al. [15], the definition of construction safety is safety which
includes 4 aspects, including worker safety (safe for labor), community safety (safe
for people), construction safety during the project development process (safe for the
property) and safe environment around the work location (safe for the environment).
Safety performance is treated as a concept that revolves around maintaining a safe
workplace. This can be achieved through the organization’s safety culture and worker
safety behaviors in the workplace environment or the ability to minimize work-related
injuries in the workplace [16].
There are 6 indicators for construction safety performance: safety awareness,
safety records, safety costs, accident costs, and the safety and productivity scores [17].
Safety performance has different indicators according to each researcher. According
to [18], the work safety performance can be measured by 5 aspects: Injury severity
rate, injury frequency rate, Project accident cost figures, average days change per
disabling injury, and the number of accidents at work.
3.3 Risk
Risk is a measure of the probability or severity of an impact detrimental to life, health,

property, or the environment [19]. Another definition of risk is an event or factors
that are uncertain, and if they occur, it can negatively impact project objectives [12].
If these risk factors occur, it will increase the likelihood of not achieving the project
objectives [20]. One way to achieve project objectives is to manage risk in the entire
project implementation process [21].
According to Hamid et al. [6], the causes of construction safety risks are influenced
by park behavior and conditions then are categorized as the important cause and
Fig. 1 Research flow chart
other causes. The among causes are (1) Use of defective or damaged equipment;
(2) Failure to apply the use of PPE from its workers; (3) Human Factor—ex: under
the influence of drugs; (4) Lack Of commitment from workers to policies safety; (5)
Poor construction site management.
4 Methodology
This research was conducted with a qualitative approach to answer research objec-
tives. The survey and discussion were conducted using a structured research instru-
ment in the form of a questionnaire to experts from railway construction work with
more than 10 years of experience. The research flow can be seen from Fig. 1.
5.1 WBS Standard of Railway Construction
In the WBS preparation process, the determination of the WBS level is from archival
analysis with the benchmarking Bill of Quantity (BOQ) carried out 3 railway projects
in Indonesia, the international railway regulatory standards, and refers to the Railways
Law No. 23 of 2007. Then the WBS arranged into 6 levels/levels, which consists of
WBS Level 1: Name of Sub-Project; WBS Level 2: Group of Work; WBS Level 3:
Type of Work; WBS Level 4: Work Package; WBS Level 5: Work Activities; WBS
Level 6: Resources (Labor, Materials, Equipment).
Experts in the field of railway construction need to validate the WBS to make
standardized WBS. 3 experts were used as resource persons with at least 25 years of
work experience in the railway construction field and understand the WBS concept in
general construction work, especially in railway construction project work. The vali-
dation of WBS is to determine whether the independent variable in the form of stan-
dard WBS (level 1–6) obtained from the archival study is under the standardization
of railway construction work.
From the results of a literature study of archival data and the validation form the
experts, it was obtained WBS Railway construction standards for 6 sub-projects,
namely: Preparatory Work; Railroad Works; Rail Bridge Works; Tunnel Work;
Station Building Works; Railway Operation Facility Works.
5.2 Identification of Risk Affecting Safety Performance
The risk identification process is carried out from the WBS, which is viewed from
the lowest level of WBS as at level 4: Work Packages, in which each work package
is detailed per activity with some literature that looks at potential risks to 4 aspects,
namely: Labor, Property (Material and Equipment), Environment, and Public Safety.
The literature review produces as many as 985 risks that affect construction safety
performance (229 risks for Railroad Works, 373 risks for Rail Bridge Works, and
383 risks for Tunnel Works).
Then on this potential risk, a validation process is carried out with an expert in
the field of safety or an expert in railway projects by discussing whether if the risk
is relevant to the work package.
By the validation results, those identified risks are admissible and there are 19
eliminated risks eliminated so that in total, there are 966 risks associated with
construction safety performance (229 risks for Railroad Works and 365 risks for
Rail Bridge Works, and 372 risks for Tunnel Works). The risk variables that experts
have validated are then carried out by pilot survey testing to test whether all project
workers easily understand the identified risks.
5.3 Risks Assessment
The validated risk variable is then calculated the risk value to determine the level
of risk belonging to the ‘high’ category based on the assessment matrix in PMBOK
and through respondents’ opinions and expert validation. The number of respondents
used in the study was 30 people with the criteria of respondents who filled out the
frequency and impact questionnaires based on at least 5 years of experience working
in railway construction. The risk assessment is obtained by multiplying the frequency
rates and severity rates of each risk.
Risk = Pr obabilit y × I mpact (1)

Table 1 Probability rate scale indicators

Scale Rating Remarks Weight
1 Very low Very unlikely to occur, only under certain conditions 0.1
2 Low Unlikely to occur 0.3
3 Moderate Possible occur in every condition 0.5
4 High Likely to occur 0.7
5 Very high Very likely to occur 0.9
Table 2 Severity rate scale indicators

Scale Rating Remarks Weight
1 Very low Has no impact on construction safety performance 0.1
2 Low Little impact on construction safety performance 0.3
3 Moderate Sufficient impact on construction safety performance 0.5
4 High Impact on construction safety performance 0.7
5 Very high Very impact on construction safety performance 0.9
This risk assessment is carried out by distributing questionnaires using a Likert

scale of 1–5 for frequency rate and severity rate with the following indicators scale
(Tables 1 and 2).
Then, the multiplication result between the frequency level and the impact level
is converted according to the risk assessment matrix referring to the PMBOK 6th
edition and generate risk category, as follows (Table 3).
Based on the risk assessment results that has been carried out, there are 37 risks
identified as high-risk levels. The method for assessing high risk is based on data
processing results generated from 30 qualitatively analyzed respondents who refer
to the PMBOK assessment matrix, with the following details (Table 4).
From the identified high risks, initial control can be carried out against high risks;
the results of discussions with experts are that several preventive actions can prevent
these risks: including: Completing workers’ PPE, installing K3 signs, completing
work safety equipment, conducting safety morning talks before work, Make Work
Instructions for each job and tool operation, and Ensure safe worksite conditions.
Table 3 Risk category

Risk Score Risk Category
0.18-0.72 High Risk
0.06-0.17 Moderate Risk
0.01-0.05 Low Risk
Table 4 Highest risks

Work package Activity Risk affect OHS Risk score Risk rank
performance
Railroad works
Soft rock Mechanical excavation The fall of heavy 0.228 13
excavation equipment due to
unstable soil
A landslide occurred in 0.210 18
the excavation
Stony soil disposal Traffic accident due to 0.229 12
scattered land on the
arterial road
Rock excavation Stone blasting Crushed or hit by 0.249 3
blasting debris
Noise resulting from 0.228 13
explosion
Asphalt pavement Mechanical excavation People buried by 0.248 5
excavation landslides
Structure Stony soil disposal Lack of waste 0.195 24
excavation control/project
material waste
Precast concrete Precast concrete Precast concrete fell on 0.186 27
drainage drainage the workers
manufacturing
Ballast work Stone sowing Heavy equipment 0.182 36
crashed into nearby
workers/facilities
Sub ballast work There was a landslide 0.199 21
(soil, rocks, etc.)
Rail bridge works
Land clearing and Carry out land clearing Accidents during 0.190 25
stripping and stripping equipment
mobilization
Rock excavation Stone blasting Contaminated 0.230 8
underground water
Precast concrete Erection of precast Workers are hit by 0.216 16
piles concrete piles piles during lifting
Cast concrete drill Wash drilling Dirty/slippery roads 0.203 20
piles due to scattered soil
Abutment Formwork work Formwork collapse 0.213 17
work/column/pier Concrete works Broken sling crane 0.181 37
head (concrete cast
in situ) Fall from a height 0.232 7
Elastomer bearing Bearing pad Fall from a height 0.233 6

pad installation
(continued)
Table 4 (continued)
Work package Activity Risk affect OHS Risk score Risk rank
performance
Post-tension Setting girder Crashed by precast 0.184 30
prestressed steel material
girder Stressing the bridge Got electric shock 0.331 1
girder
Transportation and Precast concrete Crashed by precast 0.184 32
storage of precast delivery units material
concrete Storage of precast Crashed by precast 0.184 32
concrete units material
Erection using the The placement of the A worker is hit by a 0.189 26
double crane crane on two different crane
method sides of the bridge Crane falls 0.182 35
Girder placement on Girder fell and hit the 0.207 19
bearing pad worker
Diaphragm work Stressing the bridge Got electric shock 0.326 2
girder
Implementation of Clamped beam 0.183 34
segmental diaphragm segments
beams installation
Floor plate work Concrete works Fall from a height 0.184 31
Tunnel works
Regular excavation Land disposal Traffic accident due to 0.221 15
scattered land on the
arterial road
Rock excavation Stone blasting Ear disorders due to 0.230 10
noise
Stone removal Material splattered on 0.185 29
the road
Tunnel excavation Stone blasting Crushed or hit by 0.249 3
blasting debris
Ear disorders due to 0.230 10
noise
Contaminated 0.230 8
underground water
Tunnel curb Formwork works Formworks collapse 0.197 22
concrete
Longitudinal side Precast drainage Precast concrete fell on 0.186 27
drainage installation the workers
Crossing drainage Precast concrete is 0.196 23
damaged/cannot be
used again
6 Conclusions
From these results of the study, it can be concluded that 966 risks have been identified
in railway construction projects. There are 37 highest risks in sub-project of railway,
with details, 10 Railroad Works, 18 Rail Bridge Works, and 9 Tunnel Works, which
are the most dangerous risks in the stressing process of bridge girders in Girder
Installation Works. The identified risk factors can be used as a reference in preparing
the safety plan document. With identifying these potential hazards, preparing a safety
plan for contractors and other stakeholders to determine risks, and evaluating the
project can be shortened because of the detailed explanation.
References
1. Kementrian Perhubungan DJP (2018) Reviu rencana induk perkeretaapian nasional (RIPNas).
Kementrian Perhubungan 134
2. Li Q, Liu R, Zhang J, Sun Q (2014) Quality risk management model for railway construction
projects. Proc Eng 84:195–203. https://doi.org/10.1016/j.proeng.2014.10.426
3. Yu QZ, Ding LY, Zhou C, Luo HB (2014) Analysis of factors influencing safety management
for metro construction in China. Accid Anal Prev 68:131–138. https://doi.org/10.1016/j.aap.
2013.07.016
4. Tsang YT, Fung IWH, Tam VWY, Sing CP, Lu CT (2017) Development of an accident
modelling in the Hong Kong construction industry. Int J Constr Manage 17(2):124–131. https://
doi.org/10.1080/15623599.2016.1222664
5. Migiro G (2018) The 10 deadliest construction projects in the world—WorldAtlas.com.
[Online]. Available: https://www.worldatlas.com/articles/the-10-deadliest-construction-pro
jects-in-the-world.html. Accessed 18 April 2020
6. Hamid ARA, Majid MZA, Singh B (2008) Causes of accidents at construction sites. Malays J
Civ Eng 20(2):242–259
7. Hamzaoui F, Taillandier F, Mehdizadeh R, Breysse D, Allal A (2015) Evolutive risk breakdown
structure for managing construction project risks: application to a railway project in Algeria.
Eur J Environ Civ Eng 19(2):238–262. https://doi.org/10.1080/19648189.2014.939416
8. Elsye V, Latief Y, Sagita L (2018) Development of work breakdown structure (WBS) standard
for producing the risk based structural work safety plan of building. MATEC Web Conf 147.
https://doi.org/10.1051/matecconf/201814706003
9. Satrio G, Latief Y (2018) Development of risk-based standardized WBS (work breakdown
structure) for safety planning of cable stayed bridge project. In: Proceedings of the international
conference on industrial engineering and operations management, vol 2018, pp 2694–2710
10. AREMA (2003) Practical guide to railway engineering. AREMA COMMITEE
11. Undang-Undang Republik Indonesia Nomor 23 Tahun 2007. Perkeretaapian 3:108 (2007)
12. PMBOK (2017) A guide to the project management body of knowledge (PMBOK Guide), 6th
edn. Project Management Institute, Newtown Square, Pennsylvania
13. PMI (2006) Practice standard for work breakdown structures. 2nd edn.
14. Wardahni NI, Latief Y, Machfudiyanto RA (2020) Development of safety plan to improve
OHS (occupational health and safety) performance for construction of dam (supporting infras-
tructure) based on WBS (work breakdown structure). IOP Conf Ser Earth Environ Sci 426(1).
https://doi.org/10.1088/1755-1315/426/1/012017
15. Endroyo B, Suraji A, Besari MS (2017) Model of the maturity of pre-construction safety
planning. Proc Eng 171:413–418. https://doi.org/10.1016/j.proeng.2017.01.351
16. Sanni-Anibire MO, Mahmoud AS, Hassanain MA, Salami BA (2020) A risk assessment
approach for enhancing construction safety performance. Saf Sci 121:15–29. https://doi.org/
10.1016/j.ssci.2019.08.044
17. Lu M, Cheung CM, Li H, Hsu SC (2016) Understanding the relationship between safety
investment and safety performance of construction projects through agent-based modeling.
Accid Anal Prev 94:8–17. https://doi.org/10.1016/j.aap.2016.05.014
18. de la Garza JM, Hancher DE, Decker L (1998) Analysis of safety indicators in construction
by Jesus M. de la Garza; Associate Member, ASCE, Donn E. Hancher, 2 and Lisa Decker.
Construction 81(Cii):2–4
19. Kementrian PUPR (2019) Permen PUPR No. 21, no. 1 1–25
20. Yao Y, Peng Y, Li X, Zhang A (2018) Research on safety risk management of civil construction
projects based on risk matrix method. IOP Conf Ser Mater Sci Eng 392(6). https://doi.org/10.
1088/1757-899X/392/6/062080
21. Sharma SK, Swain N (2011) Risk management in construction projects. Asia Pacific Bus Rev
Identification Factors of Safety Climate,
Awareness, and Behaviors to Improve
Safety Performance
in Telecommunication Tower
Construction at PT X
B. F. Athaya, L. S. Riantini, and R. A. Machfudiyanto
Abstract Telecommunication infrastructure has an important role in economic

development in Indonesia. Along with the development of the times, the need for a
network to access the internet increases in demand. This has resulted in increased
demand for telecommunications tower infrastructure development as the infras-
tructure for telecommunications operators in Indonesia to develop their businesses.
However, with the high development of infrastructure development, there is a possi-
bility that there will be risks to worker safety. This study aims to describe the safety
climate in the construction of PT X Telecommunication Tower which consists of, a
form of work safety awareness among business owners and service providers, forms
of work safety behavior from workers to business owners and service providers, as
well as knowing, and overcoming obstacles or obstacles to the concept with the hope
of reducing the number of work accidents. The research method used is literature
review and case study review to obtain variables and indicators that show the rela-
tionship and influence of safety climate, safety behavior, and safety awareness to
improve safety performance in PT X telecommunication tower construction. Then
the selected variables and indicators will go through the expert validation process.
The results of data collection from experts and interviews that have been conducted,
obtained 6 indicators of safety climate, 5 indicators of safety behavior, 3 indicators
of safety awareness, and 3 indicators of safety performance.
Keywords Telecommunication · Safety climate · Safety behavior · Safety

awareness · Safety performance
B. F. Athaya (B) · L. S. Riantini · R. A. Machfudiyanto

L. S. Riantini
R. A. Machfudiyanto
652 B. F. Athaya et al.
1 Introduction
Cellular telecommunication tower/BTS tower (Base Transceiver Station) is a device

that functions to place a signal transmitting antenna (access network) to provide
services to subscribers around the tower. The need for telecommunication towers
that are getting faster has encouraged people to always be creative by creating new
technology [1].
In line with the high development of telecommunication tower construction in
Indonesia, the level of work safety must also be increased to reduce the rate of
work accidents in construction projects [2], especially at Telecommunication Towers.
Statistics show that the construction industry still faces safety issues, which are
worth studying. Most accidents are caused by humans, which means that the risk of
accidents and injuries can be reduced with safety behaviors [3].
Meanwhile, individual safety behavior can be improved by increasing self-
awareness of safety [4]. In other words, individual behavior may be the result
of certain environmental factors in the organization, such as the safety climate
[5]. Therefore, to reduce construction accidents, both individual and organizational
factors must be considered. From an individual and organizational perspective, it is
important to analyze the relationship between safety climate, safety awareness, and
safety behavior [6].
This study aims to identify and analyze the work safety climate in the construction
of PT X Telecommunication Tower which consists of forms of work safety awareness
among business owners and service providers, forms of work safety behavior from
workers to business owners and service providers, as well as forms of work safety
behavior. From workers to business owners and service providers, as well as knowing,
and overcoming conceptual barriers or obstacles in the hope of reducing the number
of work accidents.
2 Theoretical Study
2.1 Safety Climate
The safety climate describes the perception of the value of safety in the work envi-
ronment and can be distinguished by attitudes, namely individual beliefs and feelings
about certain objects or activities [7]. Safety climate as a picture that is felt or related
to workers’ perceptions of the importance of safety and how this can be determined
in the organization [8].
Identification Factors of Safety Climate, Awareness … 653
2.2 Safety Behavior
Safety behavior is an approach designed to directly change safety-related behavior

through the application of behavioral principles and strategies such as peer observa-
tion of safe behavior, goal setting, and performance feedback [9]. Individual behavior
may be the result of certain environmental factors in the organization, such as
safety climate. Therefore, to reduce construction accidents, individual factors and
organizational factors must be considered [6].
2.3 Safety Awareness
Safety awareness is a concept that contains several dimensions of work safety. Aware-
ness is a process of internalizing the information obtained and becoming the values
that are embraced and can be manifested every day [10]. Safety awareness is a reac-
tion to work safety-related events that occur outside of the individual, including the
experience of a condition that can affect worker activities [11].
2.4 Safety Performance
Safety performance shows the results of safe work records for future periods, where
safety performance has a positive effect on project performance. Safety performance
is a function of many factors including safety culture [12]. Safety performance is
work behavior that is relevant to safety that can be conceptualized the same as other
work behaviors that are the result of work, so it must be determined by the knowledge
and skills needed to improve a safety performance [5].
3 Methodology
This research uses a qualitative approach. Through this approach, it is hoped that
the development of a concept to improve safety performance in the construction of
PT X Telecommunication Tower by using the parameters of Safety Climate, Aware-
ness, and Behavior. This qualitative approach is also supported by quantitative data
obtained from survey results through a questionnaire instrument to obtain data that is
more relevant to related research. A qualitative approach is taken to explore further
the existing conditions and the form of the relationship model of the climate safety
variable, awareness, and behavior in the construction of the Telecommunication
Tower PT X. Before conducting further research, expert validation will be carried
out using the Delphi Method to validate the variables to be used in this study. The
Fig. 1 Research flow line (input-process-output)
qualitative research strategy used in this research is a case study. That is, choosing
an event or symptom to be investigated by applying various methods. The method
used is in the form of interviews, observation and document analysis. The choice
of strategy is related to the purpose of this study, namely explanative, this study
aims to improve safety performance in the construction work environment of PT X
Telecommunication Tower.
The survey questionnaire allows the researcher to answer questions related to
the form of content analysis and the construct of this research. Thus, a validated
questionnaire with several modifications was adopted for this study to collect and
analyze data on Safety Climate, Safety Behavior, Safety Awareness, and Safety
Performance. The questionnaire was also examined as a form of content validation
and construct reliability by 1 academic and 4 industry practitioners as a form of
validation by experts using the Delphi method.
Figure 1 shows that the research flow begins with literature review and case studies
analysis to obtain research variables, then continues with expert validation (content
and construct), and ends with the results of validated factors that can affect safety
performance in Telco construction.
Research variables are used to deepen research methods based on factors related
to research. In this case, the safety performance is the dependent variable and the
independent variables are sought to facilitate research. A variable is defined as a
symbol or concept that is assumed to be a set of values [13].
Based on previous studies and discussions with several experts, it has been deter-
mined and agreed that the three variables between safety climate, safety awareness
and safety awareness are related to each other and play a role in improving safety
performance. This linkage is supported by the existence of several additions to content
and construction according to suggestions and input from experts to increase the range
of research that will be carried out next.
Based on Fig. 2, it can be seen that there is a relationship between safety climate,
safety awareness, safety behavior and safety performance. The safety climate is a
predictor of accidents, although this relationship is bridged by the Safety Behavior,
where in the safety climate there are two aspects, namely organizational policies
and procedures as well as the work and implementation environment [14]. There is
Fig. 2 Conceptual framework of this study
a relationship between safety climate and unsafe behavior which is a reference for
measuring safety behavior and climate safety for work accidents [15]. The safety
awareness element is about the causes of work accidents that most occur in construc-
tion work related to human errors, namely detecting and anticipating situations
around when doing work. Safety behavior has been proven to prevent accidents
from occurring in European and American countries [16].
After going through expert validation with 1 academic and 4 industry practitioners
on this research instrument, the content and construct variables were obtained which
will be used in the next research stage. Table 1 is the content and construct variables
obtained based on expert validation as the result of this study.
Table 1 Variable and indicators that have been validated by experts

No. item Variable Indicators
X1 Safety climate (SC)
X1.1 Communication support 1. Problems at work
2. Opinions of workers on policies
3. Discussion of the problem
4. The effect of changing procedures
5. Communication of changes
6. Concern between workers
7. Anticipate risks and consequences
X1.2 Job pressure 1. Time for planning
2. Time to do work
3. Time to get work done
4. Project schedule
5. Workload
6. External factor problem
X1.3 Work relationship 1. The future of workers
2. Work relationship
3. Company Morale
(continued)
Table 1 (continued)
X1.4 Adequacy of procedures 1. Completeness of work procedures
2. Technically accurate job procedures
3. Work procedures are clearly available
4. Procedure compliance
5. Worker’s understanding of procedures
6. Documentation management system
X1.5 PPE (Personal protective equipment) 1. PPE use is monitored
2. PPE users consulted
3. Findings from PPE monitoring
4. The use of PPE is enforced
X1.6 Safety rules 1. Understanding of safety rules
2. Implementation of safety rules
3. Safety rules constraints
X2 Safety behavior (SB)
X2.1 Leadership 1. Superior support
2. Team engagement for a purpose
3. Superior involvement
4. Team involvement for decisions
5. OHS activity leadership
6. Responsibility
7. Role models in the team
X2.2 Management commitment 1. OHS organization
2. Making safety decisions
3. Safety policy
4. Check and investigate safety reports
5. Safety system audits
6. Information gathering mechanism
7. Safety management system
X2.3 Safety training 1. Safety training
2. OHS competency requirements
3. Efficiency of the safety training program
4. Job specific training
5. Worker monitoring
6. The relevance and effectiveness
X2.4 Allocation of resources 1. Availability of resources
2. Effectiveness of resource allocation
3. Adequacy of fund allocation
X2.5 Rewards 1. Recognition and praise
2. Worker performance
3. Incentive
(continued)
Table 1 (continued)
X3 Safety awareness (SA)
X3.1 Safety risk management 1. Possible danger
2. Engagement determines the likelihood of
harm
3. Department of occupational safety and
health
4. Hazard identification
5. Risk assessment procedures
6. Avoid or minimize the risk of harm
7. HIRADC guidelines
8. Management decision
X3.2 Safety promotion 1. Safety induction
2. Safety sign
3. Social security for workers
4. Work safety operational standards
X3.3 Workers safety awareness 1. Safety level awareness
2. Awareness of the use of PPE
3. Knowledge of safety
4. Employer assertiveness
5. Giving penalties to offenders
6. Worker culture
7. Worker safety culture
Y Safety performance (SP)
Y1 Safety compliance 1. Work safety procedures
2. Application of work safety procedures
3. The effectiveness of safety procedures
Y2 Safety participation 1. Importance of safety compliance
2. Safety participation of coworkers
3. Safety meeting
Y3 Number of accident and near-miss 1. Near-miss incident
2. Accident incident
3. Short-term lost time injury
4. Long-term lost time injury
5. Number of accident and near-miss
5 Conclusion
Safety climate, awareness, and behavior are the main factors that affect the work
safety of a construction project. Aspects of safety, climate, awareness, and good
behavior are expected to reduce and prevent disturbances to occupational health and
safety, both on the part of the project owner and the service provider.
Based on literature review analysis and expert validation by five experts on safety
climate, safety awareness, safety behavior, and safety performance to obtain research
content and constructs, it is concluded that there are 6 content on climate safety:
Communication Support, Job Pressure, Work relationship, Adequacy of Procedures,

PPE (Personal Protective Equipment), and Safety Rules. In the Safety climate there
are 5 contents: Leadership, Management Commitment, Safety Training, Alloca-
tion of Resources, Rewards. There are 3 contents in Safety Awareness: Safety Risk
Management, Safety Promotion, Workers Safety Awareness. Then the last one is the
perception indicator of safety performance: Safety Compliance, Safety Participation,
Number of Accident and Near-miss.
Acknowledgements The author is grateful to University of Indonesia for giving support in this
research.
References
1. Ismail et al (2015) Analysis of BTS (base transceiver station) development planning based on
earth curvature and Fresnel areas in the Southern Sumatra regional project
2. Asanka, Malik (2015) Study on the impact of accidents on construction projects
3. Hu K et al (2011) Factors influencing the risk of falls in the construction industry: a review of
the evidence
4. Cooper (1994) Implementing the behavior based approach to safety: a practical guide
5. Neal et al (2000) The impact of organizational climate on safety climate and individual behavior
6. Mudan et al (2018) Relations between safety climate, awareness, and behavior in the Chinese
construction industry: a hierarchical linear investigation
7. Guldenmund (2000) The nature of safety culture: a review of theory and research
8. Cooper et al (2004) Exploratory analysis of the safety climate and safety behavior relationship
9. Wirth, Sigurdsson (2018) WITHDRAWN: reprint of “when workplace safety depends on
behavior change: topics for behavioral safety research
10. Yudhanto, Budiarto (2007) Education and behavior. Rineka Cipta, Jakarta, Indonesia
11. WHO—Regional Office for the Eastern Mediterranean (2001) Occupational health: a manual
for primary health care workers. Cairo
12. Rossy, Yusuf (2017) A conceptual framework to development of construction safety culture in
Indonesia
13. Goundar (2012) Chapter 3—Research methodology and research method
14. Neal, Griffin (2006) A study of the lagged relationships among safety climate, safety motivation,
safety behavior, and accidents at the individual and group levels
15. Clarke (2006) The relationship between safety climate and safety performance: a meta-analytic
review. J Occup Health Psychol 11(4):315–327
16. Ceren (2014) Examining safety behaviour with the safety climate and the theory of planned
behaviour
Cost Structure of Construction Safety
on High Residential Buildings
in Indonesia
Ratih Fitriani, Yusuf Latief, and Rossy A. Machfudiyanto
Abstract The plan for the Cost implementation of Construction Safety Management
System in the Public Works and Housing Ministry sector is part of the Construction
Safety Plan, which is agreed upon and approved at the preparatory meeting for the
implementation of construction work (Pre Construction Meeting). This study aims
to identify and analyze the cost structure of safety in high residential buildings in
Indonesia. The methodology used in this study is to use Bill of Quantity data from
16 projects in various regions in Indonesia and then analyzed using the Monte Carlo
software to determine the optimal cost of safety. The results of this study indicate
that the component of expert consultation related to construction safety is the least
priority component, followed by the insurance and licensing component, then the
construction safety personnel component which is not prioritized. For the construc-
tion safety personnel component, according to the construction cost budgeting, the
personnel cost has been included, so that it is no longer budgeted for construction
safety costs.
Keywords Cost structure · Construction safety · High residential buildings
1 Introduction
The high intensity of accidents in the construction sector in Indonesia is caused by

a weak construction safety management system, namely lack of supervision [1],
weak quality of human resources and a lack of understanding of external issues
[2]. The Construction Safety Committee stated that the results of the investigation
in general were that the contractor did not prioritize construction safety, in other
R. Fitriani · Y. Latief · R. A. Machfudiyanto (B)

Department of Civil and Environmental Engineering, Faculty of Engineering, Universitas
Indonesia, Depok, Indonesia
Y. Latief
660 R. Fitriani et al.
words the Construction Safety Management System had not been implemented opti-
mally [3]. In connection with the target of building one million Flats listed in the
National Medium Term Development Plan 2020–2024, the Indonesian Public Works
and Housing Ministry stated that the implementation of the management system.
Construction safety is the key to infrastructure development [4]. However, most
of the construction works in the public sector in many countries use the competitive
tendering system. The implementation of competitive tendering which is carried out
to select the contractor with the lowest price, this condition makes the contractor
press down the price which results in a low budget for construction safety costs [5].
According to a survey conducted by the Public Works and Housing Ministry in
2017, the implementation of a construction safety management system in infras-
tructure development work in Indonesia is only 34% [3]. The main factor causing
the low implementation of the construction safety management system is related to
construction safety costs. This study aims to analyze the cost structure of safety in
high residential buildings in Indonesia.
2 Theoretical Study
High Residential Buildings is a building built in an environment divided into sections

that are functionally structured in horizontal and vertical directions and are units that
each can be owned and used separately, which functions as a dwelling place, which
is equipped with common parts, common objects and common ground. From the
research results [5] the application of safety costs is one dimension in building a
construction safety culture.
The Construction Expenses Index, especially in 2020, is a price index that
describes the level of costly construction of a district/city compared to the refer-
ence city, namely Semarang City. The Construction Expenses Index data is
obtained from the results of the Construction Expensive Price Survey especially for
building/construction materials, heavy equipment rental, and construction service
fees which are carried out in all districts/cities in Indonesia. The data is calculated
based on quarterly price data for July 2019, October 2019, January 2020, and April
2020. The calculation of the Construction Expenses Index weigh diagram uses data
from the Work Unit Price Analysis Book of the Public Works and Housing Ministry,
Bill of Quantity (BoQ) and revenue budget realization data. And Regional Expendi-
ture (APBD). The 2020 IKK is one of the main components used for calculating the
General Allocation Fund (DAU) for the 2021 Fiscal Year [6].
In the Public Works and Housing Ministry Regulation 21/2019, it is stated that the
cost components of the construction safety management system consist of at least
[7]:
• preparation of a safety plan (comp. 1);
• outreach, promotion, and training (comp. 2);
• work protective equipment and personal protective equipment (comp. 3);
Cost Structure of Construction Safety on High Residential … 661
• insurance and licensing (comp. 4);

• construction Safety personnel (comp. 5);
• facilities, infrastructure and medical devices (comp. 6);
• signs and necessary traffic equipment or traffic management (comp. 7);
• consultation with experts related to Construction Safety (comp. 8), and
• activities and equipment related to the control of Construction Safety risk
including the cost of environmental testing/inspection (comp. 9).
In the safety plan preparation component, there are subcomponents for preparing
forms, work instructions as well as daily and monthly reporting. In the socializa-
tion, promotion and training component, there are subcomponents of socialization,
communication and training related to construction safety. The third component,
namely personal protective equipment and work protective equipment, consists of
several subcomponents, namely safety helmet, safety shoes, safety harness, lifeline,
safety deck, safety net, etc. The fourth component contains the construction insur-
ance (CAR) subcomponent as well as heavy equipment testing. The fifth component
consists of the subcomponent of construction safety experts, emergency respon-
ders, security officers, etc. Meanwhile, the sixth component consists of the initial
health test sub-component for construction, fogging, clinic and ambulance workers
if needed. The seventh component deals with the signs in the project such as warning
signs, directions, prohibitions, etc. The eighth component is a facilitation related
to consultation with construction safety experts, which can be structural experts,
building engineers, geologists and others who ensure design standards have been
met. The last component is activities and equipment related to construction safety
risk control, including costs for environmental testing and inspection in accordance
with the recommendations in environmental documents [5].
3 Methodology
In this study, a mapping of the construction safety cost structure will be carried out
by the contractor implementing the construction of flats at the Public Works and
Housing Ministry in 2020. The construction work of the flats in this study is a 3–6
storey building with a moderate construction safety hazard risk level. Meanwhile,
the construction sites are grouped into 2 (two) parts, namely the western and eastern
parts to be able to see the difference in the amount of costs in the two groups of
locations, as shown in Fig. 1.
Furthermore, in the structure, the amount of construction safety costs will be
seen and simulated using the Monte Carlo analysis. The analysis was carried out
to obtain the optimal amount of the cost of implementing construction safety in the
construction of flats in Indonesia.
Fig. 1 Research methodology
Table 1 Project name

No. Project name
West Indonesia
1 High residential building in Batam
2 High residential building Indragiri Hulu
3 High residential building in Sarolangun
4 High residential building in BWS Kalimantan
5 High residential building in East Waringin City
6 High residential building in Surakarta City
7 High residential building in UGM student dormitory
8 High residential building in Pesisir Barat
9 High residential building in Brebes
10 High residential building in ASN PUPR
East Indonesia
11 High residential building in ASN Region Sulawesi 1
12 High residential building in ASN Region Sulawesi 2
13 High residential building in TNI AL Sorong
14 High residential building in Maluku
15 High residential building in Tambrauw
16 High residential building in Tanimbar Island
The data used in the calculation of this simulation is the Bill of Quantity (BoQ)
document on 16 (sixteen) construction works of the Public Works and Housing
Ministry flats built in 2019–2020 as seen in Table 1. The location of the apartment
construction is spread throughout Indonesia, which is grouped into 2 (two) namely
west and east.
The Result of the study have mapped the structure of the existing construction
safety costs for 16 flat projects. In the regulations, construction safety costs must
be included in every infrastructure development in Indonesia [6] however, there are
several components in construction safety costs that are left blank or not budgeted. In
the expert validation process, information regarding the components of construction
safety costs that are not budgeted can be obtained because there is no need for control
using these components, or the items in these components are already owned by the
contractor (obtained from previous projects), or the cost of these components has
been obtained. Included in other costs. The following is an explanation of Structure
of Construction Safety Costs in the High Residential Buildings as shown in Table 2.
Monte Carlo simulation
The calculation of construction safety costs in this study was carried out by simulating
the data obtained from the BoQ of construction work at the Public Works and Housing
Ministry in 2020. The simulation was carried out using the Monte Carlo method with
Crystal Ball software. The data will be grouped into 2 (two) groups, namely based
on the location of the western and eastern parts. This is done to be able to prove
the difference in the amount of construction safety costs based on the construction
location.
The first step is to compile a simulation design. For each project that will be
simulated, construction safety costs are taken which consists of 9 components. The
lowest and highest values for each component of construction safety costs are taken
to calculate the estimated cost using the Monte Carlo analysis, as shown in Table 3.
Monte Carlo simulations are carried out by identifying the lowest and highest
costs of each construction safety cost component as shown in the Table 4.
From the mapping results in Table 2, it is found that the components that are
not budgeted for in many apartment constructions projects. These components are
insurance and licensing, construction safety personnel and expert consultation related
to construction safety. As for the three components, the component of consultation
with construction safety experts was the component that was not budgeted at most,
namely from 16 projects only 3 projects budgeted for this component. Next, the
licensing insurance component was budgeted for only 7 projects, and the construction
safety personnel component was budgeted for 8 projects.
There are two projects that do not budget for traffic signs and equipment compo-
nents and each one project that does not budget for the safety plan preparation
component, health infrastructure facility components and activity components and
equipment related to construction safety control. All projects budget for socialization,
promotion and training components and PPE and WPE components.
664
Table 2 Structure of construction safety costs in the high residential buildings

No. Construction cost Safety cost Comp. 1 Comp. 2 Comp. 3 Comp. 4 Comp. 5 Comp. 6 Comp. 7 Comp. 8 Comp. 9
1 Rp. 17,542,751,000 Rp. 303,380,000 Rp. 3,250,000 Rp. 27,950,000 Rp. 74,180,000 Rp. 7,500,000 Rp. 176,000,000 Rp. 3,900,000 Rp. 6,300,000 Rp. – Rp. 4,300,000
2 Rp. 18,835,689,900 Rp. 22,475,000 Rp. 800,000 Rp. 4,110,000 Rp. 11,055,000 Rp. – Rp. – Rp. 2,350,000 Rp. 250,000 Rp. – Rp. 3,910,000
4 Rp. 45,515,773,600 Rp. 376,005,000 Rp. 5,000,000 Rp. 77,800,000 Rp. 251,105,000 Rp. – Rp. – Rp. 7,500,000 Rp. 10,000,000 Rp. – Rp. 24,600,000
5 Rp. 21,553,900,000 Rp. 96,452,500 Rp. 2,700,000 Rp. 2,150,000 Rp. 57,325,000 Rp. 15,000,000 Rp. 187,500 Rp. 11,500,000 Rp. 850,000 Rp. – Rp. 6,740,000
7 Rp. 13,455,651,000 Rp. 128,015,000 Rp. 5,500,000 Rp. 10,065,000 Rp. 37,165,000 Rp. 17,050,000 Rp. 5,000,000 Rp. 10,250,000 Rp. 5,000,000 Rp. 5,000,000 Rp. 21,550,000
9 Rp. 16,735,965,000 Rp. 48,500,000 Rp. 2,350,000 Rp. 1,400,000 Rp. 31,050,000 Rp. – Rp. – Rp. 11,500,000 Rp. – Rp. – Rp. 833,496
14 Rp. 20,621,891,078 Rp. 67,012,000 Rp. – Rp. 7,315,000 Rp. 5,137,000 Rp. – Rp. 24,200,000 Rp. 22,825,000 Rp. – Rp. – Rp. 7,535,000
16 Rp. 25,123,473,271 Rp. 104,243,100 Rp. 75,230,900 Rp. 1,126,250 Rp. 25,363,150 Rp. - Rp. – Rp. – Rp. 2,522,800 Rp. – Rp. –
R. Fitriani et al.
Table 3 The lowest and highest costs of 9 (nine) components of construction safety costs for 16
flat projects by location
No. Cost structure construction safety Lowest cost Highest cost
West Region
1 Preparation component 800.000 5.500.000
2 Outreach and promotion 1.400.000 77.800.000
3 PPE 11.055.000 251.105.000
4 Insurance and licensing 0 27.000.000
5 Construction safety personnel 0 176.000.000
6 Health infrastructure facilities 2.350.000 13.650.000
7 Signs 0 10.000.000
8 Consultation with experts 0 5.000.000
9 Others related to hazard control 833.496 24.600.000
Total cost of West SMKK 22.475.000 376.005.000
East Region
1 Preparation component 0 75.230.900
2 Outreach and promotion 1.126.250 18.865.000
3 PPE 5.137.000 351.778.921
4 Insurance and licensing 0 13.943.000
5 Construction safety personnel 0 45.500.000
6 Health infrastructure facilities 0 23.830.950
7 Signs 0 4.101.955
8 Consultation with experts 0 16.500.000
9 Others related to hazard control 0 19.500.000
Total cost of East SMKK 67.012.000 420.274.370
In the application of safety costs, in addition to paying attention to the components

of safety costs, stakeholders must also be involved in its application so that it is
optimal and effective [8]. The results of the Monte Carlo simulation can be seen in
Table 4, where each project with a western and eastern location is analyzed and the
amount is obtained according to 9 (nine) components of the construction safety cost
structure in the ministerial regulation Public Works and Housing Ministry 21/2019.
Table 4 Simulations of high residential building construction safety costs in Indonesia by location
No. Cost structure construction safety Lowest cost Highest cost Optimal cost
West Region
1 Preparation component 2.870.767 3.416.474 3.153.907
2 Outreach and promotion 35.793.495 43.311.460 39.542.540
3 PPE 119.309.513 143.488.280 131.182.954
4 Insurance and licensing 13.346.654 16.056.672 14.739.141
5 Construction safety personnel 7.299.301 8.705.269 7.984.271
6 Health infrastructure facilities 79.691.411 96.501.026 88.367.483
7 Signs 4.637.082 5.607.474 5.127.883
8 Consultation with experts 4.572.545 5.391.783 5.008.059
9 Others related to hazard control 11.488.262 13.875.758 12.720.852
Total cost of construction safety 181.024.062 218.060.777 198.865.898
East Region
1 Preparation component 35.179.322 42.389.405 38.675.792
2 Outreach and promotion 9.048.662 10.937.392 10.009.704
3 PPE 161.294.869 194.801.835 178.527.189
4 Insurance and licensing 11.454.381 13.448.432 12.479.709
5 Construction safety personnel 11.484.629 13.731.628 12.591.237
6 Health infrastructure facilities 31.817.964 37.524.153 34.870.913
7 Signs 2.348.325 2.789.993 2.571.673
8 Consultation with experts 9.080.267 10.794.800 9.952.653
9 Others related to hazard control 10.804.404 12.866.524 11.842.686
Total cost of construction safety 222.439.408 263.697.106 243.536.924
5 Conclusion
From this research it can be concluded that in budgeting the construction safety costs
for the construction of high-rise flats up to 6 floors or moderate risk, the component
of expert consultation related to construction safety is the least priority component,
followed by the insurance and licensing component, then the construction safety
personnel component which is not prioritized.
For the construction safety personnel component, according to the construction
cost budgeting, it has entered the personnel cost, so that it is no longer budgeted
for construction safety costs. The Monte Carlo simulation clearly shows that the
difference in the amount of construction safety costs for a flat construction project
located in the west is smaller, compared to similar developments carried out in eastern
Indonesia.
Acknowledgements The Authors would like to thank the financial support provided by Ministry
of Research and Technology/National Research and Innovation Agency through PDUPT Grant
2021 with contract number: NKB-219/UN2.RST/HKP.05.00/2021 managed by the Directorate for
Research and Community Engagement (DRPM) Ministry of Research and Technology/National
Research and Innovation Agency (BRIN).
References
1. Rachman FF (2018) Menteri PUPR Buka Suara Penyebab Maraknya Kecelakaan Konstruksi
2. Bahfein S (2020) Kualitas SDM Rendah, Penyebab Kecelakaan Konstruksi
3. Fitriani R, Latief Y (2019) A conceptual framework of cost of safety model in infrastructure
project based on work breakdown structure (wbs) to improve safety policy in indonesia. IOP
Conf Ser Mater Sci Eng 650. https://doi.org/10.1088/1757-899X/650/1/012003
4. Choi TNY, Chan DWM, Chan APC (2011) Potential difficulties in applying the pay for safety
scheme (PFSS) in construction projects. Accid Anal Prevent 48:10. https://doi.org/10.1016/j.
aap.2011.04.015
5. Machfudiyanto RA, Latief Y, Arifuddin R, Yogiswara Y (2017) Identification of safety culture
dimensions based on the implementation of OSH management system in construction company.
Proc Eng 171:405–412
6. Peraturan Presiden Nomor 18 Tahun 2020 Tentang Rencana Pembangunan Jangka Menengah
Nasional Tahun 2020–2024, Peraturan Presiden Nomor 18 Tahun 2020 C.F.R. (2020)
7. Pedoman Penerapan SMKK, 21 C.F.R. (2019)
8. Machfudiyanto RA, Latief Y, Suraji A, Sagita L (2020) Main task and functions stakeholders
institutional to improve safety culture in construction sector. IOP Conf Ser Mater Sci Eng
771(1):012063
Adaptive Traffic Signal Control Using
Fuzzy Logic Under Mixed Traffic
Conditions
Budi Yulianto
Abstract The effectiveness of the intersection control performance is very influen-

tial on the occurrence of vehicular delays and traffic accidents. One of the systems
for controlling vehicle movements at intersection is a traffic signal control. Traffic
signal control widely used in Indonesia is Fixed Time Control (FTC). The deficiency
of the FTC is that it cannot accommodate high traffic flow fluctuations, which causes
excessive vehicle delays. The purpose of this study is to evaluate the Fuzzy Logic
Traffic Signal Control (FLTSC) algorithm on the VISSIM simulation model for an
isolated four-way intersection with two-stage signal type 42 and compare its perfor-
mance with the optimized FTC under different traffic conditions. The simulation
results showed the performance of the proposed FLTSC is generally better than the
optimized FTC, especially in time-varying traffic cases.
Keywords Fuzzy logic · Signal control · Mixed traffic
1 Introduction
One of the most prominent problems in big cities of the world is traffic congestion,
especially in developing countries where there is a significantly higher gap between
growing vehicle ownership and new road construction. The traffic congestion is
mainly associated with the use limited road space by high number of people at the
same time. This is observed during peak periods in the road network of urban areas
in most big cities in Indonesia, thereby, causing excessive delays and increment in
air pollution. The inability to find a major solution is, however, projected to make the
situation much worse in the years ahead due to the continuous growth of traffic. One
of the efforts to overcome traffic congestion is by controlling the vehicles movement
at intersections. Intersections are part of the road network which is a point of conflict
for vehicle movements that have the potential to cause vehicular delays and traffic
accidents as well as vehicle exhaust emissions that cause air pollution. In order to
B. Yulianto (B)
Department of Civil Engineering, Sebelas Maret University, Surakarta, Indonesia
670 B. Yulianto
minimize traffic congestion and accidents at the intersection, therefore, the vehicle
movements at the intersection are regulated by using traffic signal control.
The performance of signalized intersection depends on type of traffic signal
control in optimizing the signal setting parameters. The signalized traffic control
at isolated intersections in Indonesia mostly uses FTC and this method allows the
master controller to set the splits and cycle length based on predetermined values
usually from historical traffic data which are kept constant during certain period. It
was discovered that FTC produces good results when the condition of the traffic is
normal but unable to deal the period when it is complex and time-varying [1–4]. The
classic problem with the FTC is that updating of signal timing based on changing
traffic conditions is rarely done so that signal timing is not optimized based on the
latest conditions, because it is expensive and tiring works. In connection with this, to
improve the performance of signalized intersections, it is necessary to use demand
responsive traffic signal control, where signal timing based on the real time traffic data
fluctuation according to information from the detector. Moreover, Adaptive Traffic
Signal (ATS) control is discovered to have the ability of adjusting the green time
length at a specific stage using the traffic demand associated with the stages having
both right of way and red. A complex conventional mathematical model is, however,
usually used to develop the controller algorithm.
The complexity associated with the determination of the appropriate signal setting
parameters through the use of conventional mathematical models led scholars to
apply artificial intelligence methods. An example of this is the fuzzy logic which
was reported to be an interest mathematical method with prospects to model traffic
and transportation processes described by subjectivity, uncertainty, and imprecision
[5]. The first study observed to have applied fuzzy logic to the control of traffic
signal was by Pappis and Mamdani [6] and this was followed by its wide usage
in developing a theoretical and practical ATS controller [4, 7–12]. Yulianto [4] also
applied this technique to design traffic signal to be used in mixed traffic conditions or
when the traffic is untidy due to the fact that it does not follow lane-based operations.
The fuzzy logic controller performs much better than optimized FTC in time varying
traffic conditions. The purpose of this study is to evaluate the fuzzy logic traffic
signal control algorithm proposed by Yulianto [4] to be applied in an isolated four-
way intersection with two-stage signal type 42 [13] and compare its performance
with the optimized FTC under two case studies namely constant traffic flow and
fluctuated traffic flow during one-hour period.
2 Traffic Signal Control Strategies
2.1 Fixed Time Control
Webster [14] was reported to have developed the signal timing method which is
mostly applied in FTC. The process includes the proposal of mathematical equations
Adaptive Traffic Signal Control Using Fuzzy Logic Under Mixed … 671
for the calculation of the signal setting plan to minimize the average vehicle delay for
all vehicles through the use of an intersection. Meanwhile, the signal timing plan is
usually designed in Indonesia using the methods provided by the Indonesian Highway
Capacity Manual [13]. Moreover, the cycle length and green time were calculated
using the equations adapted from the Webster model. It is important to note that
the cycle length (seconds) was determined in this study based on the Indonesian
Highway Capacity Manual (IHCM) using the following equation:
1.5L T I + 5
C= (1)
1− F RC R I T
where L T I is the lost time per cycle (seconds), F R is the ratio of flow to saturation
flow (Q/S), F R C R I T n is the maximum value of F R among the signal group for the
same stage, and F R C R I T is the summation of F R C R I T n for all N stages in the
cycle which is written as

N
F RC R I T = F RC R I T n . (2)
n=1
The determination of the nth stage green time was through the following equation.
F RC R I T n
gn = (C − L T I ) (3)
F RC R I T
It is, therefore, possible for the intersection to perform optimally by setting the
signal timing based on the traffic currently flowing through.
2.2 Adaptive Traffic Signal Control
The aim of the ATS control is to ensure the maintenance of the best desirable flow of
traffic through the automatic adaptation of the individual controller characteristics
including the cycle times, splits, and possibly stage sequences in order to satisfy the
traffic conditions which are changing heavily [15]. Real-time traffic data on the traffic
conditions are also set to the controller through the vehicular detectors installed on
each approach lane upstream of the intersection. It is important to note that the ATS
control logic is most times designed to be based on optimization process and this
means it has the ability to optimize the flow of traffic using online method without
confinement to particular cycle time. Meanwhile, it is possible for the signal timing
plan to change at any time depending on the optimizing algorithm.
The major focus of recent research is on the application of artificial intelligence
techniques such as fuzzy logic on intersection control. In general, the fuzzy logic
traffic signal control structure consists of three main modules, namely fuzzification,
672 B. Yulianto
fuzzy inference and defuzzification modules. Fuzzification is a process where the

crisp input values for a fuzzy logic traffic signal control are converted into fuzzy
values, which are represented by a degree of membership. The input data from
vehicular detectors provide precise information on the current conditions of traffic
in real time. Meanwhile, fuzzy inference has the ability to evaluate the control rules
stored in the fuzzy rule base in order to served as the representation of the knowledge
of experts through If–Then rule structure. The fuzzy rule base is, therefore, a set of
fuzzy rules used in mapping the inputs to outputs. However, defuzzification involves
the conversion of the fuzzy inference output values to crisp output values which are
applied in determining the following cycle signal status through the signal controller
program.
The fuzzy logic traffic signal control (FLTSC) strategies that have been developed
can be categorized into three groups namely extension level, extension-termination,
and urgency concept. Extension Level, the controller decides whether or not to give
an extension time to the current green stage at the end of each intervention [6,
9]. Extension-Termination, the controller decides at regular time intervals to either
extend or terminate the current green stage [3, 10]. Urgency Concept, the output
of this controller is similar to the extension-termination strategy, but the urgency
concept has the ability to determine what stage, among the possible ones, should be
the next [4, 7, 8, 12].
3 Research Methods
This research was conducted using a simulation modeling approach. The FLTSC
algorithm [4] which is designed for an isolated four-way without turning move-
ments is applied to an isolated four-way intersection with two-stage signal type 42.
Figure 1 shows the geometric intersection with its phase sequence. To determine
effectiveness of the proposed FLTSC performance in minimizing delays, therefore,
it is compared with the optimized FTC performance. An isolated four-way intersec-
tion with two-stage signal type 42 with both types of control is modeled using the
VISSIM simulation program.
VISSIM simulation program has a traffic simulator to simulate traffic flow.
FTC modeling uses the default signal controller program in the traffic simulator,
while FLTSC uses the signal controller program created by using Vehicle Actuated
Programme (VAP) language [4]. This signal controller program is then connected to
the traffic simulator via the signal stage generator. The traffic simulator simulated
the traffic operation during the simulation period and output values were obtained in
the form of average delay of vehicles at the end.
Various case studies were carried out with different traffic conditions at the inter-
section under constant and fluctuating conditions to evaluate the effectiveness of
the proposed FLTSC and optimized FTC performance. Traffic signal control that
produces the smallest average delay of vehicles at intersection is the most effective
traffic signal control in providing a good intersection level of service.
Fig. 1 Geometric intersection with its phase sequence
4 Proposed Fuzzy Logic Traffic Signal Control
4.1 Fuzzy Control Strategy
Development of the proposed FLTSC process is carried out in the VISSIM simulation
program. The controller was developed by using Vehicle Actuated Programming [4].
The proposed FLTSC applied the urgency concept used in Sayers et al.’s research
[7] which was selected as a suitable strategy due to its flexibility and computational
simplicity. Moreover, traffic information on the maximum queue lengths and average
occupancy rates retrieved from the previous cycle were also applied in estimating
the green time length needed by each set of signal groups or stages in the next cycle
based on the stage weight values. The process used in calculating the next cycle’s
new green time values is, therefore, presented in Fig. 2.
4.2 Fuzzification and Membership Function
There was an adjustment of the configuration for the input and output fuzzy vari-
ables membership functions through the analysis of the performance of controller at
various traffic conditions to improve its performance. These membership functions
including the maximum queue length and average occupancy rate were observed to
have the linguistic labels of Low, Medium, High, and Very High while the weight
has Very-Very Low, Very Low, Low, Medium, High, Very High, and Very-Very High
respectively as indicated in Figs. 3, 4 and 5.
674 B. Yulianto
Fig. 2 Process of calculation of new green time values
Fig. 3 Input maximum queue length
Fig. 4 Input average occupancy rate

Fig. 5 Output weight
Table 1 Fuzzy rule base

Maximum queue length
L M H VH
Average Occupancy Rate L VVL L M H
M VL L H VH
H L M H VVH
VH M H VH VVH
4.3 Fuzzy Inference Rule Base
There are two inputs with each having four membership functions to produce sixteen
fuzzy rules in the proposed FLTSC. Moreover, there was an adjustment of the fuzzy
rule base through the use of the trial-and-error method [6] and the fuzzy rules config-
uration in the proposed FLTSC matrix structure is presented in Table 1 while the
fuzzy inference system used was the Mamdani max–min composition method [16].
4.4 Defuzzification
This is the process of transforming the output value generated from the fuzzy infer-
ence system into a real crisp value. Moreover, the Centre-of-Sum method is usually
applied in this process to determine the output variable Weight due to its ability to
provide better performance in relation to disambiguate, continuity, and computer
complexity.
676 B. Yulianto
This research used two case studies with different traffic volume to evaluate the
effectiveness of controller. Case 1 shows the traffic flow is constant during one-hour
period while Case 2 involves the variation of the traffic flow every 15 min and the
set of traffic volumes for the case studies are indicated in Table 2. Traffic turning
proportions for all cases is to turn right 10%, straight 75% and turn left 15%.
The IHCM method was used in calculating the optimized FTC signal timing using
the cycle and green times presented in Table 3. The comparison between the proposed
FLTSC and optimized FTC is only valid if the calculations of cycle time and green
time using the IHCM method gives rise to optimum time settings. To find out whether
the IHCM method produces the optimum signal timing, therefore, a sensitivity test
is performed using the TRANSYT program. The results of the sensitivity test show
that the IHCM method indeed produces the optimum signal timing [17].
There was imposition of the lowest and highest values of total green time at
10 s and 52 s respectively in the cycle of the proposed FLTSC to avoid the cycle
length from being too long or too short. Meanwhile, the amber and all red periods
were 3 s and 1 s respectively for both controllers. Identical simulations were run for
approximately one-hour on both controllers to produce the output value which was
the average delay at the intersection. The values for all cases were later compared
for both controllers and the results are presented in Fig. 6.
Table 2 Set of traffic volumes (vehicle/hour)

Case Traffic volume Case Traffic volume Case Traffic volume
E–W N–S E–W N–S E–W N–S
1a, 2a 1.080 1.080 1e, 2e 2.520 1.080 1i, 2i 1.800 1.440
1b, 2b 1.440 1.080 1f, 2f 1.440 1.440 1j, 2j 2.160 1.440
1c, 2c 1.800 1.080 1g, 2g 1.800 1.440 1k, 2k 2.160 1.800
1d, 2d 2.160 1.080 1h, 2h 2.160 1.440
Table 3 Cycle time and green time used for FTC

Case Green time (s) Cycle time (s) Case Green time (s) Cycle time (s)
E–W N–S E–W N–S
1a, 2a 14 15 45 1g, 2g 27 21 64
1b, 2b 19 15 50 1h, 2h 35 23 74
1c, 2c 25 15 56 1i, 2i 47 26 89
1d, 2d 32 16 64 1j, 2j 29 29 74
1e, 2e 42 17 75 1k, 2k 39 33 88
1f, 2f 20 20 56
40
35.2
32.9
29.6 30.6
Average Delay (seconds)
28.5 27.7 29.6

30 27.4
22.8 23.2 25.7 25.4
20.3 22.4 20.3 22.8
18.9 20.1 19.7
20 17.4 18.5
16.2
10
0
1a 1b 1c 1d 1e 1f 1g 1h 1i 1j 1k
Cases FTC FLTSC
(a)
50
44.2 44.2
40 36.2 37.4
Average Delay (seconds)
33.4
32.2
27.8 29.6 28.8 28.9 29.4
30 24.1 27.5
23.3 23.5 26.0 24.3
23.9
19.2 21.8 22.5
20 18.8
10
0
2a 2b 2c 2d 2e 2f 2g 2h 2i 2j 2k
Cases FTC FLTSC
(b)
Fig. 6 Average delay of vehicles of the optimized FTC and proposed FLTSC, a case 1 b case 2
The simulation for Case 1 showed the optimized FTC generally produced an
average delay higher than the proposed FLTSC with the proposed FLTSC observed to
have reduced the average delay by 1.9–15.8% against the optimized FTC as indicated
in Fig. 7. Meanwhile, the simulation for Case 2 indicated the proposed FLTSC is
superior to the optimized FTC for all cases, except case 2a. The average delay saving
24
20 18.1
15.8 15.8
Improvement (%)
16 13.6 13.9
12
6.8 6.4 6.4 7.2 7.1 7.2
8 6.3
4.2
2.9
4 2.3 1.9 1.0 2.0
0.1
0
-4 -2.1 -1.2
-3.9
-8
1a, 2a 1b, 2b 1c, 2c 1d, 2d 1e, 2e 1f, 2f 1g, 2g 1h, 2h 1i, 2i 1j, 2j 1k, 2k
Cases Case 1 Case 2
Fig. 7 Percentage improvement of the proposed FLTSC against the optimized FTC
678 B. Yulianto
by the proposed FLTSC is between 1.0% and 18.1% against the optimized FTC, see
Fig. 7. Overall, the proposed FLTSC performs much better than optimized FTC. It is
due to the proposed FLTSC’s ability to cause an adjustment of the green time length
due to the fluctuation of the traffic flow.
The main concern of this type of signal control is the driver behavior of right
turning movements. For opposed approaches, the queue discharge is influenced by
the fact that right turning vehicles do not respect the right of way of straight through
traffic. This may reduce the proposed FLTSC’s effectiveness, where the length of
green time to discharge vehicles cannot be used efficiently. However, this behavior
does not affect the performance of the controller significantly, since the number of
right turning vehicles is low (less than 200 pcu/hr).
6 Conclusion
The development of an adaptive FLTSC for an isolated four-way intersection with

two-stage signal type 42 suitable for mixed traffic has been presented. A wide range
of traffic flow rates was applied to compare the controller’s performance in terms of
average delay at an intersection experienced with the optimized FTC and proposed
FLTSC. The simulation showed the performance of the proposed FLTSC is generally
better than the optimized FTC, especially in time-varying traffic cases.
References
1. Kell JH, Fullerton IJ (1991) Manual of traffic signal design—chapter 7: detectors. Institute of
Transportation Engineers, Englewood Cliffs, New Jersey
2. Lee JH, Lee KM, Leekwang H (1994) Fuzzy controller for intersection group. In: IEEE confer-
ence on industrial automation and control emerging technology applications, pp 376–382.
Taipei, Taiwan
3. Trabia MB, Kaseko MS (1996) A fuzzy logic controller for a traffic signal. In: IASTED
international conference on applications of control and robotics, pp 117–122
4. Yulianto B (2003) Application of fuzzy logic to traffic signal control under mixed traffic
conditions. Traffic Eng Control 44(9):332–336
5. Zadeh LA (1965) Fuzzy sets. Inf Control 8(3):338–353
6. Pappis CP, Mamdani EH (1977) A fuzzy controller for a traffic junction. IEEE Trans Syst Man
Cybern 7(10):707–717
7. Sayers T, Bell MGH, Mieden T, Busch F (1996) Traffic responsive signal control using fuzzy
logic—a practical modular approach. In: Proceedings of the fourth European congress on
intelligent techniques and soft computing, pp 2159–2163. Aachen, Germany
8. Sayers T, Anderson J, Clement S (1999) The multi-objective optimisation of a traffic control
system. In: Transportation and traffic theory. Papers presented at the abbreviated presentation
sessions, pp 153–176
9. Niittymäki J, Pursula M (2000) Signal control using fuzzy logic. Fuzzy Sets Syst 116(1):11–22
10. Trabia MB, Kaseko MS, Ande M (1999) A two-stage fuzzy logic controller for traffic signals.
Transp Res Part C Emerg Technol 7(6):353–367
11. Murat YS, Gedizlioglu E (2005) A fuzzy logic multi-phased signal control model for isolated
junctions. Transp Res Part C Emerg Technol 13(1):19–36
12. Yulianto B (2012) Setiono: traffic signal controller for mixed traffic conditions. IOSR J Mech
Civ Eng 4(1):18–26
13. Indonesian Highway Capacity Manual (1997) Directorate general Bina Marga, Indonesia
14. Webster FV (1958) Traffic signal settings. Road Research Technical Paper No 39
15. Strobel H (1982) Computer controller Urban transportation. John Wiley & Sons, UK
16. Mamdani EH, Assilian S (1975) An experiment in linguistic synthesis with a fuzzy controller.
Int J Man–Mach Stud 7(1):1–13
17. Yulianto B (2019) Sensitivity test of IHCM 1997 traffic signal timings using TRANSYT
program and VISSIM. AIP Conf Proc 2114(1):1–10
Application of Fuzzy Inference System
Mamdani at Pelican Crossing
Salsabila Naura Putri
Abstract Fuzzy logic is one of the soft computing methodologies which aims to
analyze the system that contains fuzziness. Fuzzy inference system (FIS) defines as a
decision-making process based on fuzzy set theory, fuzzy rules, and fuzzy reasoning.
In this research, we use FIS Mamdani which has been used in several applications,
especially transportation. Pedestrians are the weakest mode of transportation that
requires an infrastructure that is safe and comfortable when crossing the road. Pelican
crossing is one of the pedestrian crossing types. Pelican crossing in Indonesia is
based on fixed time where the pedestrian time is fixed. It is not effective because
the pedestrian time is not based on the volume of traffic conditions. To overcome
this problem, the FIS Mamdani method is used. Mamdani method is composed
of three steps to obtain an output, namely pedestrian time. Fuzzification with two
input variables, namely pedestrian and vehicle volumes, and the output variable is
pedestrian time. Inference engine which executes all logic with the if–then rule.
Apply defuzzification with the Centroid method to obtain the output result. The
purpose of this research is to determine the pedestrian time on pelican crossing using
FIS Mamdani. The result shows the pedestrian time change according to the volume
of pedestrians and vehicles. FIS Mamdani can adjust the traffic light signal time to
the traffic conditions that occur on a pelican crossing.
Keywords Fuzzy logic · Fuzzy inference system · Pelican crossing
1 Introduction
Problem-solving in the real world requires an intelligent system that has human-
like skills, able to utilize knowledge, and learns to adapt to its environment. Soft
computing is a collection of methodologies that aim to exploit the tolerance for impre-
cision and uncertainty to achieve tractability, robustness, and low solution cost [1].
Soft computing consists of several computing paradigms, including neural networks,
S. N. Putri (B)
Department of Mathematics, Sebelas Maret University, Surakarta, Indonesia
682 S. N. Putri
fuzzy set theory, approximate reasoning, genetic algorithm, and simulated annealing
[2].
Fuzzy logic was first established by Lotfi A. Zadeh of California University in
1965 [3]. Fuzzy logic defines as a logic that has value fuzziness between true and
false with a membership degree between 0 and 1. This logic proved to be a powerful
and effective tool for dealing with uncertainty to fuzzy systems which could support
decisions in several fields, from control theory to artificial intelligence.
Fuzzy Inference System (FIS) is a formulation process from input to output with
fuzzy logic [4]. Generally, three methods may be used in FIS which are Mamdani
method [5], Sugeno method [6], and Tsukamoto method [7]. Mamdani method
is often applied in fuzzy models. Several applications have been used using FIS
Mamdani, such as making decisions in transportations. The uses of FIS Mamdani in
the transportation field help solve various complex traffic and transportation prob-
lems with many objectives, many constraints, unclear input information, and vague
decision criteria. An example of application is the setting of traffic signal lights which
can adjust to the density of traffic flow on the path arranged [8].
Pedestrians are an important form of transportation in urban areas. It is the weakest
position if they are mix with vehicles, therefore pedestrian crossing is needed. The
pedestrian crossing is a special area on the roadway that is allocated for pedestrians
to cross the road [9]. Pelican crossing is a type of pedestrian crossing which provided
traffic lights and push buttons to activate traffic lights.
The facility of pelican crossing provides safety and comfort for the pedestrians.
However, the timing of pelican crossing signal lights should be concerned because
pelican crossing determines the service level of the roads. The service level of the
roads describes the traffic conditions in which the service level of a good road has a
small delay for vehicles and pedestrians and vice versa. Poor timing of signal lights
will have an impact on traffic congestion on the road and or traffic violations by
pedestrians because it takes too long to wait for the time to cross [10].
Pelican crossings in Indonesia are based on pedestrian actuated control (PA-C)
with fixed time operational. In this case, the pedestrian phase will activate when
the pedestrian presses the button. The length of pedestrian time remains under the
calculations established by several standards [11]. The existing pelican crossing is
not effective because the pedestrian time is not based on the traffic conditions which
may cause time-wasting by pedestrians and vehicles.
Turky et al. [12] have successfully simulated the performance between the control
of pedestrian crossing using fixed-time and fuzzy logic. It found that the fuzzy
logic pedestrian crossing controller performs better than the fixed-time controller.
Hunsanon et al. [13] proposed a fuzzy logic control (FLC) strategy for analyzing the
optimal traffic signal timing setting for pedestrian crossing signals in the urban area.
The results showed that the operational performance of fuzzy logic at the crossing
guard is better than PA-C. Makarova et al. [9] researched fuzzy logic to improve
safety at road crossings. Vehicle queue length, pedestrian queue length, and time
after the completion of the pedestrian phase were used as the input variable while
waiting time for the pedestrian phase as the output variable. The results showed
that the use of fuzzy logic can normalize the traffic parameters. In this research, we
Application of Fuzzy Inference System Mamdani at Pelican Crossing 683
proposed FIS Mamdani to calculate pedestrian time based on traffic control, namely
the pedestrian and vehicle volumes.
2 Fuzzy Inference System (FIS) Mamdani
Fuzzy inference system (FIS) is a decision-making process based on fuzzy sets,

fuzzy rules, and fuzzy reasoning [14]. This system is compatible with solving real-
world problems which are complicated to obtain a mathematical model, whereas it
is uncomplicated to describe the solution linguistically with if–then rules [15].
FIS Mamdani provides understandable results with a clear structure, also for the
basic nature of rules that are intuitive and explainable [16]. FIS Mamdani is composed
of three parts namely fuzzification, fuzzy inference engine, and defuzzification [14].
3 Pelican Crossing
Pelican crossing is a type of pedestrian crossing which provided traffic signs and
traffic lights [17]. It is generally featured with a push button to activate the traffic
lights. On pelican crossing the pedestrian time is calculated using the following
equation,

L N
PT = + 1.7 −1 (1)
vt W
where P T is pedestrian time (second), vt is pedestrian speed (1.2 m/s), L is crossing

length (meter), N is the pedestrian volume per cycle, and W is crossing width (meter)
[18].
The phasing sequence on pelican crossing can be shown in Fig. 1 where Y is
yellow, FG is flashing green, and R is red.
Fig. 1 Phasing sequence on a pelican crossing

684 S. N. Putri
The duration of flashing green is not always 3 s, but it depends on the crossing
length and the average pedestrian speed (1.77 m/s) [11]. The flashing green can be
written as
L
FG = (2)
1.77
where F G is flashing green (second) and L is crossing length (meter).
4 Research Methods
This research is applied research that uses pedestrians and vehicles data located on
the pelican crossing in Jalan M. H. Thamrin, Jakarta [19]. The data were taken in
April 2019 at 12.00–1.00 PM and 5.00–6.00 PM. The procedures that apply to this
research are: (1) to explore related data, (2) describe the logic overview of pelican
crossing with fixed-time and fuzzy logic, (3) describe the model FLC, (4) analyze
the applied results of pedestrian time, (5) conclusions.
5.1 Data Description
The data of pedestrians and vehicles on the pelican crossing in Jalan M. H. Thamrin
(Plaza Indonesia) are shown in Table 1. Pelican crossing Plaza Indonesia consists of
two pelican crossings as presented in Fig. 2, where the traffic signal control is based
on fixed-time.
Table 1 Data of pedestrians and vehicles

Time (PM) Volume of pedestrian Time (PM) Volume of vehicle
East West Total East West Total
12.00–12.15 146 158 304 12.00–12.15 915 751 1666
12.15–12.30 135 152 287 12.15–12.30 964 786 1750
12.30–12.45 111 123 234 12.30–12.45 1097 779 1876
12.45–1.00 154 199 353 12.45–1.00 1034 819 1853
5.00–5.15 232 197 429 5.00–5.15 928 1001 1929
5.15–5.30 184 255 429 5.15–5.30 856 935 1791
5.30–5.45 224 316 540 5.30–5.45 886 992 1878
5.45–6.00 190 312 502 5.45–6.00 712 720 1432
Fig. 2 The layout of the

pelican crossing
5.2 Logic Overview of Pelican Crossing
Pelican Crossing Existing (Fixed Time) Pelican crossing has two-phase which are
the pedestrian phase and vehicle phase. In this research, the cycle time on the pelican
crossing Plaza Indonesia is 90 s. The minimum green time for pedestrian time with
Eq. (1) during the day is 16 s and during the evening is 19 s. Pelican crossing signal
diagram is shown in Fig. 3 with flashing green time using Eq. (2) is 8 s.
Fuzzy Logic Pelican Crossing In the fuzzy logic system, the pedestrian time
changes according to the pedestrian and vehicle volumes. In this case, the cycle
time is fixed but the vehicle time changes according to the changes in the pedestrian
time.
Pedestrian time is calculated every cycle. Data from the volume of pedestrians
and vehicles in the current cycle are calculated by FIS Mamdani to get the pedestrian
time and vehicle time for the next cycle. The explanation is shown in Fig. 4.
In Fig. 4, vehicle time is obtained from the equation,
V T = CT − I G − R A − PT
686 S. N. Putri
Fig. 3 Pelican crossing plaza Indonesia signal diagram during a day and b evening
Fig. 4 Signal diagram fuzzy logic
where C T is cycle time (90 s), I G is intergreen (6 s), and R A is red all (10 s).
Maximum green time for pedestrian time is calculated by using the equation G max =
1.5 × P T [20]. During the day, the maximum green time is 24 s while evening the
maximum green time is 28 s.
5.3 Fuzzy Logic Control
Fuzzification A FIS was constructed to output the pedestrian time with two input
variables, namely the volume of pedestrians and vehicles. This research consists
of two FIS which represent during the day and the evening. In fuzzification, the
input and output variables are divided into three fuzzy sets. The plots of membership
functions are shown in Figs. 5, 6, 7 and 8.
Inference System Based on Fuzzy Rules The general format of the rules is the
following:
IF V ol Pedestrian is X 1 AND V ol V ehicle is X 2 THEN Ped T ime is y
Where X 1 , X 2 , and y are natural language expressions of the condition of respective

variables. The numbers of rules are combinations of fuzzy sets for X 1 and X 2 . The
sets of rules used are shown in Table 2.
Fig. 5 Volume of pedestrians
Fig. 6 Volume of vehicles

688 S. N. Putri
Fig. 7 Pedestrian time during the day
Fig. 8 Pedestrian time during the evening
Table 2 Fuzzy rules

Volume of vehicles
Low Average High
Volume of pedestrians Low Short Short Short
Average Medium Medium Medium
High Long Long Medium
Defuzzification Defuzzification is defined as a transformation process in the output

from FIS into a single numerical value. The centroid method is used in defuzzification
to determine the pedestrian time.
5.4 Result of Pedestrian Time
The fuzzy sets are used to determine the pedestrian time, namely pedestrian and
vehicle volumes as input and pedestrian time as output. An experiment of pelican
Fig. 9 Rule viewer of 23 pedestrians and 78 vehicles
crossing with FIS Mamdani based on the research data in Table 1 is given. The
snapshot of the rule viewer with 23 pedestrians and 78 vehicles is shown in Fig. 9.
Figure 9a represent the membership degree of 23 pedestrians in the pedes-
trian volume variable, (b) represent the membership degree of 78 vehicles in the
vehicle volume variable, and (c) represents the output results of the inference system
according to the conditions and the last row is combined fuzzy solution area of each
rule. The red line on the pedestrian time variable shows that the pedestrian time for
23 pedestrians and 78 vehicles is 17 s. Figure 10 represents the signal diagram of
Fig. 9.
Other experiments of pelican crossing with FIS Mamdani during the day and
evening are shown in Table 3. Based on Table 3, it appears that when the pedestrian
volume is low, the pedestrian time is short. This shows that pedestrians have priority
Fig. 10 Signal diagram of 23 pedestrians and 78 vehicles

690 S. N. Putri
Table 3 Pedestrian time experiment

No. Vol. Ped. Vol. Veh. Pedestrian time No. Vol. Ped. Vol. Veh. Pedestrian time
D E D E
1 8 35 16 19 9 32 42 21 25
2 12 59 16 19 10 34 74 22 26
3 14 87 16 19 11 37 80 23 27
4 19 110 16 19 12 38 113 20 24
5 22 30 17 20 13 44 24 24 28
6 25 60 17 21 14 48 64 24 28
7 28 90 19 22 15 50 83 23 27
8 29 108 19 23 16 53 115 20 24
on pelican crossing because pedestrians are a vulnerable mode of transportation

which must be prioritized when crossing the road at the pedestrian crossing.
6 Conclusion
Pelican crossings in Indonesia are based on PA-C with fixed-time operational. It

controls signal with pre-set timing however this controller has some limitations. For
example, when many pedestrians are crossing the pelican crossing, the pedestrian
time is terminated when reaches the maximum green time. Even when the pedestrian
volume is low, the vehicles must wait until the pedestrian time ends which can
cause congestions. To outcome this problem, FIS Mamdani is used to determining
the pedestrian time of pelican crossing since it helps to control the traffic signal
at the pelican crossing. In FIS Mamdani, the volume of pedestrians and vehicles
was selected as input while the output is the pedestrian time. The results showed
that when the pedestrian volume is low and vehicle volume is high, pedestrian time
will be short. While, when the pedestrian volume is high and vehicle volume is low,
pedestrian time will be long. This indicates that the pedestrian time changes according
to the pedestrian and vehicle volumes. The use of FIS Mamdani for determining the
pedestrian time able to adjust the traffic signal control to the traffic conditions that
occur at the pelican crossing.
References
1. Zadeh LA (1994) Soft computing and fuzzy logic. IEEE Softw 48–56
2. Jang JSR, Sun CT, Mizutani E (1997) Neuro-fuzzy and soft computing: a computational
approach to learning and machine intelligence. Prentice-Hall Inc., USA
3. Pamuji A (2017) Fuzzy logic inference system for determining the quality assessment of
student’s learning ICT. Sci J Inform 4(1):57–65
4. The MathWorks Inc (1998) Fuzzy logic toolbox user’s guide, 4th edn. The MathWorks Inc.,
USA
5. Pourjavad E, Shahin A (2018) The application of Mamdani fuzzy inference system in evaluating
green supply chain management performance. Int J Fuzzy Syst 20(3):901–912
6. Yulianto T, Komariyah S, Ulfaniyah N (2017) Application of fuzzy inference system by Sugeno
method on estimating of salt production. AIP Conf Proc 1867(1):020039
7. Hastono T, Santoso AJ (2017) Honey yield prediction using Tsukamoto fuzzy inference system.
In: 4th international conference on electrical engineering, computer science and informatics
(EECSI), pp 1–6
8. Sumiati, Sigit HT, Kapuji A (2014) Mamdani fuzzy inference system application setting for
traffic lights. Int J Fuzzy Logic Syst (IJFLS) 10(3):56–62
9. Makarova I, Shubenkova K, Mavrin V, Buyyol P (2018) Improving safety on the crosswalks
with the use of fuzzy logic. Transp Prob 13(1):97–109
10. Yulianto B, Sugiyarto (2019) Evaluation of midblock pedestrian crossing facility in the
Surakarta City, Indonesia. ICETsAS J Phys Conf Ser 1376(1):1–9 (2019).
11. Malkhamah S (2004) Evaluasi penetapan periode pada pelican di Indonesia. Media Teknik
26:74–81
12. Turky AM, Ahmad MS, Yusoff MZM, Sabar NR (2008) Fuzzy logic application for pedestrian
crossing. Int Symp Inf Technol 1:1–6
13. Hunsanon T, Kronprasert N, Upayokin A, Songchitruksa P (2017) Control strategy for vehicular
and pedestrian midblock crossing movements. Transp Res Proc 25:1672–1689
14. Putri SN, Saputro DRS (2021) Construction fuzzy logic with curve shoulder in inference system
Mamdani. KNPMP V J Phys Conf Ser 1776:1–8
15. Castellano G, Castiello C, Pasquadibisceglie V, Zaza G (2017) FISDet: fuzzy inference system
development tool. Int J Comput Intell Syst 10(1):13–22
16. Kovac P, Rodic D, Pucovsky V, Savkovic B, Gostimirovic M (2014) Multi-output fuzzy infer-
ence system for modeling cutting temperature and tool life in face milling. J Mech Sci Technol
28(10):4247–4256
17. Departemen Pekerjaan Umum (1999) Pedoman Perencanaan Jalur Pejalan Kaki pada Jalan
Umum. PT Mediatama Septakarya, Jakarta
18. Dirjen Perhubungan Darat (1997) Keputusan Direktur Jenderal Perhubungan Darat Nomor:
SK.43/AJ 007/DRJD/97, Perekayasaan Fasilitas Pejalan Kaki di Wilayah Kota. Departemen
Perhubungan, Jakarta
19. Widyaningsih N, Daniel O (2019) Analisis karakteristik dan perilaku penyeberangan orang
pada fasilitas penyeberangan zebra cross dan pelican cross (Studi Kasus Ruas Jalan M. H.
Thamrin). J Pengembangan Rekayasa dan Teknologi 15(1):27–32
20. Roess RP, McShane WR, Prassas ES (1998) Traffic engineering, 4th edn. Prentice-Hall Inc.,
New Jersey
Determining the Maximum Speed Limit
in Residential Area
Gito Sugiyanto , Fadli Wirawan, Eva Wahyu Indriyati, Yanto,
and Mina Yumei Santi
Abstract The residential area which is located on the right or left side of the arterial
or collector road is one of the areas with high potential to traffic accidents. An effort
to reduce the number of traffic accidents is applying speed limits. This is important
as most traffic accidents are caused by speeding factors. The aim of this research
is to determine the value of the maximum speed limit in residential areas using
the 85th percentile and stopping sight distance (SSD). The study locations are in
three collector roads in Purbalingga Regency namely Padamara Street, Letnan Yusuf
Street, and Letkol Isdiman Street. Based on the analysis result, the maximum speed
limit of vehicles based on the 85th percentile method varies between 42 up to 44 kph.
The maximum speed limit based on the SSD method is 39.32 kph for Padamara
Street, 38.10 kph for Letnan Yusuf Street, and 38.40 kph for Letkol Isdiman Street.
The maximum speed limit of vehicles proposed for collector roads with residential
area functions on the right or left side is 40 kph.
Keywords Accident · 85th Percentile · Residential area · Speed limit · Stopping

sight distance
G. Sugiyanto (B) · F. Wirawan · E. W. Indriyati · Yanto

Department of Civil Engineering, Jenderal Soedirman University, Purbalingga, Central Java,
Indonesia
E. W. Indriyati
Yanto
M. Y. Santi
Health Polytechnic Ministry of Health of Yogyakarta, Mangkuyudan Street MJ III/304,
Yogyakarta, Indonesia
694 G. Sugiyanto et al.
1 Introduction
Residential areas that are located on the right or left side of arterial or collector
roads are areas that have high potential for traffic accidents. According to the Traffic
Corps Republic of Indonesia, 75% of traffic accident locations are on roads with
land use around the residential areas [1]. Traffic accidents are caused by several
factors such as vehicles, negligence of road users (human error), road conditions
and the environment [2–5]. From the four factors, human error such as an increase
in vehicle speed is a factor that often results in traffic accidents [6, 7], which can
even reach one-third of fatal accidents [8]. The highest human error as a cause of
accidents is speeding [9, 10]. Speeding increases the chance of accidents occurring
[8] and the severity of traffic accident victims [11, 12]. Speed limit was introduced
to reduce the severity of crashes [7, 13], accident cost [14, 15], driving behavior [10,
16, 17], air pollution [18], travel time [19], generalized cost [20, 21], mobility [22],
and increased efficiency in transportation cost [23–26]. Limiting the vehicle speed
is one of strategies to improve traffic safety [27]. The purpose of speed limiting is
to balance mobility and safety by ensuring safe speed limits that are suitable for
road conditions [11, 28]. For this reason, a vehicle speed limit value is required in
accordance with road and traffic conditions in residential areas.
Higher vehicle speeds will cause noise and increase the potential for traffic acci-
dents. Based on the theory of movement, speed plays an important role as one of
the causes of traffic accidents [29]. Higher vehicle speed will have an effect on the
increase in reaction time and braking distance required for the vehicle to stop. Also,
it can increase the likelihood of mistakes being made by the driver. The higher speed
of the vehicle, the higher death rate due to accidents, both involving pedestrians and
other vehicles [30]. The ideal and safe speed is less than 30 kph to avoid conflicts that
could involve pedestrians and bicycles, less than 45 kph for conflicts with vehicles
from the side-impact collision, and less than 70 kph to avoid possible conflicts with
vehicles from the front (head-on collision). According to Baguley, traffic conflict is a
condition in which one or more road users approach each other or approach another
object at a certain time and place so that it causes a risk of collision if the move-
ment cannot be stopped or changed [31]. Traffic conflicts occur because of various
types of traffic flow movements. Xie et al. introduced an improved crash frequency
model for signalized intersections in urban road networks using conditional autore-
gressive model [32]. Road network patterns signal density and access densities are
used to analyze the safety impacts [33]. The other study, Tapp et al. analyze the lack
of compliance with 20 mph limits in Great Britain and 30 kph limits in Europe is
concerning [34].
Nowadays, the maximum speed limit setting is determined based on the vehicle
speed in the 85th percentile method [35, 36]. The 85th percentile method has disad-
vantage because it only uses vehicle speed as the reference parameter. For this reason,
another method is needed as a comparison in order to produce a more valid speed
limit value. The method to be used is stopping sight distance (SSD). In this method,
the determination of maximum speed limit refers to SSD, free-flow speed, and road
Determining the Maximum Speed Limit in Residential Area 695
friction coefficient between tire and road pavement surface. Kagaya et al. identified
the traffic management specific to setting of speed limits implemented in Hokkaido.
Based on the results of the questionnaire, it was found that more than half of the
respondents indicated that the vehicle speed limit for roads in Hokkaido is too low
and must be distinguished between summer and winter. The vehicle speed limit is
proposed depending on the characteristics and road conditions. The free-flow speed
was introduced to calculate the maximum speed limit of vehicles in summer and free-
flow speed and SSD are used in winter [37]. The aim of this research is to determine
the maximum speed limit in residential areas using the 85th percentile and stopping
sight distance method.
2 Method
2.1 Study Location
The research was conducted on three collector roads in Purbalingga Regency, namely
Padamara Street (access road to Abdi Negara Permai Housing), Letnan Yusuf Street
(access road to Griya Perwira Asri Housing), and Letkol Isdiman Street (access road
to Griya Abdi Kencana Housing). On the left or right side from three roads are
residential areas as shown in Fig. 1.
Fig. 1 Study location on three collector roads in Purbalingga Regency, Central Java [38]
2.2 Analysis Approach
There are several methods to determine the maximum speed limit value of a vehicle.
The method that is most commonly used is the 85th percentile. Another method that
can be used is the stopping sight distance method.
The 85th Percentile Method
In this method, the determination of the maximum speed limit refers to 85th percentile
speed. The 85th percentile speed is a traffic speed where 85% of the driver driving
the vehicle on the road without being affected by traffic speeds lower or bad weather
[39]. The purpose of this method is to determine the ideal speed limit on the reviewed
road sections based on the average speed of the vehicle. The 85th percentile method
only uses one parameter namely the vehicle speed. Speed is one of the main risk
factors in traffic safety [40].
Stopping Sight Distance Method
In this method, the determination of the maximum speed limit value refers to the
vehicles stopping sight distance [41], free-flow speed, and road friction coefficient
between tire and road pavement surface. The coefficient of friction between tire and
road pavement surface is influenced by vehicle speed and roughness. The pavement
type affects the road roughness value [42, 43]. Stopping sight distance is the distance
the driver has to stop this vehicle. For comfort and safety in driving, every length
of the road section should be filled with the minimum visibility along the minimum
stopping sight distance [44]. Two parameters in stopping sight distance method are
distance traveled during reaction time and braking distance. The sum of these two
parameters becomes the minimum stopping sight distance. Formula for minimum
stopping sight distance is:
d = d1 + d2 (1)
To find the distance traveled during the reaction time (d1 ), we must first know the
total reaction time, which is the total time it takes the driver from seeing the obstacle
to applying the brakes. The total reaction time is taken as 2.5 s which is the sum of
1.5 s of perception, identification, emotion, and volition or PIEV time (the time it
takes the driver from the moment he realizes an obstacle until he makes a decision)
and 1.0 s for the time it takes the driver to hit the brake pedal. So, d1 is the distance
traveled in 2.50 s. The equation for calculating the value of the distance traveled
during the reaction time (d1 ) is:
d1 = 0.278 · vt (2)
With v is vehicle speed (kph) and t is the total reaction time (seconds).
The braking distance (d2 ) is the distance a vehicle will travel from the point when
its brakes are fully applied to when it comes to a complete stop. The braking distance
is calculated due to the effect of friction between the tires and the road surface. The
equation for calculating the braking distance (d2 ) is:
v2
d2 = (3)
2 · g · ( f ± G) · (3.6)2
where v is the vehicle speed (kph), g is the acceleration due to gravity (m/s2 ) and
f is the factor of the road friction coefficient between tire and road pavement surface.
The value of road coefficient friction between tire and road surface is formulated:
f = −0.00065V + 0.192 f or V value between 40 and 80 kph (4)
f = −0.00125V + 0.24 f or V value between 80 and 112 kph (5)
Furthermore, the reaction time value of 2.50 s and the acceleration due to gravity
of 9.81 m/s2 are added to Eqs. (2) and (3) to become:
v2
d = 0.694 · v + (6)
254.016 · ( f ± G)
To solve the polynomial function, Eqs. (7) and (8) are used. The equation for
calculating maximum speed limit using the SSD method can be seen in Eq. (9).
ax 2 + bx + c = 0 (7)
√
−b ± b2 − 4ac
x= (8)
2a

V c2
−0.694 + 0.6942 + 4 · 254.016·(
1
f ±G)
· 0.694V c + 254.016·( f c±G)
v= (9)
2 · 254.016·(1
f ±G)
where V is the maximum speed limit of the vehicle (kph), Vc is free-flow speed (kph),
f is the value of road friction coefficient between tire and road pavement surface based
on the design speed, fc is the value of road friction coefficient between tire and road
pavement surface based on free-flow speed and G is the road slope (%).
Table 1 Road geometric conditions and side friction

Name of Road type Road Left Right Side Design Terrain
road width (m) shoulder shoulder friction speed condition
width (m) width (m) (kph)
Padamara 2/2 UD 7.5 1.0 1.0 201.5 60 Flat
street (low)
Letnan 2/2 UD 7.0 1.2 1.2 299 60 Flat
Yusuf (low)
Letkol 2/2 UD 12.0 1.5 1.5 916.3 60 Flat
Isdiman (very
high)
3.1 The Geometric Conditions and Side Friction of the Road
The geometric data of the roads collected include the number of lanes, the lane width,
the width of the shoulders and the conditions of the road terrain. The road geometric
conditions include road type, road width, road shoulder width, design speed, and
side friction on the three roads are shown in Table 1.
3.2 Traffic Flows During Peak Hour Time
The traffic flow data during peak hour time on the three collector roads in Purbalingga:
Padamara Street, Letnan Yusuf Street, and Letkol Isdiman Street are shown in Table 2.
Based on Table 2, the peak hours’ time for Padamara Street and Letkol Isdiman Street
are at 06.15–07.15 with a total traffic flow of 4049 vehicles/h and 3287 vehicles/h.
Table 2 Traffic flow during peak time

Name of Peak hour Vehicle type Total
road time Motor Passenger Pick Trucks Light Bus Minibus (veh./h)
cycles cars up, trucks
City
trans
Padamara 06.15–07.15 3768 246 19 1 12 1 2 4049
street
Letnan 06.30–07.30 2985 224 20 1 20 11 26 3287
Yusuf
Letkol 06.15–07.15 2802 287 18 0 1 0 8 3116
Isdiman
The peak hour time for Letnan Yusuf Street at 06.30–07.30 with a total traffic flow
of 3116 vehicles/h.
3.3 Determination of the Maximum Speed Limit Using 85th

Percentile Method
S-Curve of Vehicle Speed

The S-curve of vehicle speed presents the relationship between vehicle speed and
the percentage of the cumulative frequency of vehicle speeds. Vehicle speeds of
motorcycle (MC), passenger car (PC), pick up and city transport (PU&CT), light
truck (LT), minibus (MB), bus, truck (TR), obtained in the field using Speed Gun.
S-curve of vehicle speed is used in the determination of the 85th percentile speed.
The S-curves of vehicle speed on Padamara Street, Letnan Yusuf Street, and Letkol
Isdiman Street are presented in Figs. 2, 3 and 4.
Maximum Speed Limit based on 85th Percentile Method
The maximum speed limit for each type of vehicle and all vehicles using 85th
percentile method are shown in Table 3. The maximum speed limit of vehicles is
44 kph on Padamara Street and 42 kph to Letnan Yusuf and Letkol Isdiman.
Fig. 2 S-curve of vehicle speed on Padamara Street, Purbalingga

Fig. 3 S-curve of vehicle speed on Letnan Yusuf Street, Purbalingga
Fig. 4 S-curve of vehicle speed on Letkol Isdiman Street, Purbalingga
Table 3 Maximum speed limit of each vehicle using the 85th percentile method
Name of Maximum speed limit value of vehicles (veh/h)
road Motor Passenger Pick Truck Light Mini Big 4/ more All
cycles cars up, truck bus buses wheels vehicles
City
trans
Padamara 44 40 37 38 23 35 38 38 44
street
Letnan 43 35 39 42 N.A 34 38 37 42
Yusuf
Letkol 42 34 44 33 N.A 30 N.A 37 42
Isdiman
3.4 Determination of the Maximum Speed Limit Using SSD

Method
Free-flow Speed
Free-flow speed is influenced by the value of basic current speed (Co ), correction
factor of road width (FCw ), correction factor of side friction (FCsf ), correction factor
of separation (FCsp ) and correction factor of citizen size (FCcs ). The value of the
correction factor for speed refers to the Indonesian Highway Capacity Manual 1997
[45]. Data on road width and side friction in each road section refers to Table 1,
while the correction factor of citizen size is determined based on the population in
Purbalingga Regency in 2019 amounting to 933,989 peoples so the value of FCcs is
0.95. The free-flow speed for Padamara Street at 40.50 kph, Letnan Yusuf Street at
39.26 kph, and Letkol Isdiman Street is 39.57 kph.
Road Friction Coefficient between Tire and Road Pavement Surface
Road friction coefficient between tire and road surface is obtained from calculations
based on Eqs. (4) or (5). The value of road friction coefficient between tire and road
pavement surface is calculated for two conditions, the first based on the design speed
and second based on free-flow speed. Road friction coefficient for three collector
roads in Purbalingga Regency with a design speed of 60 kph is 0.153.
Maximum Speed Limit based on Stopping Sight Distance Method
The maximum speed limit based on the SSD method is determined using Eq. (9).
Vehicle speed limit for the three collector roads in Purbalingga are as follows:
Padamara Street, Purbalingga
The free-flow speed on Padamara Street is 40.5 kph. The road friction coefficient
between tire and road pavement surface with a design speed of 60 kph is 0.153 and
based on free-flow speed is 0.1657. Slope of the road is 0%. Obtained the maximum
speed limit on Padamara Street is 39.32 kph.
Letnan Yusuf Street, Purbalingga
The free-flow speed on Letnan Yusuf Street is 39.26 kph. The road friction coefficient
with a design speed of 60 kph is 0.153 and based on free-flow speed is 0.166. The
maximum speed limit on Letnan Yusuf Street is 38.10 kph.
Letkol Isdiman Street, Purbalingga
The free-flow speed on Letkol Isdiman Street is 39.57 kph. The road friction coef-
ficient with a design speed of 60 kph is 0.153 and based on free-flow speed is
0.166. The road slope is 0%. The maximum speed limit on Letkol Isdiman Street is
38.40 kph. The recapitulation of data from the analysis of the maximum speed limit
using stopping sight distance method is shown in Table 4.
Road safety strategies potentially are improved by applying system theory and
safety models [46]. Implementation of speed limit in arterial and collector roads
Table 4 Road friction coefficient and maximum speed limit using SSD method
Name of Design speed Friction Free-flow Friction Maximum
collector road (kph) coefficient speed (kph) coefficient speed limit
based on based on (kph)
design speed free-flow
(f) speed (fc)
Padamara street 60 0.153 40.50 0.1657 39.32
Letnan Yusuf 60 0.153 39.26 0.1660 38.10
Letkol Isdiman 60 0.153 39.57 0.1660 38.40
is one of the strategies to increase road safety. Operating speed has a significant
contribution to the total number of crashes [47]. The maximum speed limit of vehicles
proposed for collector roads with residential land functions on the right or left side is
40 kph. The proposed maximum speed limit value is in line with Ministerial Decree
PM 111/2015 [48] where the speed limit for vehicles in residential areas is 40 kph.
The result of this study relevant with study by Soriguera et al. that a stable flow of
1942 veh/h/lane was measured with a 40 kph speed limit in force [49].
4 Conclusion
Based on the results of study to determine the maximum speed limit in residential
area, the following conclusion can be drawn:
a. The maximum speed limit based on the 85th percentile method is greater than
that obtained from stopping sight distance method.
b. The maximum speed limit of vehicles based on the 85th percentile method
between 42 up to 44 kph and based on the stopping sight distance method
between 38.10 up to 39.32 kph.
c. The maximum speed limit of vehicles proposed for collector roads with
residential land functions on the right or left side is 40 kph.
Acknowledgements The authors fully acknowledged the Ministry of Education, Culture, Research
and Technology and Jenderal Soedirman University for the approved fund through Research Grant
Basic Research in the fiscal year 2021 which makes this important research viable and effective.
References
1. Traffic Corps Republic of Indonesia (Korlantas Polri) (2014) Kecelakaan Lalu Lintas dalam
Angka 2013. Traffic Corps Republic of Indonesia. Korlantas Polri, Jakarta
2. Treat JR, Tumbas NS, McDonald ST, Shinar D, Hume RD, Meyer RE (1977) Tri-level study
of the causes of traffic accidents. NTHSA, Washington
3. Sugiyanto G, Mulyono B, Santi MY (2014) Karakteristik kecelakaan lalu lintas dan lokasi
black spot di Kabupaten Cilacap. Jurnal Teknik Sipil 12(4):259–266
4. Sugiyanto G, Fadli A, Santi MY (2017) Identification of black spot and equivalent accident
number using upper control limit method. ARPN J Eng Appl Sci 12(2):528–535
5. Astarina L, Sugiyanto G, Indriyati EW (2018) Karakteristik kecelakaan lalu lintas dan analisis
lokasi black spot di Kabupaten Bogor. Dinamika Rekayasa 14(2):65–76
6. Nilsson G (1991) Speed limits, enforcement and other factors influencing speed. Enforcement
and rewarding: strategies and effects. Leidschendam, Netherlands, pp 46–50
7. Hoareau E, Newstead S, Oxley P, Cameron M (2002) An evaluation of the 50 kph speed limit
in Southeast Queensland. Journal Report No.264, Monash Univ. Accident Research Centre
8. Organization for Economic Cooperation Development (OECD) (2006) Speed management.
Transport Research Centre, OECD and European Conference of Ministers of Transport, Paris
9. Elvik R (2010) A restatement of the case for speed limits. Transp Pol 17(3):196–204
10. Schmid-Mast M, Sieverding M, Esslen M, Graber K, Jäncke L (2008) Masculinity causes
speeding in young man. Accid Anal Prev 40:840–842
11. Haglund M, Aberg L (2000) Speed choice in relation to speed limit and influences from other
drivers. Transp Res Part F: Traffic Psychol Behav 3(1):39–51
12. Sugiyanto G, Jajang, Santi MY (2019) The impact of lowering speed limit on mobility and the
environment. AIP Conf Proc 2094(1):020019, 1–8
13. Russo BJ, Savolainen PT, Schneider WH, Anastasopoulos PC (2014) Comparison of factors
affecting injury severity in angle collisions by fault status using random parameters bivariate
ordered probit model. Anal Methods Accid Res 2:21–29
14. Sugiyanto G (2017) The cost of traffic accident and equivalent accident number in developing
countries (case study in Indonesia). ARPN J Eng Appl Sci 12(2):389–397
15. Sugiyanto G, Santi MY (2017) Road traffic accident cost using human capital method (case
study in Purbalingga, Central Java, Indonesia). J Teknol (Sci Eng) 79(2):107–116
16. Tarko AP (2009) Modeling drivers’ speed selection as a trade-off behavior. Accid Anal Prev
41:608–616
17. Fuller R (2005) Towards a general theory of driver behavior. Accid Anal Prev 37(3):461–472
18. Delhomme P, Chappé J, Grenier K, Pinto M, Martha C (2010) Reducing air pollution: a new
argument for getting drivers to abide by the speed limit? Accid Anal Prev 42:327–338
19. Fuller R, Gormley M, Stradling S, Broughton P, Kinnear N, O’Dolan C, Hannigan B (2009)
Impact of speed change on estimated journey time: failure of drivers to appreciate relevance
of initial speed. Accid Anal Prev 41(1):10–14
20. Sugiyanto G (2016) The impact of congestion pricing scheme on the generalized cost and speed
of motorcycle to city of Yogyakarta, Indonesia. J Eng Appl Sci 11(8):1740–1746
21. Sugiyanto G (2018) The effect of congestion pricing scheme on the generalized cost and speed
of motorcycle. Walailak J Sci Tech (WJST) 15(1):95–106
22. Schepers P, Hagenzieker M, Methorst R, Wee BV (2014) A conceptual framework for road
safety and mobility applied to cycling safety. Accid Anal Prev 62:331–340
23. Sugiyanto G, Santosa PB, Wibowo A, Santi MY (2016) Hub and spoke airport networks in
Indonesia based on Herfindahl-Hirschmann Index. J Eng Appl Sci 11(8):1804–1810
24. Sugiyanto G, Santosa PB, Jajang, Fadli A, Santi MY (2018) Evaluation of hub-spoke airport
networks in Sumatra Island, Indonesia to increase efficiency of air transportation. MATEC Web
Conf 195(04009):1–9
25. Sugiyanto G, Santosa PB, Wibowo A, Santi MY (2015) Analysis of hub-and-spoke airport
networks in Java Island, Indonesia based on cargo volume and freight ratio. Procedia Eng
125:556–563. https://doi.org/10.1016/j.proeng.2015.11.061
26. Sugiyanto G, Santosa PB, Santi MY (2020) Hub and spoke airport networks in Sulawesi Island,
Indonesia based on freight ratio. ARPN J Eng Appl. Sci. 15(10):1201–1209
27. Sugiyanto G, Suciningtyas R, Jajang, Santi MY (2020) Strategies for handling black spot area
to increased road safety. ARPN J Eng Appl Sci 15(9):1101–1110
28. Wooley J (2005) Recent advantages of lower speed limits in Australia. Proc Eastern Asia Soc
Transp 6:3562–3573
29. Global Road Safety Partnership (GRSP) (2008) Speed management: a road safety manual for
decision-makers and practitioners, Geneva
30. Depart. of Planning, Transport, Infrastructure (2017) Speed limit guideline for South Australia
31. Baguley CJ (1984) The British traffic conflict technique. Transport and Road Research
Laboratory, NATO ASI Series Vol. F5
32. Xie K, Wang X, Ozbay K, Yang H (2014) Crash frequency modeling for signalized intersections
in a high-density urban road network. Anal Methods Accid Res 2:39–51
33. Wang X, Yuan J, Schultz G, Fang S (2018) Investigating the safety impact of roadway network
features of suburban arterials in Shanghai. Accid Anal Prev 113:137–148
34. Tapp A, Nancarrow C, Davis A, Jones S (2016) Vicious or virtuous circles? Exploring the
vulnerability of drivers to break low urban speed limits. Transp Res Part A 91:195–212
35. Sugiyanto G, Malkhamah S (2018) Determining the maximum speed limit in urban road to
increase traffic safety. J Teknol (Sci Eng) 80(5):67–77
36. Sugiyanto G, Fadli A, Suciningtyas R, Indriyati EW, Santi MY (2094) Road safety audit at black
spot area (Case study in Tlahab Lor, Karangreja, Purbalingga). AIP Conf Proc 2094(1):020020,
1–10
37. Kagaya S, Thanesuen S, Uchida K (2005) Traffic management in terms of speed limit in
Hokkaido Area. J Eastern Asia Soc Transp Stud 6:2247–2262
38. https://www.google.com/maps/search/Perumahan+Griya+Abdi+Kencana (2021)
39. Abraham J (2001) Analysis of highway speed limits. Faculty of Applied Science and
Engineering University Toronto, Canada
40. Pauw ED, Daniels S, Thierie M, Brijs T (2014) Safety effects of reducing the speed limit from
90 kph to 70 kph. Accid Anal Prev 62:426–431
41. Shaheed MS, Gkritza K, Carriquiry AL, Hallmark SL (2016) Analysis of occupant injury
severity in winter weather crashes: a fully Bayesian multivariate approach. Anal Methods
Accid Res 11:33–47
42. Sugiyanto G (2017) Characterization of asphalt concrete produced from scrapped tire rubber.
Eng J 21(4):193–206
43. Sugiyanto G (2017) Marshall test characteristics of asphalt concrete mixture with scrapped tire
rubber as a fine aggregate. J Teknol (Sci Eng) 79(2):55–64
44. Sukirman S (1994) Dasar-Dasar Perencanaan Geometrik Jalan. NOVA, Bandung
45. Ministry of Public Work (1997) Manual Kapasitas Jalan Indonesia (Indonesian highway
capacity manual) 1997. Jakarta
46. Hughes BP, Anund A, Falkmer T (2015) System theory and safety models in Swedish, UK,
Dutch and Australian road safety strategies. Accid Anal Prev 74:271–278
47. Vayalamkuzhi P, Amirthalingam V (2016) Influence of geometric design characteristics on
safety under heterogeneous traffic flow. J. Traffic Transp Eng (Engl Ed) 3(6):559–570
48. Ministry of Transportation (2015) Peraturan Menteri Perhubungan Republik Indonesia Nomor
PM 111 Tahun 2015 tentang Tata Cara Penetapan Batas Kecepatan. Jakarta
49. Soriguera F, Martinez I, Sala M, Menendez M (2017) Effects of low speed limits on freeway
traffic flow. Transp Res Part C: Emerg Technol 77:257–274
Identification of Traffic Accident
Hazardous Location and Cost
of Accidents in Developing Country
(Case Study of Tabanan Regency,
Bali-Indonesia)
Putu Alit Suthanaya and Made Oka Sugiana
Abstract Road traffic accidents have long become a major issue in Indonesia. The
number of accident fatalities has increased following the increase in population and
motor vehicle ownership. The increase in traffic accidents will be followed by an
increase in traffic accident costs. As limited funds are available, one important step
required is to identify traffic accident hazardous locations. Previous studies on black
spots in developing countries such as Indonesia mainly did not consider the frequency
of each type of accident, the impact on mitigation measures such as shortcuts devel-
opment in reducing accidents, and the cost of traffic accidents. The objectives of this
study were to identify black sites and black spots, calculate economic loss due to
road traffic accidents, and evaluate the impact on the shortcut’s development. This
study applied a z-score to identify black sites and CUSUM to identify black spots.
The cost incurred due to the accident was analyzed by using the gross output method.
Three road sections were found to be the locations of accident-prone areas (black
sites). By comparing conditions before and after three shortcuts were built on black
spot location, it was found that the existence of the shortcuts has reduced the number
of traffic accidents by 70%. The refinement of the road infrastructure has reduced
the number of traffic accidents significantly. The average annual cost of accidents in
the Tabanan Regency was IDR 10,115,918,496/year.
Keywords Accident costs · Black-sites · Black-spots · CUSUM · Z-score
1 Introduction
Motor vehicle accidents cause more than 1.35 million deaths worldwide. The global
rate of road traffic deaths is 18.2 per 100,000 population and regional rates of road
traffic deaths in Africa and South-East Asia are 26.6 and 20.7 deaths per 100,000
population respectively [1].
P. A. Suthanaya (B) · M. O. Sugiana

Civil Engineering Department, Engineering Faculty, Udayana University, Kampus Bukit Jimbaran
Street-Bali, Kuta Selatan, Indonesia
706 P. A. Suthanaya and M. O. Sugiana
In Asia, such as India, a study has analyzed traffic accidents at the national, state,
and metropolitan city levels and found that road accidents vary according to age,
gender, month, and time. It was also found that the age group of 30–59 years is the
most vulnerable population group. Males tended to have higher fatality rates than
females, and higher accidents occurred in extreme weather and during working hours
[2].
Similar to India, Indonesia also experienced a high rate of traffic accidents. The
traffic fatalities from traffic accidents in 2014 were predicted to be 28,000 with a
rate of 12 per 100,000 population compared to only 4.8 in Singapore and 5.2 in
Australia. It is estimated that traffic fatality in Indonesia will reach 65,000 in 2035.
The Indonesian government has targeted to reduce these numbers by 50% in 2020
and 80% in 2035 [3]. The increase in the fatality rates in Indonesia was correlated
with the increase in the number of vehicles on the road, which is dominated by
motorcycles [4].
Road traffic accidents are a major issue in Indonesia that contribute to economic
and social losses. Strategies need to be developed to reduce road crashes. One
important field in road traffic accident studies is the identification of traffic acci-
dent hazardous locations. The accuracy of the black spots identification contributes
to the effectiveness of the traffic management implemented [5]. Suitable criteria for
identifying black spots are different from country to country and are also influenced
by financial capabilities [6]. In developing countries like Indonesia, due to limited
funds available, the treatments need to be prioritized. Therefore, it is required to
develop an approach to identify the black spot and also a priority location to be
treated so that it is economically effective.
Various research on black spots has been conducted. A study has employed Kernel
Density through the GIS environment to identify black spot locations and also used
the SI (severity index) method to generate a black site map to determine accident-
prone zones and take appropriate measures to be implemented [7]. An empirical
Bayes technique was used to identify black spots of two small-sized cities, in India,
i.e. Patiala and Rajpura. It was found that the black spots mainly lie on the national
highways, state highways, and major district roads passing through the cities [8].
Another study has identified the accident-prone zones within Cochin City using
road accident data for three years (2014, 2015, and 2016) [9]. The Accident Severity
Index (ASI) method was used to rank the accident locations, while other studies also
applied the index [10, 11].
Much research on traffic accidents has been conducted in Indonesia to identify
appropriate traffic management measures to be implemented [12]. A study focused on
identification of the factors that influence fatal traffic accidents involving motorcycles
[13]. Another study investigated types of and factors that influence traffic violations
by young motorcyclists [14]. The cost of traffic accidents has been analyzed based
on a gross output method [15]. A black spot analysis has been conducted on Siak
highway, Riau Province, Indonesia [16]. There were some limitations on the previous
studies such as only considering one section of the road, did not analyze black site
areas in the region [16], and only focused on the traffic accident cost analysis [15]
based on the gross output method developed by [17]. Mainly, previous studies on the
Identification of Traffic Accident Hazardous Location … 707
black spots have not considered the frequency of each type of traffic accident, the
impact on mitigation measures such as shortcuts development in reducing accidents,
and the cost of traffic accidents. The objectives of this study were to identify the traffic
accident hazardous locations (black sites and black spots), investigate the frequency
of accidents at black spot locations, evaluate the impact on shortcut development,
and analyze the cost of traffic accidents in developing countries with the case study
of Tabanan regency in Bali, Indonesia, where the mode of transport was dominated
by motorcycle with the average share of about 70% on the road.
This study applied a z-score to identify black sites and CUSUM to identify black
spots. The cost incurred due to the accident was analyzed by using the gross output
method. Unlike previous study [15], this study applied the gross output method
developed by the Department of Public Works [18]. To analyze the impact on the
shortcut development at the same time, this study used road traffic accident data
over a 5-year time period from 2013 to 2017 as three shortcuts were developed in
2016. The period from 2013 to 2015 was analyzed as the time before the shortcuts
developed and the period from 2016 to 2017 was the time after the development.
2 Theoretical Background
2.1 Black Sites and Black Spots
In an accident analysis, it is required to determine the highest rate of traffic accident

locations based on the black-spot analysis [19]. Prioritization of hazardous locations
can be done based on real-time data and traffic accident records from the police
station. The black-spot is characterized by a high number of traffic accidents. Crash
black site analysis is usually focused on the road segments or junctions. There is no
general method to determine black sites. Thus, continued review of all methods is
necessary to identify the most successful way of identifying black sites [20].
The definition of black spots has changed over time. Several statistical definitions
of road traffic accident black spots in different studies have been discussed. The
black spots has been defined as the “number of accidents (crashes), or accident
consequences by kind and severity, expected to occur on the entity during the specified
period on an entity (an intersection, a road segment, a driver, a bus fleet, etc.)” [21].
Another researcher defined black spots as the identification of transportation system
locations bearing treatable safety problems which are observed through the risen
frequency of crashes relative to similar locations [22].
2.2 CUSUM
CUSUM (cumulative sum) can be used to determine changes in a series of quantities

over time. CUSUM method has been used in various studies, including in the medical
and financial fields. CUSUM can also be used to identify black spots. CUSUM graph
is a standard statistical procedure as quality control to detect changes in the mean
value. The calculation to find the next year’s CUSUM accident value is as follows
[23]:
S = [S0 + (X1 − W)] (1)
where S is the accident CUSUM value, S0 is the first-year accident CUSUM value,
X1 is the number of accidents every year, and W is the mean value.
2.3 Z-Score
The Z-score is a widely used standardized score. Z-score and describes the position
of a score from the mean. It can be used if the mean and standard deviation of the
population is known. A positive z-value explains a standard score above the mean
and, on the other hand, a negative z value describes a standard score below the mean.
The Z-score is very useful in explaining the probability of a value occurring within
a normal distribution. It can be calculated as follows [24]:
Z i = Xi − μσ (2)
where Z i is the Z-score accident at i location, σ is the standard deviation, Xi is the

number of accidents for location i, and μ is the mean value.
2.4 The Cost of Traffic Accident
To calculate the cost of traffic accidents, this study uses the Guidelines for Calculating
the Cost of Traffic Accidents [18]. The approach used in this guideline is the gross
output (human capital) approach. The cost of traffic accidents (CTA) in year n is
calculated as follows [18]:

k
CTA = (NTAi × UCTAi (Tn )) (3)
i =1
where CTA is the amount of traffic accident costs on the road or intersection or area
in year n (IDR/year), NTAi is the number of traffic accidents for each accident class
(accidents/year), UCTAi (Tn ) is the unit cost of traffic accidents in year ‘n’ for each
accident class (IDR/accident), and i is the traffic accident class.
3 Research Methods
Figure 1 shows the study location in the Tabanan Regency, Bali Province, Indonesia.
The population of the Tabanan Regency was 445,700 people with a total area of
839.33 km2 and a total road length of 1,058.379 km [25]. The traffic accident data
used were for a 5-year time period from 2013 to 2017 [26]. This study period was
selected based on the development of three shortcuts in the study area. The period
2013–2015 was the time before the shortcuts were built and the period 2016–2017
was the time after the shortcuts were built. Therefore, the impact on the development
of shortcuts to reduce traffic accidents could be measured by comparing conditions
before and after they were built.
Fig. 1 Study location in the Tabanan Regency, Bali-Indonesia

The first stage of the research was analyzing the location of black sites. The Z-
score values of each location were calculated to determine the growth of the number
of accidents per year for a 5-year time period (2013–2017) and to determine the
value of the Z-score based on the latest data (2017’s data). The black site location was
identified with the graph of the Z-score. Next, the second step was to analyze the black
spot location. Black spot analysis was performed using the CUSUM (cumulative
sum) method. The road was divided into stations every 1 km. The third stage was
to analyze the frequency of accidents at each black spot location. Finally, the fourth
stage was to analyze the costs incurred due to accidents that occurred in the Tabanan
Regency using the Gross Output method [18].
4.1 Traffic Accident Characteristics
Table 1 shows that the number of traffic accidents on the roads in the Tabanan Regency
from 2013 to 2017 varied and tended to decrease every year. The highest number of
accidents occurred in 2013 with 299 events, while the lowest occurred in 2017 with
128 incidents.
Table 2 shows that during 2013–2017, the number of human victims due to acci-
dents in the Tabanan Regency reached 1428 people, where the highest number of
human victims occurred in 2013, as many as 445 people.
Table 1 The number of

No. Year Number of accidents
traffic accidents on the roads
in Tabanan Regency 1 2013 299
(2013–2017) 2 2014 234
3 2015 167
4 2016 129
5 2017 128
Table 2 The number of traffic victims

No. Year Fatal Serious Minor Total
1 2013 47 112 286 445
2 2014 37 81 243 361
3 2015 44 44 161 249
4 2016 40 20 108 168
5 2017 42 25 138 205
Total 210 282 936 1,428
Table 3 Traffic accident rates

No. Road section Number of accidents Total
2013 2014 2015 2016 2017
1 Denpasar-Gilimanuk 2000 1649 1370 645 860 6524
2 Ir. Soekarno 702 401 281 273 234 1891
3 Ahmad Yani 344 253 244 76 189 1106
4 Perean Tengah 366 202 104 80 128 880
5 Batukaru 92 192 272 64 196 816
6 Baturiti-Mekarsari 236 177 88 171 73 745
7 Wisnu Marga 279 137 69 0 16 501
8 Marga-Apuan 48 324 52 20 16 460
9 Raya Baturiti 64 120 69 104 100 457
10 Raya Tanah Lot 136 132 97 32 32 429
4.2 Traffic Accident Rates
The top ten ranking results of the traffic accidents on the roads in the Tabanan Regency
can be seen in Table 3. It was found that the Denpasar-Gilimanuk road segment had
the highest accident rate compared to the accident rate for the other road sections
in the Tabanan Regency. The total number of accidents on the Denpasar-Gilimanuk
road during 2013–2017 was 6524. The highest number of accidents occurred in 2013
(2000 accidents).
4.3 Analysis of Black Sites
Figure 2 shows the graph of the Z-score. The three roads with the highest Z-score
value were Denpasar-Gilimanuk, Ir. Soekarno, and the Ahmad Yani road section.
4.4 Analysis of Black Spots on Denpasar-Gilimanuk Road

Section
Black spot analysis on the Denpasar-Gilimanuk road section was divided into two,
namely black spot analysis before the shortcuts were built (during 2013–2015) and
black spot analysis after the shortcuts were built (during 2016–2017). Before the three
shortcuts were built, the Denpasar-Gilimanuk road section had a length of 34.84 km
with 35 stations (station per 1 km). After the shortcuts were built, the Denpasar-
Gilimanuk road section had a length of 33.10 km, with 34 stations. Figure 3 shows
Fig. 2 Z-scores based on the accident growth rates and accident rate in 2017
the graph of the relationship between the year of the accident and the value of
CUSUM.
From Fig. 3, it can be seen that the station identified as the location of the black
spot was at Sta 27-Sta 28 (total 20 accidents) located in the Lalanglinggah Village
area with the largest CUSUM value of 13,914. The accident happened in a flat area
with slight corners in areas with several shops and schools. Several factors have
caused the accidents such as a shaky vehicle, mistakes in overtaking, and crashing
into the vehicle from behind. The type of accidents was dominated by collisions
when overtaking (sideswipe) by 50%, head-on collisions, and out of control, each
by 18.75%, and front-rear collisions (rear end) by 12.5%. The accident occurred in
the morning 37.5% of the time during the day 25% of the time, and at night 18.75%
of the time. Collisions were dominated by collisions between motorbikes and cars
by 50%, motorbikes with motorbikes by 31.25%, and cars with cars by 18.75%.
After three shortcuts were built in 2016, the Denpasar-Gilimanuk road section
had a length of 33.1 km. It was found that the station identified as the location of the
accident or black spot was at Sta 10–Sta 11 (km 10,600 to km 10,900) which was
located in the Megati area with the largest CUSUM value of 4,176. There were six
accidents identified at this location. The cause of the accident, based on police data
in the area, was a shaky vehicle, an error in overtaking and crashing into a vehicle
from behind. The type of accidents was dominated by collisions when overtaking
(sideswipe) and front-rear collision (rear end) each by 33.33%, head-on collision, and
out of control by 16.67%. Traffic collisions were dominated by collisions between
motorbikes-cars (83.33%) and motorbikes-motorbikes (16.67%).
Fig. 3 Black spot on Denpasar-Gilimanuk road section (2013–2015) (before the shortcut) based
on CUSUM
4.5 Analysis of the Amount of Traffic Accident Costs (TAC)
Table 4 shows the total amount of traffic accident costs in the Tabanan Regency for
the period of the year 2013–2017.
The highest traffic accident costs were obtained in 2013 where the total costs of
fatal, severe, and minor accidents reached IDR 12,774,884,608/year. The region’s
gross domestic product (GDP) was IDR 6,452,645.72 million in 2013 [27]. Therefore,
the total accident cost of the Tabanan regency was about 0.20% of the GDP, which
was lower if compared to 0.38% in the Purbalingga Regency [15]. One way to reduce
Table 4 The total amount of traffic accident costs (2013–2017)

No. Year Accident Cost of Cost of Economic Total economic loss
accidents on accidents at loss (IDR/Year)
the road intersection (IDR/Year)
section (IDR/Year)
(IDR/Year)
1 2013 Fatal 7,833,677,574 – 7,833,677,574 12,774,884,608
Serious 2,367,084,333 – 2,367,084,333
Minor 2,574,122,701 – 2,574,122,701
2 2014 Fatal 5,960,338,731 – 5,960,338,731 9,845,880,419
Serious 1,396,012,643 25,382,048 1,421,394,691
Minor 2,447,269,279 16,877,719 2,464,146,998
3 2015 Fatal 6,857,404,771 – 6,857,404,771 9,378,780,377
Serious 836,104,968 – 836,104,968
Minor 1,685,270,638 – 1,685,270,638
4 2016 Fatal 7,244,775,957 – 7,244,775,957 8,971,414,324
Serious 295,860,969 – 295,860,969
Minor 1,412,892,680 17,884,717 1,430,777,398
5 2017 Fatal 7,840,222,236 – 7,840,222,236 9,608,632,751
Serious 332,246,489 55,374,415 387,620,903
Minor 1,380,789,613 – 1,380,789,613
Average 10,115,918,496
the traffic accident costs was by improving the road geometry, such as developing
shortcuts in the identified black spot location. As limited funds are available, the
strategy to be implemented in reducing traffic accidents needs to be preceded by the
identification of the black site and black spot. The unit cost of the traffic accident
that has been developed by the Department of Public Work [18] needs to be updated
to accommodate present conditions.
5 Conclusions
As limited funds are available, it is important to rank the location of accident-prone

areas to prioritize the mitigation measures. This study applied a Z-score to rank black
site location and CUSUM to determine black spot location. Three roads were found
to be the locations of black sites. By comparing conditions before and after three
shortcuts were built on the black spot location (Gilimanuk-Denpasar road section), it
was found that the existence of the shortcuts has reduced the number of traffic acci-
dents by 70%. This indicated that the refinement of road infrastructure can reduce the
number of traffic accidents significantly. The highest total cost of accidents occurred
in 2013 with the cost of fatalities and injuries reaching IDR 12,774,884,608/year or
0.20% of GDP. The average annual cost of accidents in the Tabanan Regency during
the year 2013–2017 was IDR. 10,115,918,496/year. The unit cost of the traffic acci-
dent that has been developed by the Department of Public Work [18] needs to be
updated to accommodate present conditions.
References
1. World Health Organization (2018) Global status report on road safety 2018 summary. Avail-
able at: https://www.who.int/publications-detail/global-status-report-on-road-safety-2018.
Accessed 26 March 2019
2. Singh SK (2017) Road traffic accidents in India: issues and challenges. Transp Res Procedia
25:4708–4719
3. Directorate General of Land Transportation, Ministry of Transportation Republic of Indonesia
(2011) National plan on road safety 2011–2035, Jakarta
4. Jusuf A, Nurprasetio IP, Prihutama A (2017) Macro data analysis of traffic accidents in
Indonesia. J Eng Technol Sci 49(1):132–143
5. Elvik R (2006) New approach to accident analysis for hazardous road locations. Transp Res
Record J Transp Res Board 1953:50–55
6. Jafri SF (2013) An effective and way forward approach for road safety acquisition of knowledge
correlative study for Pakistan. Eng J 17:41–48
7. Shad R, Rahimi S (2017) Identification of road crash black-sites using geographic information
system. Int J Traffic Transp Eng 7(3):368–380
8. Dhanoa KK, Tiwari G, Malayath M (2017) Method to identify black spots with minimal
available police data. J Eastern Asia Soc Transp Stud 12:2054–2073
9. Fayaz MM, Mrudula SP, George SJ, Yoyak SP, Roy SS (2018) Black spot identification using
accident severity index method. Int J Curr Eng Sci Res ((IJCESR) 5(3):2394–8374
10. Katre N, Pitale NH, Bobade S (2019) Analysis of black spots on NH-3 and its rectification.
MAT J Transp Syst 4(2):16–32
11. Matthew NA, Silven A, Mathew EE, Dominic S, Joy T (2019) Identification of accident black
spots along NH-85. Int J Sci Res Rev 8(6):222–226
12. Soehodho S (2017) Public transportation development and traffic accident prevention in
Indonesia. IATSS Res 40:76–80
13. Suthanaya PA (2016) Analysis of fatal accidents involving motorcycles in low income region
(case study of Karangasem Region, Bali-Indonesia). Int J Eng Res Afr 19:112–122
14. Joewono TB, Susilo YO (2017) Traffic violations by young motorcyclists on Indonesian urban
roads. J Transp Saf Secur 9:236–261
15. Sugiyanto G (2017) The cost of traffic accident and equivalent accident number in developing
countries (case study in Indonesia). ARPN J Eng Appl Sci 12(2):389–397
16. Sandhyavitria A, Zamrib, Wiyono S, Subiantoro (2017) Three strategies reducing accidents
rates at black spots and black sites road in Riau Province, Indonesia. Transp Res Procedia
25:2153–2166
17. Silcock R, Transport Research Laboratory (TRL) (2003) Guidelines for estimating the cost of
road crashes in developing countries. Department of International Development. Project R7780
18. Department of Public Works (2003) Calculation of the cost of traffic accidents using the gross
output (human capital) method. Department of Public Work, Jakarta
19. Raut UM, Dhumal RK, Nagne SD, Kale KV (2015) GIS contribution for identification of
accident black spots—a review. Int J Comput Sci Eng 3(7):75–80
20. Anderson T (2007) Comparison of spatial methods for measuring road accident ‘hotspots’: a
case study of London. J Maps 3(1):55–63
21. Hauer E (1996) Identification of sites with promise. Transp Res Record J Transp Res Board
1542:54–60
22. Cheng W, Washington SP (2005) Experimental evaluation of hotspot identification methods.
Accid Anal Prev 37(5):870–881
23. Austroads (1992) Road crashes, guide and traffic engineering practice part 4. Australia, Sydney
24. Base C (2018) Understanding basic statistics. Cengage Learning
25. Tabanan Bureau of Statistics (2020) Tabanan regency in figures 2019. Bureau of Statistics,
Tabanan. Bali
26. Tabanan Police Department (2018) Annual traffic accident report. Tabanan
27. Tabanan Bureau of Statistics (2014) Tabanan regency in figures 2013. Bureau of Statistics,
Tabanan, Bali
Characteristics of Foreign Motorcyclists
in Tourism Areas in Bali

Abstract Along with the development of tourism, the number of foreign motor-
cyclists also increases. Meanwhile, it is widely known and empirically found that
foreign riders are prone to greater risks than domestic riders in most countries around
the world. This is due to different traffic cultures. Therefore, safety measures of
foreign riders are highly important. This study examined the characteristics of foreign
motorcyclists in tourism areas in Bali, as a preliminary study on the behaviour of
foreign motorcyclists in tourism areas in Bali. This study was conducted through
questionnaire survey on 30 foreign motorcyclists. The results obtained were analyzed
using SPSS. Based on the survey results respondents’ characteristics were catego-
rized based on gender, age, education, marital status, and countries of origin, driver’s
license ownership, purpose of riding and the riding distance traveled.
Keywords Characteristics · Motorcyclist · Foreigner
1 Introduction
Foreigners coming to Indonesia come from many countries with varying character-
istics, both in terms of age, purposes of arrival, education and cultural background.
These differences in characteristics lead to differences in the choice of transportation
modes in Indonesia. The mode of transportation used varied from large buses with
C. P. Wirasutama (B)
Mahasaraswati University, Denpasar, Indonesia
P. A. Suthanaya · D. M. P. Wedagama · A. A. G. A. Yana
Udayana University, Denpasar, Indonesia
D. M. P. Wedagama
A. A. G. A. Yana
718 C. P. Wirasutama et al.
a capacity of 40–50 passengers, medium buses with a capacity of 20 passengers

minibuses with a capacity of dozens of passengers, vehicles like MPV (multipurpose
vehicle) with a capacity of 7 or 8 passengers, and some also prefer bikes as mode
of transportation. Apart from using these modes, foreigners often use two-wheeled
motorcycles as a means of transportation. The behaviour of foreigners who ride
two-wheeled motorcycles will be examined in this study.
Foreign motorcyclists are considered to be prone to greater risks than domestic
drivers in most countries around the world. Several empirical findings have been
reported, including by [1]. It is said that the traffic cultures of different countries
largely explain the differences of some specific technical parameters.
Previously, the risks faced by foreign motorcyclists were not thoroughly analyzed
in detail, or at least not as thoroughly documented. The same applies, however, in
most countries around the world [2]. Lack of knowledge of the road network, lack of
understanding of local traffic regulations, and insufficient driving skills in unfamiliar
conditions, can lead to an increase in the number of accidents, severity, and the risks
faced by foreign motorcyclists.
Variations in general road safety attitudes expressed in riding behaviour might
further lead to a greater risk of accidents for foreign motorcyclists compared to
locals. A study on the characteristics of foreign motorcyclists in Bali tourist regions
was done as a preliminary study on the behaviour of foreign motorcyclists in Bali
tourist areas. The method used in this study is SPSS. The intended contribution of
this study is to know the characteristics of foreign motorcyclists in tourist areas.
Subsequently, it can be utilised as input for future study on foreign motorcyclists.
2 Research Method
See Fig. 1.
2.1 Research Location and Time
The research location was conducted in tourism destination areas in Bali, such as in
Kuta, Ubud, Sanur, and Canggu. The research was conducted from September 2020
to December 2020.
2.2 Scope of Study
The study was carried out in tourism destination areas in Bali, such as in Kuta,
Ubud, Sanur, and Canggu. Therefore, the questionnaires were made based on the
characteristics that need to be observed from foreign motorcyclists in those areas.
Characteristics of Foreign Motorcyclists in Tourism Areas in Bali 719
Fig. 1 Research design

Preliminary studies
Observation
(Photos of Violations / Behaviour of Foreign
Motorcyclists)
Background
Research Questions
Study Objectives
Selecting Locations for Questionnaire

Distribution
Questionnaire Design
• Question
• Respondents
Distribution of Questionnaires
Characteristics Analysis with SPSS
Results and Discussion

Conclusion and Suggestions
2.3 Determination of Data Sources
Sources of data used are from the collection of observations or recordings in the field
and the results of questionnaires about the characteristics of foreign motorcyclists in
tourism areas in Bali. The questions given were designed in such a way as to get the
desired results without being disturbance to the respondents. A structured question-
naire was developed to collect information on foreign motorcyclists’ characteristics
in tourism areas in Bali.
There are 30 respondents examined in this study, according to Bailey’s statement
[3]. The questionnaires were distributed as independent questionnaires. On this occa-
sion, if the respondent does not want to write, the surveyor asks the question as a
face-to-face interview. This implies that the sample selection process may not be
completely random as respondents are approached based on respondents’ availability.
The questionnaires were given only to respondents who claimed to be motorcyclists.
The questionnaire survey per respondent took about 20 min to complete. The data
sources in this study are as follows:
1. Primary data is data obtained directly from primary data sources. The primary
data source in this study is the result of direct field observations on foreign
Table 1 Data, data types and data sources

Data Type of data Data source
Primary Secondary
Photos of foreign motorcyclists Qualitative Observation
Characteristics of foreign motorcyclists Quantitative Questionnaire Literature
motorcyclists’ behaviour (field observation). Results of the questionnaires given

to foreign motorcyclists are also included as primary data source.
2. Secondary data is obtained from secondary data sources which are not directly
obtained by researchers, such as through relevant journals, books, and archives
related to the discussions. These sources can be used as guidelines or references
in this study, including data on Bali’s foreign visitors (Tables 1 and 2).
3 Results
3.1 Respondents Characteristics Based on Gender
The results show that the questionnaires had been distributed to 30 respondents
consist of 14 male respondents (46.7%), and 16 female respondents (53.3%) (Fig. 2
and Table 3).
3.2 Based on Age
The results showed that the questionnaire distributed to 30 respondents with the
following age groups: 18–25 years (5 people) with a percentage of 16.7%, 26–
35 years (7 people) a percentage of 23.3%, age 36–45 years (6 people) a percentage
of 20.0%, age 46–55 years (8 people) a percentage of 26.7% and age > 50 years (4
people) a percentage of 13.3% (Fig. 3 and Table 4).
3.3 Based on Country of Origin
The results showed that the questionnaires distributed to 30 respondents from

different country of origin, such as African countries (1 person) with a percentage of
3.3%, America (1 person) 3.3%, Australia (1 person) 3.3%, Canada (1 person) 3.3%,
German (2 people) 6.7%, Holland (2 people) 6.7%, India (1 person) 3.3%, Italy (1
person) 3.3%, Japan (1 person) 3.3%, Malaysia (1 person) 3.3%, Poland (1 person)
3.3%, France (2 people) a percentage of 6.7%, Romania (1 person) a percentage of
Table 2 Number of foreign tourist visits to Indonesia in 2019 by entrance
Entrance Number of visits Changes in Dec Changes in Des 2019 Changes in Jan-Des
Dec 2018 Nov 2019 Dec 2019 Jan-Dec 2018 Jan-Dec 2019 2019 from 2018 (%) from Nov 2019 (%) 2019 from 2018
A. Air 825,635 777,244 838,978 10,088,781 9,834,706 1.62 7.94 −2.52

1. Ngurah Rai 495,641 492,904 544,726 6,025,760 6,239,543 9.9 10.51 3.55
2. 224,989 183,759 186,723 2,814,586 2,419,196 −17.01 1.61 −14.05
Soekarno-Hatta
3. Juanda 26,609 20,780 20,546 320,529 243,899 −22.79 −1.13 −23.91
4. Kualanamu 19,966 20,798 22,431 229,586 244,530 12.35 7.85 6.51
5. Husein 17,311 14,616 14,951 155,566 157,833 −13.63 2.29 1.46
Sastranegara
6. Adi Sucipto 10,813 9218 9795 138,822 113,028 −9.41 6.26 −18.58
7. Bandara Int. 3151 4635 5499 79,807 57,763 74.52 18.64 −27.62
Lombok
8. Sam Ratulangi 8845 9261 10,743 122,104 130,285 21.46 16 6.7
9. Minangkabau 4801 5013 5180 54,383 61,131 7.89 3.33 12.41
10. Sultan Syarif 3013 2917 4042 29,776 34,419 34.15 38.57 15.59
Characteristics of Foreign Motorcyclists in Tourism Areas in Bali
Kasim II
11. Sultan 1315 2417 4012 29,213 28,303 205.1 65.99 −3.12
Iskandar Muda
12. Ahmad Yani 1930 2296 2627 23,592 23,977 36.11 14.42 1.63
13. Supadio 2033 1811 1527 23,050 22,500 −24.89 −15.68 −2.39
14. Hasanuddin 1406 1595 1821 14,126 17,771 29.52 14.17 25.8
15. Sultan 1431 1218 762 13,862 15,846 −46.75 −37.44 14.31
Badaruddin II
16. Others 2381 4006 3593 14,019 24,682 50.9 −10.31 76.06
721
(continued)
Table 2 (continued)
722
Entrance Number of visits Changes in Dec Changes in Des 2019 Changes in Jan-Des
Dec 2018 Nov 2019 Dec 2019 Jan-Dec 2018 Jan-Dec 2019 2019 from 2018 (%) from Nov 2019 (%) 2019 from 2018
B. Sea 370,629 358,264 388,495 3,214,420 4,160,632 4.82 8.44 29.44

1. Batam 233,153 167,288 190,232 1,887,284 1,947,943 −18.41 13.72 3.21
2. Tanjung Uban 61,258 52,374 59,340 522,399 634,735 −3.13 13.3 21.5
3. Tanjung 16,020 12,378 16,426 140,603 169,364 2.53 32.7 20.46
Pinang
4. Tanjung Balai 9020 9209 10,916 84,718 112,753 21.02 18.54 33.09
Karimun
5. Tanjung Benoa 3178 5021 7677 31,062 29,456 141.57 52.9 −5.17
6. Tanjung Emas 2315 2202 5135 19,907 17,761 121.81 133.2 −10.78
7. Others 45,685 109,792 98,769 528,447 1,248,620 116.2 −10.04 136.28
C. Land 209,290 145,273 149,594 2,507,104 2,111,616 −28.52 2.97 -15.77
1. Jayapura 3202 5006 6155 104,075 56,287 92.22 22.95 -45.92
2. Atambua 13,915 9824 12,561 85,914 103,905 -9.73 27.86 20.94
3. Entikong 3641 1691 3177 23,213 23,284 −12.74 87.88 0.31
4. Aruk 1125 1009 1708 12,329 16,508 51.82 69.28 33.9
5. Nanga Badau 1826 954 1585 12,148 14,472 −13.2 66.14 19.13
6. Others 185,581 126,789 124,408 2,269,425 1,897,160 −32.96 −1.88 −16.4
Total (A + B + 1,405,554 1,280,781 1,377,067 15,810,305 16,106,954 −2.03 7.52 1.88
C)
aIncluded as entrance via air (Hang Nadim airport)
Source BPS, 2016 [4]
C. P. Wirasutama et al.
Fig. 2 Percentage based on

gender
Gender
Male 47%
53%
Female
Table 3 Respondents
Gender Total Percentage (%)
characteristics based on
gender Male 14 46.7
Female 16 53.3
30 100

age
Table 4 Characteristics of
Age Total Percentage (%)
respondents by age
18–25 years old 5 16.7
26–35 years old 7 23.3
36–45 years old 6 20
46–55 years old 8 26.7
> 50 years 4 13.3
30 100
3.3%, Russia (8 people) a percentage of 26.7%, Serbia (1 person) a percentage of

3.3%, Singapore (1 person) a percentage of 3.3%, Spain (1 person) a percentage of
3.3%, the UK (2 people) a percentage of 6.7%, USA (1 person) a percentage of 3.3%,
and Welsh (1 person) a percentage of 3.3% (Fig. 4 and Table 5).
Country of Origin 3% 3% 3% 3%
Africa American 3%
Aussie Canada 7% 3%
3%
German Holland 7% 3%
India Italy 3%
Malaysia Polish 7%
French Romanian 3%
Russia Serbian 27%
Singapore Spanish 3%
UK USA 3%
3%
Welsh 7% 3%
Fig. 4 Percentage based on country of origin
Table 5 Respondent
Country name Total Percentage (%)
character based on country of
origin Africa 1 3.3
America 1 3.3
Australia 1 3.3
Canada 1 3.3
German 2 6.7
Holland 2 6.7
India 1 3.3
Italy 1 3.3
Malaysia 1 3.3
Poland 1 3.3
France 2 6.7
Romania 1 3.3
Russia 8 26.7
Serbia 1 3.3
Singapore 1 3.3
Spain 1 3.3
UK 2 6.7
USA 1 3.3
Welsh 1 3.3
30 100
3.4 Based on Marital Status
Based on marital status, there are single (12 people) with a percentage of 40.0%,
married/de facto (13 people) with a percentage of 43.3%, and separated, divorced,
widowed (5 people) with a percentage of 16.7% (Fig. 5 and Table 6).

marital status Marital Status
Single
17%
40%
Married / de facto
43%
Separated, divorced,
widowed
Table 6 Respondent
Marital Status Total Percentage (%)
character based on status
Single 12 40.0
Married/de facto 13 43.3
Separated, divorced, widowed 5 16.7
30 100
3.5 Based on Education
Based on education, there are Senior High School (4 people) with a percentage of
13.3%, completed tertiary education (26 people) with a percentage of 86.7%, while
for Elementary school and Junior High School (0%) (Fig. 6 and Table 7).

education
Education 13%
Senior High School
Completed Tertiary
87%
Education
Table 7 Characteristics
Education Total Percentage (%)
based on education
Senior high school 4 13.3
Completed tertiary education 26 86.7
30 100
10%
Driver's License Ownership
Less than 1 year
1-5 years 33%
6-10 years
11-15 years
44%
16-20 years
3%
More than 20 years
7% 3%
Fig. 7 Based on driver’s license ownership
Driver’s license ownership Total Percentage (%)
respondents based on driver’s
license ownership Less than 1 year 3 10.0
1–5 years 13 43.3
6–10 years 2 6.7
11–15 years 1 3.3
16–20 years 1 3.3
More than 20 years 10 33.3
30 100
3.6 Based on Driver’s License Ownership
Based on Driver’s License ownership, there are those with less than 1-year ownership
(3 people) with a percentage of 10.0%, 1–5 years (13 people) 43.3%, 6–10 years (2
people) 6.7%, 11–15 years (1 person) 3.3%, 16–20 years (1 person) 3.3%, and more
than 20 years (10 people) 33.3% (Fig. 7 and Table 8).
3.7 Based on the Distance Traveled
Based on the distance traveled, there are those who ride less than 20 km (18 people)
with a percentage of 60.0%, 20–49 km (11 people) 36.7%, 50–99 km (1 person)
3.3%, and the riding distance beyond 100 km (0%) (Fig. 8 and Table 9).
Fig. 8 Based on the distance

traveled
3%
Driving Distance
Less than 20 kms
37%
Between 20 to 49
Kms 60%
Between 50 to 99 kms
Riding distance Total Percentage (%)
respondents based on riding
distance traveled Less than 20 km 18 60
Between 20 to 49 km 11 36.7
Between 50 to 99 km 1 3.3
30 100
3.8 Based on the Purpose of Riding
Based on the purpose of riding, there are recreational purposes with 24 people or
a percentage of 80.0%, while for commuting or chain trips were 6 people or a
percentage of 20.0% (Fig. 9 and Table 10).
Fig. 9 Based on the purpose

of riding Purpose of Riding
20%
Recreational
Commuting 80%
Riding purpose Total Percentage (%)
respondents based on riding
purpose Recreational 24 80.0
Commuting 6 20
30 100
4 Conclusion
1. The results shows that the questionnaires distributed to 30 respondents consisted

of 18–25 years (5 people) with a percentage of 16.7%, 26–35 years (7 people)
a percentage of 23.3%, age 36–45 years (6 people) a percentage of 20.0%, age
46–55 years (8 people) a percentage of 26.7% and age > 50 years (4 people) a
percentage of 13.3%.
2. Characteristics of respondents based on status are single (12 people) 40.0%,
married/de facto (13 people) 43.3%, separated, divorced, and widowed (5
people) 16.7%.
3. Based on education, there are Senior High School (4 people) 13.3%, completed
tertiary education (26 people) 86.7%, while for Elementary school and Junior
High School (0%).
4. Based on driver’s license ownership, there are those with less than 1-year owner-
ship (3 people) 10.0%, 1–5 years (13 people) 43.3%, 6–10 years (2 people)
6.7%, 11–15 years (1 person) 3.3%, 16–20 years (1 person) 3.3%, and more
than 20 years (10 people) 33.3%.
5. Based on the riding distance, there are those who drive less than 20 km (18
people) 60.0%, 20–49 km (11 people) 36.7%, 50 - 99 km (1 person) 3.3%, and
the riding distance beyond 100 km (0%).
6. Based on the purpose of riding, there are recreational purposes with 24 people
or a percentage of 80.0%, while to ride for commuting or chain trips, there are
6 people or a percentage of 20.0%.
References
1. Yannis G, Golias J, Papadimitriou E (2007) Accident risk of foreign drivers in various road
environments. J Saf Res 38:471–480
2. Leviäkangas P (1997) Finnra’s VIKING-programme. Finnra internal publications 24/1997.
Finnish National Road Administration, Helsinki
3. Bailey KD (1994) Methods of social research, 4th ed. Free Press, Maxwell Macmillan, Canada
4. BPS (2016) Jumlah Kunjungan Wisatawan Mancanegara Tahun 2016 ke Indonesia menurut
Pintu Masuk. Departemen Pariwisata, Badan Pusat Statistik, Jakarta
Analysis of the Behaviour Model
of Foreign Motorcyclists in Tourism
Areas in Bali

Abstract Foreign motorcyclists are prone to greater risks than the locals in most
countries around the world. Some of the risk parameters that can be identified, such as
lack of knowledge of traffic regulations, inadequate driving skills, different seasons,
and general attitudes towards traffic safety that are reflected in driving behaviour.
Along with the development of tourism, the number of foreign motorcyclists also
increases. Therefore, the safety measures are essential. In this research, we study
the influence of local road and traffic conditions, the influence of local motorcyclists
and law enforcement on the behaviour of foreign motorcyclists around tourism areas
in Bali. The method used is Structural Equation Modeling (SEM). The expected
contribution is a model that finds the dominant factors influencing foreign motor-
cyclists’ behaviour, which can significantly reduce the risk of accidents involving
foreigners. The results are the local road and traffic factors affect foreign motorcy-
clists’ behaviour with a bootstrapping value of 2.426. And local motorcyclists and
law enforcement also affect with a bootstrapping number of 3.874. Both factors posi-
tively affect foreign motorcyclists’ behaviour with a coefficient of 0.217 and 0.371.
The research concluded that the behaviour of foreign motorcyclists in tourist areas in
Bali is highly influenced by the behaviour of local motorcyclists and law enforcement
by authorized officials. It shows that the more stringent enforcement of the law, the
better the behaviour of foreign motorcyclists. Additionally, road and traffic conditions
also affect the behaviour of foreign motorcyclists. The better foreign motorcyclists’
behaviour will reduce the risk of accidents.
C. P. Wirasutama (B)
Mahasaraswati University, Denpasar, Indonesia
P. A. Suthanaya · D. M. P. Wedagama · A. A. G. A. Yana
Udayana University, Denpasar, Indonesia
D. M. P. Wedagama
A. A. G. A. Yana
Keywords Motorcyclist · Foreigner · Behaviour · Model
1 Introduction
People who come to Indonesia come from many countries with varying charac-
teristics, in terms of age, purposes of arrival, education, and cultural background.
These differences in characteristics lead to differences in the choice of transporta-
tion modes in Indonesia. Some of them prefer motorcycle as a mode of transportation.
The behaviour of foreigners who ride motorcycles will be examined in this study.
Foreign riders are prone to greater risks than domestic drivers in most countries
around the world, proven by several empirical findings, including [1]. It is said that
different countries’ traffic cultures largely explain the differences in some specific
technical parameters. Lack of knowledge of the road network, local traffic regulations
and insufficient driving skills in unfamiliar conditions, can lead to an increase in the
number of accidents, severity, and the risks imposed on foreign motorcyclists.
Variations in general attitudes towards road safety which are reflected in riding
behaviour can further contribute to a higher accident risk for foreign motorcyclists
than locals. In this research, we want to study the influence of local road and traffic
conditions, the influence of local motorcyclists, and the influence of law enforcement
on the behaviour of foreign motorcyclists around tourism areas in Bali. The method
used is Structural Equation Modeling (SEM). Research on this field, especially on
the effect of local motorcyclists and law enforcement on the behaviour of foreign
motorcyclists has not been specifically addressed. The expected contribution from
this study results is to provide a model that finds the dominant factors influencing
foreign motorcyclists’ behaviour, which can significantly reduce the risk of accidents
involving foreigners.
The problems to be researched are:
• How the local road and traffic factors influence foreign motorcyclists’ behaviour
in Bali’s tourism areas?
• How does the behaviour of local motorcyclists and law enforcement affect the
foreigner motorcyclists’ behaviour in tourism areas in Bali?
The benefit of this research is that it is expected to help develop corrective
measures in foreign motorcyclists’ behaviour to reduce the rate of motorcycle acci-
dents involving foreign motorcyclists. Expectantly, the solution to reduce accidents
against foreign motorcyclists will not only be ’hard’ solutions such as infrastructure
development or law enforcement but it is also hoped that the solution will be more
directed at ‘soft’ improvements such as how to modify the motorcyclists’ behaviour.
Analysis of the Behaviour Model of Foreign … 731
2 Theory of Planned Behaviour (TPB)
According to the Theory of Planned Behaviour (TPB), human behaviour is guided

by three types of considerations: beliefs about possible consequences of behaviour
(behavioural beliefs), beliefs about the normative expectations of others (norma-
tive beliefs), and beliefs about the existence of factors that can facilitate or hinder
behavioural performance (control beliefs). In their respective proportions, beliefs
result in behaviour that is favorable or unfavorable to that behaviour.
The intention is assumed to be the precursor to behaviour. However, since many
behaviours present difficulties that may limit control over will, it is necessary to
consider feeling control over behaviour other than the only intention. If the feeling
that control over behaviour is sincere, it can serve indirectly as a solid control and
can contribute to the prediction of the behaviour that is being examined [2].
The risk factors associated with injuries related to motor vehicles should be
described and framed appropriately in the socio-cultural context. A thorough under-
standing of socio-cultural factors related to motorcycling is essential to develop
targeted and efficient measures and interventions [3].
Road infrastructures, damages, and the maintenance condition apparently are
partly responsible for the phenomenon of motorcyclists paying less attention to traffic
regulations. According to statistics around the world, the number of traffic fatalities
and injuries is estimated to have increased by about 65% between 2000 and 2020 [4,
5]. The outlook is devastating in low and middle-income countries where the number
of traffic fatalities is estimated to increase by as much as 80% [6]. Road accidents
have become big problems in many countries especially in urban areas [6, 7], and
Indonesia did not escape this trend.
The characteristics and consequences of road accidents that result in death and
injury in developing countries are different from those of high-income countries. In
the United States, 80% of the victims who died from road collisions are motorcyclists
and passengers of four-wheeled motor vehicles [8].
In low-income countries, most victims are pedestrians, cyclists, motorcyclists,
and even passengers on public transportation systems [6].
What tourists do or not do contributes to many of these occurring accidents.
Thus, the study of human behaviour is very important for road safety studies [9,
10]. Human behaviour factor is an important factor that should be considered when
designing strategies to reduce traffic accidents and fatalities.
Furthermore, they also argued that motorcyclists’ perceptions of the road and
traffic environment were correlated with the rider’s age, driving experience, fatigue,
stress, and emotions. While there has been much research of this kind in developed
countries, there has been little research into traffic violations committed by motor-
cyclists in developing countries, where the use of motorcycles has been growing
rapidly recently. Motorcycles hold a large share of the mode of transportation choice
in several Southeast Asian countries, such as Indonesia, Malaysia, Vietnam, and
Thailand [11]. Meanwhile, motorcycles are also the main reason for traffic accidents
pointed out by the researchers. For example, in Indonesia, there are 105.1 million
motorcycles from 129.3 million units of motor vehicles in 2016 [12]. In Vietnam,
80–90% of households in Ho Chi Minh City have access to motorcycles [13].
The presence of motorcycles significantly impacts road performance and accident
rates [14, 15]. It is widely believed that motorcyclists have relatively poor safety
records compared to other groups of road users [16]. Motorcyclists are often seen
as road users who trigger many collisions. Besides, they are seen to be more risk-
taking and to have more speeding tendencies than car drivers [10]. In Indonesia,
motorcyclists have the highest share of recorded accidents [17].
Most of the traffic accidents are related to human factors. Riders regularly deviate
from safe driving conditions for various reasons [18]. Disobedience, driving errors,
and road traffic violations are the main reasons for traffic accidents pointed out by
many researchers such as by [18–22]. O’Connell [23] has stated that many undesir-
able aspects of rider behaviour happened due to intentional or unintentional actions,
which means that the mistakes are committed by themselves and not being forced by
others to make mistakes. A study by [24] in Australia, showed that a motorcyclist’s
intention to high-risk activities is primarily initiated by the desire to seek thrills.
In contrast, the intention of safety is based on behavioural control in the form of
personal discipline.
Many studies have shown that the attitudes of road users can be explored by
studying noncompliance and violations. A detailed discussion related to this topic is
provided by [21, 22, 25–27]. It is said that driver performance is influenced by the
tendency to conduct driving errors and traffic violations, attitudes towards their own
and other road users’ driving style, attitudes towards the vehicles they control, and
personal characteristics [26]. In other words, violations can occur due to the driver’s
attitude towards the behaviour, social norms regarding the action, behaviour control,
and moral norms [27].
3 Hypothesis
The hypotheses of this study are:

• H1: The behaviour of foreign motorcyclists is influenced by local conditions,
which are the road environment and local traffic conditions.
• H2: The behaviour of foreign motorcyclists is also influenced by the behaviour
of local motorcyclists and law enforcement.
• H3: local motorcyclists behaviour factors and law enforcement moderate local
condition factors
• H4: The behaviour factor of local motorcyclists is influenced by local conditions,
which are the road environment and local traffic conditions
4 Methodology
Initial observations were made by observing the behaviour of foreign motorcyclists in

tourist areas in Bali. In this observation, it was found that many foreign motorcyclists
violated the traffic regulations. This phenomenon is what will be examined by initially
determining the causal variables. Then a questionnaire was made to measure these
variables. After the questionnaire was distributed and re-collected, the SEM method is
used to measure reliability, validity and how much these variables affect the behaviour
of foreign motorcyclists in tourist areas in Bali (Fig. 1).
The research will be carried out in tourism destination areas in Bali, which are in
Kuta, Ubud, Sanur, and Canggu. The research was conducted from September 2020
to December 2020.
Fig. 1 Research design

5 Determination of Data Sources
Sources of data used are from collecting field observations or recording of foreign
motorcyclists’ behaviour. Then a survey was conducted using questionnaires to the
respondents selected as samples of foreign motorcyclists. A structured questionnaire
was developed to collect motorcyclists’ information, motorcycle riding behaviour,
and attitudes in tourism areas.
Three types of accidents were examined in this questionnaire, which are minor
accidents, serious injury accidents and fatal accidents. The target number of respon-
dents is based on [28], who recommends a sample size of at least 10 times the number
of research variables for research using multivariate analysis. The variables in this
study are 3, so the sample required is 30 respondents. If the respondent does not
want to write, the surveyor asks the question as a face-to-face interview. This implies
that the sample selection process may not be completely random as respondents are
approached based on respondents’ availability. The questionnaire was given only to
respondents who claimed to be motorcyclists. The data sources in this study were as
follows:
1. Primary data is data obtained directly from primary data sources. The primary
data source in this study is the result of direct field observations on foreign
motorcyclists’ behaviour (field observation). And the results of questionnaires
for foreign motorcyclists are also the primary data source.
2. Secondary data is obtained from secondary data sources, which are not directly
obtained by researchers, namely through related journals, books, and archives
related to the discussions. It can be used as guidelines or references in this study,
including data on Bali’s foreign visits (Table 1).
Table 1 Data, data types and data sources

Data Type of data Data source
Primary Secondary
Photos of foreign motorcyclists Qualitative Observation
Theory and conceptual basis Qualitative Literature
– Theory of planned behaviour Quantitative
– Structural equation modelling (SEM)
Research Question 1
The influence of local road and traffic factors Quantitative Questionnaire Literature
Research Question 3
The influence of local motorcyclists and law Quantitative Questionnaire Literature
enforcement factors
Fig. 2 Research variables
6 Research Variables
The variables used in this study were obtained from previous studies related to
this research. The variables of Local Motorcyclist Behaviour (LB) and Local Law
Enforcement (LE) are variables that are augmented by the hypothesis that the
behaviour of foreign motorcyclists is influenced by local motorcyclists and local
law enforcement (Fig. 2).
7 Findings
Outer Model Evaluation

The outer model’s assessment is carried out to determine the validity and reliability
of the indicators and research latent variables. Validity is known by using the values
Table 2 Convergent validity

Outer loading AVE
test
LB 0.741 0.731
LE 0.956
RC 0.921 0.804
TC 0.872
MC 0.981 0.981
IC 0.996
FC 0.994
of convergent validity and discriminant validity. Meanwhile, reliability is known by

using the indicator reliability value and value of internal consistency reliability.
7.1 Validity Test
The validity test can be seen from the results of convergent validity and discriminant
validity.
Convergent Validity Test
The convergent validity of the model is known from the loading factor value and
the AVE (average variance extracted) value. Table 2 shows that the loading factor
value is between 0.741 and 1.172, which means that it exceeds 0.5 according to the
recommendations of [29]. This figure shows that all measurement indicators meet
the requirements of convergent validity, and each indicator is valid in measuring the
variable in question.
Table 2 also shows the AVE construct values between 0.731 and 0.981. This value
is greater than 0.5 [29] which means that the construct in this study can measure the
corresponding latent variable.
Discriminant Validity
To calculate data validity, discriminant validity was used through the value of cross-
loadings listed in Table 3. This indicator has a high correlation with its latent vari-
able compared to other latent variables. This shows that all valid indicators and
discriminant validity are met.
7.2 Reliability Test
The reliability value shows the consistency of the results even though the measure-
ments are repeated. Therefore, reliability can be interpreted as a calculation that is free
Table 3 Cross loading

LB RT Y
LB 0.741 0.053 0.134
LE 0.956 0.227 0.204
RC 0.161 0.921 0.249
TC 0.192 0.872 0.143
MC 0.160 0.228 0.981
IC 0.229 0.222 0.996
FC 0.217 0.220 0.994
LB Local Motorcyclist Behaviour, LE Local Law Enforcement, RC
Road Conditions, TC Local Traffic Conditions, MC Minor Crash,
IC Injury Crash, FC Fatal Crash
Table 4 Reliability Test

Cronbach’s alpha Composite reliability
LR 0.776 0843
RT 0.759 0.891
FB 0.991 0.994
LR Local Motorcyclist Factor, RT Local Road and Traffic Condi-
tions Factor, FB Foreign Motorcyclist’s Behaviour
from random errors [30]. Testing reliability is calculated using PLS through internal
consistency reliability. For internal consistency reliability, the value of composite
reliability and Cronbach’s alpha must be higher or equal to 0.7. Table 4 shows that
the results of composite reliability and Cronbach’s alpha are greater than 0.7, which
means that these latent variables have consistent and reliable results.
7.3 Evaluation of the Inner Model (Structural Model)
This evaluation step is carried out to confirm the theoretical model outlined in
the research structural model [31]. Here, the structural evaluation model uses two
indicators: R2 and predictive relevance (Q2) [31].
R2 value
The R2 value in this PLS calculation represents the amount of variance contained in
the model [31]. The R2 value in this study is listed in Table 5 and Fig. 3.
From Table 5 above it is known that:
• The LR variable (local motorcyclists behaviour) can be explained by the LB
variable (behaviour of local motorcyclists), LE (local law enforcement), and RT
(environmental condition factors for roads and local traffic) by 13.30% while the
Table 5 R square
Variable R square
LB
LE
RT
LR 0.133
FB 0.283
Fig. 3 PLS ALOG
remaining 86.70% is the contribution of other variables that are not included in
this model.
• The FB variable (foreign motorcyclists behaviour) of 0.283 or 28.30% indicates
that the behaviour of foreign motorcyclists (FB) is explained by the variables of
RT (road conditions and local traffic factors) and LR (local motorcyclists) with
28.30%. While the remaining 71.70% is the contribution of other variables not
included in this model.
Predictive Relevance
The Q2 predictive relevance value in a structural model is used to measure how well
the observation value generated by the model and also its parameter estimation.
The value of Q2 in this study is calculated using the following formula:
Q2 = 1 − (1 − R12)(1 − R22)
= 1 − (1 − 0.133)(1 − 0.283)
= 0.376
This value shows in general the contribution of the variables of road conditions
and local traffic (RT), as well as the behaviour of local motorcyclists (LR), to the
variable of the behaviour of foreign motorcyclists (FB) was 37.60%. The remaining
62.40% is the contribution of other variables that are not included in this model.
7.4 Summary of the Outer Model and Structural Model

Evaluation
From the results of the outer model evaluation in this study, it is known that the
constructs used in this study are valid and reliable so that the evaluation of the
structural model can be continued. From the results of the structural model evaluation,
it is also known that the proposed research model is appropriate, and hypothesis
testing can be carried out.
The results of this analysis indicate that the behaviour of foreign motorcyclists
in tourist areas in Bali is strongly influenced, especially by the behaviour of local
motorcyclists and law enforcement by authorized officials. It shows that the better
the behaviour of local motorcyclists, the better the behaviour of foreign motorcy-
clists, and vice versa. This also applies to law enforcement against traffic violations.
Wherein the more stringent enforcement of the law, the better the behaviour of foreign
motorcyclists. Road and traffic conditions also affect the traffic behaviour of foreign
motorcyclists. The better the road conditions and the more organized traffic will
result in better foreign motorcyclists’ behaviour too.
8 Hypothesis Testing
Hypothesis testing is done to evaluate the relationship between latent variables,

whether it is significant or not. Hypothesis testing can be calculated from the results
of the path coefficient and the significance of the model based on the T-statistic value.
The path coefficient value is significant if the T-statistic value is greater than 1.96.
The results of this study indicate that RT has a significant positive relation to FB
(t-value 2.426). This significant result means that the better the local road and traffic
conditions, the better the foreign motorcyclists’ behaviour is. These results support
Hypothesis 1.
The results of this study also indicate that LR has a significant positive relation with
FB (t-value 3.874), which means that the better the behaviour of local motorcyclists,
the better the foreign motorcyclists’ behaviour is. Hypothesis 2 is proven.
In contrast to the RT and LR variables as latent variables, the results of this study
indicate that the LR variable does not moderate the RT variable (t-value 0.455).
This means that the behaviour of local motorbike riders (LR) does not play a role
in improving the road conditions and local traffic (RT) factors on their effect on FB
(behaviour of foreign motorbike riders). Thus hypothesis 3 is not proven.
Fig. 4 Bootstrapping
Furthermore, this study found that the RT variable (road conditions and local traffic
factors) did not have a significant relation with LR (local motorcyclists’ behaviour),
with a t-value of 1.339. Therefore, hypothesis 4 is not proven (Fig. 4).
9 Conclusion
From the results of the study above, it can be concluded that the local road and
traffic factors influence the behaviour of foreign motorcyclists with a bootstrapping
value of 2.426. Influence from local motorcyclists and law enforcement also affects
foreign motorcyclists’ behaviour with a bootstrapping number of 3.874. Both factors
positively affect foreign motorcyclists’ behaviour with a coefficient of 0.217 and
0.371. In conclusion, the behaviour of foreign motorcyclists in tourist areas in Bali
is very much influenced, especially by the behaviour of local motorcyclists and law
enforcement by authorized officials. This also applies to law enforcement against
traffic violations. Wherein the more stringent enforcement of the law, the better the
behaviour of foreign motorcyclists. Road and traffic conditions also affect the traffic
behaviour of foreign motorcyclists. Where the better the road conditions and the
more regular traffic will result in the better foreign motorcyclists’ behaviour too.
Better foreign motorcyclists’ behaviour will reduce the risk of accidents for foreign
motorcyclists, passengers, and other traffic users.
References
1. Yannis G, Golias J, Papadimitriou E (2007) Accident risk of foreign drivers in various road
environments. J Safety Res 38:471–480. https://doi.org/10.1016/j.jsr.2007.01.014
2. Ajzen I (1991) The theory of planned behavior. Organ Behav Hum Decis Process 50:179–211.
https://doi.org/10.1016/0749-5978(91)90020-t
3. Njå O, Nesvåg SM (2007) Traffic behaviour among adolescents using mopeds and light
motorcycles. J Safety Res 38:481–492. https://doi.org/10.1016/j.jsr.2007.03.012
4. Murray CJL, Lopez AD (1996) The global burden of disease, a comprehensive assessment of
mortality and disability from diseases, injuries, and risk factors in 1990 and projected to 2020.
Harvard School of Public Health. https://doi.org/10.1016/s0140-6736(96)07495-8
5. Kopits E, Cropper M (2003) Traffic fatalities and economic growth. Policy Res. Work. Pap.
6. Peden M, Scurfield R, Sleet D, et al (2004) World report on road traffic injury prevention.
Geneva
7. Zietlow DG (2006) The road safety cent-management and financing of road safety in low-
income countries
8. TRB (2006) Improving road safety in developing countries. Opportunities for US Cooperation
and Engagement
9. Fuller R, Santos JA (2002) Psychology and the highway engineer. Hum Factors Highw Eng,
1–10
10. Musselwhite C, Avineri E, Susilo Y (2013) Legitimising risk taking: articulating dangerous
behaviour on the road. Transp Plan Technol 37:62–82. https://doi.org/10.1080/03081060.2013.
844905
11. Kaltheier RM (2002) Urban transport and poverty in developing countries : analysis and
options for transport policy and planning. Dtsch Gesellschaft für Tech Zusammenarbeit GmbH,
Eschborn 49
12. BPS (2016) Number of motorized vehicle 1987–2009. Jakarta
13. PADECO (1986) Study on urban transport development, p 183
14. Department of the Environment T and the R (2000) Tomorrow’s roads: safer for everyone, pp
1–7
15. Hossain M, Iamtrakul P (2007) Medical investigation of motorcycle accidents in. J East Asia
Soc Transp Stud 7:2770–2785
16. Clarke DD, Ward P, Bartle C, Truman W (2004) Road safety research report No. 54 : In Depth
Study of Motorcycle Accidents In-depth Study of Motorcycle Accidents, pp 1–67
17. Indriastuti AK, Sulistio H (2010) Influencing factors on motorcycle accident in urban area of
Malang, Indonesia. Int J Acad Res 2:252–255
18. Rothengatter T (1998) Errors and violations as factors in accident causation, in Traffic and
transport psychology. Theory and application. Rech - Transp - Sécurité 58:70–71. https://doi.
org/10.1016/s0761-8980(98)80022-8
19. Yagil (1998) Gender and age-related differences in attitudes toward traffic laws and traffic
violations. Transp Res Part F 1
20. Rimmö PA, Åberg L (1999) On the distinction between violations and errors: sensation seeking
associations. Transp Res Part F Traffic Psychol Behav 2:151–166. https://doi.org/10.1016/
S1369-8478(99)00013-3
21. Forward S (2006) The intention to commit driving violations—a qualitative study. Transporta-
tion. Transp Res Part F 9
22. Forward S (2009) An assessment of what motivates road violations. Transp Res Part F 12
23. O’Connell M (2008) Social psychological principles: the group inside the person. In: Fuller R,
Santos JA (eds) Human factors for highway engineers. Emerald, Bingley
24. Watson B, Tunnicliff D, White K et al (2007) Psychological and social factors influencing
motorcycle rider intentions and behaviour. ATSB Res Anal Rep 152
25. Rothengatter T (1997) Errors and violations as factors in accident causation. In Rothengatter
T, Vaya EC (eds) Traffic and transport psychology: theory and application. Elsevier Science
Ltd, Oxford
26. Underwood G, Chapman P, Wright S, Crundall D (1997) Estimating accident liability.
In Rothengatter, T. and Vaya, E.C. (Eds.) Traffic and Transport Psychology: Theory and
Application. Elsevier Science Ltd, Oxford (1997).
27. Rothengatter T, Manstead AR (1997) The role of subjective norm in predicting the intention to
commit traffic violations. In: Rothengatter T, Vaya EC (eds) Traffic dan transport psychology:
theory and application. Pergamon, Oxford
28. Roscoe (1982) Research methods for business. McGraw Hill, New York
29. Fornell C, Larcker DF (1981) Evaluating structural equation models with unobservable
variables and measurement error. J Mark Res 18:39. https://doi.org/10.2307/3151312
30. Malhotra NK, Peterson M (2006) Basic marketing research: a decision-making approach.
Prentice Hall, New Jersey
31. Chin WW (2010) How to write up and report PLS analyses
Knowledge and Practice of Helmet Usage
Among Senior High School Students
in Klaten Regency
Dewi Handayani, Putri Dewi Prasetianingrum,

and H. M. Amirotul Musthofiah
Abstract Indonesia is one of the countries with the most significant number of
motorcycle users in Southeast Asia. Because of the heavy traffic and high price of
cars, most Indonesians choose to use two-wheeled vehicles. The accidents in Klaten
Regency in 2019 to early 2020 increased by 30% from 2018 and were dominated by
teenage riders (16–25 years old), with one of the leading causes, driving behavior.
This research aims to obtain the correlation between the characteristics of respondents
and their practice in using helmets—only four characteristics from 9 characteristics
related to the habit of using a helmet. One of the characteristics is the sanctions if
the driver does not use the helmet correctly while driving. The results of the data
analysis obtained no correlation between knowledge about helmets and the practice
of helmet usage while driving. This research has never been done in detail by linking
the respondent’s characteristics with the habit of respondents in using a helmet.
Therefore, it is worth doing further research. This research applied to the proportion
of respondents by 75% female & 25% male senior high school students in Klaten
Regency.
Keywords Chi-square test · Cross tabulation · Helmet USAGE · Knowledge

practice attitude (KPA)
1 Introduction
Indonesia is one of the countries with the most significant number of motorcycle
users in Southeast Asia. This is due to heavy traffic and high car prices, making
most Indonesians use two-wheeled vehicles. In 2018 the Central Bureau of Statistics
D. Handayani · P. D. Prasetianingrum (B) · H. M. Amirotul Musthofiah

Civil Engineering, Sebelas Maret University, Surakarta, Indonesia
D. Handayani
H. M. Amirotul Musthofiah
744 D. Handayani et al.
released data related to the number of motorcycles registered with the Indonesian
police numbered 120 million units and data related to accidents in Indonesia that
resulted in fatalities of 19.016 cases with a death toll of 4.115 people, severe injuries
of 97 people, and minor injuries of 21.967 people with 74.49% involvement of motor-
cycles in accidents. A helmet is a mandatory attribute when driving a motorcycle.
It is intended to protect the head from impacts in a traffic accident to avoid the user
from minor to severe head injuries. In 2018 the World Health Organization (WHO)
stated that proper use of helmets could reduce 42% of the risk of fatal traffic acci-
dents and 69% can protect against head injuries. Therefore, research is needed on
compliance studies for senior high school students in Klaten Regency to know the
level of knowledge about helmets and the habits in using a helmet that is expected to
minimize the risk in the traffic accident and know whether or not there is a correlation
between the habits of using a helmet and the characteristics of respondents.
This research was conducted basically to see the correlation between the knowl-
edge level of helmets and the habit of using helmets of senior high school students in
the Klaten Regency. Similar to previous researches that used many interview methods
and shared questionnaires in data collection as conducted by Rizky [1], Ambak and
Hashim [2], Guntur [3], Pramono et al. [4], and Fadilah et al. [5], the data collection
in this research was also conducted by sharing questionnaires online on google-form
because the research was conducted when Indonesia and the world were affected
by COVID-19 pandemic. To get the value of the correlation between knowledge
about helmets and the habit of using helmets while driving cross-tabulation method
is used. The specialty of the crosstab method is the ability of the method to analyze
the correlation between rows and columns. This method has been used in previous
research, but the previous research didn’t link the characteristics of respondents and
their habit in using helmets. So this research is expected can provide good results in
transportation science.
Previous researchers have used cross-sectional methods and univariate-bivariate
correlation analysis. Cross sectional methods conducted by Iskandar [6], Muryatma
[7], Widiyawati [8], Dwinanda and Wijaya [9] and Fadilah et al. [5] as well as
univariate-bivariate correlation methods have been conducted by Ambak and Hashim
[2], Pramono et al. [4], Widyastuti and Trisnawati [10], and Putri et al. [11]. From
previous researches with significant subjects in the adolescent age group, the majority
only examined in general such as the relationship of knowledge with driving safety,
the relationship of knowledge & attitude of using helmets with driving safety, driving
behavior, compliance factors in driving, and the relationship of safety factors with
safety riding. From this research, more in-depth research can be done on the charac-
teristics of drivers as well as analysis of knowledge about helmets and their habit in
using a helmet when driving a motorcycle, especially on senior high school students
in Klaten Regency to know the relationship between knowledge about helmets and
their habit in using a helmet that expected to minimize the risk of traffic accident.
Knowledge and Practice of Helmet Usage … 745
2 Basic Theory
Safety riding means putting safety first, in personal safety or the safety of other road
users. Drivers should never forget that the road is not our own private but rather a
shared public facility. Thus, carelessness can cause others to be wretched [12]. Based
on this idea, the drivers have to be aware of the importance of driving safety. The
application of safety riding has been regulated in Law No. 22 of 2009 concerning
Traffic and Road Transport in Chapter XI Article 203 Paragraph (2), which reads
“To ensure Traffic Safety and Road Traffic Control as referred to in paragraph (1), a
national general plan of Traffic Safety and Road Transport is established, including
the preparation of a national program of Traffic Safety and Road Transport activities.”
Referring to the paragraph, the National Traffic Safety and Road Transport program
is among others about the procedures of driving safety. Therefore, the government
recommends the use of safety riding equipment, one of which is a helmet. The
government made a regulation written in Law No. 22 of 2009 on Traffic and Road
Transport Article 106 paragraph (8) to protect the public from potential head injuries
in the accident. To discipline the public to use a helmet, the government also issued
criminal laws contained in Law No. 22 of 2009 on Traffic and Road Transport Article
291 paragraphs (1) and (2), i.e.:
Every person who drives a motorcycle that does not wear an Indonesian National Standard
helmet as referred to Article 106 paragraph (8) shall be penalized with a maximum imprison-
ment of 1 (one) month or a maximum fine of Rp 250,000.- (two hundred and fifty thousand
rupiahs).
Every person who drives a motorcycle that allows his passengers not to wear a helmet as
referred to Article 106 paragraph (8) shall be penalized with a maximum imprisonment of 1
(one) month or a maximum fine of Rp 250,000.- (two hundred and fifty thousand rupiahs).
KPA (Knowledge, Practice, Attitude) survey method is one of the methods to know
precisely about education and knowledge, the tendency of attitudes and practices
of respondents [13]. This research used the KPA survey method to determine the
behavior of current drivers and set future targets by making policies that are expected
to change the behavior of motorcyclists to be safer when driving.
3 Methods
Before conducting the research, problem identification & literature study were carried
out to get an overview of the research. To support the validity of the data, secondary
data was collected from local government that can be accessed through the website.
Then the survey form and the pilot survey form were compiled. The survey results
were analyzed using reliability and validity tests to obtain a suitable survey form
for further research. Then the primary survey was conducted after determining the
number of samples to be taken. Finally, the data obtained from the primary survey
were analyzed using cross-tabulation analysis so that the conclusion could be taken
from the research.
The research was conducted in several of Klaten Regency’s senior high schools,
especially SMAN 1 Jogonalan, SMAN 1 Klaten, SMAN 2 Klaten, and SMAN 3
Klaten. The samples count in this research was taken based on Krejcie and Morgan’s
table (1970) with the results of 266 respondents [14]. Online sampling techniques
are assisted by google-form that contains questions related to respondent’s charac-
teristics, knowledge of helmets, and practices of helmet usage while driving. The
data analysis method is done by cross-tabulation method and chi-square statistical
test to see a correlation between variables with the Statistical Product and Service
Solutions (SPSS) tool. Cross-tabulation is a method of analysis by presenting data
with two or more variables into one matrix presented in rows and columns [15]. The
Chi-square test used is as follows:
( f 0 − f h )2
x2 = (1)
fh
where:
x 2 = Chi-square
f 0 = observed frequency
f h = expected frequency.
Hypothesis tests conducted are
H0 = no correlation between rows and columns
H1 = there is a correlation between rows and columns.
Or
If x 2 count < 0.05, then the data does not support rejecting H0 .
If x 2 count > 0.05, then the data support to reject H0 .
Analysis of respondent’s characteristics is one of the essential things in research.

By analyzing the respondent’s characteristics, researchers can see the characteristics
that can affect the respondent’s habit of using helmets while driving. In this research,
respondent’s characteristics include gender, age, ownership of the driving license,
knowledge of the obligation to use Indonesian Standard Nasional helmet, knowl-
edge of sanctions if the driver does not use the helmet while driving, knowledge of
sanctions if the driver let the passenger not use the helmet properly while driving,
knowledge of Indonesia Standard National helmets, frequency of correctly angled
helmet, and ensuring passengers use helmet correctly. Then, all of these characteris-
tics are correlated with the respondent’s habits in using helmets while driving using
cross-tabulation and chi-square tests. These correlations can be seen in Tables 1, 2,

and 3.
Table 1 shows that the respondent’s characteristics, namely gender, age, and
ownership of the driving license, have a chi-square value of more than 0.05, which
is no correlation or relationship with the respondent’s habit in using helmets.
In the category of knowledge about helmets are made several categories to facil-
itate grouping the level of respondents’ knowledge, such as sound, sufficient, and
less, by summing the value of statements in questionnaires. As in the question of
the obligation to use a helmet, respondents who had a score of 3 to 4 had “good”
knowledge, a value of 2 had “sufficient” knowledge, and a value of only 1 had “less”
knowledge. Table 2 shows that only 3 out of 4 characteristics of respondents who
have a chi-square value below 0.05 which correlates with the habit of using helmets,
namely knowledge about sanctions if not using a helmet when driving both drivers
and passengers and knowledge about Indonesian Standard National helmets.
Table 3 shows that one of the characteristics of respondents has a chi-square value
below 0.05 which is the habit of ensuring passengers use the helmet correctly, which
means that characteristic correlates with the respondent’s habit in using a helmet.
After analyzing the relationship between characteristics, an analysis of the rela-
tionship between knowledge about the helmet and the habit of using helmets while
driving, in this variable, the categorization of variables becomes “good,” “sufficient,”
and “less” derived from the recapitulation of the values of each segment in a single
variable. The results of the analysis can be seen in Table 4.
Table 4 showed that the results of 20% of respondents who practiced “less” in using
helmets while driving had “good” knowledge about helmets. Still, on the contrary,
1.74% of respondents who practiced “good” in using helmets while driving had “less”
knowledge about helmets. The statistical chi-square test showed the value of 0,611
Table 1 Correlation between characteristics and respondent’s habits of using helmet

Variables Habits in using helmet Chi-square
Always Sometimes Never Total
n % n % n % n %
Gender Girls 138 69.0 60 30.0 2 1.0 200 100.0 0.373
Boys 41 62.1 25 37.9 0 0.0 66 100.0
Age 14 3 50.0 3 50.0 0 0.0 6 100.0 0.248
15 35 60.3 21 36.2 2 3.5 58 100.0
16 55 66.3 28 33.7 0 0.0 83 100.0
17 73 73.7 26 26.3 0 0.0 99 100.0
18 10 58.8 7 41.2 0 0.0 17 100.0
19 3 100.0 0 0.0 0 0.0 3 100.0
Ownership Have 46 74.2 16 25.8 0 0.0 62 100.0 0.344
of driving Don’t 133 65.2 69 33.8 2 1.0 204 100.0
license have
Table 2 Correlation between knowledge regarding helmet and respondent’s habit in using helmet
Variable Habits in using helmet Chi-square
n % n % n % n %
Knowledge Good 82 64.5 34 28.8 2 1.7 118 100.0 0.387
of the Sufficient 42 61.7 26 38.2 0 0.0 68 100.0
obligation
to use Less 55 68.8 25 31.3 0 0.0 80 100.0
Indonesian
Standard
National
Helmet
Knowledge Know 109 75.2 35 24.1 1 0.7 145 100.0 0.011
of sanctions Don’t 70 57.9 50 41.3 1 0.8 121 100.0
if the driver know
doesn’t use
helmet
Knowledge Know 83 79.1 22 20.9 0 0.0 105 100.0 0.003
of sanctions Don’t 96 59.6 63 39.1 2 1.2 161 100.0
if the driver know
allows the
passenger
not to use
the helmet
properly
Knowledge Good 165 69.6 71 29.3 1 0.4 237 100.0 0.000
of Sufficient 12 50.0 12 50.0 0 0.0 24 100.0
Indonesian
Standard Less 2 40.0 2 40.0 1 20.0 5 100.0
National
helmet
Table 3 Correlation between practice of helmet usage and respondent’s habits of using helmet
Variable Habits in using helmet Chi-square
n % n % n % n %
Frequency Always 129 73.3 47 26.7 0 0.0 176 100.0 0.179
of angled Sometimes 18 50.0 18 50.0 0 0.0 36 100.0
helmet
correctly Never 32 59.3 20 37.0 2 3.7 54 100.0
Ensuring Always 134 80.7 32 19.3 0 0.0 166 100.0 0.000

passengers Sometimes 19 48.7 20 51.3 0 0.0 39 100.0
using
helmets Never 26 42.6 33 54.1 2 3.3 61 100.0
correctly
Table 4 Cross-Tabulation between respondent’s knowledge regarding helmet and respondent’s

practice of using helmet while driving
Variable Respondent’s practice of using helmet Chi-square
Good Sufficient Less
n % n % n %
Respondent’s Good 69 40.12 28 31.46 1 20.00 0.611
knowledge Sufficient 100 58.14 59 66.29 4 80.00
regarding
helmet Less 3 1.74 2 2.25 0 0.00
Total 172 100.0 89 100.0 5 100.0
(> 0.05) so that H0 was received, which means there is no correlation between the
respondent’s knowledge of helmets and the respondent’s practice of using helmets
while driving.
This research is in line with the research of Nova (2017), which states that there
is no correlation between knowledge and driving behavior in 2nd-grade students of
Vocational High School 4 Surabaya with a p-value of 1.921 and research by Novita
(2018), which states that there is no correlation between the level of knowledge
and safety riding behavior in online motorcycle and taxi drivers in Surakarta with a
p-value of 0.507.
5 Conclusion
Based on the results of research on the study of helmet use compliance among
Senior High School students in Klaten Regency with the cross-tabulation method
and chi-square test, the following conclusions can be drawn:
1. The results of the analysis of the relationship between habits in using a helmet
and the characteristics of respondents were found that only 4 out of 9 charac-
teristics correlated, namely knowledge about sanctions if not using helmet both
the driver and passenger, knowledge of Indonesian Standard National helmets,
and the habits of the driver in ensuring his passengers use the helmet correctly.
2. From the results of the analysis of the relationship between knowledge regarding
helmets and the practice of helmet usage, there is no correlation between those
variables.
Recommendations that can be given so the further research can give better
results are:
1. From the conclusions can be recommended the material for further socialization
on driving safety guidance is emphasized more on age, ownership of the driving
license, the reason why it is mandatory to use a helmet while driving, and how
to use helmets correctly.
2. This research was conducted in the COVID-19 pandemic, and data collection
was collected with the google-form tool. However, it is possible if the data
collection results will be ineffective if respondents do not understand the intent
of the questions asked. Therefore, it is recommended to conduct further research
under normal conditions with a questionnaire/live interview method.
3. In the following research, some respondents who know helmet regulation while
driving can be further followed up or analyzed in more detail.
References
1. Rizky EDE (2019) Analisa Perilaku Pengendara Sepeda Motor Tanpa Helm pad Ruas
Palembang-Banyuasin
2. Ambak K, Hashim H, Yousff I, David B (2006) An evaluation on the compliance to safety
helmet usage among motorcyclists in Batu: an evaluation on the compliance to safety helmet
usage among motorcyclists In Batu Pahat, Johor. Int J Integr Eng (Issue on Civil and Environ
Eng) 2(October):45–51
3. Guntur M (2015) Studi Perilaku Pengendara Sepeda Motor Di Kota Makassar. Retrieved from
https://core.ac.uk/download/pdf/77620427.pdf
4. Pramono TD, Subekti AT, Atmoko D (2017) Hubungan pengetahuan dan sikap dengan praktik
keselamatan berkendara sepeda motor pada siswa smk semesta bumiayu. Jurnal Ilmu Kesehatan
(JIK) Bhamada 9(1):8
5. Fadilah D, Supriyanto, Ginanjar R (2018) Faktor-Faktor Yang Berhubungan Dengan Perilaku
Pemakaian Helm Pengendara Sepeda Motor Pada Pelajar Kelas X ( Sepuluh). Promotor: Jurnal
Mahasiswa Kesehatan Masyarakat 1(1):29–36
6. Iskandar NM (2015) Keselamatan Berkendara Sepeda Motor Siswa Smkn 3 Pekalongan Tahun
Ajaran 2015/2016
7. Muryatma NM (2017) Hubungan Antara Faktor Keselamatan Berkendara Dengan Perilaku
Keselamatan Berkendara. Jurnal PROMKES 5(2):155
8. Widiyawati N (2018) Hubungan Tingkat Pengetahuan dengan Perilaku Safety Riding pada
Pengendara Ojek Online di Surakarta
9. Dwinanda D, Wijaya O (2019) Hubungan Antara Pengetahuan Milenial Road Safety, Sikap
Berkendara Dan Fasilitas Berkendara Dengan Perilaku Berkendara Pada Mahasiswa FKM
UAD Kota Yogyakarta
10. Widyastuti SD, Trisnawati N (2017) Gambaran pengetahuan dan sikap tentang keselamatan
berkendara (safety riding) pada mahasiswa STIKES Indramayu. J Chem Inf Model 53(9):1689–
1699
11. Putri EPP, Nurhidayat S, Sukamto FI (2019) Hubungan pengetahuan dengan sikap remaja
dalam keselamatan berkendara. In: 1st seminar nasional and call for paper, pp 176–183
12. Kusmagi MA (2019) Pengertian safety riding. Retrieved from https://pengertianmenurutpara
ahli.org/pengertian-safety-riding/. Last accessed 3–4 June 2020
13. Yapeka R (2018) Pelaksanaan Survei KAP Desa Braja Yekti Tahun 2018. Retrieved from
https://yapeka.or.id/survei-kap-desa-braja-yekti-tahun-2018/. Last accessed 5 June 2020
14. Basri S (2012) Cara Menentukan Jumlah Sampel Penelitian Menggunakan Tabel dan
Berdasarkan Taraf Keyakinan Penelitian. Retrieved from https://setabasri01.blogspot.com/
2012/04/cara-menentukan-sampel-penelitian.html?m=1. Last accessed 9 June 2020
15. Savi (2018) Analisis Crosstab. Retrieved from https://savinotes.wordpress.com/2018/05/22/
analisis-crosstab/. Last accessed 3 June 2020
Sustainability Aspects in Transportation
Engineering
Investigation of CO2 Emissions on Two
Local Streets by Means of IPCC
and Direct Method
Florentina Pungky Pramesti, Annisa Tri Utami, Iva Yenis Septiariva,

and Fajar Sri Handayani
Abstract One of the most widely used modes of transportation is passenger cars.
The increasing use of passenger cars can have an impact on the environment. The
residue from combustion in a vehicle engine can cause CO2 emissions. Excessive
CO2 Emissions leads to greenhouse gas emissions. The purpose of this study is to
calculate the CO2 emissions produced by passenger cars using the Tier 1 of IPCC
method and a so-called direct method, then comparing the two results afterwards.
The case study in this research is two local roads with an average annual daily traffic
of 3305 and 2673 vehicles/day. The IPCC method analyzes emissions based on the
vehicle’s fuel consumption, while the direct method identifies and analyzes emissions
based on the measurement of actual gases emitted by the passenger cars based on
its production year. Exhaust gas emission data is obtained from an emission survey
conducted by the local Environmental Agency. The calculation of emissions using
the direct method on both roads gave higher results than the IPCC method. This is
because the IPCC method uses the tier 1 approach while the direct method is closer
to the tier 3 approach.
Keywords CO2 emissions · Passenger cars · IPCC method · Direct method
F. P. Pramesti (B) · A. T. Utami · I. Y. Septiariva · F. S. Handayani

Civil Engineering Department, Faculty of Enginering, Universitas Sebelas Maret,
I. Y. Septiariva
F. S. Handayani
754 F. P. Pramesti et al.
1 Introduction
As the economy grows, it stimulates the opportunity to invest more in many other
sectors including transportation infrastructure. Vice versa, investments in infras-
tructure are prescribed to trigger economical growth. In the short term, it creates
construction work employment and in the long term it generates a more efficient and
reliable transportation system, which later will drive the growth of the economy.
Road infrastructure has developed very well in Indonesia. It has lowered the
cost of moving people and goods, which in turn, increases economic productivity.
Improved productivity leads to higher prosperity, marked by the growing ownership
of a private car. Nowadays it can be seen that road transportation is dominated by
the use of private vehicles. Typical light passenger vehicles nowadays are commonly
seen operating in city street networks as people enjoy driving within and between
cities.
On the other hand, an increase in the number of light vehicles may have a nega-
tive impact on the environment. The combustion engine emits exhaust gases such
as carbon monoxide (CO), carbon dioxide (CO2 ), nitrogen oxides (NOx), sulphur
dioxides (SOx) and hydrocarbons (HC) that may also lead to public health issues.
Nurdjanah [1] estimated that the average CO2 emission on 14 streets of Denpasar,
Bali, in 2012 was as much as 245.08 kg/hour.km [1]. This amount was calculated
using 2011 annual average daily traffic (AADT) and based on a 2012 fuel usage
survey as well as local vehicle emission factors. CO2 emissions were predominantly
contributed by light passenger cars and motorbikes. Muziansyah [2] reported that the
largest annual emission of CO2 on the 0.2 km long Pasar Bawah Ramayana Street
terminal section in the city of Bandar Lampung reached 12.46 tons [2]. Due to the
enclosed terminal building located at the ground floor of the central business district,
the pollutant may affect the people working in this packed environment.
This ever-increasing greenhouse gas (GHG) may hamper public health directly.
On a larger scale, this situation may contribute to the global warming and climate
change. Yuliana [3] investigated the level of GHG emitted in Indramayu [3], where
the impact attributed to climate change was profound, such as the local fishermen’s
inability to catch fish consistently due to disrupted seasonal patterns, the difficulty
of local traditional farmers in maintaining rice production as unusual pests and other
disturbing organisms attack their paddy field, and the rising sea level which disrupts
socio-economic activity through floods.
Identification and estimation of the amount of GHG emissions is one of the first
step to take in order to mitigate the problem and propose strategies for sustainable
development.
One of the common methods used to identify and estimate the GHG emissions is
the IPCC method [4–7]. The IPCC method was developed by the Intergovernmental
Panel on Climate Change more than 10 years ago by using a database of emission
factors on each activity. Although frequently updated, many reports have raised
questions about the accuracy of this method, especially when it is used in certain
Investigation of CO2 Emissions on Two Local Streets … 755
conditions where traffic and vehicle engine maintenance are not ideal [5]. This non-
ideal situation may lead to identification errors, which in turn can lead to inaccurate
emission analysis results.
Furthermore, recent development in science has led environmental researchers to
develop new methods for measuring and estimating emissions [5, 8, 9]. More data has
been collected from field observation. This data collection is stored and maintained
by local municipal environmental agencies in Indonesia. However, data analysis has
not yet been carried out to provide broader dan deeper insight that in turn will provide
the authority with science base policy making. This situation leads to some question;
Can a large number of vehicle emission data points be used as actual shift factors for
determining vehicle emissions?
This paper aims to identify and estimate the emission of GHGs due to light
passenger cars by means of two methods: the Intergovernmental Panel on Climate
Change (IPCC) method [4, 10, 11] and the so-called Direct method. In this case, a
passenger car is a light vehicle defined as a two-axle motorized vehicle with four
wheels and an axle distance of 2.0–3.0 m.
2 IPCC Method and Direct Method
Greenhouse Gases may be defined as gases contained in the atmosphere, both due
to natural and anthropogenic activities, which absorb and re-emit infrared radiation.
Some of the solar radiation in the form of short waves that is received by the earth’s
surface is reflected to the atmosphere in the form of long-wave radiation. This long-
wave radiation is absorbed (entrapped) by GHGs in the lower atmosphere (near the
earth’s surface) and causes a heating effect known as the ‘Greenhouse Effect’. The
larger the volume of GHGs in the atmosphere, the more intense the warming effect
of earth’s surface can be observed on the whole world. This phenomenon is known
as global warming.
Two methods of calculating GHGs will be employed in this paper; (1) IPCC
method, (2) Direct method. IPCC method is the approach developed by the Inter-
governmental Panel on Climate Change by using a database of emission factors on
each activity. While direct method uses direct measurement data to estimate total
emission.
2.1 IPCC Method
In its 2006 report, the IPCC suggested three categories for calculating the GHG
emission based on data, approach, and accuracy. These are what are known as tier
1, tier 2, and tier 3. Tier 1 estimates GHG emission based on activity in general and
default emission factors proposed by IPCC. In tier 2, estimation is based on more
accurate activity data and default IPCC emission factors or country/plant specific
factors. Meanwhile in tier 3, estimation of GHG emission is based on specific methods

established in the country with accurate data of activities (direct measurement at
country/plant specific emission factors) [4].
Tier determination in the GHG inventory is largely determined by the availability
of data and the level of progress country or factory has made in terms of researching
country/factory specific factors that are specific and applicable to that. In Indonesia
and non-Annex 1 countries, the emission sources for key sectors/activities in the
GHG inventory use Tier-1, which is based on activity data and IPCC default emission
factors [11].
This paper analyses GHG emission on 2 streets, namely RA Kartini which will
later be referred to as street 1, and HOS Cokroaminoto which will later be referred
to as street 2, in Sragen, a regency with nearly 900 thousand inhabitants [12].
Based on the IPCC method, the amount of emissions, in this case CO2 emissions,
is calculated using the following formula (1) [11]

kg TJ kg
G H G emission = Energy consumption × Emmision f actor (1)
day day TJ
which is adopted from Eq. (2) below [5]:
G H G emission = Data activit y × Emmision f actor (2)
whereas emission factor is the emission shift factor of each activity. It has been modi-
fied and updated by the Ministry of the Environment from the IPCC database, taking
into account the types of fuel available in Indonesia [10]. This database contains
emission factors for calculating the GHG emission emitted by equipment and/or
activities consuming certain types of fuel [11].
The energy consumption, as an independent variable in Eq. 1, is calculated based
on Eq. 3

TJ liter
Energy consumption = Fuel consumption
day day

TJ
× Calori f ic value (3)
liter
where the calorific value of premium fuel in gasoline type combustion is

0.000033TJ/Liter [11]. The fuel consumption (KBBMi in liter/km) in Eq. 3 is
calculated through Eq. 4 [13].
K B B Mi

β1
α+ + β2 · V R2 + β3 · R R + β4 · FR + β5 · FR2 + β6 · DTR + β7 · A R + β8 · S A + β9 · B K + β10 · B K · A R + β11 · B K · S A
=
VR
1000
(4)
where:
α = Constant
β1-β11 = Parameter coefficients
VR = Average speed
RR = Average road inclination (uphill)
FR = Average road declination (downhill)
DTR = Average degree of road curve
AR = Average acceleration
SA = Standard deviation of acceleration
BK = Vehicle weight.
Equation (4) shows that said data (average speed, road inclination and declination,
acceleration, standard deviation of acceleration and vehicle weight) needs to be
collected prior to emission determination.
2.2 Direct Method
Meanwhile, in determining the amount of vehicle greenhouse gas emissions using

the direct method (actual multiplier), a direct vehicle emission measurement survey
was conducted using the Emission Analyzer tool. This measurement was carried
out in collaboration with the local district’s environmental office. The measurement
produces data on CO2 emission levels which are classified by years of vehicle produc-
tion. The parameter required is the average emission, which can be calculated using
Eq. 5.

CO2 emission
Average CO2 emission = (5)
N umber o f vehicle
After obtaining the average CO2 emission, the multiplier factor of CO2 per year of
vehicle manufacture can be determined by multiplying said average by the emission
sampling discharge being tested. The discharge is obtained from the specification of
the Emission Analyzer tool.
Multi plier f actor o f CO2 Emission = Average C O2 emission

× Discharge (6)
Afterwards, the actual CO2 emission can be calculated using Eq. 7:
CO2 emission = multi plier f actor o f CO2 emission

× T ra f f ic volume × T raveling time (7)
The emission of CO2 on the two streets were carried out using two methods; the IPCC
method and the direct method. Table 1, 2, and 3, below recapitulates the results of
these methodologies.
For the direct method, it is necessary to first identify the emission data per year
of vehicle production. See Fig. 1.
It was decided to classify emissions for every 5 years of production, with the
consideration that in those 5 years the types of vehicle engines commonly used in
Indonesia did not experience significant design changes, and the emissions produced
by vehicles in that range were also almost the same. The value is transformed into
Kg/L and shows in Table 2 column 3. These values can be understood as the shift
factor to estimate the daily CO2 emission in the street having the average daily traffic.
Table 1 Recapitulation of CO2 emission calculated using the IPCC method for light vehicle (fuel
type: gasoline)
Street Fuel Fuel Average Calorific Energy Emission CO2 Emission
consumption consumed daily value consumed factor of (kg/day)
l/km (Liter) traffic TJ/Liter (TJ/day) CO2 Kg/TJ
Veh./day
(1) (2) (3) = (4) (5) (6) = (7) (8) = (6)×(7)
(2)× (3)×
street (4)x(5)
length
1 0,0949 0,1044 3305 0,000,033 0,0114 69,300 789,1792
2 0,0937 0,0515 2673 0,000,033 0,0045 69,300 315,0219
Table 2 Recapitulation of CO2 emission calculated using the direct method on street 1
Vehicle Average CO2 emission Discharge Emission Travelling CO2 emission
production daily (Kg/L) (L/m) multiplier CO2 time (Kg/day)
year range traffic (Kg/hour) (hour)
(Veh./day)
(1) (2) (3) (4) (5) = (3)×(4) (6) (7) = (2)×
× 60 (5)×(6)
>1989 0 0,1066 5 31,9714 0,0342 0,0000
1989–1993 20 0,1077 5 32,3000 0,0342 22,0907
1994–1998 109 0,1003 5 30,1033 0,0342 112,2065
1999–2003 237 0,1220 5 36,5850 0,0342 296,5026
2004–2008 663 0,1343 5 40,2872 0,0342 913,3925
2009–2013 1266 0,1410 5 42,3086 0,0342 1831,6383
2014–2018 1010 0,1419 5 42,5823 0,0342 1469,9831
Total 4645,8138
Table 3 Recapitulation of CO2 emission calculated using the Direct method on street 2
Vehicle Average CO2 emission Discharge Emission Travelling CO2 emission
production daily (Kg/L) (L/m) multiplier CO2 time (Kg/day)
year range traffic (Kg/hour) (hour)
(Veh./day)
(1) (2) (3) (4) (5) = (3)×(4) (6) (7) = (2)×
× 60 (5)×(6)
> 1989 0 0.1066 5 31.9714 0.0164 0.0000
1989–1993 0 0.1077 5 32.3000 0.0164 0.0000
1994–1998 1 0.1003 5 30.1033 0.0164 0.2462
1999–2003 13 0.1220 5 36.5850 0.0164 7.4798
2004–2008 128 0.1343 5 40.2872 0.0164 84.3438
2009–2013 659 0.1410 5 42.3086 0.0164 456.0271
2014–2018 1873 0.1419 5 42.5823 0.0164 1304.1487
Total 1852.2456
14.5 3.5
14.0 3.0
13.5
2.5
13.0
2.0
CO (%)
12.5 Carbon dioxide (%)
CO2 (%)
12.0 Carbon monoxide (%) 1.5

11.5
1.0
11.0
10.5 0.5
10.0 0.0
>2004
2018
2017
2016
2015
2014
2013
2012
2011
2010
2009
2008
2007
2006
2005
Vehicle Production Year
Fig. 1 Diagram of average CO2 emission per vehicle production year tested in the beginning of
2019
Table 1 shows the amount of CO2 emission for street 1 and street 2 being
789.2 kg/day and 315 kg/day, respectively. This IPCC method uses emission factors
issued by the Ministry of the Environment which are the average number of vehicles
per day and energy consumption. The emission factors of CO2 used for gasoline-
fueled passenger cars is 69300 kg/TJ. As seen from the tables above, CO2 emissions
on street 1 are greater than street 2. This can be easily explained, as street 1 is longer
and has a higher average number of vehicles per day than street 2.
From Table 2, the amount of CO2 emission calculated using the direct method
give values of 4645.8 kg/day for street 1 and 1852.2 kg/day for street 2. The direct
method is carried out by identifying the average amount of CO2 emitted by passenger
cars which are categorized according to the year of vehicle’s manufacture. It can be
observed from Table 2 and Table 3 that the number of vehicles on street 1 is greater
than street 2, hence the amount of CO2 emission calculated on street 1 is greater than
that of street 2.
By observing Fig. 1, we can summarize that the older the engine’s production
year, the lower the average value of CO2 emission of said engine is. This is because
vehicles with the latter manufacturing year usually have good engine performance
so that complete combustion may occur as often as possible. This can be attributed to
the chemical reaction in the internal combustion engine in which complete combus-
tion will yield more CO2 and less Carbon monoxide (CO), Meanwhile incomplete
combustion engine produces more CO and less CO2 . The newer engine production
performs complete combustion therefore yield more CO2 . Also confirmed by Dineen
is the fact that the average amount of CO2 emissions of all new private cars purchased
in 2013 in Ireland was 162.4 g/km which is lower than the average of cars purchased
in 2014 which shows an average value of 170.3 g/km [14, 15]. It should be noted
that CO2 gas is the result of complete combustion of fuel, hence why the engine’s
production year may influence the amount of CO2 it produces. This is different from
the results of Kusumawati’s research [16], which states that the amount of CO2 emis-
sions is not closely related to the year of engine production but strongly influenced
by the engine’s performance. The performance of the engine itself is influenced by
its level of maintenance. In other words, instead of the production year, the level of
machine maintenance has more influence on the amount of CO2 emissions produced
[16].
From Fig. 2, we can see that the amount of CO2 emissions calculated by the direct
method are approximately 6 times higher than that of the IPCC method. We can see
that, as the calculation of the IPCC method carried out in this paper is classified as
tier 1 which only uses activity data in the form of fuel consumption and IPCC default
emission factors, the IPCC method’s level of accuracy is the lowest. All the while, the
direct method belongs to tier 3 wherein the CO2 emission test is carried out directly
on the vehicle so that its accuracy is higher than that of the IPCC method.
5000
CO2 Emission (Kg/day)
4000
3000
2000
1000
0
Street 1 Street 2
IPCC Method Direct Method
Fig. 2 Diagram of CO2 emissions on street 1 and street 2

Considering the equations used by the IPCC method and the direct method, there
are several variables that greatly affect the calculation’s results. These variables
are travel distance, average speed, and traffic volume. These variables are directly
proportional to the amount of emissions produced. The travel distance and the average
speed affect the required energy consumption, thus the greater each variable is, the
greater the emissions produced. The difference between the two methods is the
emission factor used. The IPCC method uses emission factors issued by the Ministry
of the Environment, which are sourced from IPCC 2006, while the direct method
uses data from emission tests conducted by the Environment Agency.
4 Conclusion
Investigation of CO2 emission on two local streets has been carried out and presented
in this paper by means of the IPCC tier 1 method and the direct method. It appears
that the results of calculating CO2 emissions using the direct method give 6 times
larger results than the IPCC method. This is because the calculated emissions in the
direct method are basically the result of the traffic volume times a CO2 emission
multiplier. This multiplier is determined from the measurement of passenger cars’
exhaust emissions in the field. Meanwhile, in the IPCC method, the CO2 emission
factor used is atmospheric CO2 gas that has either accumulated or interacted with
other gases. In addition to this, the IPCC method used in this research is a tier 1
approach, while the direct method is more likely a tier 3 approach.
Acknowledgements The authors express their gratitude and high appreciation to the Institute
for Research and Community Service (LPPM) UNS which has provided research grants from
non-APBN UNS funds to carry out this research.
References
1. Nurdjanah N (2014) Emisi CO2 akibat kendaraan bermotor di Kota Denpasar. J Penelitian
Transp Darat 16(4):189–202
2. Muziansyah D, Sulistyorini R, Sebayang S (2015) Model emisi gas buangan kendaraan
bermotor akibat aktivitas transportasi (studi kasus: terminal pasar bawah ramayana kota bandar
lampung). JRSDD 3(1):57–70
3. Yuliana DK (2018) Tingkat emisi gas rumah kaca di kabupaten indramayu. Jurnal Sains dan
Teknologi Mitigasi Bencana 12(2):1. https://doi.org/10.29122/jstmb.v12i2.2098
4. Eggleston HS, Buendia L, Miwa K, Ngara T, Tanabe K (2006) 2006 IPCC guidelines for
national greenhouse gas inventories.
5. Borgen SK, Grønlund A, Andrén O, Kätterer T, Tveito OE, Bakken LR, Paustian K (2012)
CO2 emissions from cropland in Norway estimated by IPCC default and Tier 2 methods. GHG
Measure Manage 2(1):5–21
6. Ramírez-Melgarejo M, Reyes-Figueroa AD, Gassó-Domingo S, Güereca LP (2020) Anal-

ysis of empirical methods for the quantification of N2 O emissions in wastewater treatment
plants: comparison of emission results obtained from the IPCC Tier 1 methodology and the
methodologies that integrate operational data. Sci Total Environ 747:141288
7. Kim KD, Ko HK, Lee TJ, Kim DS (2011) Comparison of greenhouse gas emissions from
road transportation of local government by calculation methods. J Korean Soc Atmos Environ
27(4):405–415
8. Oh Y, Park J, Lee JT, Seo J, Park S (2017) Estimation of CO2 reduction by parallel hard-type
power hybridization for gasoline and diesel vehicles. Sci Total Environ 595:2–12
9. Seo J, Park J, Park J, Park S (2021) Emission factor development for light-duty vehicles based
on real-world emissions using emission map-based simulation. Environ Pollut 270:116081
10. Hidup KL (2012) Pedoman Penyelenggaraan Inventarisasi Gas Rumah Kaca Nasional [Buku
I Pedoman Umum]. KLH, Jakarta
11. Hidup KL (2012) Pedoman Penyelenggaraan Inventarisasi Gas Rumah Kaca Nasional [Buku
II Volume 1, Metodologi Penghitungan Tingkat Emisi Gas Rumah Kaca]. KLH, Jakarta
12. BPS Statistics of Sragen Regency (2020) Sragen regency in figures, bilingual report, pp 29–31
13. DPU (2005) ‘Pedoman Perhitungan Biaya Operasi Kendaraan (Pd T-15–2005- B)’, pp 1–32
14. Dineen D, Howley M, Holland M (2014) Energy in transport 2014 report, Sustainable Energy
Authority of Ireland
15. Alam MS, Duffy P, Hyde B, McNabola A (2017) Improvement in the estimation and back-
extrapolation of CO2 emissions from the Irish road transport sector using a bottom-up data
modelling approach. Transp Res Part D: Transp Environ 56:18–32
16. Kusumawati PS, Tang UM, Nurhidayah T (2013) Hubungan Jumlah Kendaraan Bermotor,
Odometer Kendaraan dan Tahun Pembuatan Kendaraan dengan Emisi CO2 di Kota Pekanbaru.
Ilmu Lingkungan 7(1):49–59
Reducing the Release of Greenhouse
Gases in the Rigid Pavement Material
Transportation Process Unit
Fajar Sri Handayani, Florentina Pungky Pramesty, and Ary Setyawan
Abstract Material transportation of the road construction project has had some
negative impact on the environment, one of which is the release of greenhouse gas
emissions. In an attempt to develop sustainable strategy, there is a need to identify
and quantitatively evaluate the GHG emission reduction opportunities that exist in
material transportation of the rigid pavement life cycle. The aims of this study is to
estimate and evaluate the reduction of greenhouse gas emissions in the rigid pavement
material transportation process unit. The life cycle assessment was used to estimate
the greenhouse gases emissions. The evaluation of greenhouse gas emission reduc-
tions is not only on the percentage of the reduction, but also on the cost effectiveness
of the reduction. The cost effectiveness analysis (CEA) was used to calculate cost
effectiveness. The result shows that a reduction in greenhouse gas emissions in the
variation of the use of dump trucks in the rigid pavement material transportation
process unit ranged from 2.89% to 6.24%. Each alternative to reducing emissions
has the potential to reduce greenhouse gas emissions at a cost comparable to reduced
emission.
Keywords Reducing · Greenhouse gases · Rigid pavement
1 Introduction
Greenhouse gas (GHG) emissions due to construction projects in Indonesia in the

period of 2009–2019 was estimated at 29,941,737 tons of carbon dioxide equiva-
lent (CO2 e) of which 20% was contributed by toll road construction projects, 39%
was issued by national road construction projects, 17% was issued by projects of
F. S. Handayani (B) · F. P. Pramesty · A. Setyawan

Civil Engineering Department, Faculty of Engineering, Sebelas Maret University, Surakarta,
Indonesia
F. P. Pramesty
764 F. S. Handayani et al.
provincial road construction, and 14% was generated from regional road construc-
tion projects [1]. Road construction activities also have a significant impact on the
environment. To anticipate this, the government responds by issuing important strate-
gies in infrastructure development, namely sustainable development strategies or
development that applies the principles of sustainability.
Sustainable development is defined as the development of meeting the needs of
the present without limiting the ability of future generations to meet their needs.
This form of development is a stable relationship between activity and nature so
that future generations obtain a good quality of life. Sustainable road development is
development that is designed and built with reference to the concept of sustainability,
which is a system that reflects the capacity to support nature and human values
(ecology, economy, and equity) [2].
One strategic method to reduce the potential for global warming on pavement is
development based on the concept of increasing environmental performance most
effectively by focusing on components with high potential impacts. This system
takes advantage of the fact that small changes in high-impact components will have
a greater effect of change than low-impact components [3].
In the rigid pavement activities, the transportation process contributes 26.5% of
greenhouse gas emissions from the total emissions. The transportation of aggregate
to the Batching Plant contributes 9.48% of greenhouse gas emissions from the total
emissions [4]. Therefore, the efforts to reduce greenhouse gas emissions in this
article focus on opportunities for reduction in the pavement construction material
transportation process unit.
Reducing greenhouse gas emissions on pavement requires a complete under-
standing of how the life cycle of a construction project impacts the environment. Like
other products, road pavement also produces greenhouse gas emissions throughout
its service life, which starts from the material production, construction, operational
and maintenance stages.
Life cycle assessment (LCA) is a method used to estimate the environmental
impact resulting from the use of materials, distribution/transportation, construction,
maintenance and congestion, during the service life of the road construction [5].
Cost-effectiveness analysis (CEA) is a form of economic analysis that compares
the relative costs and outcomes (effects) of different courses of action. In several
references, efforts to reduce emissions have not yet carried out cost-effectiveness
analysis. A comparison of the additional costs incurred for emission reduction activ-
ities with a number of reduced greenhouse gas emissions is necessary [6]. In this
paper, applying the concept of cost effectiveness to the opportunities for reducing
greenhouse gas emissions at the road pavement construction stage, alternative reduc-
tions can be evaluated not only on the potential for reduction, but also on the relative
costs incurred for efforts to reduce these emissions.
An analysis of greenhouse gas emissions in variation dump truck (DT) needs
to be done to determine the percentage of emissions reductions that can be done.
Determination of alternative options for reducing greenhouse gas emissions is not
only in the percentage of the reduction, but also in the relative costs of the reduction.
Therefore, the purpose of this article is to estimate the percentage of reduction in
Reducing the Release of Greenhouse Gases in the Rigid Pavement … 765
greenhouse gas emissions in the material transport process unit and an analysis of
its cost effectiveness, on rigid pavement.
2 Methods
Stages of calculation of greenhouse gas emissions in this research are the determi-
nation of the activities of each stage of construction carried out, the determination of
emission factors per unit of activity, and the calculation of greenhouse gas emissions
which are the results of multiplication of activity data and emission factors per unit of
activity. Emission factors per unit of activity are based on the Eco-Invent data base and
the results of previous research [7–10]. Emission factors of transportation equipment
and construction equipment in units kilogram of carbon dioxide equivalent (kgC02e)
per hour use of the tool is based on an Eco-Invent database [11]. Greenhouse gases
conversion into equivalent carbon dioxide equivalent (CO2e) units used TRACI 2.1
data base. TRACI is an environmental impact assessment tool. It provides charac-
terization factors for Life Cycle Impact Assessment (LCIA), industrial ecology, and
sustainability metrics. Characterization factors quantify the potential impacts that
inputs and releases have on specific impact categories in common equivalence units.
The emissions analysis stages reviewed in this research are the material production
stage and the construction stage. The material production stage consists of cement,
aggregate, sand, and steel production processes. The construction phase consists of
pavement material transportation activities to the Batching Plant (BP), the process
of the Batching Plant, transportation of concrete to the project site, and the spread
of concrete at the project site.
Calculation of GHG emission reduction and cost was done on 3 alternative varia-
tions in the capacity of the used dump truck bouquet, alternative 1 used dump truck
DT 10, alternative 2 used dump truck DT 20, and alternative 3 used dump truck
DT 24. The existing mode of transportation of the material uses DT 7. The calcu-
lation of the CE was done by comparing the magnitude of the cost difference for
reducing greenhouse gas emissions with a reduction in greenhouse gas emissions in
each alternative.
Material production process unit is a material production process that exists in nature.
This process looks at the CO2e gas produced from each material used in the produc-
tion process. Material transportation process unit is a process of transportation of
a number of pavement materials needed to reach BP. Greenhouse gas emissions
calculated in this process are vehicle emissions resulting from the transportation
of material to BP. The rigid pavement construction which is the case study in this
research is the construction of rigid pavement with the volume of material usage
Table 1 The material usage

Material Usage (kg)
Cement 884,502
Gravel 2,137,602
Sand 1,499,699
Steel 13,538
Table 2 GHG emission in the transportation material to BP

Process unit Distance to BP Repetition GHG emission GHG emission
(km) factor (kgCO2 e)
(kgCO2 e/t.km)
The transportation of 31.2 286 0.353 47,257
gravel
sand
cement
presented in Table 1. The mixing process unit is the process of mixing raw materials
at BP. In this activity, it was found that the amount of greenhouse gas emitted due to
the mixing process at BP.
In the material transportation process unit to BP, the calculation of emissions is
done by multiplying the time of use of the transportation mode (hours) with the
emission factor (kgCO2 e/hour). The time of use of the transportation mode depends
on the volume of material, the distance of quarry to the BP, the volume of the bouquet
of transportation modes (dump truck), and the average speed of the transportation
mode. Calculation of greenhouse gas emissions in the process unit for transporting
material to BP is presented in Table 2.
The recapitulation of the calculation of greenhouse gas emissions released in the
rigid pavement construction process case study using material dump truck DT 7
means of transportation in the material transport process unit is presented in Table 3.
In this case study of rigid pavement construction, the LCA process calculation
results in the scope from raw material to the gate before the operation process (cradle
to gate), the material production process unit releases the largest greenhouse gas
emissions of 1,395,089.58 kgCO2 e or 86.56%, while the transportation process unit
releases greenhouse gas emissions by 178,945.62 kgCO2 e or 10.97% of total emis-
sions. However, when viewed from the LCA process in the scope of construction
(gate to gate), 75.99% of emission releases is generated by the process of transporting
pavement material.
The strategy to reduce greenhouse gas emissions in this case study was carried
out by looking at how much greenhouse gas emissions will be reduced by using
a dump truck transportation with a larger bouquet capacity, that is DT 10, 20, and
24. The assumption was established that with a larger dump truck bouquet capacity,
Table 3 Recapitulation of
Stage Process unit GHG emission
GHG emissions in rigid
(kgCO2 e)
pavement
Material production The production of 943,763.46
cement
The production of 6,840.34
gravel
sand
steel
Transportation to The transportation of 36,046.40
BP cement
The transportation of 47,256.01
gravel
sand
Mixing process Mixing process in BP 19,454.01
Transportation to The transportation of 22,616.57
site concrete
steel
Concrete spreading 2,326.83
Sum of emission 1,630,590.26
the number of repetitions and time of dump truck usage will be reduced, so that
greenhouse gas emissions can be reduced.
The reduction effort was evaluated not only on the percentage of the reduction, but
also on the relative costs incurred for the emission reduction effort. The simulation
results of the calculation of greenhouse gas emissions in the process of transporting
materials to BP with alternative means of transportation 1, 2, and 3, the difference in
emissions, the percentage of emission reduction, cost difference, and effective cost
effectiveness, are presented in Table 4.
The results of the calculation of Table 4 show that the variation of dump trucks
usage in the aggregate and sand transportation unit causes a reduction in greenhouse
gas emissions by 2.89%, 4.73%, and 6.24%, respectively. Alternative 3 results in the
highest reduction in greenhouse gas emissions. With the same volume, the use of
large capacity means of transportation will be able to reduce repetition and reduce
the time of its use, so that the resulting emissions will be reduced. The emission
reduction strategy with the use of large capacity dump trucks will reduce the biggest
emissions as well, but this strategy requires the biggest effective cost. Therefore,
when viewed from the value of cost effectiveness (CE), alternatives 2 and 3 have a
lower cost effectiveness value (CE) compared to alternative CE value 1. In this case
study, when evaluated from its CE value, alternative 1 is a greenhouse gas emission
Table 4 Recapitulation of GHG emission, percentage of reductions and cost effectiveness

Process unit Case study GHG emission Alt. 1 (kgCO2 e) Alt. 2 Alt. 3
The production of 943,764 943,764 943,764 943,764
cement
The production of 6840 6840 6840 6840
gravel
The production of sand 4799 4799 4799 4799
The production of steel 439,687 439,687 439,687 439,687
The transportation of 36,046 36,046 36,046 36,046
cement
gravel
sand
concrete
steel
Concrete spreading 2327 2327 2327 2327
GHG emission 1,630,590 1,626,460 1,623,831 1,528,841
(kgCO2e)
emission (kgCO2e) 4130 6759 8917
Percentage of 2.89 4.73 6.24
reductions (%)
cost (IDR*1000) 7000 13,000 21,000
CE 1.69 1.92 2.36
(IDR*1000/kgCO2e)
reduction strategy in the most effective pavement material transportation process

unit.
4 Conclusion
The reduction of greenhouse gas emissions in the variation of the use of dump trucks
in the material transportation process unit to BP ranged from 2.89% to 6.24%. Based
on the value of cost effectiveness, the use of dump truck DT 10 is the most effective
strategy to reduce greenhouse gas emissions in the material transportation process
unit. Each alternative to reducing emissions has the potential to reduce greenhouse
gas emissions and requires a comparable cost for emissions that can be reduced.
References
1. The Word Bank Group (2011) Green gas emission mitigation in road construction and
rehabilitation. A toolkit for developing countries, Washington DC
2. Muench ST, Anderson JL, Hatfield JP, Koester JR, Soderlund M (2011) Greenroads manual.
V1.5. University of Washington
3. Santero NJ, Horvath A (2009) Global warming potential of pavements. Environ Res Lett
4(034011):1–7
4. Handayani FS, Pramesti FP, Wibowo MA, Setyawan A (2019) Estimating and reducing the
release of greenhouse gases in local road pavement contructions. Int J Adv Sci Eng Inf Technol
9(5):1709–1715
5. Yu B, Qin L (2012) Life cycle assessment of pavement: methodology and case study. Transp
Res Part D 17(5):380–388
6. Santero NJ, Loijos A, Ochsendorf J (2013) Greenhouse gas emissions reduction opportunities
for concrete pavements. J Ind Ecol 17(6):1–10
7. Marceau ML, Nisbet MA, VanGeem MG (2007) Life cycle inventory of portland cement
concrete SN3011. Portland Cement Association Skokie Illinois
8. Zapata P, Gambatese JA (2005) Energy consumption of asphalt and reinforced concrete
pavement materials and construction. J Infrastruct Syst 11(1):9–20
9. Hughes B, Hare W (2012) LCI Data for steel product report produced. World Steel Association
10. Loijos A (2011) Life cycle assessment of concrete pavements: impacts and opportunities.
National Ready Mixed Concrete Association (NRMCA) San Diego
11. Wernet G, Bauer C, Steubing B, Reinhard J, Moreno-Ruiz E, Weidema B (2016) The ecoinvent
database version 3 (part I): overview and methodology. Int J Life Cycle Assess 21(9):1218–1230
Cost Effectiveness Analysis
of Greenhouse Gas Emissions Reduction
in the Flexible Pavement Material
Transportation Process Unit
Fajar Sri Handayani, Florentina Pungky Pramesti, and Ary Setyawan
Abstract The activity of material transportation regarding the project of road

construction indicated several negative impacts on the environment including the
greenhouse gas emissions release. In developing sustainable development strategy
especially in transportation infrastructure, it is necessary to find a strategy in reducing
greenhouse gas emissions of road construction cycle. This study aimed to estimate
and evaluate the greenhouse gas emissions reduction in the flexible pavement mate-
rial transportation process unit. The life cycle assessment was used to estimate the
emissions produced. The greenhouse gas emissions evaluation did not only underline
the reduction percentage, but also on the relative cost of the reduction. The approach
of cost effectiveness analysis concept was implemented to estimate the cost effective-
ness. The results showed that the greenhouse gas emissions in the dump trucks use
variation in the flexible pavement material transportation process unit ranged from
2.99% to 6.23%. Every alternative in emissions reduction showed the potential to
decrease the greenhouse gas emissions at a cost comparable to emissions that could
be reduced.
Keywords Cost effectiveness · Greenhouse gases · Flexible pavement
1 Introduction
Pavement construction activities have mostly used 22% of global energy consump-
tion, 25% of fossil fuels and 30% of global air pollution, namely greenhouse gases
[1]. Greenhouse gas emissions due to the projects of construction in Indonesia in the
period of 2009–2019 showed an estimate of 29,941,737 ton CO2 e of which 20% was
donated by the project of toll road construction, 39% was issued by the project of
F. S. Handayani (B) · F. P. Pramesti · A. Setyawan

Civil Engineering Department, Faculty of Engineering, Sebelas Maret University, Surakarta,
Indonesia
F. P. Pramesti
national road construction, 17% was issued by provincial road construction project,
and 14% was resulted from the project of local road construction [2].
A significant impact on the environment has emerged regarding road construc-
tion and improvement activities. In anticipation, the government has given responses
by issuing a strategy in infrastructure development that is necessary as a strategy
of sustainable development or development implementing the principles of sustain-
ability.
The definition of sustainable development refers to meeting the present needs
development without restricting the future generations’ ability in fulfilling their
needs. This development form is a stable relationship between activity and nature
and therefore there is a good quality of life provided for the future generations.
Sustainable development is designed and built referring to the sustainability concept
which is a system indicating the capacity in order to support nature and the values
of human (ecology, economy, and equity) [3]. A sustainable road construction is a
road construction starting from the design, construction, maintenance, and rehabili-
tation stages using environmentally friendly, environmentally friendly measures and
applying sustainable principles [4].
One of the strategic methods in reducing the global warming potential on pavement
includes the development based on the environmental performance increase most
effectively by underlining the high potential impacts components. The benefit taken
in this system is that the fact showing small changes in high-impact components will
get a greater effect than low-impact components [5].
In the flexible pavement material production stage, the material extraction process
contributed 92.11% of greenhouse gas emissions from the total emissions during the
material production stage. In the flexible pavement construction phase, the material
transportation process to the project site accounted for 70.09% of the total emis-
sions generated during the construction phase [6]. Thus, the approach used to reduce
greenhouse gas emission in this article underlined the reduction opportunities in the
process unit of construction material transportation.
In order to reduce the greenhouse gas emissions on pavement, it needs a deep
understanding about how the cycle of life of a construction project affects the envi-
ronment. Similar to other products, road pavement also results in greenhouse gas
emissions throughout its service life starting from the stages of material production,
construction, operational and maintenance.
Life cycle assessment (LCA) is a method used to estimate the environmental
effects resulting from the application of materials, distribution/transportation,
construction, maintenance and congestion, during the road construction service life
[7].
In general, cost effectiveness analysis is applied in the fields of health and
medicine. It compares the costs and health effects of an intervention to assess the
extent to which the business can give economic value. In implementing the cost
effectiveness concept aiming to decrease the emissions of greenhouse gas at the
construction stage of road pavement, reductions alternative can be assessed not only
on the reduction potential but also on the relative costs acquired for the emissions
reduction efforts. Cost Effectiveness (CE) becomes a comparison of the additional
Cost Effectiveness Analysis of Greenhouse Gas Emissions Reduction … 773
costs incurred for the activities of emission reduction with a number of reduced
greenhouse gas emissions [8].
An analysis of greenhouse gas emissions on several material transportation alter-
native is necessary to be conducted to determine the emissions reductions percentage
that can be done. The determination of alternative choices of greenhouse gas emis-
sions is not only in the reduction percentage but also in the reduction relative costs.
Thus, this article aimed to estimate the percentage of greenhouse gas emissions reduc-
tion in the material transportation process unit and its cost effectiveness analysis, on
the road pavement construction project.
2 Methods
This research took a case study of regional road construction in Sukoharjo district.
The road construction was flexible pavement. The data used in the life cycle assess-
ment method consisted of primary data obtained from direct observations in the
field and secondary data obtained from the database of related agencies and the offi-
cial website of the database provider. Data on quarry locations for pavement material
materials, asphalt mixing plant locations, types of material transportation equipment,
types of heavy equipment and construction methods used were primary data, while
secondary data required were general project data, TRACI 2.1 conversion database,
and life cycle inventory database from previous research and the Ecoinvent database
[9].
The estimate stages of greenhouse gas emission in this research were the determi-
nation of the activities of each construction stage which was conducted, the factors
of emission per unit of activity, and the estimate of greenhouse gas emissions as the
results of multiplication of activity data and emission factors per unit of activity. The
flexible pavement construction activities of this case study included the production
process of flexible pavement material (asphalt, filler, sand, aggregate), the transporta-
tion of pavement material to the Asphalt Mixing Plant, the process at the asphalt
mixing plant, the transportation to the project site, and the laying process at the
project site. The emission data for pavement material factors in units of kgCO2 e per
kilogram-material in this calculation was based on the results of research by Chappat
[10], Blomberg [11] and Korre [12]. Energy use and specifications for asphalt mixing
plant were based on the results of Zapata’s research [13]. The emissions factor of
transportation and construction equipment in units of kgCO2 e per hour of tool use
was based on the ecoinvent database [14].
The estimate of greenhouse gas emissions reduction and costs was conducted on
three alternative variations in the capacity of dump trucks that were used, alternative
1 applied dump truck DT 10, alternative 2 applied dump truck DT 20, and alternative
3 applied dump truck DT 24. The Cost Efficiency estimate (CE) was calculated by
comparing the cost difference magnitude to reducing the greenhouse gas emissions
with greenhouse gas emissions reduction in each alternative.
The road pavement construction which is the case study in this research is the flexible
pavement construction of regency road sections with a road length of 1446 m, a road
width of 15.6 m and a pavement thickness of 0.4486 m consisting of 0.1486 m thick
base course, sub-base thickness course of 0.2 m and the thickness of the surface
course of 0.10 m.
The emission analysis stages reviewed in this study were the material production
stage and the construction stage. The material production and construction stages
were divided into several process units, namely the material production process unit,
material transportation to asphalt mixing plant (AMP), the mixing process at AMP,
transportation to the project site, and the on-site laying process.
The unit of the material production process is a material process occurring in
nature. This process sees the CO2e gas resulted from every material applied in the
process of production. The unit of material transportation process is the transporta-
tion process of a number of pavement material required to towards AMP. Greenhouse
gas emissions calculated in this process were vehicle emissions resulting from the
material transportation process to AMP. Material transportation in the form of aggre-
gate, sand and filler in this case study used a 7.5 ton dump truck with a transportation
effectiveness of 90%, which means a truck capacity of 6.75 tonnes. Transportation
of materials in the form of asphalt and binder used a tank truck with a capacity of
8.75 tons and the effectiveness of transporting was 90% meaning the capacity of the
truck was 7.785 tons. All trucks used are assumed to use a speed of 40 km/hour.
The mixing process unit is the process of mixing raw materials. In this mixing
activity, the amount of greenhouse gases emission that was released was due to the
asphalt mixing process in AMP. The transportation process unit to the project site
was the transportation activity from AMP to the project site. The distance from AMP
to the work site required a certain amount of fuel to be used for truck operations. The
use of this fuel resulted in the release of greenhouse gas emissions in the process of
moving asphalt from AMP to the project site.
The unit of the spreading process at the project site was the activity of spreading
the asphalt mixture at the project site. Heavy equipment used in the sub-base course
construction process included an asphalt finisher, motor grader, water tank and dump
truck. Heavy equipment used in the base course construction process included asphalt
finisher, pneumatic tire rollers, tandem rollers and dump trucks. Heavy equipment
used in the surface course construction process included asphalt finisher, pneumatic
tire rollers, tandem rollers and dump trucks.
The material production process in the form of aggregate, sand, asphalt, filler and
binder was calculated by multiplying the volume of material required by the material
emission factor. Material emission factors in units of kgCO2 e per ton and calculation
of emissions in the material production process can be seen in Table 1.
In the material transportation process unit to AMP, the emission calculation is done
by multiplying the time used for the transportation mode (hours) by the emissions
factor (kgCO2 e/hour). The time to use the transportation mode depends on the volume
Table 1 Greenhouse gas emissions generated in the material production process unit
Material Volume Unit Emissions factor (kgCO2 e/ton) Usage (kg)
Aggregate 1997.83 Ton 10.0 19,978.32
Sand 459.89 Ton 2.5 19,978.32
Asphalt 60.68 Ton 285.0 1149.71
Filler 9.35 Ton 10.0 93.51
Asphalt binder 2.13 Litre 221.0 470.60
of material, the quarry distance to AMP, the volume of the dump truck, and the average
speed of the mode of transportation. The calculation of greenhouse gas emissions in
the material transportation process unit to AMP is presented in Table 2.
Material processing at AMP is in the form of mixing asphalt, aggregate, sand
and filler into a flexible pavement material mixture. The calculation of emissions in
the processing unit at AMP was done by multiplying the amount of time spent in
AMP by the AMP emission factor per hour. Based on the Ecoinvent 3.3 and TRACI
2.1 database, the emission factor in the processing unit at AMP was 119.0141 kg
CO2 e/hour. At an AMP capacity of 60 tonnes/hour with the use time in AMP in this
case study of 16.6 h , and therefore the greenhouse gas emissions produced in the
processing unit at AMP were 1975.63 kgCO2 e.
The calculation of greenhouse gas emissions in the asphalt spread process unit
at the project site is divided into 3 sub-process units, namely the base course, sub-
base course, and surface course. The calculation of greenhouse gas emissions at the
Table 2 Greenhouse gas emissions in the material transportation process to AMP

The unit of Distance to AMP The time length GHG emission GHG emission
material (km) (hour) factor (kgCO2 e)
transportation (kgCO2 e/hour)
process to AMP
The 33 488.4 5.06 463.99
transportation of
aggregate
The 33 113.85 5,06 2471.30
transportation of
sand
The 238 91.70 24.83 576.08
transportation of
asphalt
The 33 3.30 24.83 60.21
transportation of
filler
The 238 11.9 5.06 16.70
transportation of
a. binder
Table 3 Greenhouse gas emissions in the pavement construction process unit

Tool capacity Volume Tool use Emissions factor Greenhouse
(ton/hour)-(l/hour) (ton) (hour) (kgCO2 e/hour) gas emission
(kgCO2 e)
Motor grader 234.77 1015.20 4.32 24.80 107.25
Tandem roller 73.94 1015.20 13.73 24.80 340.54
Water tank 71.14 1015.20 14.27 5.06 72.20
Asphalt 4800.00 2248.13 0.47 3.58 1.68
distributor
Compressor 4800.00 2248.13 0.47 5.06 2.37
Asphalt finisher 72.79 994.80 13.67 24.80 338.98
Tandem roller 73.94 994.80 13.45 24.80 333.70
Pneumatic tire 172.34 994.80 5.77 24.80 143.17
roller
construction stage was carried out by multiplying the total time used by the heavy
equipment at each construction stage by the heavy equipment emission factor. Heavy
equipment used in the sub-base course construction process includes an asphalt
finisher, motor grader, water tank and dump truck. Heavy equipment used in the base
course construction process includes asphalt finisher, pneumatic tire rollers, tandem
rollers and dump trucks. Heavy equipment used in the surface course construction
process includes asphalt finisher, pneumatic tire rollers, tandem rollers and dump
trucks. The results of calculating the emission of construction heavy equipment in
the flexible pavement construction process unit are presented in Table 3.
The recapitulation of the greenhouse gas emissions calculation released in the flex-
ible pavement construction process using the DT 8 dump truck material transportation
unit in the material transportation process unit is presented in Table 4.
In this case study of flexible pavement construction, the estimate results of LCA
process in the scope from raw material to gate before the process of operation (cradle
to gate), the unit of material production process released greenhouse gas emissions
of 38,986.74 kgCO2 e or 79.52%, while the unit of transportation processes released
greenhouse gas emissions of 6727.51 kgCO2 e or 13.72% of total emissions. However,
when it is viewed from the process of LCA in the construction scope (gate to gate),
66.99% of emissions release was produced by the pavement material transportation
process towards the site location.
The greenhouse gas emissions reduction strategy in this case study was conducted
by observing how much greenhouse gas emissions would be reduced by applying
dump truck transportation tool with bigger bucket capacity, such as DT 10, 20, and
24. The assumption built that the bigger the dump truck bucket capacity, the lower
the repetitions number and the dump truck use time so the greenhouse gas emissions
can be reduced.
The efforts of reduction evaluated not only on the reduction percentage but also
on the relative costs incurred for the efforts of emissions production. The estimate
Table 4 The recapitulation

Process unit Process sub unit GHG emission
of greenhouse gas emissions
(kgCO2 e)
in flexible pavement
Material production Asphalt 17,294.60
production
Aggregate 19,978.32
production
Sand production 1149.71
Filler production 93.51
Asphalt binder 470.60
production
Transportation to Asphalt 463.99
AMP transportation
Aggregate 2471.30
transportation
Sand 576.08
transportation
Filler 60.21
transportation
Asphalt binder 16.70
transportation
Mixing process Mixing process in 1975.63
AMP
Transportation to site Mixed asphalt 3130.88
transportation
Asphalt binder 8.35
transportation
Pavement Base course 520.01
construction Sub-base course 4.05
Surface course 815.85
Sum of emission 49,029.29
simulation results of greenhouse gas emissions in the material transportation process

unit to AMP by using means of transportation of alternatives 1, 2, and 3, the difference
in emissions, the emissions decrease percentage, the cost difference, and the effective
cost effectiveness are presented in Table 5.
The estimate results presented in Table 5 indicated that the dump trucks use
variation in the unit of sand and aggregate transportation resulted in greenhouse
gas emissions reduction by 2.99%, 4.74%, and 6.23% respectively. Alternative 3
produced the highest greenhouse gas emissions reduction. Having similar volume,
the use of means of transportation in large capacity will be able to lower repetition
and reduce the time of its use, and therefore the emissions that are produced will be
reduced.
Table 5 Recapitulation of GHG emission, percentage of reductions and cost effectiveness

Process unit Existing GHG emission (kgCO2 e) Alt. 3
Alt. 1 Alt. 2
Asphalt production 17,294.60 17,294.60 17,294.60 17,294.60
Aggregate production 19,978.32 19,978.32 19,978.32 19,978.32
Sand production 1149.71 1149.71 1149.71 1149.71
Filler production 93.51 93.51 93.51 93.51
Asphalt binder production 470.60 470.60 470.60 470.60
Asphalt transportation 463.99 463.99 463.99 463.99
Aggregate transportation 2471.30 2229.183 2087.25 1962.02
Sand transportation 576.08 517.64 484.242 459.20
Filler transportation 60.21 60.21 60.21 60.21
Asphalt binder transportation 16.70 16.70 16.70 16.70
Mixing process in AMP 1975.63 1975.63 1975.63 1975.63
Mixed asphalt transportation 3130.88 3130.88 3130.88 3130.88
Asphalt binder transportation 8.35 8.35 8.35 8.35
Base course 520.01 520.01 520.01 520.01
Sub-base course 4.05 4.05 4.05 4.05
Surface course 815.85 815.85 815.85 815.85
GHG emission (kgCO2 e) 49,029.79 48,729.23 48,553.90 48,403.63
emission (kgCO2 e) 300.56 475.89 626.16
Percentage of reductions (%) 2.99 4.74 6.23
cost (IDR*1000) 7000 13,000 21,000
CE (IDR*1000/kgCO2 e) 23.29 27.32 33.54
The strategy of emissions reduction by using large capacity dump trucks will also
reduce the biggest emissions. However, this strategy needs the biggest effective cost.
Therefore, when it is seen from the cost effectiveness value (CE), alternative 3 has
highest cost effectiveness value. In this case study, if evaluated from its CE value,
alternative 1 is the most effective alternative strategy for reducing greenhouse gas
emissions in the pavement material transportation process unit.
4 Conclusion
The greenhouse gas emissions in the dump trucks use variation in the unit of material
transportation process towards AMP ranged from 2.99% to 6.23%. Based on the
cost effectiveness value, the dump trucks DT 10 use was the most effective strategy
alternative of greenhouse gas emissions reduction in the unit of pavement material
transportation. Every emissions reduction alternative has the potential in lowering
greenhouse gas emissions and need a cost which is comparable with the emissions
reduced.
References
1. Resmi CS, Satyakumar M (2011) Sustainable pavement: the green pavement for future.
National Technological Congress, Kerala, College of Engineering Trivandrum
2. The Word Bank Group (2011) Green gas emission mitigation in road construction and
rehabilitation. A toolkit for developing countries, Washington DC
3. Muench ST, Anderson JL, Hatfield JP, Koester JR, Soderlund M (2011) Greenroads manual.
V1.5. University of Washington
4. Santero NJ, Horvath A (2009) Global warming potential of pavements. Environ Res Lett
4(034011):1–7
5. Handayani FS, Pramesti FP, Wibowo MA, Setyawan A (2017) Strategy toward sustainable
local road network infrastructure. MATEC Web of Conf EACEF 138(07007):1–8
6. Handayani FS, Pramesti FP, Wibowo MA, Setyawan A (2019) Estimating and reducing the
release of greenhouse gases in local road pavement contructions. Int J Adv Sci Eng Inf Technol
9(5):1709–1715
7. Yu B, Qin L (2012) Life cycle assessment of pavement: Methodology and case study. Transp
Res Part D 17(5):380–388
8. Santero NJ, Loijos A, Ochsendorf J (2013) Greenhouse gas emissions reduction opportunities
for concrete pavements. J Ind Ecol 17(6):1–10
9. USEPA (2012) Tool for the reduction and assessment of chemical and other environmental
impacts (TRACI) users guide tool for the reduction and assessment of chemical and ther
environmental impacts
10. Chappat M, Bilal J (2003) The environmental road of the future: life cycle analysis, Energy
Consumption and Greenhouse Gas Emissions. Colas Group
11. Blomberg T (2011) Life cycle inventory: bitumen. (Second Edi). Brussels, European Bitumen
Association
12. Korre A, Durucan S (2009) Life cycle assessment of aggregates. EVA025 Final Report
13. Zapata P, Gambatese JA (2005) Energy consumption of asphalt and reinforced concrete
pavement materials and construction. J Infrastruct Syst 11(1):9–20
14. Wernet G, Bauer C, Steubing B, Reinhard J, Moreno-Ruiz E, Weidema B (2016) The ecoinvent
database version 3 (part I): overview and methodology. Int J Life Cycle Assess 21(9):1218–1230
A Probabilistic Model of Container Port
Demand in Java Concerning the Port
Hinterland Connectivity
Lydia Novitriana Nur Hidayati , Gerard de Jong , Anthony Whiteing ,

and Munajat Tri Nugroho
Abstract The port demand could be modelled from the port choice of the ports’
actors. From the actors’ perspectives, factors influencing port choice do not solely
depend on the port performance itself. As a port is an interchange location of land and
sea transport, the port selection is closely associated with the inland mode choice.
This study applies a multinomial logit (MNL) model to predict the probability of a
joint port and inland mode choice in Java, Indonesia. Six factors classified into port
performance and port hinterland connectivity are applied to estimate the probability
of port demand. The model also allows us to estimates the port hinterland boundaries
as a corollary of port choice made by the actors. Further from the simulation on several
transport strategies, introducing Patimban port as a new container port in Java greatly
affects the adjacent port demand. Reducing rail transport time and improving port
performance also leads to substantial port and mode shifting for container transport
in Java.
Keywords MNL model · Port choice · Inland mode choice · Port hinterland
connectivity
L. N. N. Hidayati (B) · G. de Jong · A. Whiteing

Institute for Transport Studies, University of Leeds, Leeds, UK
G. de Jong
A. Whiteing
L. N. N. Hidayati
Civil Engineering Department, Universitas Sebelas Maret Surakarta, Surakarta, Indonesia
G. de Jong
Significance, Amsterdam, The Netherlands
M. T. Nugroho
Industrial Engineering Department, Universitas Muhammadiyah Surakarta, Surakarta, Indonesia
782 L. N. N. Hidayati et al.
1 Introduction
This paper presents a study on forecasting probability of container port demand using
a discrete choice model. The demand of a port could be estimated through the choice
of port made by the ports’ actors (e.g., shippers and carriers), in which their loyalty in
choosing the port depends on many factors. In this paper, six variables classified into
two main categories, port-hinterland connection, and port performance, are exam-
ined. The model simulation results can be used for predicting the port hinterland
boundaries and predict the demand volatility of port and inland mode if particular
transport strategies are applied.
Port hinterland connectivity provides port authorities information about factors
that may affect their port market share. Wang et al. argue that efficiently managing
transport connectivity of the port and its hinterland areas could build customer devo-
tion and potentially attracting more customers hence expand the port market share [1].
For example, Nugroho et al. [2] demonstrated that a reduction of 20% in hinterland
mode time could raise port market share by 2% [2]. Secondly, Wang et al. stated that
the merit of analyzing port hinterland is to identify competitiveness with surrounding
ports, thus knowing the major port rivals or opening up a chance to collaborate with
other ports to develop transport strategies to guarantee port sustainability [3].
A surge of research effort has been directed to utilize discrete choice models to
analyze port choice. Malchow and Kanafani [4] discussed applying the logit-based
disaggregate model to predict choice made by shipper over the route and port choices.
Zondag et al. [5] demonstrate the MNL model to estimate the probability of using a
specific transport logistics chain which included port within the transport chain deter-
mined by carriers. Tang et al. [6] offered an approach to improve the MNL model’s
use to predict shipping companies’ port choice by integrating a transport network
attributes into the choice explanatory variables. Tapia et al. [7] perform an advanced
discrete choice model, a Multiple Discrete Extreme Value Model (MDCEV), to esti-
mate the mode and port choice in Argentina and further use the model to evaluate
impact of rail transport investment.
The model discussed in this paper is the multinomial logit (MNL) model on
port and inland mode choice. This model combines both the port choice and inland
mode choice into one choice alternative and examines the actor decision over the
possible alternatives. Nugroho et al. [2] stated that not many literatures investigated
these merged choices compared to abundance of literature analyzing the choices
separately. Port is a midway interchange hub connecting land and sea transport;
further, joint estimation of port and inland mode seems complicated. It may generate
a massive number of possible alternatives that create difficulties in data collection.
Thus, some studies either estimate port choice mode choice separately or treat port
choice as part of intermodal transport routes to transport goods from origin to final
destination location [4, 5, 8]. Applying the joint method, particularly for Java island,
Indonesia, is possible as there are only three major container ports. Moreover, by
this joint choice model, the impact of managing port hinterland connectivity to port
demand could be identified. In the model, factors affecting port and inland mode
A Probabilistic Model of Container Port Demand in Java … 783
choice categories as port performance variables are port cost and port ship calls. On
the other hand, inland transport cost, transport time, reliability and GHG emission
are the relevant variables for the port hinterland connection.
The rest of this paper is arranged as follows. The following section elaborates
on the modelling framework applied. It starts with the intermodal transport builder
to generate the value of explanatory variables of port hinterland connectivity and
then continues with the MNL model’s explanation. The third section summarizes the
model application, and the final conclusions are in section four.
2 Modelling Framework
2.1 Intermodal Transport Builder
We begin this section by introducing the model of an intermodal transportation

network in Indonesia, which we will denote as the transport chain builder (TC
builder), to generate the explanatory values for the variables of port hinterland
connectivity (i.e., inland mode cost, time, and GHG emission). The TC builder model
portrays the interaction between transportation network and possible transport chain1
combinations and the level of service connecting origin and destination location. In
the TC builder, a node could represent a TC zone (i.e., the centroid of a group of
subdistrict areas), transport terminal, and road junction. The link acts as a connector
between nodes that depict the road, railways, sea routes, and flight routes. For this
paper, only road and rail transport network within Java considered.
The mode characteristics applied in TC builder are provided in Table 1. The
transport cost function used here is derived from the cost function applied in Zukhruf
et al. [9], while the speed values are obtained from the data of the ministry of transport
and Nugroho [10]. The transport chain’s determination for only truck mode and
truck-train mode combination follows the shortest path algorithm based on the total
transport cost. From this path, then the total transport time and total GHG emission
will be calculated. Apart from the mode cost function and speed variable, the total
Table 1 Mode characteristics

Mode Cost function (Rp/TEU per Time function (hour)b Emission function (KgCO2 e)
km)a
Truck distance*846.96*11.6 distance/37.00 distance*62*11.6
Rail distance*686.33*11.6 distance/36.24 distance*22*11.6
a In this paper, the cost value of the TC builder will be converted into TEU/Ton by multiply the
initial value of 1 TEU = 11.6 Ton
b Time function is the calculation of distance divided by speed in km/hour
1 A transport chain defines as a sequence of modes employed in the process of shipping goods from
the point of origin to the destination location.
transport cost and total transport time also consider the handling cost and waiting
time at the mode interchange location (i.e., railway station/terminal).
2.2 Probabilistic Model in Port and Inland Mode Choice

in Java
Probabilistic model using discrete choice modelling usually roots on observed behav-
ioral data that better reflect the reality of decision-making. An individual (in this case
a shipper or exporter firm)
chooses the alternative based on its attractiveness, char-
acterized by a utility V pm , in comparison to other alternatives. In this paper, the
utility of each alternative is derived from the research performed in Nugroho [10]
which estimates its model based on the stated preference data. The utility function
of port and inland mode choice in Java can be expressed by Eq. (1).

V pm = ASC pm − 0.257 PC p − 0.817 P SC p + 1.93 I MCopm

− 0.863 I M Topm − 0.406 I M Ropm + 0.683 I M G opm (1)
V pm The observed utility of the alternative p using mode m

ASC pm Alternative specific constant for alternative port p using inland mode m
PC p Port cost for 1 TEU in port p (thousand Rp).
P SC p Ship calls of international container vessel per week in port p.
I MCopm Inland mode cost for transporting 1 TEU FCL from origin o to port p
using inland mode m (thousand Rp).
I M Topm Inland mode time for transporting 1 TEU FCL from origin o to port p
using inland mode m (hours).
I M Ropm Inland mode reliability for transporting 1 TEU FCL from origin o to port
p using inland mode m (%).
I M G opm Inland mode GHG emission for transporting 1 TEU FCL container from
origin o to port p using inland mode m (Kg CO2 e).
As inferred from Eq. (1), the alternative i defined as alternative choosing port
p using inland mode m, consequently the choice probability of alternative i for
individual n, for a Multinomial Logit (MNL) model (Pin ) can be calculated as in
Eq. (2):
e Vin
Pin = J (2)
V jn
j=1 e
3 A Case Study
In a case study of port hinterland connectivity in Java island, the proposed models will
be used to calculate the probability of port and inland mode choice from surrounding
zones. Java island has three main container port: Tanjung Priok, Tanjung Emas, and
Tanjung Perak, which hold about 60% of total container transport in Indonesia [8].
In an attempt to meet the growth of container transport, the government has a plan to
build the biggest container Port in Java, namely Patimban Port. This port expected
to support the container traffic existing in Tanjung Priok port which currently faces a
capacity problem and reduces the logistics cost due to its strategic location in between
industrial areas of Bekasi, Karawang, and Purwakarta [2]. Referring to the Eq. (1)
and Eq. (2), the probability of port and inland mode choice is depended on some
variables related to the port and inland mode characteristics. The port hinterland
connectivity variables except for inland mode reliability (no data available, thus it is
assumed to be 80% for all alternatives) have a specific value for each zone area. The
port performance variables applied are refer to Munajat [10] as provided in Table 2.
3.1 Port Hinterland Boundaries
From the probability of port and inland mode choice value, we can also infer the
information about the port hinterland boundaries. A high probability of a particular
port indicates that the respective area is the hinterland area of the port. The hinterland
boundaries of the Tanjung Priok Port presumed by the probability value can be seen
as in Fig. 1. Further, by the operation of Patimban Port, Tanjung Priok Port and
Patimban port’s hinterland boundaries are quite overlapping due to the proximity
between these two ports that reduce the demand of Tanjung Priok (see Fig. 2).
Table 2 Port performance

Port Port cost (thousand Rp) Port ship call
Tanjung Priok 1127 82
Tanjung Emas 1098 12
Tanjung Perak 1154 23
Patimbana 1126 82
a Patimban port is not yet operated. However, as Patimban port
is expected to be the biggest port in Java, hence all of the port
performance values are assumed to be the same as Tanjung Priok
Port which currently is the biggest port.
Probability to
Tj Priok Port
Fig. 1 Hinterland boundaries of Tanjung Priok port prior to Patimban port operation
Probability to
Tj Priok Port
Fig. 2 Hinterland boundaries of Tanjung Priok port after Patimban port operated
3.2 Elasticities
The effect of changes in each variable on the probability of alternative could be

expressed through the elasticity value. Elasticities can be calculated as percent
changes of probability in response to one percent changes in the observed vari-
ables. Table 4 depicts the elasticities of probability choosing Tanjung Priok port as a
service port and truck as the port hinterland transport mode from three zones, namely
Jakarta, Cirebon, and Surabaya. Those three zones are selected to represent zones
with high, medium, and low probability. Probability value close to 1 is assumed
as high probability, and close to 0 is assumed as low probability, while medium
probability is around 0.5.
Table 3 shows that the probability of choosing Tanjung Priok port by using truck
mode for Jakarta area is inelastic demand (10% reduction of inland mode time cause
only a 0.76% increase in probability). It means almost all containers from Jakarta
will be choosing Tanjung Priok via road transport even if there are big changes in the
observed variables. Unlike Jakarta, which has inelastic probability, the probability
form Cirebon area is elastic (i.e., 10% increase in inland mode time could drop
the probability of choosing Tanjung Priok port using truck by 9.30%). It means
that the container demand from this area is potentially shifting to other ports if the
port hinterland connectivity variables or the port performance are getting worse. For
Surabaya, even though its probability elasticity is high, as the initial probability is low;
hence the probability changes will not be very substantial. This result is in line with
the derivative theory of MNL modelling. The impact of observed variable changes
Table 3 Elasticities of probability choosing Tanjung Priok port using Truck as inland mode respect
to the one percent increase of observed variables in Jakarta, Cirebon, and Surabaya
Jakarta Cirebon Surabaya
The probability to Tanjung Priok port using Truck 93.67% 43.78% 0.33%
Inland mode cost −0.003 −0.055 −0.195
Inland mode time −0.076 −0.930 −2.772
Inland mode GHG emission −0.005 −0.693 −4.058
Inland mode reliability 0.098 0.868 1.539
Port cost −0.029 −0.251 −0.467
Port ship call 0.035 0.046 0.157
Table 4 MNL simulation result: variation of probability demand as respond to transport policy
scenario in Cirebon, Kebumen, and Trenggalek
Port Inland Probability (%)
mode Cirebon Kebumen Trenggalek
Scenario Scenario Scenario
0 1 2 3 0 1 2 3 0 1 2 3
Tj Priok Truck 43.8 38.9 43.5 31.7 15.6 13.8 15.3 13.4 1.4 1.3 1.3 1.4
Train 11.2 15.5 11.1 8.1 7.5 11.4 7.3 6.4 0.9 1.7 0.8 0.9
Tj Emas Truck 33.5 29.8 33.2 24.2 59.0 52.1 57.7 50.4 32.8 29.9 29.7 32.3
Train 7.4 10.4 7.4 5.4 5.0 7.6 4.9 4.3 2.5 4.0 2.2 2.4
Tj Perak Truck 1.8 1.6 2.1 1.3 6.9 6.1 7.9 5.9 50.4 45.9 53.3 49.5
Train 2.3 3.7 2.7 1.7 6.0 9.1 6.8 5.1 11.9 17.2 12.6 11.7
Patimban Truck 0.0 0.0 0.0 22.9 0.0 0.0 0.0 9.7 0.0 0.0 0.0 1.0
Train 0.0 0.0 0.0 4.7 0.0 0.0 0.0 4.9 0.0 0.0 0.0 0.7
is higher when the degree of uncertainty from the probability is high (P around 0.5).
Meanwhile, the impact of changes will be diminished when the probability is almost
certain (P close to 1 or 0) [11].
3.3 Transport Policies and Estimation Result
In this section, the multinomial logit (MNL) model will be used to estimate the
probability of port and inland mode demand in Java. For this simulation, three zones
with probability of around 50%, namely Cirebon, Kebumen, and Trenggalek will be
examined. Further, the effect of two transport policiy scenarios and the operation of
Patimban port on the port and inland mode demand will also be discussed. Scenario
0 represents the existing condition without transport policy implementation. The
transport policy scenarios that will be simulated are:
• Scenario 1: improving the transshipment process of rail transport is expected to

reduce train transport time by 20%.
• Scenario 2: investment in Tanjung Perak port is expected to double the port ship
call.
• Scenario 3: Patimban port starts to operate.
The effect of these transport policy scenarios on port and inland mode demand is
presented in Table 4. Improvement of rail transport as applied in scenario 1 is expected
to encourage mode shifting from road to rail transport. Based on the simulation,
reducing rail transport time by 20% has a high impact on boosting the demand for
rail inland mode in all zones. However, the probability of using a train as the inland
mode is still lower than road transport, indicating the low initial market share of
rail transport. Comparing the base mode cost of road and rail transport; train is
cheaper than truck. However, unlike trucks that offer cost efficiency as its cost only
consider the cost from origin to destination (port), the cost component in rail transport
requires an additional cost of loading and unloading process and the short-distance
trucking cost that could scale up the total cost considerably. Moreover, inadequate
law enforcement of overloading trucks also impacted the low market share of rail
transport.
The second scenario is expanding the port performance, for example, by planning
investment on Tanjung Perak port. By developing the Tanjung Perak port’s infras-
tructure, it is expected that enlarging the pier will pull the number ship calls from
23 to 46. Referring to Table 5, the effect of increasing port performance positively
impacts the port choice. The probability to Tanjung Perak port increases in all zones
for both road and rail inland modes. It is quite explicit that by providing better perfor-
mance on Tanjung Perak port, the shipper will be more attracted to shipping their
containers through this port.
The last scenario is the operation of Patimban port which is expected to be fully
operated in 2027. Cirebon has the highest port demand shifting from Tanjung Priok
port to Patimban port, followed by Kebumen. In contrast, Trenggalek has slight
port shifting from Tanjung Perak port to Patimban (about 1%). The high impact of
Patimban port on Cirebon, which is located in West Java area is apparently due to
close proximity to this port, and the assumption that both ports, Patimban and Tanjung
Priok, are similar in port performance make the hinterland boundaries hazier (see
Figs. 1 and 2).
4 Conclusion
This paper has demonstrated the use of a multinomial logit model to identify the port
hinterland boundaries and the impact of changes in observed factors on the probability
demand of port and inland mode. The transport time and reliability of inland mode
are variables from the port hinterland connectivity category that greatly affects the
joint demand of port and inland mode. At the same time, the number of port ship
calls has the highest impact from the port performance category. From the simulation
on several transport strategies, introducing Patimban port as a new container port in
Java will significantly impact Tanjung Priok port’s demand. Reducing rail transport
time and improving port performance also impacted in encouraging port and mode
shifting for container transport in Java.
Acknowledgements The authors wish to thank the Directorate of Higher Education, Ministry of
Education and Culture, the Republic of Indonesia for financial support for this research. The authors
also wish to thank the Universitas Sebelas Maret Surakarta for the support on this publication. The
authors are also grateful to anonymous reviewers for their suggested improvements to this paper.
References
1. Wang X, Meng Q, Miao L (2016) Delimiting port hinterlands based on intermodal network
flows: Model and algorithm. Transp Res Part E: Logistics Transpo Rev 88:32–51
2. Nugroho MT, Whiteing A, de Jong G (2016) Port and inland mode choice from the exporters’
and forwarders’ perspectives: case study—Java, Indonesia. Res Transp Bus Manag 19:73–82
3. Wang G et al (2016) Port connectivity in a logistic network: The case of Bohai Bay, China.
Transp Res Part E: Logistics Transp Rev 95(C):341–354
4. Malchow MB, Kanafani A (2004) A disaggregate analysis of port selection. Transp Res Part
E: Logistics Transp Rev 40(4):317–337
5. Zondag B et al (2010) Port competition modeling including maritime, port, and hinterland
characteristics. Marit Policy Manag 37:179–194
6. Tang LC, Low J, Lam S-W (2011) Understanding port choice behavior—a network perspective.
Netw Spat Econ 11:65–82
7. Tapia RJ et al (2020) Application of MDCEV to infrastructure planning in regional freight
transport. Transp Res Part A: Policy Practice 133:255–271
8. Frazila R, Zukhruf F, Burhani J (2018) Developing a probabilistic model for constructing
seaport hinterland boundaries. IOP Conf Series: Earth Environ Sci 158:012023
9. Zukhruf F, Frazila R, Burhani J (2017) A stochastic discrete optimization model for designing
container terminal facilities 1903:060007.
10. Nugroho MT (2015) Stated preference study of port and inland mode choice for containerized
export from java, in institute for transport studies. The University of Leeds, United Kingdom
11. Train K (2009) Discrete Choice Methods with Simulation. Cambridge University Press
Evaluation of Logistics System
Performance-Based on Indonesian
Government Policy
Muhammad Rizky Prakoso, Mohammed Ali Berawi, and Gunawan
Abstract Currently, Indonesia’s logistics system has not been maximized, as

evidenced by the high cost of logistics that account for about 23.5% of GDP nation-
ally and still far from developed countries that are only 10% of GDP even less. This
paper evaluates the performance of logistics systems in Indonesia based on govern-
ment policy. The cost component that is a problem in Indonesia’s logistics system
will be mentioned in this paper based on journals, reports and related regulations. The
data is taken by distributing questionnaires to the logistics delivery service providers.
The data that has been obtained is analyzed using matrix paired comparison. This
method is applied to rank the cost components presented. The result was that the
logistics cost component that most affects the cost of logistics in Indonesia are the
cost of sea and land transportation, due to too much tax and levy that needs to be
spent and inadequate infrastructure.
Keywords Logistics system · Government policy · Logistics · Indonesia
1 Introduction
Indonesia is the largest island nation in the world, with more than 17,000 islands.
Logistics distribution is a challenge for Indonesia, which is mostly territorial waters.
The constraints of the national logistics system are the high cost of logistics in
Indonesia. Currently, Indonesia’s logistics system is under neighbouring countries
such as Singapore, Malaysia and Vietnam. According to Logistic Parameter Index
(LPI), data in 2018 shows that Indonesia is ranked 46th with logistics costs in
Indonesia currently at 23.5% of GDP nationally [1, 2], while in developed countries,
logistics costs are only 10% of GDP and even less [3].
M. R. Prakoso (B) · M. A. Berawi · Gunawan

M. A. Berawi
792 M. R. Prakoso et al.
Logistics is defined as a process of planning, implementing, and monitoring effi-

ciently and effectively the flow and storage of goods, services and related information
from the point of origin to the destination point of goods to meet the user’s need [4].
Logistics aims is to ensure the proper flow of finished goods and various materials
in usable circumstances to the location where they are needed and at the lowest total
cost [5]. High logistics costs impact the competitiveness of the domestic industrial
sector with other countries’ product and service. Moreover, it also causes people’s
purchasing power to decrease.
Inter-island logistics delivery has been mainly through ship due to its large size. It
can accommodate a larger volume than other transportation, and shipping intra-island
still relies on truck because of its advantages that can do door to door delivery from
warehouse/factory to consumers. According to previous research, the performance
of logistics systems at the port is still not optimal, high logistic costs also occur as
long as the truck enters the port until it is loaded onto the ship and vice versa [6–10].
While the performance of logistics systems on land is still low, based on previous
studies it is due to inadequate road infrastructure, causing increased logistics delivery
costs [11].
The price disparity between regions is still relatively high, indicating that there
are problems in the logistics system. Based on previous studies, improved connec-
tivity is needed to reduce logistics costs. Although connectivity is closely related to
infrastructure, government policy is necessary so that various relevant stakeholders
can use existing infrastructure to be more optimal [12]. Therefore, this paper aims to
evaluate the performance of the logistics system in Indonesia based on government
policy.
2 Methodology
This paper will use a paired comparison analysis method. The Paired comparison
method selects one of the two available options or which objects are more suitable
or preferred according to the specified characteristics [13]. This method is applied to
rank the cost components presented. The cost component as a variable in this study is
taken from relevant reports, regulations and journals and obtained as follows: goods
transportation operation permit cost (Biaya Izin Penyelenggaraan Angkutan Barang)
[14, 15], Cross-district/city permit cost (Biaya Izin Lintas Antar Kabupaten) [16],
port entry fees [16, 17], trucking costs [11], port service costs [11, 16, 18–21] and
shipping costs [11, 12, 16]. The data was taken by sharing questionnaires to observe
the highest cost components that led to Indonesia’s high logistics cost. There was
30 sample in this study with a population of employees in the transportation and
warehousing sector according to Central Bureau of Statistics, as much 2,296,109
people. The steps taken in this study are: first, conducting literature studies related
to logistics systems in Indonesia based on reports, regulations and journals; second,
determine the variables of the literature studies that have been conducted; third, the
preparation of questionnaires; fourth, distribute questionnaires of expedition service
Evaluation of Logistics System Performance … 793
Fig. 1 Cost component ranking
companies/Freight forwarders and the like; lastly analyzed the data using matrix
paired comparison and generated a ranking order for logistics costs.
3.1 Identification of Logistics Cost Components
Based on the variables that have been provided in the previous section, given six
(6) cost components, namely: (1) goods transportation operation permit cost; (2)
Cross-district/city permit cost; (3) port entry fees; (4) trucking cost; (5) port service
cost; (6) shipping cost. Thirty respondents filled out this questionnaire, consisting
of expedition services, freight forwarders and shipping freight expeditions that send
logistics to all regions in Indonesia; they were asked to choose what costs are the
most expensive of the two cost options presented. The result is provided in Fig. 1.
The picture above shows that the highest cost component ratings were the shipping
cost after the analysis method paired comparison obtained. Shipping costs are an
essential part of the shipping cost by sea or river. The high cost of shipping is caused
by several factors that cause shipping companies to charge high fees, such as service
fees to be paid by ships at the port and ship operating costs, followed by trucking
cost, port service cost, port entry fees, goods transportation operation permit cost and
Cross-district/city permit cost. Each cost component will be described in the next
section.
3.2 Factors Affecting Logistics System Performance Based

on Logistics Cost
The government plays a vital role in the performance of the logistics system through
its policies and regulations. Logistics systems performance is still lacking attention
from the government with proven high logistics costs. The previous section’s data,
shipping costs, trucking costs and dock service fee-fees are considered by logistics
delivery service companies as the most expensive costs, followed by port entry fees,
goods transportation operation permit cost and Cross-district/city permit cost. The
costs are interrelated with each other. Here is a discussion on the logistics cost
component based on government policy.
Shipping costs
Shipping costs are considered expensive due to some things such as service fees to be
paid by ships at the port and ship operating costs. The results align with the previous
study that shipping costs are the most expensive logistics delivery [22]. Services to
be paid at the port by shipping companies are port dues, tug services, wharfage and
pilotage services included in the Non-Tax State Revenue (PNBP) [17]. The longer
the ship is in the port, the more handling done by the port, the cost of paying for
services at the port will be expensive. So the cost of shipping is considered the most
expensive cost, among others.
Trucking costs
Trucking costs become expensive due to several factors, including inadequate infras-
tructure, the existence of some areas that impose a ban on passing for container trucks
so that it is necessary to move logistics transported to smaller trucks, as well as the
presence of local governments that issue various permits and impose various levies
[11, 23]. According to LPI 2018 data, Indonesia’s infrastructure performance is still
low with a score of 2.89, still below 3.00 where the infrastructure performance score
in Indonesia is below other neighbouring countries such as Vietnam’s infrastructure
performance with a score of 3.01, Malaysia with a score of 3.15, and Singapore with
a score of 4.06 [1]. Additionally, currently in Indonesia, most logistics shipments
within the island use truck mode. This is because there is still no complete regulation
on the concept of multimodal transportation. The regulation on multimodal trans-
portation in Government Regulation No. 8 of 2011 has not regulated the procedures
and documents for goods that switch modes of transportation.
Port service cost
Port service cost consists of services for trucks at the port and services for ships.
The service cost for the ship has been explained in advance regarding the high cost
of the ship. According to the Ministry of Transportation Directorate General of Sea
Transportation, port services for trucks consist of three groups of costs; namely, the
cost of Loading Port Fee (LPF)/Destination Port Fee (DPF) Stevedoring, LPF/DPF
Cargodoring cost, and LPF/DPF Receiving/Delivery fee. LPF/DPF Stevedoring costs
are costs incurred for loading and unloading services of goods/containers from the
hatch/deck to the dock/truck/barge or loading from the dock/barge/truck on board by
using a ship crane. If interference with the ship’s crane is not used, it uses a mechan-
ical device owned by the port and is charged a certain fee above the stevedoring
tariff. LPF/DPF Cargodoring costs are costs incurred for the work of removing cargo
from the network (sling /ex tackle) above the dock, transporting from the dock, trans-
porting to the warehouse and arranging in the warehouse or vice versa. LPF/DPF
Receiving/Delivery fee is the cost charged for taking cargo from the warehouse and
delivering it to the top of the vehicle docked in the warehouse or vice versa until
making its preparation in the warehouse. The number of services to be paid at the
port makes the cost of port services considered expensive. However, the Republic
of Indonesia has just issued Presidential Instruction No. 5 of 2020 on Structuring
the National Logistics Ecosystem. The Existence of this Presidential Instruction will
simplify the process of activities at the port that will shorten the time at the port, but
whether it can lower the costs that service users must incur.
Port entry fee
Trucks entering the port must pay the entrance fee as Non-Tax State Revenue (PNBP)
following Government Regulation No. 15 of 2016. This rate is paid each time service
user enter the port. If the service user sends several logistics and requires in and out
of the port, the costs incurred to pay the port entrance fee will increase.
Goods transportation operation permit cost
According to the regulation of Law No. 22 of 2009 concerning Traffic and Road
Transport, the permit for the implementation of freight transport is issued by the
Minister responsible for traffic facilities and infrastructure and road transport. This
permit must be owned by the public transport company that organizes people’s
transportation and goods.
Cross-district/city permit cost
Cross-District permit fees are charged for the delivery of goods whose functional
area exceeds provincial limits. An example is the cost of escorting at a particular
location to ensure that the goods transported are avoided from criminality and the
cost of passing through the City district that applies the rules prohibited crossing at
certain hours [11, 14].
If logistics costs can be lowered, it can reduce the disparity in the price of prod-
ucts/services between regions. The purchasing power of the community will be the
price of products/services offered is higher. Therefore, the Indonesian government
must make logistics costs fall for long-term economic development purposes. There
is an evaluation of factors causing high logistics costs, it is expected that further
research can model the development of existing national logistics systems to lower
logistics costs.
4 Conclusion
This paper aims to provide input to the Government of Indonesia on logistics systems
in Indonesia based on government policy. The logistics system’s evaluation improves
Indonesia’s logistics system, especially by paying attention to the existing logistics
costs. The questionnaire results with paired comparison analysis method can be seen
as the ranking of logistics cost components considered expensive by logistics delivery
businesses in Indonesia. From the study literature and the results of questionnaires
with paired comparison analysis methods can be seen that the ranking of logistics cost
components that are considered expensive by logistics delivery business in Indonesia,
namely shipping costs, related to expensive port services such as port dues, tug
services, wharfage and pilotage services; and trucking costs due to several factors
such as inadequate infrastructure, prohibition of crossing by local regulations and the
number of permits and levies, followed by port service cost; port entry fee; Goods
transportation operation permit cost, and Cross-district/city permit cost. The paper’s
results provide direction for the Indonesian government to pay more attention to
the logistics cost components described above to improve the logistics system in
Indonesia.
Acknowledgements The authors would like to thank Universitas Indonesia for the support given
to this research.
References
1. The World Bank (2018) LPI aggregated rankings. In: World Bank. https://lpi.worldbank.org/
international/global. Accessed 14 Sep 2020
2. Kemenkeu (2020) National logistic ecosystem, Upaya Indonesia Perbaiki Daya Saing Logistik.
In: Kementeri. Keuang. Republik Indones
3. Zhao X, Tang Q (2009) Analysis and strategy of the chinese logistics cost reduction. Int J Bus
Manag 4:188–191. https://doi.org/10.5539/ijbm.v4n4p188
4. Council of Logistics Management (1991) Careers in logistics. Oak Brook, IL 60521, 2803
Butterfield Road, Suite 380
5. Agustini ED, Yarlina L, Murtadho A (2015) Effectiveness and efficiency of air transport logistic
in soekarno-hatta international airport. War Ardhia 41:97. https://doi.org/10.25104/wa.v41i2.
148.97-110
6. Iswanto (2015) Simplifikasi Sistem dan Prosedur di Pelabuhan dapat mengurangi celah
penyimpangan dan Ekonomi Biaya Tinggi. XIII:1–20
7. Kadarisman M, Suharto Y, Majid A (2016) Policy formulation of sea transportation system. J
Manaj Transp Logistik 03:161–183
8. Rumambi CC, Sendouw TK, Manoppo MRE (2016) Evaluasi kinerja operasional pelabuhan
bitung. 14:0215–9617
9. Prasetyo SA, Wicaksono A, Anwar MR et al (2014) Evaluasi Sistem Logistik Di Pelabuhan
Tanjung Perak Surabaya 8:113–120
10. Sirajuddin, Sunaryo, Zagloel TY (2018) Effect of concession model and deregulation logistics
policy for increasing port customer satisfaction in Indonesia. MATEC Web Conf 204. https://
doi.org/10.1051/matecconf/201820406008
11. Kemendag (2014) Kajian Kinerja Logistik Perdagangan Antar Pulau: Studi Kasus Baja
12. Sandee H (2016) Improving connectivity in Indonesia: the challenges of better infrastructure,
better regulations, and better coordination. Asian Econ Policy Rev 11:222–238. https://doi.org/
10.1111/aepr.12138
13. Lawless HT, Heymann H (1999) Sensory evaluation of food: principles and praactices, 2nd
edn. Kluwer Academic/Plenum Publisher, New York
14. Safitri RD (2020) Implementation of samarinda mayor regulation No. 23 of 2010 concerning
The determination of freight transportation tracks in Samarinda city area against containers. J
Law 1:86–101
15. Undang-Undang RI (2009) Undang-Undang Republik Indonesia Nomor 22 Tahun 2009
Tentang Lalu Lintas dan Angkutan Jalan
16. Kemendag (2013) Kajian Pengembangan Indikator Kinerja Logistik Indonesia
17. Peraturan Pemerintah RI (2016) Peraturan Pemerintah Republik Indonesia Nomor 15 Tahun
2016 Tentang Jenis Dan Tarif Atas Jenis Penerimaan Negara Bukan Pajak Yang Berlaku Pada
Kementerian Perhubungan
18. Brooks MR, Frost JD (2004) Short sea shipping: a Canadian perspective. Marit Policy Manag
31:393–407. https://doi.org/10.1080/0308883042000304881
19. Miftach Z, Pasek P (2016) Analysis of Port Costs and Loading and Unloading Costs on Income
PT. Abdi Nusantara Indonesia Line branch of Gresik. Surabaya
20. Direktorat Jenderal Perhubungan Laut K (2020) Penyelenggaraan dan Pengusahaan Bongkar
Muat Barang di Pelabuhan. Jakarta, Indonesia
21. Peraturan Pemerintah RI (2010) Peraturan Pemerintah Republik Indonesia Nomor 20 Tahun
2010 Tentang Angkutan di Perairan
22. Nascimento DLM, Alencastro V, Quelhas OLG et al (2019) Exploring Industry 4.0 technologies
to enable circular economy practices in a manufacturing context: a business model proposal. J
Manuf Technol Manag 30:607–627. https://doi.org/10.1108/JMTM-03-2018-0071
23. Mulyono AT (2020) Implementation of national transportation infrastructure. The role and
challenges of future development. Jakarta
Analysis of the Influence of Region
Development Factors, Individual
and Activity, Internal Operator
and External Operator on the Demand
for the Jakarta Bandung High-Speed Rail
Samijan, Mohammed Ali Berawi, and Andyka Kusuma
Abstract For the last few decades the High-Speed Rail has been growing quite
rapidly in various parts of the world. However, many High-Speed Rail Projects
were not as successful as expected. Most of the contributing factors were over-
estimate demand and cost-overrun. This study aims to formulate a strategy in order
to increase the demand for the Jakarta-Bandung High-Speed Rail by investigating
the factors of region development, individuals and activities, internal operators and
external operators. In conducting this research, the primary data collection was done
by using a stated preference questionnaire and the data analysis used the structural
equation modeling (SEM) method with the help of Lisrel software. The analysis
results concluded that the strong factors influencing the demand for the Jakarta-
Bandung High-Speed Rail are region development and external operators. This result
is useful as a reference for planners in planning the next similar project.
Keywords High-speed rail · Demand · Region development · Individuals and

activities · Internal operators · External operators · Structural equation modeling
(SEM)
1 Introduction
During the period of half a century, the development of the High-Speed Rail in the
world was quite rapid. Since Japan introduced the world’s first High-Speed Rail in
1964 with its Shinkansen [1, 2], many European countries such as France in 1981
Samijan (B) · M. A. Berawi · A. Kusuma

Department of Civil and Environmental Engineering, Faculty of Engineering,
Universitas Indonesia, Depok, Indonesia
A. Kusuma
M. A. Berawi
Faculty of Engineering, Center for Sustainable Infrastructure Development,
Universitas Indonesia, Depok, Indonesia
800 Samijan et al.
[1, 2] and Germany (1988–1998) [2] had followed suit. The next countries were Italy,
Spain, the Netherlands, Belgium, South Korea, Taiwan, China, England, Turkey and
Saudi Arabia [2]. Whereas in Southeast Asia the first High-Speed Rail to be built
was in Indonesia, namely the Jakarta-Bandung corridor, whose construction began
in 2019 [3].
However, not all of the High-Speed Rail Projects were successful, even many
of which were just wishful in their success, due to cost overruns and demand over-
estimation [4]. Demand is the key factor in the High-Speed Rail Project so that a
very large investment is economically feasible [4–6]. This paper aims to investigate
four factors that have the potential to affect the demand for the Jakarta-Bandung
Indonesian High-Speed Rail. The four factors are region development, individuals
and activities, internal operators, and external operators. Those factors were selected
due to refer to the literature review below. The Jakarta-Bandung High-Speed Rail
was chosen as the case because there were several controversies related to investment
break even and demand [7].
2 Literature Review
The construction cost of the High-Speed Rail infrastructure is very high because
there are many underground constructions [8], bridges, overpasses, and underpasses
to avoid level meeting with other modes of transportation. Thus, high cost of construc-
tion needs a large income in order to achieve a breakeven and be economically viable.
Passenger volume is the key factor for investment in High-Speed Rail [6, 8]. Demand
levels below 10 million per year, such as Madrid-Seville and Madrid-Barcelona in
Spain, shows a surprising fact in terms of return of capital or break even [6].
There is a relationship between region development factors such as transit-
oriented development with High-Speed Rail passenger demand [5, 9–13]. Mean-
while, high-rise buildings and mixed-use are also the key factors to improve demand
for transportation passengers [13].
Individual factors such as income [14, 15], activity [16], the purpose of the trip
[14, 17] and the number of members or cohort [14, 17, 18] can influence the choice
of mode and transportation demand decisions.
Internal operator factors that affect the demand for High-Speed Rail are usually
related to service levels [19]. The variables cost, travel time and distance have a
significant effect on the choice of transportation modes [9]. The fare per mile and the
time spent between the origin and the destination also influence demand [20]. The
utility function is defined as the attribute level of service: travel cost, the different
components of total travel time and service frequency [21]. The selection of alter-
native time and cost options is very significant for transport users [22]. The security
factors of the station also turns out to be a barrier for people to use the train [5, 23].
Other factors that affect demand are safety [5], comfort [21, 23], frequency [15, 24]
and reliability [5].
Analysis of the Influence of Region Development Factors … 801
Fig. 1 The design SEM

model Internal Operator Individual &
Activity
External Operator
Region Develop-
ment Demand
Accessibility was found as a significant factor that can induce higher mobility [18,
25]. The time consumed to get to and out of stations and the lack of integration of
local transportation also greatly reduce the number of passengers on the High-Speed
Rail [26]. People using trains depend on the quality of integration into local transport,
and the number of passengers is predominantly constrained by accessibility [27]. The
availability of parking facilities and shuttle transportation can also affect people in
using public transportation modes.
In England, a friendly environment is one of the variables and strategies to
attract passengers [23]. But this conclusion did not explain what the environmental
specifications were meant to be.
Based on the literature review above, the author made a structural equation model
(SEM) hypothesis as shown in the following Fig. 1. Of the five latent variables,
three of them are independent latent variables (exogenous), while the other two are
dependent variables (endogenous). Internal operator, external operator and region
development are exogenous latent variables. Whereas, individual and activity and
demand are endogenous latent variables.
3 Methodology
The primary data collection was carried out by means of a questionnaire survey, both
offline and online. Selected respondents were the regular users of several modes
of transportation to the Jakarta-Bandung route and vice versa, namely the users of
the Argo Parahyangan executive train, executive travel cars, executive buses and
private car users. These respondents are potential because they use these modes
routinely so that they can be encouraged to shift to High-Speed Rail. The primary data
collection was carried out from November 24 to December 18, 2020, on weekdays
and weekends.
The data analysis was performed by structural equation modeling (SEM) and
using Lisrel 8.7 software. The reason for using SEM is because SEM has several
advantages, such as using confirmatory factor analysis (CFA), capability for test
several dependent variables, model modification feature, etc. [28].
802 Samijan et al.
The data were collected from the questionnaire survey with a total number of 212
respondents. From the data collected, many respondents did not fill in questions about
personal data such as occupation, work location and especially income. This perhaps
because the questions were too personal. As a result, Lisrel could not execute those
data. So those indicators were excluded from the model. Then the individual and
activity latent variable was renamed into group activity to match the indicators that
compose them, namely the number of members in group activities. Table 1 shows
the indicators (manifest variables) of each latent variable.
The final result of running the above model on Lisrel 8.7 is shown in Fig. 2. Then
performed tests the validity of the instrument, construct reliability (CR) and variance
extracted (AVE). The validity of the instrument was measured by the value of λ
(standardized loading factor) > 0.5, CR ≥ 0.5 and AVE ≥ 0.5 (Hair et al., 2010 in
Refs. [29, 30]).
Table 2 shows the results of these measurements, in which CR and AVE are
calculated using the formulas (1) and (2) below [29]. The formula (1) was also stated
by Bagozzi and Yi [30]. Where, λ is the loading factor of each indicator for each
latent construct, n is the number of indicators for each latent construct, and ε is the
error of each indicator. So it can be concluded that all instruments are valid.
n 2
i=1 λ yi
C R = n 2 n (1)
λ yi
i=1 + i=1 ∈i
n
λ2
AV E = i=1 i (2)
n
The next test is goodness of fit (GOF) to the whole model. The result can be seen
within the output of Lisrel, written in Table 3. Most of the tests met the thresholds
so that the model was concluded fit.
Table 1 Latent variables and indicators

Latents variables Indicators
Internal operator (INT) Security, safety, amenity, cleanliness, friendliness, frequency,
simplicity and reliability
External operator (EXT) Accessibility, availability of parking facilities (P&R), availability
of shuttle transportation, Integration of the transportation and the
environment
Region development (REG) New city, Transit-Oriented Development (TOD)
Group activity (GRO) Group 1 (with 2 members or more), Group 2 (with 4 members or
more)
Demand (DEM) New region, TOD1, TOD2, TOD3, TOD4
Fig. 2 The result of running Lisrel 8.7
Table 2 λ, CR and AVE

Latent variable λ CR AVE Conclusion
Criteria Result Criteria Result Criteria Result
INT >0.5 min 0.62 ≥0.5 0.939 ≥0.5 0.664 Reliable
EXT >0.5 min 0.70 ≥0.5 0.870 ≥0.5 0.603 Reliable
REG >0.5 min 0.82 ≥0.5 0.872 ≥0.5 0.769 Reliable
GRO >0.5 min 0.81 ≥0.5 0.906 ≥0.5 0.828 Reliable
DEM >0.5 min 0.89 ≥0.5 0.978 ≥0.5 0.897 Reliable
The significance test of the structural relationship among latent variables is

measured by the T-value. It is considered significant if the T-value is more than
1.96 [29]. T-value describes the value of γ which is the level of relationship between
exogenous latent variables and endogenous latent variables or shows the value of β
namely the relationship between endogenous latent variable with one another. The
804 Samijan et al.
Table 3 Goodness of fit test

Test Threshold Result Conclusion
Chi-square/df ≤2 [29] 1.267 Fit
P-value >0.05 [31] 0.00855 Less fit
RMSEA ≤0.05 [31] 0.036 Fit
NFI >0.9 [29] 0.98 Fit
NNFI >0.9 [29] 0.99 Fit
PNFI >0.5 [29] 0.77 Fit
CFI ≥0.95 [30] 0.99 Fit
IFI >0.9 (Bollen in [29]) 0.99 Fit
RFI >0.9 [29] 0.97 Fit
CN >200 (Hoetler in [29]) 200.12 Fit
output structural model of Lisrel 8.7 can be seen in Fig. 3. It appears that there
are three significant structural relationships, namely the relationship between EXT
to GRO, EXT to DEM and REG to DEM, with T-values are 2.11, 2.56 and 7.95,
respectively.
The structural equation coefficient can be seen in Figs. 4 and 5. The significant
direct effect coefficients, according to the T-value test above, are 0.24 (EXT against
GRO) with the standard error is 0.11, the error variance is 0.89 and the determination
coefficient R2 is 0.11; 0.22 (EXT against DEM) with the standard error is 0.086,
the error variance is 0.54 and the determination coefficient R2 is 0.46; and 0.54
(REG against DEM) with the standard error is 0.068, the error variance is 0.54 and
the determination coefficient R2 is 0.46. Meanwhile, the other coefficients are not
significant. Based on the result, there is no significant indirect effect. The direct
effect, the indirect effect and the total effect among latent variables can be seen more
easily in Table 4.
Fig. 3 “T-value” mode, output of Lisrel 8.7

Fig. 4 Structural equations, output of Lisrel 8.7
Fig. 5 The “structural model” mode, output of Lisrel 8.7
Table 4 Direct effect, indirect effect and total effect

Path Direct effect Indirect effect through GRO Total effect Remark
INT → GRO 0.07 0 0.07 Not significant
INT → DEM 0.02 0.07 × (−0.00) = 0 0.02 Not significant
EXT → GRO 0.24 0 0.24 Significant
EXT → DEM 0.22 0.24 × (−0.00) = 0 0.22 Significant
REG → GRO 0.07 0 0.07 Not significant
REG → DEM 0.54 0.07 × (−0.00) = 0 0.54 Significant
Table 5 shows the significant influences among the latent variables, their indica-
tors, and their respective statistical attributes. The consecutively of influence coeffi-
cients starting from the most powerful variables are TOD3, TOD4, new region, TOD,
TOD2, TOD1, environment, integration, new city, accessibility, P&R and shuttle
transit.
806 Samijan et al.
Table 5 Significant variables and statistical attributes

Variables Coefficient Standard error T-value Error variance R2
External 0.22 0.086 2.56 0.540 0.46
Accessibility 0.70 0.028 11.11 0.100 0.49
P&R 0.70 0.029 11.23 0.110 0.49
Shuttle transit 0.70 0.029 10.95 0.110 0.48
Integration 0.84 0.029 14.57 0.077 0.70
Environment 0.84 0.031 14.42 0.081 0.71
Region development 0.54 0.068 7.95 0.540 0.46
New city 0.82 0.034 13.52 0.100 0.68
TOD 0.93 0.035 15.91 0.046 0.87
Demand
New region 0.97 0.036 0.94
TOD1 0.89 0.035 19.32 0.120 0.80
TOD2 0.90 0.033 19.57 0.100 0.81
TOD3 0.99 0.033 22.27 0.012 0.98
TOD4 0.98 0.030 23.46 0.020 0.96
According to the questionnaire, TOD1 describes respondents living at TOD

Karawang and working in Jakarta and no shuttle transit facility, and TOD2 is vice
versa. And then TOD3 describes respondents living at TOD Karawang and working
in Jakarta and there is shuttle transit facility, and TOD4 is vice versa. While TOD
shows a residential and business area that is built with a transit-oriented concept, as
referred to in the definition of TOD in general. Although EXT has a significant effect
on GRO (see Table 4), but GRO has no effect on DEM (see Fig. 5), then in Table 5
the relationship between the influence of the EXT variable on the GRO variable is
not included.
In practice in the real world, it makes sense that a fast train-based TOD which has
shuttle facilities as a means of transportation between stations and workplaces has
the high influence on the demand because of the many advantages such as saving
time, fast and convenience. Meanwhile, TODs without shuttle facilities have lower
influence. The concept of region development such as new city and TOD in general
also has a strong direct effect on demand. And the external operator factors namely
environment, integrated transportation, shuttle transportation, parking facilities and
accessibility have a strong influence on the demand and on the group activity. Yet,
because the group activity has a very little effect on the demand so the indirect effect
of the external operator on the demand through the group activity factor can be
ignored. In general, the influence of the external operator factors on the demand is
lower than the region development factors.
Thus, the accepted hypothesis is that the external operator factors (with its indica-
tors) and the region development factors (with its indicators) strongly influence the
demand. The influence of internal operator on demand is very small (coefficient is
0.02). With this result, to increase the demand of the Jakarta-Bandung High-Speed
Rail, the first strategy that must be taken is to promote region development around
the High-Speed Rail stations by building TODs and or new cities and then equip-
ping shuttle facilities. The second strategy is building access infrastructure to the
High-Speed Rail station from the surrounding areas, providing park & ride facility,
providing shuttle transit and integrating transportation system with fast, safe, cheap
and comfortable facilities. The road to the station must smooth and free from conges-
tion, with the result that reduces the resistance for people to drive to the station and
park their vehicles at the parking lot close the station and then take the High Speed
Train to their destination.
5 Conclusion
The analysis result shows that there are two factors that are strong enough to influence
the demand for the Jakarta-Bandung High-Speed Rail, namely external operator
factors and region development factors. The external operator factors consist of
access to and from the High-Speed Rail Station, the availability of parking facilities
close to the station, the availability of shuttle transportation, the integration of the
High-Speed Rail with other transportation systems, and the environment (minimal
CO2 emissions). And region development factors include TODs and new cities or
new areas. TODs which have shuttle transit from/to High-Speed Rail Station have
very powerful impact.
The strategies that must be taken in order to increase the demand for the Jakarta-
Bandung High-Speed Rail are by promoting the development of region around the fast
train stations namelyTODs, new cities, new business centers and equipping its with
shuttle transit; building access infrastructure to and out of stations, providing parking
facilities close to the stations, providing shuttle transportation between residential
and activity centers and stations, and connecting other transportation modes with the
fast train mode.
Acknowledgements We will appreciate editors and reviewers very much for positive and
constructive comments and suggestions for our paper.
References
1. Givoni M, Dobruszkes F (2013) A review of ex-post evidence for mode substitution and induced
demand following the introduction of high-speed rail. Transp Rev 33(6):720–742. https://doi.
org/10.1080/01441647.2013.853707
2. Palacin R, Raif L, Deniz Ö, Yan N (2014) High speed rail trends, technologies and operational
patterns: a comparison of established and emerging networks. Transp Probl 9:123–129
3. Anwar MC (2019) Akhir 2019, Mimpi RI Punya Kereta Cepat Bisa Separuh Terwujud. www.
cnbcindonesia.com
808 Samijan et al.
4. Beria P, Albalate D, Bel G (2018) Delusions of success: costs and demand of high-speed rail
in Italy and Spain. Transp Policy 68:63–79. https://doi.org/10.1016/j.tranpol.2018.03.011
5. Blainey S, Hickford A, Preston J (2012) Barriers to passenger rail use: a review of the evidence.
Transp Rev 32(6):675–696. https://doi.org/10.1080/01441647.2012.743489
6. Preston J, Albalate D, Bel G (2013) The economics of investment in high speed rail. Public
Adm Rev 2013–30(3):336–349. https://doi.org/10.1111/j.1540-6210.2011.02492.x
7. Afrianto D (2016) Pembangunan Kereta Cepat Diprediksi Timbulkan Kerugian.
economy.okezone.com. https://economy.okezone.com/read/2016/02/02/320/1302855/pemban
gunan-kereta-cepat-diprediksi-timbulkan-kerugian. Accessed 02 Feb 2016
8. Nash C (2014) When to invest in high-speed rail, pp 45–72. https://doi.org/10.1787/978928
2107751-3-en
9. Asensio J (2002) Barcelona’s CBD. Urban Stud 39(10):1881–1895. https://doi.org/10.1080/
004209802200000300
10. Loo BPY, Chen C, Chan ETH (2010) Rail-based transit-oriented development: lessons from
New York City and Hong Kong. Landsc Urban Plan 97(3):202–212. https://doi.org/10.1016/j.
landurbplan.2010.06.002
11. Pan H, Li J, Shen Q, Shi C (2017) What determines rail transit passenger volume? Impli-
cations for transit oriented development planning. Transp Res Part D Transp Environ
57(September):52–63. https://doi.org/10.1016/j.trd.2017.09.016
12. Duncan M (2019) Would the replacement of park-and-ride facilities with transit-oriented
development reduce vehicle kilometers traveled in an auto-oriented US region? Transp Policy
81(August 2017):293–301. https://doi.org/10.1016/j.tranpol.2017.12.005
13. Nyunt KTK, Wongchavalidkul N (2020) Evaluation of relationships between ridership demand
and transit-oriented development (TOD) indicators focused on land use density, diversity, and
accessibility: a case study of existing metro stations in Bangkok. Urban Rail Transit 6(1):56–70.
14. Ranjbari A, Chiu YC, Hickman M (2017) Exploring factors affecting demand for possible future
intercity transit options. Public Transp 9(1–2):463–481. https://doi.org/10.1007/s12469-017-
0161-3
15. Yunpeng Z, Zhuojun W, Xuemeng G (2016) Study on the influencing factors of passengers’
demand for high-speed rail. Proc Third Int Symp Manage Innov Dev Bks One Two, 29:648–653.
https://doi.org/10.5503/J.ME.2017.29.013
16. Rui F, Wei W, Min Y (2010) Analysis of individuals’ activity-travel behavior based on a
structural equation model. In: Proceedings of 2010 International conference on optoelectronics
and image processing, ICOIP 2010, vol. 1, pp. 247–250. https://doi.org/10.1109/ICOIP.201
0.100
17. Tyrinopoulos Y, Antoniou C (2013) Factors affecting modal choice in urban mobility. Eur
Transp Res Rev 5(1):27–39. https://doi.org/10.1007/s12544-012-0088-3
18. Weis C, Axhausen KW (2009) Induced travel demand: evidence from a pseudo panel data
based structural equations model. Res Transp Econ 25(1):8–18. https://doi.org/10.1016/j.ret
rec.2009.08.007
19. Zhou Z, Wang W, Hu Q (2015) An application of hierarchical structure model for trip mode
choice forecasting in China. Math Probl Eng 2015. https://doi.org/10.1155/2015/925963
20. Iseki H, Liu C, Knaap G (2018) The determinants of travel demand between rail stations:
a direct transit demand model using multilevel analysis for the Washington D.C. Metrorail
system. Transp Res Part A Policy Pract 116(March):635–649. https://doi.org/10.1016/j.tra.
2018.06.011
21. Román C, Martín JC (2010) Potential demand for new high speed rail services in high dense
air transport corridors. Int J Sustain Dev Plan 5(2):114–129. https://doi.org/10.2495/SDP-V5-
N2-114-129
22. Liu R, Li A (2012) Forecasting high-speed rail ridership using a simultaneous modeling
approach. Transp Plan Technol 35(5):577–590. https://doi.org/10.1080/03081060.2012.
701816
23. Harvey J, Thorpe N, Caygill M, Namdeo A (2014) Public attitudes to and perceptions of high
speed rail in the UK. Transp Policy 36:70–78. https://doi.org/10.1016/j.tranpol.2014.07.008
24. Diao M, Zhu Y, Zhu J (2017) Intra-city access to inter-city transport nodes: the implications
of high-speed-rail station locations for the urban development of Chinese cities. Urban Stud
25. Chang J, Lee JH (2008) Accessibility analysis of Korean high-speed rail: a case study of
the Seoul metropolitan area. Transp Rev 28(1):87–103. https://doi.org/10.1080/014416407014
21495
26. Wong B, Habib KMN (2015) Effects of accessibility to the transit stations on intercity travel
mode choices in contexts of high speed rail in the Windsor-Quebec corridor in Canada. Can J
Civ Eng 42(11):930–939. https://doi.org/10.1139/cjce-2014-0493
27. Shen Y, Zhao J (2017) Capacity constrained accessibility of high-speed rail. Transportation
(Amst) 44(2):395–422. https://doi.org/10.1007/s11116-015-9660-8
28. Sarwono Y (2010) Pengertian Dasar structural equation modeling (SEM). J Ilm Manaj Bisnis
Ukrida 10(3):98528
29. Riadi E (2018) Statistik SEM structural equation modeling. ANDI
30. Bagozzi RP, Yi Y (2012) Specification, evaluation, and interpretation of structural equation
models. J Acad Mark Sci 40(1):8–34. https://doi.org/10.1007/s11747-011-0278-x
31. Barrett P (2007) Structural equation modelling: adjudging model fit. Pers Individ Dif
42(5):815–824. https://doi.org/10.1016/j.paid.2006.09.018
Investigating the Factors Influencing
the Demand of School Bus
Willy Kriswardhana , Syamsul Arifin, and Ainal Akbar
Abstract Mode choice to school is frequently influenced by both student and parent
responses. However, the school bus is not a popular mode of school trip in Indonesia.
Escalating the demand for the school bus is essential as reducing the use of the private
motorised vehicle is a must. A stated preference survey and binary logistic regres-
sion method were performed to understand which factor influencing the willingness
to ride the school bus. This study aims to investigate which characteristic of both
children and parents significantly influence the school bus ridership. Factors that
significantly influence students’ willingness to ride school buses are gender, car
ownership, and travel time to school. Furthermore, in terms of parents’ permission,
household factors, namely car ownership and income are significant. Direction for
future study is presented.
Keywords School bus · Demand · Influencing factors · Stated preference
1 Introduction
The school bus is not a popular mode of school trip in Indonesia. Escalating the
demand for the school bus is essential as reducing the use of the private motorised
vehicle (for example, 2-wheeled scooter, which is very popular in Indonesia) is a
must. In the United States, children coming from no vehicles owned household tend
to use the school bus, while those from households with two vehicles are associated
with a lower likelihood of riding a school bus for school travel [1]. Furthermore, a
study mentioned that a good service scheduled bus could provide a proper service
for school travels [2]. A study explained that if the walking conditions are unsafe, the
W. Kriswardhana (B) · S. Arifin · A. Akbar

Universitas Jember, Jember 68121, Indonesia
S. Arifin
812 W. Kriswardhana et al.
use of the bus is essential. Nevertheless, the initial cost of operating the school bus
is not cheap [3]. Moreover, the demand for the school bus in Indonesia is declining.
Mode choice to school is frequently influenced by both student and parent
responses. Some studies have identified several significant factors related to travel
modes for school trips [4]. Socio-demographic factors included age and gender, influ-
ence the school mode choice [5]. Meanwhile, parent’s attitude is the strongest factor
influencing the transport mode of school preferences for children [6]. The leading
factor considered important to parents in choosing school trips is safety and security
[7]. Furthermore, distance also plays a role as a consideration in modal choice for
school. A study in Helsinki, Finland, implied that walking and cycling to school are
favourable if the travel distance is less than 1 km. On the contrary, students choose to
use or ride motorised vehicles when the distance exceeds 2 or 3 km [3, 8]. However,
some studies stated that travel distance does not influence the mode choice of the
school trip. For instance, variables of the built environment could affect the will-
ingness to walk [4, 9]. When the distance exceeds the students’ willingness to walk
or ride bicycles, and the need to minimise the use of private motorised vehicles is
essential, a school bus could be the possible alternative mode to school.
In Surabaya, Indonesia, the number of students riding school bus declined contin-
uously (approximately 45% in the period of 2010–2013). In 2010, there was 25,200
students/year using a bus for their school travel. However, it decreased to 14,400
students/year in 2013 [10]. Whereas, the potential number of students who were
willing to ride a school bus from 5 schools of High School Complex Area in Surabaya
reached 4195 students [11]. Meanwhile, the vehicle’s ownership is found to be a
significant variable in influencing the use of school bus in Surabaya. The higher
level of private vehicle ownership will affect students to use private vehicles as school
transportation. Therefore, it needs to improve school bus transportation services as
a strategy to increase students’ willingness to use school buses as transportation to
schools in the city of Surabaya [12]. A study on improving the services of school
buses in Malang, Indonesia, found that two factors, namely the availability of infor-
mation boards in each bus stop and the CCTV in buses, are important to increase the
performance [13]. It can be concluded that the declining number of students using
school bus for their school travel is mainly caused by the household and bus services
factors.
Therefore, this study aims to investigate which characteristic of both children and
parents significantly influence the school bus ridership. Previous studies in Indonesia
mainly used descriptive and cross-classification analysis to obtain the demand for
school bus. Furthermore, more studies concerned with the school bus level of services.
A little attention has been paid to investigate the relevant factors influencing students’
use of school bus. Logistic regression is accommodated to understand factors related
to and probability of demand of school bus ridership.
Investigating the Factors Influencing the Demand of School Bus 813
2 Methodology
The population of this study is 3209 students. A total of 550 respondents from Senior
and Junior High Schools in Jember, Indonesia, were asked to fill the questionnaire.
A Slovin method was performed to calculate the number of samples.
2.1 Stated Preference Survey
The most leading use of stated preference (SP) methods is people’s preference and
willingness to pay (WTP) measurement for non-market goods. SP valuation has two
main approaches: contingent valuation (CV) and choice experiment (CE). Generally,
SP methods have several advantages, especially when dealing with sets of goods that
do not currently exist [14].
This study used CV approach to understand the influence of household back-
ground characteristics on school bus ridership. The characteristics of the respondents
in this study were divided into 2 (two), namely the socio-economic characteristics
of students and the characteristics of student trips. Socio-economic characteristics
consist of gender, school origin, class, age, parent’s occupation, parents’ income and
vehicles owned by the students’ parents. The characteristics of the trip consist of the
student’s home address, the distance from the house to the school, the travel time,
the time students leave for school, and the mode of transportation that is often used
(Fig. 1).
Fig. 1 Location of the study

2.2 Data Analysis
A binary logistic regression was performed to understand significant factors and the
probability of each factor influencing the willingness to ride the school bus. Various
modelling techniques have been applied for investigating household travel data. This
study used the binary logistic regression to deal with binomial choices of whether
students are willing to use the school bus or not. Statistical analysis software was
performed to understand the model estimation.
Table 1 shows the general socio-demographic characteristics of the respondents.

Generally, students are willing to ride school bus when they travel form home to
school. Furthermore, 70% of parents allows their children to ride school bus. This
finding could be promising in implementing the school bus service in Jember.
3.1 Model Estimation and Goodness of Fit Test
Two estimation modelling were performed to understand which socio-demographic

factors influence students’ willingness to ride school bus and parents’ allowing their
children to ride school bus. Because of the fact that parent’s attitude is the strongest
factor influencing the mode school preferences for children [6] (Table 2).
The significance test shows that explanatory variables which were statistically
significant at the 0.1 level are gender and car ownership. Meanwhile, an explanatory
variable called travel time to school was found to be significant at the 0.05 level.
Therefore, the general estimated equation was described as follows:

ρ
logit(ρ) = ln
1−ρ
= −0.286 + 0.330 gender − 0.233 car owner shi p
+ 0.036 travel time to school
The model indicates that women students are more likely to ride a school bus.
Furthermore, a group of people with the characteristic of having no car would be likely
to choose school bus than people who have one or more car. This finding confirmed
earlier studies [1] which reported children coming from no vehicles owned household
tend to use the school bus, while those from the household with two vehicles are
associated with a lower likelihood of riding a school bus for school travel. In terms
of travel time, students experiencing a longer travel time to school also found to be
more likely to use school bus.
Table 1 Socio-demographic characteristics

Characteristics Percentage
Gender Male 36
Female 64
Parents’ income <IDR 1000,000 17
IDR 1000,000–1999,999 19
IDR 2000,000–2999,999 18
IDR 3000,000–3999,999 14
IDR 4000,000–4999,999 16
>IDR 5000,000 16
Parents’ motorcycle ownership 0 2
(2-wheeled scooter) 1 32
2 44
3 17
4 3
5 2
Parents’ car ownership 0 47
1 43
2 8
3 2
4 0
Frequently used transport modes Motorcycle 80
Car 8
Online taxibike/taxi 4
Bicycle 4
Walking 4
Students’ willingness to ride school bus Yes 67
No 33
Parents’ allowing their children to ride school bus Yes 70
No 30
Meanwhile, based on the parameter of estimation on parents’ characteristics, the

significance test shows that an explanatory variable which was statistically significant
at the 0.1 level is parents’ income. Meanwhile, an explanatory variable called car
ownership found to be significant at the 0.05 level. Therefore, the general estimated
equation was described as follows:

ρ
Logit(ρ) = ln = 0.505 − 0.295 car owner shi p
1−ρ
+ 0.121 par ents ear ning
Table 2 Parameter estimation

Variables Student Parent
Coeff. S.E. Sig. Coeff. S.E. Sig.
Gender* 0.330 0.192 0.085* 0.233 0.197 0.236
Age 0.008 0.018 0.678 −0.044 0.071 0.535
Motorcycle ownership 0.042 0.096 0.665 0.076 0.103 0.462
Car ownership* −0.233 0.132 0.077* −0.295 0.134 0.027**
Parents’ earning 0.030 0.061 0.619 0.121 0.063 0.054*
Distance to school 0.000 0.000 0.171 0.000 0.000 0.250
Travel time to school** 0.036 0.014 0.009** 0.015 0.013 0.239
Constant −0.286 0.509 0.574 0.505 0.843 0.5
H–L test Sig. = 0.455 Sig. = 0.152
Classification plot 67.3% 69.8%
Coeff. is coefficients, S.E. is the standard error, *variables were statistically significant at the 0.1
level, **variables were statistically significant at the 0.05 level. H–L Test is Hosmer Lemeshow
Test (sig = 0.455 > 0.10), the model is valid
Among individual socio-demographic characteristics, car ownership and parents’

income was found to be significant and associated with parents’ willingness to allow
their children riding school bus.
3.2 Interpretation of the Model and Discussion
The probability value of the willingness to ride a school bus is obtained by changing
the odds ratio to the probability prediction. The probability of a student who is willing
to ride a school bus can be formulated using the model. If a student is male (1), has
1 car, and spending 10 min to school, then the probability model is formulated as:
logit( p) = −0.286 + 0.330 gender − 0.233 car owner shi p

+ 0.036 travel time to school
logit( p) = −0.286 + 0.330(1) − 0.233(1) + 0.036(10) = 0.171
And the probability:
e0.171
p =1− = 0.457
1 + e0.171
The interpretation of the model is that a male student whose parent has a car
and spends 10 min to school would be willing to ride the school bus by 45.7%.
Meanwhile, in terms of parents’ willingness to allow their children to ride school

bus, the probability could be formulated using the model. Car ownership and parents’
income are two socio-demographic variables found to be statistically significant;
therefore, those variables were used to obtain the probability. If the parents have 1
car and earn IDR 3000,000–3999,999. Then the probability is presented as follows:
logit( p) = 0.505 − 0.295 car owner shi p

+ 0.121 par ents ear ning
logit( p) = 0.505 − 0.295(1) + 0.121(4) = 0.694
And the probability:
e0.694
p =1− = 0.333
1 + e0.694
The result indicates that the parents whose income is IDR 3000,000–3,999,999
and have one car would be allowing their children to ride school bus by 33%.
This study shows that travel time to school significantly influences students’ will-
ingness to ride a school bus. This finding is aligned with previous studies [2, 4].
Students with longer travel time are more likely to ride the school bus. However, it
should be noted that the distance variable does not show a significant influence on
this behaviour. Furthermore, male students are found to be less likely to ride school
bus. They prefer riding a motorcycle to the school bus.
Meanwhile, parents’ attitude also has a strong role in influencing school travel
mode. Household factors found to affect students’ mode choice to school. This study
found that household income significantly influences parents’ permission to ride a
school bus. Unlike the previous studies that stated that higher income led to less
intention to ride the school bus, this study showed that parents with higher income
are more likely to permit their children to ride a school bus. Car ownership found to
be statistically significant in both students’ and parents’ categories. It indicates that
having more car could decrease the willingness to ride a school bus.
By understanding the factors influencing the willingness to ride the school bus,
the strategies regarding this issue could be prepared and implemented. This concept
would undoubtedly support the sustainable transportation concept. Decreasing the
use of private motorised vehicles could decline the level of Green House Gases
(GHG) emitted to the air. Furthermore, safer transportation to school could decrease
the number of accidents involving students.
4 Conclusion
Travel mode to school is an essential issue because factors influencing this

behaviour could range from psychological, socio-demographic characteristics, built-
environment, congestion issues, and road safety. This study performed a stated pref-
erence survey to understand the household background and the existing journey to
school characteristics influencing school bus demand ridership.
Factors significantly influence students’ willingness to ride a school bus are
gender, car ownership, and travel time to school. Furthermore, in terms of parents’
permission, household factors, namely car ownership and income are significant.
This study provides a valuable baseline for future studies exploring the demand for
other travel modes to school. Policymakers should consider the changing behaviour
of parents and students’ attitude and behaviour. The finding of this study could be
useful to make an improvement to the travel mode to school.
Future studies could explore the attitude of students to active transportation;
because school bus can only be implemented when students are willing to walk.
Because of the pandemic situation, we performed an online-based survey. Future
studies should do a direct survey to bring a better understanding of respondents.
References
1. Lidbe A, Li X, Adanu EK, Nambisan S, Jones S (2020) Exploratory analysis of recent trends
in school travel mode choices in the U.S. Transp Res Interdiscip Perspect
2. Stark J, Beyer Bartana I, Fritz A, Unbehaun W, Hössinger R (2018) The influence of external
factors on children’s travel mode: a comparison of school trips and non-school trips. J Transp
Geogr
3. McDonald NC, Steiner RL, Palmer WM, Bullock AN, Sisiopiku VP, Lytle BF Costs of school
transportation: quantifying the fiscal impacts of encouraging walking and bicycling for school
travel. Transportation (Amst)
4. Ewing R, Schroeer W, Greene W (2004) School location and student travel analysis of factors
affecting mode choice. Transp Res Rec J Transp Res Board 1895:55–63
5. Noland RB, Park H, Von Hagen LA, Chatman DG (2014) A mode choice analysis of school
trips in New Jersey. J Transp Land Use 7(2):111–133
6. Ahern SM, Arnott B, Chatterton T, de Nazelle A, Kellar I, McEachan RRC (2017) Under-
standing parents’ school travel choices: a qualitative study using the theoretical domains
framework. J Transp Heal 4:278–293
7. He SY, Giuliano G (2017) Factors affecting children’s journeys to school: a joint escort-mode
choice model. Transportation (Amst) 44(1):199–224
8. Ito K, Reardon TG, Arcaya MC, Shamsuddin S, Gute DM, Srinivasan S (2017) Built environ-
ment and walking to school: findings from a student travel behavior survey in Massachusetts.
Transp Res Rec
9. Mitra R, Buliung RN (2014) The influence of neighborhood environment and household travel
interactions on school travel behavior: an exploration using geographically-weighted models.
J Transp Geogr
10. Kusuma OWA (2015) Evaluasi Program Bus Sekolah Di Kota Surabaya. Kebijak. Dan Manaj.
Publik
11. Sekartadji R, Widyastuti H, Herijanto W (2015) Study demand and supply bus Sekolah Rute
Dukuh Menanggal—SMA Kompleks Surabaya. In: Seminar Nasional Sains dan Teknologi
Terapan III, pp 781–788
12. Ermirasari S, Handayeni KDME (2017) Keterkaitan Karakteristik Pelajar & Perjalanannya
Terhadap Kesediaan Menggunakan Bus Sekolah Di Surabaya. J Tek ITS
13. Hariyani S (2017) School bus’s level of service in Malang City. IOP Conf Ser: Earth Environ
Sci
14. Hanley N, Czajkowski M (2019) The role of stated preference valuation methods in
understanding choices and informing policy. Rev Environ Econ Policy
Sustainability Aspects in Construction
Projects
Key Performance Indicator
for Analytical Hierarchy Process Used
for Determining the Effect of Reverse
Supply Chain Toward Green Building
Projects
Hermawan, Jati Utomo Dwi Hatmoko, and Jovana Neilkelvin
Abstract The Indonesian state also assesses that construction is an important part
of the development of a country. Concern about the negative impact of construction
waste is increasing every year and encourages the development of green or sustain-
able buildings. Reverse supply chain aims to obtain a production output from an
end-of-life process that is environmentally friendly as much as possible. The calcu-
lation of the final performance of the supply chain can be done using the Key Perfor-
mance Indicator weights obtained from the Analytical Hierarchy Process method.
Key performance indicator (KPI) is determined to measure the level of achievement
of objectives in the supply chain.
Keywords Green building · Reverse supply chain · Analytical hierarchy process ·

Key performance indicator
1 Introduction
Indonesia as a country valuing construction as an important part to the development

of countries. The growth of Gross Domestic Product (GDP) in the construction
sector is 5.4% according to data on Indonesia’s economic indicators in 2017 at the
Asia Construct Conference [19]. GDP value of the country of Indonesia estimate is
Reference to this paper should be made as follows: Sangwan, K.S., (2017) ‘key activities, decision
variables and performance indicators of reverse logistics’ and Chinda, T., and Ammarapala, V.
(2016) ‘Decision making on reverse logistics in the construction industry’.
Hermawan (B) · J. Neilkelvin

Soegijapranata Chatolic University, Semarang, Indonesia
J. Neilkelvin
J. U. D. Hatmoko
Diponegoro University, Semarang, Indonesia
824 Hermawan et al.
equivalent to USD 1015.54 million US in 2017 which represents 1.64% of the world
economy [19].
Widhiawati et al. [29] and Tazi et al. [28] state the same thing, that along with the
increasing number of construction projects, the need for materials will also increase.
At the same time, more construction waste is generated during the construction
process. Concerns about the negative impact of construction waste are increasing
every year and encouraging the development of green or sustainable building both
from the public and private sides [1].
Management of construction waste has four important aspects in waste manage-
ment, namely reduction, reuse, recycling and landfilling. According to Rogers and
Tibben-Lemke [21] Reverse supply chain is a process of planning, implementing
and controlling in reverse from raw materials, work in progress, finished goods to
consumer points to recycling points to proper disposal. Based on Rogers and Tibben-
Lemke’s explanation, it can be concluded that the reverse supply chain includes four
important things in waste management [21].
Setiawan [26] states that key performance indicators (KPI) are determined to
measure the level of achievement of objectives in the supply chain. Setiawan [26]
also states that these measures can be financial and non-financial which can be used
to measure the strategic performance of the organization. As a tool for measuring
the strategic performance of an organization, KPI shows the health and development
of the organization, the success of activities, programs or service delivery to realize
organizational targets or goals [26].
According to Kuswandi et al. [16] stated that the calculation of the final perfor-
mance of the supply chain can be done using the KPI weights obtained from the AHP
method. The purpose of weighting the KPI is to see the priority interests according
to the questionnaire filling stakeholders [16].
2 Related Literature
2.1 Supply Chain
Kusumastuti and Sugiama [15] state that in the supply chain process there are also
planned logistics activities, implementation and control of effective flow and storage
of goods, as well as services and information related from the starting point of
manufacture to the point of use to fulfill customer desires. Kusumastuti and Sugiama
[15] also state that logistics is an effort to manage goods needed for jobs related to
the supply of raw materials, equipment and equipment for the production process.
Govinand and Soleimani [8] state that in supply chain management there are
two types. The first type of supply chain management is the classic type which is
commonly used and is commonly called the forward supply chain [8]. The second
type of supply chain management is a reverse supply chain [8]. The purpose of a
Key Performance Indicator for Analytical Hierarchy Process … 825
reverse supply chain is to obtain a production result from an end-of-life process that
is environmentally friendly as much as possible [8].
2.2 Green Building
Green building define as a planning and management of construction projects

(in accordance with contract documents) to minimize the impact of the construc-
tion process on the environment so that there is a balance between environmental
capabilities and the needs of human life for present and future generations [6].
Green building or green construction is a sustainable building that leads to the
structure and use of processes that are environmentally responsible and resource-
efficient throughout the life cycle of the building, starting from site selection to
design, construction, operation, maintenance, renovation and demolition. This prac-
tice extends and complements classic building designs in terms of economy, utility,
durability and comfort [20].
The Green Building Council Indonesia [9] stipulates that a building in Indonesia
can only be considered a green building if it meets the assessment set by the GBCI.
Greenship assessment categories are appropriate site development (ASD), energy
efficiency and conservation (EEC), water conservation (WAC), material resource
and cycle (MRC), indoor health and comfort (IHC), and building environment
management (BEM) [9].
A summary of the greenship assessment criteria in the source category and mate-
rial cycle (Material Resources and Cycle-MRC) in the final assessment (FA) can be
shown in Table 1.
Table 1 MRC Greenship assessment criteria

Material resource and cycle
MRC P Fundamental refrigerant P 1 prerequisite criteria; 6 credit criteria
MRC 1 Building and material reuse 2
MRC 2 Environmentally friendly material 3
MRC 3 Non ODS usage 2
MRC 4 Certified wood 2
MRC 5 Prefab material 3
MRC 6 Regional material 2
Total 14 13.9%
Source GBCI [9]
826 Hermawan et al.
Table 2 Alternative options for RSC criteria

Citation Criteria alternative
Thierry et al. [29] Repair, refurbish, remanufacture, cannibalize and recycle
Johnson and Wang [13] Combination of remanufacture, reuse, and recycle
Rose and Ishii [22] Reuse, service, remanufacture, recycle and disposal
Guide et al. [10] Repair, remanufacturing and recycling
Ferguson and Browne [7] Reuse, remanufacture and recycle
Lee et al. [18] Reuse, remanufacture, recycle, landfill, and incineration
King et al. [14] Repair, recondition, remanufacture, and recycle
Sasikumar and Kannan [24] Direct reuse, recycling, remanufacturing, and repair
Skinner et al. [25] Destroying, recycling, refurbishing, remanufacturing, and
repackaging
Srivastava [27] Repair & refurbish, remanufacturing, and secondary market
Lambert et al. [17] Repair, reuse, remanufacture, upgrade, repackage, recycle,
reconfigure, and revaluation
Jindal and Sangwan [12] Repair, refurbishing, remanufacturing, cannibalizing, and
recycling
Source Sangwan [23]
2.3 Key Performance Indicator
Key performance indicators (KPIs) are determined to measure the level of achieve-
ment of objectives in the supply chain [26]. Setiawan [26] also states that these
measures can be financial and non-financial which can be used to measure the
strategic performance of the organization. As a tool for measuring the strategic perfor-
mance of an organization, KPI shows the health and development of the organization,
the success of activities, programs or service delivery to realize organizational targets
or goals.
Some of the product improvement processes include repair, reuse, remanufacture,
re-consumption, recycle and disposal. Sangwan [23] has summarized alternative
options in the process of repairing a product from various citation which can be
shown in Table 2.
2.4 Analytical Hierarchy Process
The Analytical Hierarchy Process (AHP) method is used to find a priority supply
chain process. The supply chain needs to be measured to ensure the success of
planning [2]. The main tool of the AHP method is the hierarchical model of the
problem. The hierarchy is created using a tree diagram which can be shown in Fig. 1.
Fig. 1 Tree diagram of analytical hierarchy process
3 Research Identification
It is necessary to identify reverse supply chain criteria related to KPI determination.

The KPI determination on the project comes from the criteria used. The criteria
used are based on the literature study that has been conducted, namely economic
and environmental criteria. Economic factors are a concern because if there is no
economic value, companies will find it more difficult to engage in reverse logistics
[5]. Meanwhile, environmental factors were chosen because they relate to green
buildings.
3.1 Key Performance Indicator for Analytical Hierarchy

Process
There are several KPIs that can be used for AHP in reverse supply chain. KPI choices
from several citations can be summarized in Table 3.
3.2 Key Performance Indicator for Green Building Research
There are fifteen key performance indicators for reverse supply chain [4]. The fifteen
KPIs were eliminated based on their relationship to the environment and economy.
The elimination of KPIs can be shown in Table 4.
828 Hermawan et al.
Table 3 Types of reverse supply chain KPI’s

Citation Criteria KPI Alternative
Barker and 1. Cost 1. Recycled product 1. Collection
Zabinsky [3] 2. Business relations 2. Test Costs 2. Sort-test
3. Scrap shipped 3. Processing
4. Original facility
5. Proprietary knowledge
6. Customer interactions
Chinda [4] 1. Economic 1. Labor cost 1. Direct reuse
2. Site constraint 2. Inventory cost 2. Remanufacturing
3. Transportation cost 3. Recycle
4. Processing cost 4. Landfill
5. Specific sorting
machine
6. Specific technology
7. Matured market
8. Landfill charge
9. Availability of landfill
10. Site space
11. Social image
12. Requirement of virgin
material
13. Limited project time
14. Environmental
concern
15. Knowledge of sorting
Hosseini et al. [11] 1. Economic gains 1. Direct reuse or resale
2. Environmental 2. Repairing
consideration 3. Refurbishing
4. Remanufacturing
5. Cannibalization
6. Recycling
7. Landfilling
Chinda and 1. Economic 1. Labor cost 1. Direct reuse
Ammarapala [5] 2. Site constraint 2. Inventory cost 2. Remanufacturing
3. Transportation cost 3. Recycle
4. Processing cost 4. Landfill
5. Specific sorting
machine
6. Specific technology
7. Matured market
8. Landfill charge
9. Availability of landfill
10. Site space
11. Green image site
12. Replacement of virgin
material
13. Limited project time
14. Legislative pressure
15. Knowledge of sorting
(continued)
Table 3 (continued)
Citation Criteria KPI Alternative
Kuswandi et al. 1. Reliability 1. On time pick up
[16] 2. Responsiveness 2. % utility of truckload
3. Agility 3. Deliver return cycle
4. Cost time
5. Assets 4. Return cycle time
management 5. Return order
reestablished and
sustained
6. Current deliver return
volume
7. Recovery rate
8. Scrap rate
Table 4 Justification of
KPI RSC Chinda (2014) Used Justification
selection of RSC research
KPI’s Labor cost Unrelated with
environment
Inventory cost Unrelated with
environment
Transportation cost Related with waste
Processing cost Related with waste
Specific sorting machine Related with waste
Specific technology Related with waste
Matured market Unrelated with
environment
Landfill charge Related with
environment
Availability of landfill Related with
environment
Site space Related with waste
Social image Unrelated with
waste
Requirement of virgin Related with
material environment
Limited project time Unrelated with
environment
Environmental concern Related with
environment
Knowledge of sorting Related with
environment
830 Hermawan et al.
4 Conclusion
This journal on the use of key performance indicators for analytical hierarchy process
becomes an additional material in knowledge about supply chain management, green
building and reverse supply chain itself. This research suggested a reverse supply
chain approach with help from key performance indicators for analytical hierarchy
process as one of many ways to determined reverse supply chain impact toward green
building.
References
1. Agyekum K, Kissi E, Andku JC (2020) Professionals’ views of vernacular building materials

and techniques for green building delivery in Ghana. Sci. Afr. 8
2. Al Farisi MS (2018) Perancangan sistem pengukuran kinerja green supply chain manage-
ment dengan metode green supply chain operation (GSCOR) and anathycal hierarchy process
(AHP). Undergraduate (S1) Thesis, University of Muhammadiyah Malang
3. Barker TJ, Zabinsky ZB (2011) A multicriteria decision making model for reverse logistics
using analytical hierarchy process. Omega 39:558–573
4. Chinda T (2014) Decision making on reverse logistics in the construction industry. Int J Civ
Struct Eng IJCSE 1(3):135–138
5. Chinda T, Ammarapala V (2016) Decision-making on reverse logistics in the construction
industry. Songklanakarin J Sci Technol 38(1):7–14
6. Ervianto WI (2012) Selamatkan bumi melalui konstruksi hijau, CV. Andi Offset, Yogyakarta
7. Ferguson N, Browne J (2001) Issues in end-of-life product recovery and reverse logistics. Prod
Plann Control 12(5):534–547
8. Govinand K, Soleimani H (2016) A review of reverse logistics and closed-loop supply chain.
J Clean Prod 142:371–384
9. Green Building Council Indonesia (2013) Greenship untuk bangunan baru versi 1.2. GBC
Indonesia
10. Guide Jr. VDR (2000) Production planning and control for remanufacturing: industry practice
and research needs. J Oper Manag 18(4):467–483
11. Hosseini MR, Chileshe N, Rameezdeen R, Lehmann S (2014) Reverse logistics for the
construction industry lesson from the manufacturing context. Int J Constr Eng Manage
3(3):75–90
12. Jindal A, Sangwan KS (2016) A fuzzy based decision support framework for product recovery
process selection in reverse logistics. Int J Serv Oper Manage. https://doi.org/10.1504/IJSOM.
2016.10000346
13. Johnson MR, Wang MH (1995) Planning product disassembly for material recovery opportu-
nities. Int J Prod Res 33(11):3119–3142
14. King AM, Burgess SC, Ijomah W, McMahon CA (2006) Reducing waste: repair, recondition,
remanufacture or recycle? Sustain Dev 14(4):257–267
15. Kusumastuti D, Sugiama AG (2017) Manajemen logistik organisasi publik. Universitas
Terbuka, Tangerang Selatan
16. Kuswandi RY, Ridwan AY, El Hadi RM (2018) Perancangan sistem monitoring reverse logistic
untuk industri penyamakan kulit dengan model SCOR. E-Proc Eng 5(3):6912–6919
17. Lambert S, Riopel D, Abdul-Kader W (2011) A reverse logistics decisions conceptual
framework. Comput Ind Eng 61(3):561–581
18. Lee SG, Lye SW, Khoo MK (2001) A multi-objective methodology for evaluating product
end-of-life options and disassembly. Int J Adv Manuf Technol 18(2):148–156
19. Pribadi KS, Soemardi BW (2018) The construction sector of Indonesia. In: 23rd Asia construct
conference 2018, Malaysia
20. Purnawan PY (2017) Pengelolaan limbah untuk pekerjaan strukturpada proyek konstruksi di
Daerah Istimewa Yogyakarta. S1 Thesis Atma Jaya Yogyakarta University
21. Roger D, Tibben-lembke RS (2002) Differences between forward and reverse logistics in a
retail environment. Suply Chain Manage Int J 7:271–282
22. Rose, CM, Ishii K (1999) Product end-of-life strategy categorization design tool. J Electron
Manuf 9(01):41–51
23. Sangwan KS (2017) Key activities, decision variables and performance indicators of reverse
logistics. Procedia CIRP 61:257–262
24. Sasikumar P, Kannan G (2008) Issues in reverse supply chains, part I: end of life product
recovery and inventory management – an overview. Int J Sustain Eng 1(3):154–172
25. Skinner LR, Bryant PT, Richey RG (2008) Examining the impact of reverse logistics disposition
strategies. Int J Phys Distrib Logistics Manage 38(7):518–539
26. Setiawan NR (2018) Perancangan sistem pengukuran kinerja green supply chain dengan metode
green SCOR berbasis AHP and OMAX (studi kasus: CV.cool clean). Undergraduate (S1)
Thesis, University of Muhammadiyah Malang
27. Srivastava SK (2008) Network design for reverse logistics. Omega 36(4):535–548
28. Tazi N, Idir R, Fraj AB (2020) Sustainable reverse logistic of construction and demolition
wastes in French regions: toward sustainable practices. Procedia CIRP 90:712–717
29. Thierry MC, Salomon M, Nunen JAEEv, Wassenhove LNv (1995) Strategic issues inproduct
recovery management. Calif Manage Rev 37(2):114–135
30. Widhiawati IAR, Astana NY, Indrayani NLA (2019) Kajian pengelolaan limbah konstruksi
pada proyek pembangunan gedung di Bali. Jurnal Ilmiah Teknik Sipil 23(1)
Development of Blockchain and Machine
Learning System in the Process
of Construction Planning Method
of the Smart Building to Save Cost
and Time
Christiantono Tedjo, Mohammed Ali Berawi, and Mustika Sari
Abstract Achieving project objectives that are effective, efficient, and safe requires
careful planning of construction methods. The construction method of a project may
vary depending on the description and design of the project. Planning for construction
methods generally takes a long time to review, involves many parties, and requires
consideration in terms of costs, resources, and safety. And the need for construction
planning characteristics for smart buildings recently. Machine learning combined
with the blockchain system is an instrument that can accommodate processes in the
construction method. Start from planning to integration-evaluation with a historical
database. This instrument is expected to be able to provide suggestions and evaluation
of methods, to be able to assist the planning process of construction methods so that
the application of work methods will be more effective and efficient.
Keywords Construction method · Smart building · Machine learning · Blockchain
1 Introduction
The construction method of a project may vary depending on the description and
design of the project. Construction method planning is generally not brief. The
construction process requires integration between disciplines and divisions. To obtain
efficient results in time, cost, ease of achievement, and a sustainable process [1].
While the more advanced and developing technology, the amount of information
in the construction process will increase, and the more dynamic it is, the larger the
knowledge network will be [2]. In interdisciplinary integration, directly or indirectly,
C. Tedjo (B) · M. A. Berawi · M. Sari

University of Indonesia, Jakarta, Indonesia
M. A. Berawi
M. Sari
834 C. Tedjo et al.
a process of transferring information is required. The knowledge transfer process,

especially by manual method, tends to have several impacts that can disrupt the
project deliverable process. One of them is the process of knowledge transfer manual
or between individuals. There is a huge potential for the information required to expe-
rience delays [3]. In addition to delays in information needs, information reduction
is also very possible because the knowledge transfer process requires good relation-
ships between individuals or stakeholders so that communication can run smoothly.
Also, the increasing intensity of work will make it difficult for each individual to
transfer information accurately and quickly [3].
Construction project clients recently need increasingly complex facilities in their
buildings, requiring a fast construction process, and low costs [4]. This is shown by
the presence of various building design concepts in the world such as green building,
healthy building, and smart building, this of course requires technological devel-
opment and an increasingly high level of complexity. Apart from the consideration
of the information transfer process, the construction project requires consideration
in terms of costs, resources, process handling, and safety. One of the criteria for a
good work method is that it can accommodate the transportation of tools, goods,
and people. Minimizing narrow areas of transportation access and accommodating
the creation of good HSE management [5]. Demand in the building sector is also
increasing with collaboration on technological advances and energy consumption
[6]. For each building to be able to manage its energy, smart building technology
needs to become a new adoption, especially in Indonesia, where construction is
currently starting to lead to the development of the smart building sector. Machine
learning combined with the blockchain system is an instrument that can accommo-
date processes in the construction method. By reducing bureaucracy and an orderly
and integrated database storage process. The study results show the potential cost
savings from using blockchain is 8.3% of the total cost of building a house, with
a standard deviation of 1.26% [4]. So that this instrument is expected to be able
to provide suggestions and evaluation of methods, to be able to assist the planning
process of construction methods so that the application of work methods will be
more effective and efficient.
2 Blockchain and Machine Learning
2.1 Blockchain
Construction is a process of collaboration and integration in every element and stage.

Collaboration and integration in transaction activity, for example in its current appli-
cation, namely economic activity, whatever the form of the transaction requires a
system that has a high level of “trust”. Most people don’t give trust free of charge,
they want verification and assurance that the other party is worthy of their trust, and
Development of Blockchain and Machine Learning System … 835
in commercial situations, they often buy that guarantee from intermediaries. There
are many intermediaries in all sectors [7].
Intermediaries need time to process some information (for example, in the process
of credit card payments, it will take a few days for the payment to arrive in our
account) and the cost of the transaction is borne by us as consumers (the vendor
only accepts part of the money charged to the card account. Credit). In general, this
process applies, intermediaries collect information about our credit cards—to verify
agreement data, to be able to convince other parties, or other intermediaries, that our
data is valid.
The process takes time, effort and money, but provides no value for the customer.
To reduce this process with the potential of blockchain as described in the previous
section, blockchain is expected to be an alternative instrument to reduce this process.
Blockchain is considered to have the potential to change many global industries
including construction. Blockchain was originally developed as a supporting tech-
nology for the world’s first cryptocurrency, Bitcoin, which was introduced in 2008
as a verification tool for the transaction process [8]. One of the breakthroughs devel-
oped by using blockchain for delivery systems in the field of construction includes
the use of smart contracts. The application of smart contracts is used by utilizing 4
characteristics of the blockchain system, namely distributed, which means that data
and functionality are distributed or not centralized (located on different servers or
users), replicated which means data can be backed up, immutable which means data
cannot be modified and non-repudiation, which means that every person who does
it can be proven validity.
2.2 Machine Learning
The application of machine learning in the world of construction has been widely
applied, one of which is the planning mapping of needs and formwork areas as done
[9] in his research journal which explains how machine learning can develop the
planning and procurement of the amount of formwork for an atypical building area.
This process saves time and resources, the identification results of machine learning
are more detailed and accurate than the identification of formwork requirements that
is done manually. There is also a journal discussed by Ahmed Gondia et al. [10] about
how machine learning is used to identify the causes of project delays. In addition to
safety identification. The application and need for machine learning are examined
for the level of need in a journal written by Ahna et al. [11]. What are the needs and
what aspects can apply machine learning in its applied in the construction world,
such as tower cranes area identification, an inspection of structural joints, prediction
of construction safety accidents, operation of construction lifts, an arrangement of
tower crane layouts? So that the areas that can be accommodated by machine learning
include mitigating risks, identifying safety, and construction planning.
836 C. Tedjo et al.
3 Blockchain Structure in Study
Blockchain in this study is used as data storage and records the approval of the work
method planning that has been approved by several related parties. In its use as data
storage to be uploaded or downloaded will have a hash. Hash on the data will serve
as a process of identifying the authenticity of the data.
3.1 Relations Between Blocks
As supervised data, the relationships between blocks are only additions. This means
that any data recorded and inputted will increase the number of data libraries. The
input data will receive verification and identification with an identity in the form of
a hash which will later function to see the approval time and data input time.
3.2 Data Record Process
Blockchain data recording is done with the IPFS platform. In this case, IPFS will
provide an Identity Document (ID) to the block. IPFS can give ID to any type of file,
but in this research, the file used is a file with the .xlsx extension, because the resource
data and project specifications are inputted in table form to facilitate grouping and
sorting data if needed. The ID given by the IPFS platform will be used by machine
learning to pull new data and download historical data which will be used as a data
train in the form of supervised data.
3.3 Use and Submit Data
Prediction data that has been evaluated and approved will be verified by machine-
learning, which will then be put back into the block as new historical data. The new
data, which already has an ID or hash, will be the train data that machine-learning
will download automatically when machine-learning will predict work methods on
the new project. To be uploaded to the blockchain platform, this data needs to be
evaluated and approved by the relevant parties. In this process, it is not necessarily
mandatory to use blockchain prediction results, but the prediction results can still
be adjusted according to productivity assumptions or other needs. Once approved,
machine learning can upload this data into the blockchain platform to obtain an ID.
4 Development of Machine-Learning Structures
In this research, machine learning acts as artificial intelligence that can identify or
predict based on existing construction data. The process carried out by machine-
learning includes data collection, identification of inputted data, prediction using the
supervised method, and giving suggestions of prediction results to engineers.
4.1 Data Identification
The data table is divided into three parts, namely the test, train, and prediction
sections. The test section functions to test the accuracy of the supervised learning
method. The data used are all historical data, both building, and resource specification
data. Data train is the development of a test data model that is used as a reference for
machine learning to predict input data. In this research, the modeling requires an input
of building and volume specifications to be able to predict the required resources.
Because the data is predictive, evaluation can be carried out by engineers, to adjust
to detailed project requirements that are not contained in the machine-learning input
process.
4.2 Supervised Learning
The data table is arranged in the form of data that can be used by machine-learning
as supervised data, meaning that the data has been grouped into columns that will
become input data and which will become predictive data. In this process, building
specification data is inputted, such as building layers, basement depth, area, and
volume of material. The input data becomes raw data which will be entered in the
historical data table, which will calculate the value of the resource requirements
which will be input data for engineers. Due to the input data structure that has
been adapted to the historical data table structure, machine-learning identifies data
according to a given table pattern without changing the format. With an appropriate
format, the .xlsx file can be connected to other tables in the file, so that the results
of the predicted data will be processed automatically in a visual display so that it is
easier for users to understand.
4.3 Manage Prediction Result Data
After the prediction results and the report are issued, the raw data of the report file
will automatically be entered into the platform block and obtain an ID, to obtain
838 C. Tedjo et al.
authentication and facilitate data tracking. Uploading data to the platform is done
with the .xlsx file extension, also without changing the structure of the files and
tables, because the data is needed to make machine-learning predictions in the next
process. To make it easier to retrieve and group data, machine learning with the flask
platform directly downloads all uploaded data and stores it in a folder in the form of
a pdf report, with the name and time of the creation of the method data.
5 Research Process
This research examines the process of defining work methods and resources in prepa-
ration for the implementation of construction projects. Based on previous data and
designs contained in literature studies, machine-learning enrichment will be carried
out to predict the characteristics of the construction planning model both in terms of
resources and construction methods.
In research, problem identification is made more conical on the machine learning

model plan. What algorithms will be used, as well as how machine learning processes
data on the blockchain.
5.2 Machine Learning Algorithm Selection
Algorithm selection is carried out after the data cleaning process. Data training is
carried out, to obtain training results whose value is close to one. This value shows
the suitability of using the algorithm for the data.
5.3 Machine Learning Model
Machine learning begins to be modeled after the appropriate algorithm is obtained.

Input definition based on the dependent variable is carried out. Defined the output
plan that will be generated by machine learning. Machine-learning development
process show in Fig. 1.
Block-
chain Data
Problem Identi- Literature Study and Algorithm Machine Learning
Start
fication Data Collection Selection Model
No Check the Suitability of Model Logic and Model
Suitable Model Evaluation
Display Program
Yes
Testing Program RQ: Machine learning schemes for No
Suitable
Result defining construction methods
Models
Yes
Finish
Fig. 1 Research process
Table 1 Project
No. Site visit form
identification data form
Data Grade Remark
1 Building type … 1 = Apartment
2 = Office
3 = School
4 = Commercial
area
5 = Smart building
2 Workforce monitoring … Schedule and list of
workers
3 Equipment monitoring … Schedule and
equipment list
4 Construction duration … Month period
5 Contract value … Amount of money
Evaluation and Model Testing. The machine learning output is evaluated, and it is
validated whether the results can be used to develop predictive resources and tools
in the construction process. All machine-learning data variables are inputted into the
form shown in Table 1.
6 Result
Based on the research stages described in the previous section, the prediction results
of machine learning are obtained with the following process.
840 C. Tedjo et al.
Fig. 2 Building specification
The data collected based on the results of the project visits were included in several
tables, including building specification tables, manpower tables, and equipment
usage tables, which are defined in the form of a schedule. Figure 2 explains the
definition of building structure specifications per floor.
6.2 Machine Learning Algorithm Selection
To confirm the accuracy of the prediction results, machine learning needs to process
the data training. Machine learning data training with a supervised learning model
is generally carried out using three analysis methods, namely linear logistic, random
forest, and decision tree, The selection algorithm is shown in the Fig. 3.
Fig. 3 Algorithm selection—Random Forrest, Logistic Regression, Decision Tree (selected)

Fig. 4 Storing, collecting, and interacting data in IPFS and machine learning result
6.3 Machine Learning Result
The process of managing input data and predictive data from machine learning is
carried out in blockchain-based clouds using the IPFS platform. Data that has been
managed and predicted is displayed in the form of a schedule table as follows Fig. 4.
7 Conclusion
Based on the formulation of the problem, research objectives, and the results of
machine learning enrichment trials based on python programming and block-chain
cloud IPFS, a research conclusion is obtained where if a new management system is
developed using machine-learning technology, which has a detailed database and is
well verified with the system blockchain, it can produce a construction method that
is more efficient and faster and can be applied, so that faster construction method
designs can be produced. By using a supervised machine-learning model, with a
data management system using blockchain, the results of machine learning are even
applied in one of the building construction projects, because they match the project
overview and evaluation of experts. This development can be particularly useful
in implementing smart building construction in Indonesia. However, this research
cannot manage the waste from the construction method used. It is hoped that later on,
resource management can also be done using machine learning. Moreover, to support
further research, it is necessary to try to also examine the process of enriching machine
learning in infrastructure projects and the natural resource industry in Indonesia.
References
1. Oti AH, Tah JH, Abanda FH (2018) Integration of lessons learned knowledge in building
information modeling
842 C. Tedjo et al.
2. Jian Sun P, Xu Ren P, Chimay J, Anumba PD (2019) Analysis of knowledge-transfer

mechanisms in construction project cooperation networks
3. Dakhli Z, Lafhaj Z, Mossman A (2019) The Potential of Blockchain in Building Construction
- Civil Engineering, Ecole Centrale de Lille, 59651 Villeneuve d’Ascq, France
4. Dakhli Z, Lafhaj Z, Mossman A (2019) The potential of blockchain in building construction
5. Ellis RD, Thomas HR (2017) Construction site management and labor productivity improve-
ment
6. Joud Al Dakheel CD (2020) Smart buildings features and key performance indicators: a review
7. Keiser J (2012) Leadership and cultural change: necessary components of a lean transformation
8. Rasoloharijaona S, Rakotosamimanana B, Randrianambinina B, Zimmermann E (2003) Pair-
specific usage of sleeping sites and their implications for social organization in a nocturnal
Malagasy primate, the Milne Edwards’ sportive lemur
9. Lee D, Lim H, Kim T, Cho H, Kang K-I (2018) Advanced planning model of formwork layout
for productivity improvement
10. Ahmed Gondia S, Siam A, Wael El-Dakhakhni F, Nassar AA (2019) Machine learning
algorithms for construction projects delay risk prediction
11. Ahn H, Lee D, Lee S, Kim T, Cho H, Kang K-I (2018) Application of machine learning
technology for construction site
The Development of Blockchain Based
Knowledge Management System Model
at EPC Projects to Improve Project Time
Performance
D. Y. Priyambodo, M. A. Berawi, and M. Sari
Abstract EPC project is one of complex construction project types which differs
from ordinary construction projects. It owns more frequent delayed occurrence than
ordinary construction projects. EPC projects tends to be labor and knowledge inten-
sive. Inputs and outputs at EPC project management involve various knowledge
and intellectual creative activities. Thus, the EPC project management process is a
knowledge management process. Knowledge Management System (KMS) aims to
improve the quality of human resources in an organization by improving commu-
nication among all parts of the organization and increasing knowledge mastery by
transferring knowledge. The application of KMS with a blockchain-based system is
expected to be a solution for the needs of distributed knowledge transfer, high-speed
transfer of knowledge and information, fast access to knowledge and high informa-
tion security. This study uses the Tiwana model framework, and for organizational
knowledge mapping uses the Zack KM model cycle, while the SECI model is used for
knowledge formation. The quality of the resulting blockchain-based KMS platform
is tested by using expert validation. The research result shows that blockchain-based
KMS model can be well implemented with excellent validation so it becomes a
solution to increase knowledge and to improve project time performance in EPC-X
project.
Keywords Knowledge management system · EPC project · Blockchain
D. Y. Priyambodo (B) · M. A. Berawi · M. Sari

Civil Engineering, University of Indonesia, Jakarta, Indonesia
M. A. Berawi
M. Sari
844 D. Y. Priyambodo et al.
1 Introduction
The era of globalization era is indicated by rapid development of science and tech-
nology as well as a high and dynamic level of competition. This condition requires
companies to innovate and to improve their performance. Due to the high-competitive
competition, it is necessary to change the company’s paradigm from being resource-
based to knowledge-based, which is based on managing data, information, and
knowledge. The condition also affects the construction industry [1]. Knowledge
management is a way to improve company capabilities by managing the knowledge
they have. EPC projects tend to be labor-intensive and knowledge-intensive. Inputs
and outputs in EPC project management involve various knowledge and intellec-
tual creative activities. The requirement for knowledge to be accessed collectively
among different project phases, stakeholders, and management functions still poses
as a challenge to be solved [1–3].
At the present era, the implementation of Knowledge Management processes
at both organizational and national levels require an innovation. Many researches
have been done to implement knowledge management in organizations, but from the
technological point of view, it is still possible to discuss the processes of knowledge
management in a fully transparent and high level style throughout the organization.
The problems of organizational knowledge storage, information, and knowledge
safety, sharing knowledge, knowledge access, intellectual property rights of scholars
are some of the challenges faced by a traditional knowledge management. Blockchain
technology is one of the IT-based tools which has a high level of innovation, which
can be used to implement knowledge management at the enterprise level or even at
the national level. It enables high speed of knowledge and information transfer, fast
access to knowledge, and high information security. When information is stored in
Blockchain, it is immutable. This design allows Blockchain to be capable of having
a permanent historical record [4].
2 Theoretical Study
2.1 Knowledge Management
Tiwana defines knowledge management broadly in the sense of managing knowledge

as “… management of organizational knowledge for creating business value and
generating a competitive advantage”. Knowledge Management provides the ability to
create, communicate and apply knowledge necessary and useful for the achievement
of all types of business goals. Tiwana also explains “Knowledge management is
the ability to create and retain greater value from core business competencies” [5].
Effective knowledge management must be started by using a strategic look at what
knowledge the organization needs. This is to ensure that the KMS will be developed
is in accordance with what is needed and the core competency of the organization.
The Development of Blockchain Based Knowledge Management … 845
This alignment can be based on the needs and the core by determining what strategies
are or will be implemented and what the organization has implemented [6, 7].
2.2 Blockchain Technology
Blockchain can act as a distributed, programmable, and encrypted database to

transfer, protect, store, and quickly access knowledge from one location to another
by using high security [8]. Blockchain also provides the potential for scientists to
share their unique knowledge with other users. These guaranteed transfers are safe,
and nothing can challenge their legality. The implementation of Blockchain’s unique
features include high reliability, privacy, accountability and data sharing, information
disorganization, responding to traditional knowledge management challenges [4].
Blockchain has become a revolutionary paradigm in record keeping systems and
has been seen as an emerging technology in the industry and research community.
Through its immutable property, this design allows blockchain to have permanent
historical records. Knowledge Management is based on management trust. As such,
blockchain trust properties can be used to manage trust in generating, storing, sharing,
protecting, and applying various knowledge [9].
3 Research Method
This research uses a qualitative descriptive approach through case studies. The
method of data collection is done by doings of interviews, historical data, and docu-
mentation. Amrit Tiwana’s 4 phased roadmap for knowledge management method is
employed for the data analysis process in this study, accompanied with some adjust-
ments. The infrastructure evaluation stage is supported by observation data, inter-
views, and document review. The system analysis and design stage include database
design, system architecture, and blockchain-based knowledge management system
interfaces using an object-oriented approach with expert validation. The expert vali-
dation related to the blockchain-based Knowledge Management System model was
carried out by 2 blockchain technology experts and 1 knowledge management expert
on the EPC project (Fig. 1).
Start Literature Data Selection & Assigning

Review Collection Case Studies
Needs Analysis, Strategies and Creating Infrastructure Design &

Knowledge Management System System Architecture and KMS
Scenario on the EPC Project Team Design
Implementation of the Blockchain

Expert
development of a based KMS End
Validation
Blockchain KMS Model Model
Fig. 1 Research stages
4 Result and Analysis
4.1 Knowledge Management System Scenario Analysis
The process of creating organizational knowledge occurs because of the interac-

tion or conversion between tacit knowledge and explicit knowledge through the
processes of socialization, externalization, combination, and internalization, which
are SECI Nonaka’s model. Every activity in the EPC-X Project to create and formu-
late knowledge needs to be carried out in a knowledge management process cycle
which consists of 4 stages in accordance with the SECI Nonaka model. Displayed
below is the relationship between the order to create and formulate knowledge
(Table 1).
4.2 Infrastructure Design and Architecture System
In developing a knowledge management system prototype, a knowledge management

system architecture design is carried out. The prototype was developed by using
a blockchain-based system. The prototype runs on the existing infrastructure and
becomes a separate module from other existing information systems at EPC-X Project
(Fig. 2).
The knowledge storage mechanism on the blockchain enables to store knowl-
edge, change the knowledge stored within, and get knowledge from inside the
blockchain. The knowledge structure on the blockchain will be adjusted to the knowl-
edge management structure. The implemented blockchain system has a feature to
record every process of the storage mechanism, add knowledge management to the
blockchain, and distribute knowledge to all blockchain members.
Table 1 Relationship between knowledge management activities and processes

No EPC-X project Knowledge SECI model stages Blockchain features
activities management
process
Project Management
1 Project contract Document Socialization, Smart contract,
sharing, Externalization Decentralized,
Discussion Immutable
2 Scope of work EPC Document Socialization, Smart contract,
project sharing, Externalization, Decentralized,
Discussion, Internalization Consensus
Sharing mechanism,
experience, Transparency
Workshop
3 Change order Document Socialization, Smart contract,
sharing, Externalization, Decentralized,
Discussion, Combination, Consensus
Sharing Internalization mechanism,
experience, Immutable
Workshop, Case
study
Engineering
4 Basic engineering Document Socialization, Smart contract,
Sharing mechanism,
Sharing idea,
Workshop
5 Detail engineering Document Socialization, Smart contract,
Discussion, Combination, Consensus
Sharing Internalization mechanism
experience,
Workshop
6 MTO and BQ Document Socialization, Smart contract,
Sharing mechanism, High
experience, Security
Workshop
7 Construction drawing Document Socialization, Smart contract,
and specification sharing, Externalization, Decentralized,
Sharing mechanism,
Workshop
(continued)
Table 1 (continued)
process
Procurement
8 Procurement plan Discussion, Combination, Smart contract,
Sharing Internalization Decentralized,
experience, Transparency, High
Internet security
browsing
Vendor
9 Vendor Document Combination, Decentralized,
databased—Product sharing, Sharing Internalization Transparency, High
Information experience, security
Internet
browsing
Material Control
10 Material tracking Discussion, Socialization, Decentralized,
Workshop, Combination, Transparency, High
Sharing Internalization security, Immutable
experience
11 Material storage Discussion, Socialization, Decentralized,
Workshop, Combination, Transparency, High
Sharing Internalization security, Immutable
experience
Fabrication—Construction
12 Fabrication procedure Workshop, Socialization, Smart contract,
Sharing Combination, Decentralized,
experience Internalization Consensus
mechanism
13 Construction Workshop, Socialization, Smart contract,
procedure Sharing Combination, Decentralized,
mechanism
14 QA-QC procedure Workshop, Socialization, Smart contract,
Sharing Combination, Decentralized,
mechanism
15 Completion certificate Document Combination Smart contract,
sharing Decentralized,
Immutable
Commissioning
16 Commissioning Workshop, Socialization, Smart contract,
procedure Sharing Combination, Decentralized,
mechanism
(continued)
Table 1 (continued)
process
17 Performance test Workshop, Socialization, Consensus
Sharing Combination, mechanism,
experience Internalization Transparency,
Decentralized,
Immutable
Fig. 2 Blockchain based knowledge management architecture
4.3 Knowledge Management Team Design
The design of the knowledge management team involved in the KMS application for
the EPC project is tailored to the existing users. Blockchain based KMS managers
consist of experts and other peers.
4.4 Knowledge Management Model
From a series of analyses based on the Tiwana framework, a blockchain based

knowledge management system model can be created as follows in Fig. 3.
The blockchain based Knowledge Management System model for the EPC-
X project consists of a data repository containing knowledge, files, and discus-
sion summaries. Then this data is verified by peers comprised of experts, content
management, consensus, and search modules.
Step 3 Step 5
Step 1
(User Earn Points) (Comparing Knowledge)
(Input Form)
Nodes earn rewards The system will provide
The system creates
from the verification recommendation for 3
blocks as transaction
process and add blocks to priorities of knowledge
representations
the blockchain according to keywords
Step 4 (Viewing and

Step 2 (Nodes Vali-
Selecting Knowledge)
date Transactions)
The system provides
The system validates
recommendation for
knowledge through smart
knowledge according to
contracts
the input keywords
Fig. 3 Blockchain based KMS model
Table 2 Entity relationship diagram blockchain-based KMS

Key partners and Key activities Key resources Cost structure
customer segments
1. EPC project contractor 1. Knowledge gathering 1. Blockchain 1. Platform
2 EPC project owner and processing 2. Big data development
3 EPC project planning 2. Platform development 3. Knowledge 2. Production cost
and supervisory verification 3. Peers acquisition
consultant cost
4 Cryptoasset service 4. Overhead cost
provider Value propositions Customer Channels
relationships
Fast, secure and 1. User friendly 1. Service excellence
distributed solution for website 2. Advertising
storing project data 2. Points as gifts 3. Media release
4.5 Implementation
Based on the needs of the system, the design of the entity relationship diagram (ERD)
for the KMS prototype. The EPC project consists of entities that are interrelated
within the system, the ERD diagram is as follows in Table 2.
Table 3 Expert validation

Validation Based on the Model Framework above, can the Model proposed by
researchers be implemented in the EPC Project? Explain Your Opinion
Blockchain expert 1 Yes. The knowledge management model proposed by researchers is very
likely to be applied in EPC projects that involve many parties
Blockchain expert 2 Yes, but it needs a little improvement on the business model
KM expert Theoretically correct, overall it will depend on the type of project and the
level of complexity but basically has the same project management flow
5 Discussion
From the Knowledge Management System model that has been created by researchers
and validated by Blockchain and Knowledge Management experts, the following
results are obtained in Table 3.
Qualitative data Analysis using Nvivo software is carried out during the
expert validation process by “identifying sentiment”. The analysis reveals that
the Blockchain based Knowledge Management System Model in the EPC Project
returns “Moderately Positive” and “Very Positive” results of 44.19% and 55.81%
respectively.
6 Conclusion
Blockchain technology implemented in the Knowledge Management System of the

EPC Project returns a “Moderately Positive” and “Very Positive” results of 44.19 and
55.81%. Analysis employing the Tiwana framework reveals designing and imple-
menting blockchain-based KMS models is able to produce prototypes that can be
applied to support the implementation of EPC projects, thereby improving knowl-
edge management and sharing among EPC project stakeholders. The developed
model carries the potential to be applied to several other projects outside of the EPC
project. In addition, Blockchain’s unique features are expected to improve project
time performance.
Acknowledgements The authors would like to thank University of Indonesia for the support given
to this research.
References
1. Zhu F, Sun X, Xu Z, Haider Z (2014) A knowledge integration framework of EPC project based
on KBS and stakeholder networks. Int J Innov Comput Inf Control 10(2)
2. Mee San K, Choy CF, Fung WP (2019) The potential and impacts of blockchain technology in
construction industry: a literature review. IOP Conf Ser: Mater Sci Eng
3. Turk Z, Klinc R (2017) Potentials of blockchain technology for construction management. In:
Creative construction conference 2017, CCC 2017, 19–22 June 2017
4. Akhavan P, Philsoophian M, Rajabion L, Namvar M (2018) Developing a blockchain knowledge
management model (BCKMM) beyond traditional knowledge management. In: 19th European
conference on knowledge management ECKM 2018
5. Tiwana A (1999) The knowledge management toolkit: practical techniques for building a
knowledge management system, 1st edn. Prentice Hall, New Jersey
6. Nazori A, Samdi, Wirda, Hilwa (2014) Prototipe knowledge management system (KMS) Prak-
tikum Perbankan Syariah Berbasis Web dan Android dengan Kerangka Kerja “Tiwana” pada
Perguruan Tinggi: Studi Kasus STEI SEBI. Jurnal Telematika MKOM 6(2)
7. Zack MH (1999) Developing a knowledge strategy, vol 41, no 3. California Management Review.
Nyame G, Qin Z, Kwame OBOA, Emmanuel BS (2020) An ECDSA approach to access control
in knowledge management systems using blockchain. MDPI J
8. Tian F (2016) An agri-food supply chain traceability system for China based on RFID &
blockchain technology. In: 13th International conference on service systems and service
management (ICSSSM), pp 1–6
9. Zheng Z, Xie S, Dai HN (2018) Blockchain challenges and opportunities: a survey. Int J Web
Grid Serv 14(4):352–375
Risk Allocation Implementation Analysis
of Public-Private Partnership
for Infrastructure Project (Case Study
of the Solo-Yogyakarta-NYIA Kulon
Progo Highway Project)
Widi Hartono, Aloysia Putri Hastari Purnomo, and Sunarmasto
Abstract Indonesia is now concentrating on infrastructure development; the

construction of roads, bridges, airports, dams, and toll roads are seemingly given
more attention in comparison to others. As the infrastructure sector rapidly grows, it
opens wider opportunities for investors. This study aims to identify and monitor the
risk allocation process and preferences along with risk response(s) applied by related
parties due to critical risk variables that might have happened in highway projects—
especially Public–Private Partnership based infrastructure projects situated in Central
Java, Indonesia; the Solo—Yogyakarta—NYIA Kulon Progo Highway Project. The
result from collected questionaries—processed with descriptive analysis—shows
that: (1) 15% of the risk variables are allocated to the Ministry of Public Works
and Housing (the Government), (2) 50% are allocated to private parties, and (3) 35%
of the variables are shared between both public and private parties. Using the Risk
Failure Mode and Effect Analysis (RFMEA), Risk Score and Risk Priority Number
of each risk variable is compared, resulting in the knowledge that 7 out of 60 risk
variables related to the project completion experienced: delays, budgetary and work
schedule amendment, project supervision and control, and risk events that occur due
to force majeure. Those are 7 of critical risks that would need to be responded or
mitigated.
Keywords Risk · Public-Private Partnership · Highway · Risk Failure Mode and

Effect
W. Hartono (B) · A. P. H. Purnomo · Sunarmasto

Civil Engineering Department, Sebelas Maret University, Surakarta, Indonesia
Sunarmasto
854 W. Hartono et al.
1 Introduction
Every construction project faces triple constraints (cost, time, and quality) as its main
project success indicators. Besides, there is no risk-free project; there are plenty of
risk possibilities that might affect project success. Well-planned risk management
is much needed for anticipating harmful risks and also to predict various risks that
significantly affect or are supposed to inhibit project success as soon as possible.
Avoidance may seem the answer to all risks, but avoiding risks also means missing
the potential gain that by accepting or retaining the risk may have allowed (Sichone
2002) [1].
Risk allocation is one of the tipping points to start risk management, especially
for Public-Private Partnership (PPP) infrastructure projects, which are conducted not
only by the government or private parties alone but both parties involved. Risks should
be allocated as precisely as possible, as PPP projects serve a wide range of goals [2].
The Public–Private Partnership process demands not only carefulness but also extra
transparency in allocating risks, as it shows not only the project’s performance but
the government’s performance as well.
This study aims to get to know the risk allocation scheme and process in a highway
project based on its Public–Private Partnership and to understand responses taken by
related parties due to critical risk variables which might occur in highway projects,
especially the Public–Private Partnership-based infrastructure project located in
Central Java, Indonesia; the Solo—Yogyakarta—NYIA Kulon Progo Highway
Project.
2 Review of Literature
In Architectural Construction and Engineering (ACE) projects, the risk allocation of

related parties has specific mechanisms, clause types, payment methods, and respon-
sibilities [1]. The risk allocation process in a PPP project always has a significant
impact on the whole project; risk allocation implementation can only be successful
when it is optimally implemented between related parties [3]. Maharani et al. [4]
pointed out that in a PPP project, risk allocation strategies do not only define which
parties are most capable of accepting risks, they also define the best timing, and
the most suitable risk solution—all of which need to be put in consideration to
achieve optimal risk allocation. To optimize the risk allocation process, Wang et al.
[5] suggested that the government has a significant role in PPP projects; he claimed
that a better government environment leads to better and more effective regula-
tion, thus attracting more powerful investors; posing less risk, increasing profitable
potentials.
In technical terms, Nasirzadeh et al. [6] stated that if risk allocation in a construc-
tion project is not done properly, the contractor party will suffer huge losses. This
may result in a project cost overrun, as contractors have to use defensive strategies to
Risk Allocation Implementation Analysis of Public-Private … 855
keep work schedules and goals on track and against one-sided risk allocation. In 2011
Rosadin [7] researched PPP Optimization, while in 2015, Nurfarida [8] researched
PPP for Performance-Based Contract (PBC) infrastructure projects.
To guide all parties related to risk allocation—especially for the PPP projects in
all sectors. The World Bank has issued a Public–Private Partnership Reference Guide
3.0 [9]. This guidebook was adapted by the Indonesian Government; they published
PPP Risk Allocation Tool 2019 Edition [9], and Acuan Alokasi Risiko 2020 [10].
These handbooks are part of the references used for this research.
3 Methodology
3.1 Data
Collected data from questionnaires are processed with descriptive statistics to find
the percentage of risk allocation preference from (1) the government as the owner
of the project, and (2) private parties as project initiators and contractors; a risk is
allocated to certain parties based on popular votes by participants.
Risk Failure Mode and Effect Analysis (RFMEA) compares the Risk Score and
Risk Priority Number with the critical value. Risk Score (RS) and Risk Priority
Number (RPN) values are processed by multiplying the probability, severity, and
detection score of every determined risk variable. The calculation is based on the
following equations [11]:
RPN = O × S × D (1)
Risk Scor e = O × S (2)
where O represents the Occurrence (Probability) value, S represents the Severity

(Impact) value and D represents the Detection value.
And to define critical value for RPN, the calculation is based on the following
equation:
ε RPN
Critical Value RPN = (3)
N cases
To add, to determine the RS’s critical value, this following risk level matrix is
used (Table 1).
Table 1 Risk level matrix

Occurrence Severity
1 2 3 4 5
5 Medium Medium High High High
4 Low Medium Medium High High
3 Low Medium Medium Medium High
2 Low Low Medium Medium Medium
1 Low Low Low Medium Medium
On Risk Failure Mode and Effect Analysis, critical risks are determined by
comparing Risk Score and Risk Priority Number with the critical value. If the RS
and or RPN value of a particular risk variable is equal to or more than the critical
value, that indicates that the risk variable is categorized as a critical risk that needs
to be responded at the early stage of the project. This method focuses on examining
risk variables based on its detection mode effectiveness; one is stated critical as it
has a less effective detection mode [11].
3.2 Risk Variables Identification
In this research, risk variables are determined based on risk factors from supporting
research and risk matrix compiled by PT. Penjaminan Infrastruktur Indonesia
(Persero)—which are adapted into research objectives. Risk variables used in this
research are identified to analyze how risk allocation between related parties is done
concerning government regulation, work scheme, construction process, operational
policy, and force majeure issues. There are 60 risk variables chosen, classified into
6 groups;
• V1-V12 (Risk variables related to contract and regulation)
• V13-V23 (Risk variables related to location and land acquisition)
• V24-V30 (Risk variables related to design issues)
• V31-V46 (Risk variables related to construction process)
• V46-56 (Risk variables related to operational policy)
• V57-V60 (Risk variables related to force majeure issues)
4 Case Study Background
Solo-Yogyakarta-New Yogyakarta International Airport (NYIA) Kulon Progo

Highway Project is considered as one of the National Strategic Projects (Proyek
Strategis Nasional) that would be part of Trans Java Highways. It is planned to be
the fastest route to connect three international airports from Yogyakarta to Central
Java. This highway sets out to be the connector between 3 big cities, the golden
triangle of Yogyakarta and Central Java; Semarang, Solo, and Yogyakarta.
This project’s construction is set out to be built in gradual steps and should be fully
operated by the end of 2024. There will be three location markers to its construc-
tion; starting from Kartasura Region (as a follow-up route from the Solo-Semarang
Highway) to Purwomartani, Yogyakarta as the first marker, then from Purwomar-
tani to Gamping, Yogyakarta as the second marker (where most of the construction
designed to be an elevated road), and the last section would start from Gamping and
lead to Kulon Progo, Bantul Regency, Yogyakarta.
As a Public-Private Partnership-based Infrastructure Project, this project plans
to have a 40 years long concession time, and costs nearly Rp 15.81 trillion (USD 1
billion) for its construction, and Rp 26,637 trillion (USD 1850 billion) for investments
value. Financially, this project is estimated to have USD 2.104 million of NPV
(Net Project Value) and 12% of IRR (Internal Rate of Return) (simpulkpbu.com).
This PPP project was proposed by PT. Gama Group; PT. Daya Mulia Turangga
as a consortium, together with PT. Adhi Karya (Persero) Tbk. and PT. Jasa Marga
(Persero) who appointed PT. Jogjasolo Marga Makmur as construction and operation
provider (Fig. 1).
Fig. 1 Planned route of Solo-Yogyakarta-NYIA Kulon Progo Highways (courtesy of PT. Jogjasolo
Marga Makmur)
5 Analysis
There were 30 respondents who have given their perspectives; respondents chosen
were from the Ministry of Public Works and Public Housing (of involved regions)
and private parties. Detailed information of respondents is presented in Table 2.
To process risk allocation implementation of the Solo-Yogyakarta-NYIA Kulon
Progo Highway Project, collected data are processed with descriptive statistics to
figure out the related parties’ preferences on how risks have or should be allocated.
Risk allocation analysis can be observed in the following table.
As shown in Table 3, 15% of risk or 9 out of 60 risk variables are allocated to the
public party, 50% of risks or 30 of them are allocated to private parties in partnership
with the Government (investors), and 35% of risks or about 21 risk variables are
shared between related parties (public and private).
To figure out which variables included in critical risk need to be responded to,
Risk Failure Mode and Effect Analysis was used. Participants was asked to score
risk variables Occurrence, Severity and Detection Methods from 1 to 5 (mild to very
serious). Analyzed data is shown in Table 4.
As shown in Table 4, 7 out of 60 risk variables are critical. Those are the ones
that need to be responded to as soon as possible. Those variables are classified as
critical because their RPN and RS values are higher than their critical values; those
determined as High-Risk Variables are RPN ≥ 26, RS ≥ 10, which is an occurrence
where high probability and impact severity is unsupported by suitable nor applicable
detection method.
Those critical variables found are related to the occurrence of delays (V40, V35,
V12), quality controls (V39), budgetary and work schedule changes (V27, V28),
and natural disaster issues (V57). Responds taken by related parties are risk transfer,
avoidance, risk sharing, and acceptance.
Table 2 Detailed
Respondents Quantity
information of respondents
Public Works Department of Central Java 3
Public Works Department of Klaten Regency 6
PT. Jogja Solo Marga Makmur 3
PT. Trans Marga Jateng 1
Public Works Department of Boyolali Regency 6
Public Works Department of Karanganyar Regency 6
PT. Jasamarga Solo-Ngawi 5
Table 3 Risk allocation analysis

Code Risk variables Risk allocation (%) Risk allocation
Public Private Shared
V1 Regulation and tax changes 60.00 20.00 20.00 Public
(Generally or specifically)
V2 Design approval delay 10.00 60,00 30.00 Private
V3 Project permission and approval 13.30 50.00 36.70 Private
delay
V4 Delays in location access 13.30 63.40 23.30 Private
permission
V5 Parastatal risk 16.60 56.70 26.70 Private
V6 Competitor risk 16.70 63.30 20.00 Private
V7 Non-toll network management 50.00 33.30 16.70 Public
V8 Road network and transportation 50.00 16.70 33.30 Public
connectivity
V9 Government support delay 30.00 33.30 36.70 Public
(subsidies, incentive, tax)
V10 Unorganized decision-making 26.70 23.30 50.00 Shared
process
V11 Expropriation risk 16.60 26.70 56.70 Shared
V12 Land acquisition delay and land 50.00 10.00 40.00 Public
price increase
V13 Land couldn’t be acquisitioned 53.30 10,00 36.70 Public
V14 Land couldn’t be used after 50.00 23.30 26.70 Public
acquisition process
V15 Obscurity of land status 66.70 13.30 20.00 Public
V16 Unanticipated location condition 23.30 33.30 43.40 Shared
V17 Complex resettlement process 46.70 16.60 36.70 Public
V18 Affected artifacts and antiquities 23.30 20.00 56.70 Shared
V19 Soil structure risk 20.00 43.30 36,70 Private
V20 Contaminated environment 10.00 43.30 46.70 Private
V21 Damaged conservational area 20.00 20.00 60.00 Shared
V22 Hindered access of public 20.00 23.30 56.70 Shared
transportation
V23 Disturbed public security and 13.30 36.70 50,00 Shared
convenience
V24 Project scope accuracy 10.00 80.00 10.00 Private
V25 Resources’ specification (tools, 6.70 80.00 13.30 Private
materials, human resources)
V26 Technology implementation 6.60 76.70 16.70 Private
V27 Project budgetary changes 10.00 43.30 46.70 Private
(continued)
Table 3 (continued)
V28 Work schedule changes 0.00 60.00 40.00 Private
V29 Undefined output specification 3.30 40.00 56.70 Shared
V30 Design correction or adjustment 3.30 33.30 63.40 Shared
V31 Project scope change after 0.00 30.00 70.00 Shared
contract signing
V32 Construction cost increase 10.00 56.70 33.30 Private
V33 contractor/sub-contractor bad 13.30 80.00 6.70 Private
performance
V34 Resources’ availability 0.00 90.00 10.00 Private
V35 Resources’ stock delays 0.00 90.00 10.00 Private
V36 Work methods discrepancy or 0.00 90.00 10.00 Private
defective tool
V37 Unpredicted geotechnical 3.30 60.00 36.70 Private
condition
V38 Unapplicable design 1000 40.00 50.00 Shared
V39 Quality control and work target 13.30 60.00 26.70 Private
control
V40 Inhibition of completion 0.00 70.00 30.00 Private
V41 Unfinished project by 3.30 76.70 20.00 Private
contractor/subcontractor error
V42 Security and safety issues 3.30 63.40 3330 Private
V43 Testing and commissioning risk 330 63.40 33.30 Private
V44 Divergent standards 3.30 60.00 36.70 Private
V45 Missed communication 3.30 40.00 56.70 Shared
V46 Project management failure 6.70 53.30 40.00 Private
V47 Facilities availability 6.60 66.70 26.70 Private
V48 Industrial action 60.70 43.30 50.00 Shared
V49 Socio-culture conflict 60.70 20.00 73.30 Shared
V50 Project control and monitoring 30.30 43.30 53.40 Shared
default
V51 Life cycle estimation error 60.60 46.70 46.70 Private
V52 Inefficient operation performance 16.60 56.70 26.70 Private
V53 Irregular facilities availability 13.30 60.00 26.70 Private
V54 Toll technology and information 26.70 46.70 26.60 Private)
system failure
V55 Accidents due to design error 20.00 30.00 50.00 Shared
(continued)
Table 3 (continued)
V56 Failure by accident 13.30 33.30 53.40 Shared
V57 Natural disasters 16.70 13.30 70.00 Shared
V58 Politic force majeure (war, 36.70 16.60 46.70 Shared
demonstration)
V59 Extreme weather 33.30 23.30 43.40 Shared
V60 Long drawn force majeure 20.00 16.70 63.30 Shared
Table 4 Critical risk analyzed with RFMEA

Code Risk variables O S D RPN RS
V40 Inhibition of completion 4.87 3.87 2,60 48.93 18.82
V39 Quality control and work target control 4.97 3.57 3.53 62.59 17.71
V28 Work schedule changes 3.57 4.53 2.53 40.96 16.17
V35 Resources’ stock delays 3.17 4.63 2.83 41.57 14.67
V57 Natural disasters 3.63 3.57 3.73 48.38 12.96
V12 Land acquisition delay and land price increase 3.77 3.67 2.73 37.75 13.81
V27 Project budgetary changes 3.57 3.67 2.77 36.18 13.08
6 Conclusion
It is known from the research that in an unsolicited PPP infrastructure project, private
parties are being responsible for most of the risk as their role as investor, designer,
contractor, and operational agent. As it is a joint cooperation of private and Ministry
of Public Works and Housing, Ministry of Public Works and Housing (government)
is giving their support as they processed supportive regulation, working on land
acquisition, and general public conditioning.
Those critical variables found are related to the occurrence of delays (V40, V35,
V12), quality controls (V39), budgetary and work schedule changes (V27, V28), and
natural disaster issues (V57). The responses taken by related parties are risk transfer,
avoidance, risk-sharing, and acceptance.
References
1. Tembo-Silungwe CK, Khatleli N (2017) A Conceptual model for risk allocation in the
construction industry. Am J Appl Sci 14(7):690–700. https://doi.org/10.3844/ajassp.2017.
690.700
2. Subprasom K, Chen A (2005) Analysis of policy and regulation on build-operatetransfer

scheme: a case study of the ban Pongkanchanaburi motorway in Thailand. J Eastern Asia
Soc Trans Stud 6(January):3883–3898
3. Valipour A, Mohammadi F, Yahaya N, Sawvari H, Noor NM (2014) Identification and evalua-
tion of risk allocation criteria and barriers: a Malaysian public private partnership project case
study. J Appl Sci 14(18):2023–2031.
4. Maharani Y, Palupie R, Yuniarto HA (2016) Alokasi Risiko Proyek Infrastruktur Dengan Skema
Kerjasama Pemerintah dan Badan Usaha (KPBU): Suatu Tinjauan Literatur. Seminar Nasional
Teknik Industri Universitas Gajah Mada 96–103
5. Wang Y, Gao HO, Liu (2019) Incentive game of investor speculation in PPP highway
projects based on the government minimum revenue guarantee. Trans Res Part A Policy Pract
125(15):20–34. https://doi.org/10.1016/j.tra.2019.05.006
6. Nasirzadeh F, Khanzadi M, Rezaie M (2014) Dynamic modeling of the quantitative risk allo-
cation in construction projects. Int J Project Manage 32(3):442–451. https://doi.org/10.1016/j.
ijproman.2013.06.002
7. Rosadin IM (2011) Optimalisasi Alokasi Risiko dalam Pembangunan Jalan Tol Studi Kasus:
Jalan Tol Bandara Juanda -Tanjung Perak. Universitas Indonesia, Jakarta
8. Nurfarida S (2015) Alokasi risiko proyek pembangunan jalan dengan sistem Performance
Based Contract. Institut Teknologi Sepuluh November, Surabaya
9. The World Bank (2017) Public-Private partnerships reference guide. PPP Knowl Lab Library
3:238. http://doi.wiley.com/10.1002/tie.5060290205
10. PT (2020) Penjaminan Infrastruktur Indonesia. Acuan PT. Penjaminan Infrastruktur Indonesia.
(2020) Acuan Alokasi Risiko 2020. JakartaAlokasi Risiko 2020. Jakarta
11. Carbone TA, Corporation FS (2004) Project risk management using the project risk FMEA.
Eng Manage J 16(4)
Risk Assessment of Construction Project
Scheduling
Zetta Rasullia Kamandang
Abstract To perform a successful construction project, accuracy in managing

project risks has to be fulfilled by project parties. Regarding time-related risk, failure
to estimate project construction duration is one of the common problems that can
affect the entire project performance during the execution. To avoid the impact of
these risks, a scheduling method using probabilistic duration has been developed,
namely Monte Carlo simulation. In this study, optimistic, most likely, and pessimistic
duration of each task and task relationship information of a real case study were
collected and were examined in several steps, (1) determining the number of itera-
tions, (2) Monte Carlo simulation, (3) scheduling analysis, and (4) probabilistic of
success analysis. Based on the calculation, the success probabilities of as-planned,
optimistic, most likely, pessimistic, and simulated durations are 0.479%, 0.054%,
2.108%, 57.912%, and 52.575% respectively. By only have 0.479% of success prob-
ability, the as-planned duration is unsuited to be submitted as project completion
duration in the contract. Furthermore, to avoid risk in scheduling, the simulation
suggests the project parties to allocate at least 332 days of project duration which
has a 100% of success probability.
Keywords Project scheduling · Monte Carlo simulation · Success probability
1 Introduction
Construction project is a part of the country’s development which is executed to

expand the infrastructure and building facilities. In its implementation, a construc-
tion project may be associated with various potential risks that can harm the project
parties and lead to loss of the project’s values [1, 2]. In literature, construction
project risks are divided into several categories. International Construction Risk
Assessment Model (ICRAM) [3] proposed three major categories of risk; risk at a
Z. R. Kamandang (B)
University of Pembangunan Nasional ‘Veteran’ Jawa Timur, Jl. Rungkut Madya No.1,
Gunung Anyar, Surabaya, Jawa Timur 60294, Indonesia
864 Z. R. Kamandang
macro level, risk in construction market, and risk at project level. Skorupka [4] also
divided risk that is linked to the stage of the construction industry investment process
including preliminary design, tender, detailed design, construction, and comple-
tion of payment. Another research [5] provided risk approaches into cost-related,
time-related, quality-related, environment-related, and safety-related aspects. For
example, the risks associated with time are unsuitable construction program planning,
inadequate program scheduling, tight project schedule, design variations, excessive
approval procedures in administrative government departments, variations by the
client, and high performance or quality expectations.
To perform a successful construction project, accuracy in managing project risks
has to be fulfilled by project parties. Risk management (RM) can be described as a
process that helps project parties to control arising risk in the project without affecting
time, cost, duration, and scope of the project. RM assists to obtain better control over
project scheduling, estimation, and quality [6], it also gives project parties beforehand
ideas about future situations that may occur and make them prepare a decision as
early as possible to avoid any risk later.
Regarding time-related risk, failure to estimate project construction duration is
one of the common problems that can affect the entire project performance during its
execution. The total project duration is one of the essential indicators representing
the efficiency of resources management to accomplish the results with the proper
quality standards [7]. Therefore, an assessment of construction project scheduling
has to be done precisely to reduce any loss possibilities.
2 Case Study
This study uses a real case project in Indonesia with more than Rp. 81 billion of the
project’s contract to assess its risk of duration estimate at the planning stage as a
case study. The project is a six-floors building constructed in Bengkulu, Indonesia,
consisting of several works: structure work, architecture work, finishing work, inte-
rior work, façade work, supporting building work, and landscape work with 294 days
of execution.
The collected data from this case is an as-planned schedule with duration and
tasks relationship information. The total duration of this project later will be used
as a decision-making baseline while the relationship of the tasks plays a significant
role in the calculation. Table 1 shows the information of case study data for structure
work.
Risk Assessment of Construction Project Scheduling 865
Table 1 The information of structure work data

ID Task Name Dur. (day) Predecessors
7 Structure work Foundation 30 3 SS
8 Semi-basement floor 28 7 FS− 14
9 1st floor 14 8 FS− 20
10 2nd floor 14 9 FS
11 3rd floor 14 10 FS
12 4th floor 14 11 FS
15 Rooftop floor and roof 14 14 FS
16 Ramp and front gardens 7 7 FS+ 13
17 Pit lift & core lift 105 7 FS
18 Stairs 63 10 SS+ 10
3 Material and Methodology
To conduct a risk assessment of construction project scheduling, this study imple-

ments Monte Carlo simulation to a case study that illustrates as a flowchart in Fig. 1.
Monte Carlo simulation itself is a mathematical technique that uses a randomizing
device to assign a value to each random variable, in accordance with its probability
distribution [8]. It helps to explain the impact of risk and uncertainty in prediction
and forecasting models.
After collecting the total duration and tasks relationship of the case study, a ques-
tionnaire was organized to collect information about the pessimistic, most likely,
and optimistic duration of each task and was distributed to the project contractor of
the case study. The main data that used in this simulation is the pessimistic, most
likely, and optimistic duration of each task and its relationship to each other. By
building models using Crystal Ball software, a Monte Carlo simulation is conducted
to find out the ideal project duration that able to cover risk in project scheduling and
is expected to calculate the probability of success of construction project duration.
Table 2 shows the questionnaire results of structure work.
866 Z. R. Kamandang
Fig. 1 The flowchart of this study
Table 2 Questionnaire result sample for structure work

ID Task name Duration (day)
Optimistic Most likely Pessimistic
7 Structure work Foundation 30 36 43
8 Semi-basement floor 28 31 38
9 1st floor 14 20 27
10 2nd floor 14 14 21
11 3rd floor 14 14 21
12 4th floor 14 14 21
13 5th floor 14 14 21
14 6th floor 14 14 21
15 Rooftop floor and roof 14 16 23
The pessimistic, most likely, and optimistic of each task duration of the project case
study are the main data that will be processed using Monte Carlo simulation by
conducting several steps; (1) determining the number of iterations, (2) Monte Carlo
simulation, (3) scheduling analysis, and (4) probabilistic of success analysis.
4.1 Determining Number of Iterations
Determining the number of iterations is necessary to conduct first to observe the

minimum number of iterations that are needed to complete the whole simulation.
Several steps in determining the number of Monte Carlo simulation are (1) calculating
the standard deviation of optimistic, most likely, and pessimistic duration (popula-
tion), (2) calculating the population average, (3) calculating the absolute error with
the expected error is less than 2%, and (4) calculating the number of iterations [9].
The results of those calculations are presented in Table 3 below.
Based on the calculation above, the main data of this study will be processed using
Monte Carlo simulation for 16500 iterations.
4.2 Monte Carlo Simulation
The main data that will be processed with 16,500 iterations of Monte Carlo simu-
lation is the optimistic, most likely, and pessimistic duration of each task and task
relationship. Starting the simulation by clicking a blank cell among the start and the
end cell of each task, this blank cell later will be categorized as define assumption
with the triangular distribution. For the first task, the start duration is 0 and the end
duration is the summation of the start duration and the blank cell. For the second task,
the start duration is the end duration of the previous task by also considering their
relationship, while the second task end duration is obtained by summing the start
duration with the blank cell of the second task. The same calculation is also applied
for the remaining tasks. The end duration of the last task later will be assumed as
define forecast and then the simulation can be conducted.
Table 3 The results of

Item Calculation results
determining iterations number
Standard deviation 37.07595807
Absolute error 0.872
Average 43.6
Iteration number 16270.25 ≈16500
868 Z. R. Kamandang
Table 4 Statistical results of

Statistic Fit: beta Forecast values
the simulated project duration
Trials – 16500
Base case – −97.00
Mean 311.71 311.71
Median 311.65 311.63
Mode 311.50 –
Standard deviation 6.89 6.89
Variance 47.53 47.53
Skewness 0.0582 0.0582
Kurtosis 2.94 2.94
Coeff. of variations 0.0221 0.0221
Minimum 255.23 285.82
Maximum 386.16 337.27
Mean Std. error – 0.05
Monte Carlo simulation provides some information including the frequency,

cumulative frequency, and statistical data (that consist of minimum, mean, median,
skewness, kurtosis, etc.). Based on the calculation result, Monte Carlo simulation
suggests 311.71 or 312 days as the project duration (mean result). Table 4 shows the
statistical results of the simulated project duration.
4.3 Scheduling Analysis
Optimistic, most likely, and pessimistic duration of each task are not only used
in Monte Carlo simulation to achieve the simulated project duration, but they
also were processed by using Microsoft Project to calculate the total duration of
the project. There are five total durations based on the schedule types that were
obtained in scheduling analysis; total duration based on as-planned schedule, opti-
mistic schedule, most likely schedule, pessimistic schedule, and simulation result
schedule. The project total duration for each type schedule is presented in Table 5
below.
Based on the scheduling analysis results above, as-planned schedule duration
is between optimistic and most likely duration. It indicates that the contractor has
considered the ideal duration of the project.
Table 5 The total duration of the project based on schedule types

Schedule type As-planned Optimistic Most likely Pessimistic Simulation result
Days 294 290 298 313 312
4.4 Probabilistic of Success Analysis
The probabilistic of success analysis for each type schedule is done by Monte Carlo
simulation. Figure 2 and 3 showed the probability analysis for as-planned duration
and for the simulated duration. As seen in Fig. 2, 294 days of as-planned duration only
has 0.479% of success probability, while simulated duration has 52.575% (Fig. 3).
Table 6 presents the probability of success of other schedule types. For optimistic,
most likely, and pessimistic duration with 290 days, 298 days, and 313 days of
project completion duration, the probabilities are 0.054%, 2.108%, and 57.912%
respectively. Furthermore, this simulation also presents its prediction of total project
completion duration with expected probability which is summarized in Table 7. For
example, to achieve 50% of success probability, the project completion duration is
312 days.
Fig. 2 Probability of success for as-planned duration

870 Z. R. Kamandang
Fig. 3 Probability of success for the simulated duration
Table 6 Probability of
Schedule type Duration (day) Probability of success (%)
success of project completion
duration based on schedule As-planned 294 0.479
types Optimistic 290 0.054
Most likely 298 2.108
Pessimistic 313 57.912
Simulation results 312 52.275
Table 7 Prediction of project

Probability (%) Duration Probability (%) Duration
completion duration based on
(day) (day)
probability of success
10 303 60 314
20 306 70 316
30 309 80 318
40 310 90 321
50 312 100 332
5 Conclusion
Based on the analysis results, the as-planned duration estimated by the contractor
(294 days) only has 0.479% of success probability, optimistic duration (290 days)
has 0.054%, most likely duration has 2.108%, pessimistic duration has 57.912% and
the simulated duration (312 days) has 52.275%, respectively. It indicates that the
as-planned duration is unsuited to be submitted as project completion duration in
the contract because according to simulation results, its success probability is much
less than 100%. Monte Carlo simulation analysis concludes that to avoid risk in
scheduling, the project parties have to allocate at least 332 days of project duration.
6 Suggestion
The writer would like to suggest other researchers to conduct further studies on this
topic. Future research may investigate the performance of Monte Carlo simulation
on assessing other risks in construction project such as risk of project productivity or
construction project emissions. The emissions topic itself is a rising one in research
regarding the sustainability construction issue. Assessing risk using Monte Carlo
could be used as a decision-making tool for construction project parties, therefore
they can avoid and reduce the risks in the future.
References
1. Gładysz B, Skorupka D, Kuchta D, Duchaczek A (2015) Project risk time management—a

proposed model and a case study in the construction industry. https://doi.org/10.1016/j.procs.
2015.08.459
2. Pant D, Srinivas K (2019) Cost estimation of construction project using monte Carlo simulation.
16(2):13–19. https://doi.org/10.9790/1684-1602031319
3. Hastak M, Shaked A (2000) ICRAM-1: model for international construction risk assessment. J
Manag Eng. https://doi.org/10.1061/(asce)0742-597x(2000)16:1(59)
4. Skorupka D (2005) The method of identification and quantification of construction projects risk.
Arch Civ Eng 51(4):647–662
872 Z. R. Kamandang
5. Zou P, Zhang G, Wang J-Y (2006) Identifying key risks in construction projects: life cycle and
stakeholder perspectives
6. Ganame P, Chaudhari P (2015) Construction building schedule risk analysis using Monte-Carlo
simulation. Int Res J Eng Technol 2(4):1402–1406
7. Koulinas GK, Xanthopoulos AS, Tsilipiras TT, Koulouriotis DE (2020) Schedule delay risk
analysis in construction projects with a simulation-based expert system. Buildings. https://doi.
org/10.3390/buildings10080134
8. Meerschaert MM (2013) Simulation of probability models. In: Mathematical modeling. Elsevier,
pp 301–358
9. Fadjar A (2008) Aplikasi Simulasi Monte Carlo dalam Estimasi Biaya Proyek. SMARTek
Investigating Construction Project Delay
Using Fault Tree Analysis Based on Its
Dominant Risk on Private Project
Jojok Widodo Soetjipto , Amalia Martha Sukmana, and Syamsul Arifin
Abstract The team project has already arranged the project planning and
scheduling, but so many projects have deviations in the schedule. Simultaneously, the
project is highly dependent on the existence of all project participants and their envi-
ronment. The previous study has resulted in many variables that cause project delays
but has no ability to predict them in the future. In this research, the author did the
survey and interviewed to solve this problem by knowing the causative factors and
predicting the probability of their occurrence. Therefore, this study uses the FTA
model to predict project delay events based on their variables and mitigate them.
Due to so many risk variables, this study used the Pareto diagram to filter them.
This diagram can reduce 16 variables into seven variables with the highest risk.
The simulation results state that the probability rate of construction project delay
on all variables is 0.1000. Still, each variable’s contribution, namely environment,
worker variables, contractor performance, owner roles, and Construction Manage-
ment (CM) consultant roles, is 0.0766 0.0138, 0.0074, and 0.0064, respectively. This
model has been validated based on existing project data and produces relatively the
same occurrence, i.e., the project has behind schedule. Finally, this research also
provides mitigation to minimize the risk of project delays.
Keywords Project delay · FTA · Simulation
1 Introduction
One of the indicators of project success is measured from the time of completion
of the work in a cost-efficient manner and meets the owner’s satisfaction. Private
J. W. Soetjipto (B) · A. M. Sukmana · S. Arifin

Universitas Jember, Kalimantan 37, Jember 68122, Indonesia
S. Arifin
874 J. W. Soetjipto et al.
projects in Indonesia generally determine the completion time as the primary indi-
cator in evaluating the project’s success. They will give a penalty if the contractor
is unable to complete the work within the defined time. Several previous studies on
the benchmarks of construction project success, namely: (i) a construction project
is declared successful if it can finish on time with a budget and quality that matches
the project owner’s expectations [1]; (ii) criteria for a successful construction project
based on time, cost, quality, safety, client’s satisfaction, employees’ satisfaction,
cash-flow management, profitability, environmental performance and learning, and
development [2]; (iii) Critical Success Factors (CSFs) in project success are the cost
(budget), time, quality and management [3]; etc.
Many construction projects are still experiencing delays even though planning
and scheduling have been prepared. Problems will arise due to a discrepancy in the
process with the actual reality [4]. The several random events caused the delay, either
intermediate or primary. These events should be able to be identified, understood,
and mitigated by the project manager.
This study aims to determine the probability of project delay based on the variables
causing the delay that have dominant risks. The detailed mitigation used only the
most dominant variable and ignoring others. This study uses a Fault Tree Analysis
to identify delay factors and reveal the cause-and-effect relationship between them
and the Pareto method to determine the variables with dominant risk.
2 Literature Review
2.1 Variables Causing Delays in Construction Projects
Many previous studies have resulted in various variables causing delays in construc-
tion projects. They used questionnaires and interviews with construction project
experts and several literature studies to obtain the variables [5–8]. Based on the
theoretical framework, the causes of construction projects’ delays are grouped into
two main factors, namely excusable delay and non-executable delay [9]. However,
the variables that cause project delays depend very much on each case studied.
The identified factors that can be the reason for delay are project identifica-
tion (impractical duration, poor contract, scope, etc.); the contractor delay (team
member, lower tender cost, improper planning, etc.); the owner delay (owner’s infer-
ence, lack of decision power, payment, etc.); consultant delay (lack of decision
power, suspension of work, late in approving, change in material, etc.); environ-
ment delay (bureaucracy of government, delay in issuing permits, change in regula-
tion, criminal activity, conflict, etc.) [5]. The project delay variables are contractor
Investigating Construction Project Delay … 875
delay, owner-related delay (payment progressively, meet-up to improve manage-

ment, etc.), consultant-related delay (communication, coordination, etc.), material-
related delay, labor-related delays, equipment-related delays, and external factor-
related delays [6]. The most significant delay causes according [7]: internal envi-
ronment (internal contractor), exogenous factors (according to the perception of all
project parties), and input factors (resources project). Result of the investigation the
causes of delays facing the Norwegian construction industry [8]: poor planning and
scheduling; slow/poor decision-making process; internal administrative procedures
and bureaucracy within project organizations; resources shortage (human resources,
machinery, equipment); poor communication and coordination between parties; slow
quality inspection process of the completed work; design changes during construc-
tion/change orders; sponsor/owner/client lack of commitment/clear demand; late
improper design, office issues, and user’s issues.
2.2 Risk Dominant Analysis of Variables Causing Delays

in Construction Projects
The risk variable that causes project delay was so many variables so that need analyze
the risk variable dominant. This dominant risk determination aims to obtain the
most influential project delay variables for mitigation in more detail. The stages of
dominant risk analysis are as follows [10]:
1. Interview the experts/practitioners to respond to the event’s probability (P) and
the impact (I) on each variable.
2. Analyzing the data’s validity and reliability, they eliminated if there are invalid
or unreliable variables.
3. Calculating the risk value using the equation below:
R=P∗I (1)
4. Determining the dominant risk uses the Pareto method based on the risk value
from the largest to the cumulative 80% of the data.
2.3 Fault Tree Analysis
FTA is a method that uses a deductive approach to describe the process of system
failure. FTA is build based on the symbols and gates to show the system behavior.
The analysis of FTA uses the rules of Boolean Algebra. FTA is useful to describe
the root cause of an accident logically. The FTA has already applied model delays
in construction projects [11], investigation on the grounds of project problem [4],
root-case analysis of construction cost over-run [6], determine delay in construction
[7], model prediction deterioration of bridge [12], etc.
The FTA can find the root cause of potential failure systems and control them.
The FTA can be modeled through the following steps:
1. Define system activity behavior
2. Define the top event
3. Define the tree of the structure
4. Expand each branch in successive levels of detail
5. Solve the fault tree for the combination of events contributing to the top event
6. Calculate the probability of failure.
3 Research Method
3.1 Background the Case Study
The case study taken in this research is the Suncity Sidoarjo Apartment Development
Project. This project is a private project which is a high-rise building located on Jalan
Pahlawan No. 01 Sidoarjo. Some of the obstacles that often occur in implementing
this project are: (i) the location of this project which is very close to community
settlement; (ii) coordination between the parties involved in this project because
there are several participants, namely the contractor, the management construction
consultant, and the owner; and (iii) a very complex scope of work.
3.2 Data Collection
The short project data was obtained from a construction company and included
written records, i.e., detailed engineering design (DED) projects, planning-
scheduling projects, a problem report and event logs, and an interview with the
project manager. DED has been used to describe the work scope; the planning and
scheduling could help determine the time frame of project work. The project’s report
and event logs give all of the project’s milestones.
To obtain response data from the contractor, CM consultant, and owner, the author
uses questionnaires and interviews. This questionnaire data is needed to: (i) confirm
the variables that occur in the project. The author determines the variables based
on both the previous study and the survey and data collecting that might occur
in the project case, then confirm whether these variables occur or not (see Table
1); (ii) obtain risk data on each variable. The author makes questions about the
probability and impact of each confirmed variable (see Table 2); and (iii) determines
the probability and dependence logic between variables that cause the project to be
late. The author interviewed all respondents to make FTA logic.
Table 1 The confirmed delay variables

Code Risk variable Yes No Relevant/no
A Owner
A1 Delay in contractors 9 75% 3 25% Relevant
payments and related
parties by the owner
A2 Delay in approving 7 58% 5 42% Relevant
change order by the
owner
A3 Delay in preparing the 9 75% 3 25% Relevant
project land by the owner
A4 Lack of communication 7 58% 5 42% Relevant
between the owner and
related parties
A5 Delay in permit by the 8 67% 4 33% Relevant
owner
B CM Consultant
B1 Lack of oversight of 8 67% 4 33% Relevant
design
B2 Lack of coordination 9 75% 3 25% Relevant
between CM and
contractors
B3 Lack of control over 10 83% 2 17% Relevant
every job
C Contractor
C1 Damage and 11 92% 1 8% Relevant
effectiveness of
equipment
C2 Planning and control has 7 58% 5 42% Relevant
no well carried out by
contractors
C3 A production system is 8 67% 4 33% Relevant
no well
C4 Lack of construction 7 58% 5 42% Relevant
materials
D Environmental and worker conditions
D1 The occurrence of work 9 75% 3 25% Relevant
accidents
D2 Problems with residents 7 58% 5 42% Relevant
around the project
D3 Late work due to 10 83% 2 17% Relevant
environmental conditions
D4 Soil conditions on 9 75% 3 25% Relevant
construction projects
Table 2 Percentage cumulative of risk value from largest to smallest

Code Risk variable P I R R Percent Pareto diagram
cum. cum.
A1 Delay in 4 3 12 12 14%
contractors
payments and
related parties
by the owner
A4 Lack of 3 3 9 21 25%
communication
between the
owner and
related parties
B3 Lack of control 3 3 9 30 36%
over every job
D1 The occurrence 3 3 9 39 46%
of work
accidents
D2 Problems with 3 3 9 48 57%
residents
around the
project
D3 Late work due 3 3 9 57 68%
to
environmental
conditions
C2 Planning and 3 3 9 66 79%
control has no
well carried out
by contractors
A3 Delay in 3 2 6 72 86%
preparing the
project land by
the owner
B2 Lack of 2 3 6 78 93%
coordination
between CM
and contractors
C4 Lack of 3 2 6 84 100%
construction
materials
3.3 The Research Stages
This study can show the research’s road map and achieve objectives step by step
(see flow chart in Fig. 1). First step: starting with identification of delay variable of
the construction project, second step: analyze the risk of project delay variable and
Fig. 1 Flowchart for the

research stage
then determine the dominant variable based on the highest risk until 80% of total
risk, the third step: construct the project delay model using FTA, fourth step: running
the model to simulate and make a scenario. Finally: this research result mitigate to
minimize the risk of project delay.
4.1 Identification Project Delay Variable
This study confirmed the literature study variables with the conditions of private
projects in Indonesia, especially in the case study of the Suncity Apartment’s
construction in Sidoarjo. These variables are subject to confirmation interviews with
contractors, CM consultants, and owners involved in the project. Confirmation of this
variable was carried out on 12 respondents with different responses. These variables
are relevant if more than 50% of respondents agreed, so that the author will analyze
further. The variables are grouped based on project participants and environmental
and worker conditions. The distribution data of respondents on the confirmation
survey is 58–92% agree these variables causing the project delays (see Table 1).
4.2 Risk Analysis of Dominant’s Project Delay Variable
This research analyzed the dominant risk of a variable to determine the risk’s primary
cause. It is done to simplify the risk calculation process but produce the right variables
for mitigation. This study obtained the variable’s dominant risk causing project delays
by giving questionnaires and interviews to respondents/experts. The author tested
the responses, both probability, and impact, to determine the variables’ validity and
reliability. If the variables are invalid or unreliable, then the variables were ignored.
The validity and reliability analysis results showed that there were 10 out of 16 valid
and reliable variables. Ten variables are analyzed, both the risk value using Eq. (1)
and the Pareto diagram (see Table 2).
The Pareto diagram is used to choose variables that have a dominant risk by
taking 80% of the cumulative value of risk. The Pareto diagram can reduce the risk
variable that causes project delays from 10 to 7 dominant variables, i.e.: (1) Delay in
contractors payments and related parties by the owner; (2) Lack of communication
between the owner and related parties; (3) Lack of control over every job; (4) The
occurrence of work accidents; (5) Problems with residents around the project; (6)
Late work due to environmental conditions; and (7) Planning and control has no well
carried out by contractors.
4.3 Modeling of Project Delay Variable Using FTA
This paper constructs the model of project delay using FTA to find the contributing
variable. This model uses the dominant variable determined in the previous process.
The FTA model was generated by several stages, i.e.:
Step 1: Define System Activity Behaviour
In this study, the causes of project delays were classified based on participants’
roles and the environmental and worker condition. The number of dominant vari-
ables causing project delays due to owner-related delay, contractor-related delay, CM
consultant-related delay, and environment and worker condition-related delay is 2
variables, 1 variable, 1 variable, and 3 variables, respectively.
Step 2: Define the Top Event
The top event is the main objective seeking a solution, namely the probability of
project delays.
Step 3: Define the Tree of the Structure
The tree of FTA structure on project delay is built based on the project delay variable
dominant. The diagram of the tree of the structure can be shown in Fig. 2.
Step 4: Expand Each Branch in Successive Levels of Detail
To expand each branch, we must search and explore as many causes as possible to
meet each fault. Each event has a gate symbol based on the relationship between each
event and successive levels of detail. In this study, consecutive levels have got by in-
depth-interview with contractor project managers and CM Consultant team leaders.
The expansion of each branch of the FTA until to find the basic event. Figure 3 is the
Fig. 2 FTA diagram of the

top event of the project delay
Fig. 3 FTA diagram of branch in successive levels of detail
detailed extension of the FTA Diagram.

Step 5: Solve the Fault Tree for the Combination of Events Contributing to the
Top Event
Problem-solving of the FTA uses the Boolean Algebra formula. Each event
contributes to the upper event based on its logic gate. An AND gate represents an
independent event relationship and unaffected by any other event to the same gate.
The probability formula for AND gate is given in Eq. 2.
P(A and B) = P(A ∩ B) = P(A) P(B) (2)
On the other hand, an OR gate has represented a dependent event or corresponds to

a set union. The probability of OR gate is given in Eq. 3.
P(A or B) = P(A ∪ B) = P(A) + P(B) − P(A ∩ B) (3)

4.4 Simulation, Validation and Scenario Model
Simulation Model
The model simulation requires data from interviews with experts/team leaders/project
managers to predict the basic event probability value. Meanwhile, to indicate each
basic event’s delay rate using data from a previous study, the delayed payment’s
research results affected the project completion [13] and revisiting critical delay
factors for construction [10]. The probability of each top event can be calculated
using Eqs. (2) and (3) using a minimum cut set in Boolean Algebra. The minimum
cut set produces the project delay variable (see Table 3).
This study obtained the model simulation results by calculating a minimum cut
based on the probability value of the basic event, giving probability rate for each
in-term mediated event, namely A1, A4, B3, D1, D2, D3, and C2 of 0.0019, 0.0074,
0.0065, 0.0020, 0.0498, 0.0102 and 0.0138 respectively. Meanwhile, this project’s
probability of experiencing delays if all basic events occur is 0.1000 (see Table 4).
Validation Model
This study validates the FTA model using data that occurred in 2020. The work plan
and milestones that the contractor has set are as follows:
1. The event log of the project obtained the following data:
• There has been a termination of work in March due to the public’s complaints
regarding the contractor’s negligence, which caused the material to fall in
the housing area.
• Many houses in the area were damaged, and the charges filed were ignored
by the contractors, so the community complained to the government to help
solve this problem.
• There is a problem of land acquisition for the expansion of the apartment
area with the community.
• The government requests the contractor to solve the public complaints.
• The use of heavy equipment has experienced many obstacles due to the
surrounding community’s complain.
• Topping off will be conducted in June 2020 but realize in September 2020.
• Project completion delayed
2. Based on these data, this research runs the FTA model (see Table 5).
3. The FTA model can calculate the probability of delays in a construction project
based on the project event log data in Table 5. The results of the running model
show that the probability of the project experiencing delays is 0.3407. This
shows that the probability of the project experiencing delays is 34%.
4. Data obtained from the project schedule shows that in the 7th month, the project
experienced a delay of 15%.
5. This shows that the FTA Model can estimate the probability of project delays
even though it has not been able to determine the percentage of the delay.
Table 3 Basic event and the minimum cut set of an intermediate event
Code Basic event Probability
A1 Delay in contractors payments and related parties by the owner
Z1A1 Regulations from banking that complicate lending 0.0019
Z2A1 There is an increase in foreign currency exchange rates 0.0019
Z3A1 The bank refused the loan from the owner 0.0019
Z4A1 The owner is unable to pay off the remaining funds 0.0019
Z5A1 The allocation of funds is not used to pay contractors 0.0074
The result from the minimum cut set 0.0019
A4 Lack of communication between the owner and related parties
Z1A4 There is a conflict between the owner and related parties 0.0019
Z2A4 Lack of understanding of contract documents 0.0019
Z3A4 The results of the meeting were not carried out properly 0.0019
Z4A4 Unclear mechanisms for providing requests for information 0.0019
B3 Lack of control over every job
Z1B3 Frequent billing of shop drawing approval by contractors 0.0009
Z2B3 Slow administration handling by CM consultant staff 0.0009
Z3B3 The results of the meeting were not held 0.0019
Z4B3 Often some disagreements cause miscommunication 0.0009
Z5B3 Slow job review by CM consultant staff 0.0009
Z6B3 Evaluation results cannot be done/executed 0.0009
D1 The occurrence of work accidents
Z1D1 A competent officer does not carry out material identification 0.0025
Z2D1 There are no proper procedures for moving and storing materials 0.0025
Z3D1 The use of heavy equipment exceeds the maximum capacity 0.0025
Z4D1 The work done is not up to standard 0.0025
Z5D1 Workers fall from a height 0.0025
Z6D1 Workers are punctured by the material and crushed by the material 0.0050
Z7D1 Workers smoke in a prohibited/dangerous place 0.0025
Z8D1 Workers are in a hurry to do work 0.0050
D2 Problems with residents around the project
Z1D2 There was a crack due to the installation of the bored pile 0.0116
Z2D2 Heavy equipment activities carried out simultaneously 0.0116
Z3D2 There is work activity in overtime hours 0.0039
Z4D2 The load exceeds the strength limit of the sling rope 0.0039
(continued)
Table 3 (continued)
Code Basic event Probability
Z5D2 The scaffolding ladder is not strong enough to withstand the weight of the 0.0010
material
Z6D2 Worker’s negligence in carrying materials 0.0077
Z7D2 Sling rope is not strong enough to withstand the weight of the material 0.0116
D3 Late work due to environmental conditions
Z1D3 There is a natural disaster 0.0026
Z2D3 Rain 0.0026
Z3D3 Hot 0.0051
Z4D3 Wind 0.0026
C2 Planning and control has no well carried out by contractors
Z1C2 The specified time duration does not match 0.0046
Z2C2 The action plan doesn’t work 0.0046
Z3C2 The method of doing work is not appropriate 0.0023
Z4C2 Incomplete job identification 0.0023
Hasil minimal cut set 0.0138
Note Probability = rate of the event probability per month
Table 4 The minimum cut set (probability) of project delay

Code Variable Probability
A1 Delay in contractors payments and related parties by the owner 0.0001
A4 Lack of communication between the owner and related parties 0.0074
B3 Lack of control over every job 0.0065
D1 The occurrence of work accidents 0.0200
D2 Problems with residents around the project 0.0498
D3 Late work due to environmental conditions 0.0102
C2 Planning and control has no well carried out by contractors 0.0138
The Result Minimal Cut Set (Probability Project Delay) 0.1000
Note Probability = rate of the event probability per month
6. The FTA model can also be used to predict probability with the following
requirements: (i) monthly event prediction data is required as shown in Table
5; (ii) the event data is entered at the rate of probability to determine the basic
event probability value; (iii) The prediction of project delay probability can be
determined by the FTA Model by calculating the basic event recapitulation.
Table 5 Recapitulation of data variable based on project event log
Code Risk variable Event on month to Occurrence log
1 2 3 4 5 6 7
A1 Delay in contractors payments and related parties by the N N N N N N N The contractor makes payment on time
owner
A4 Lack of communication between the owner and related Y Y Y Y Y Y Y Many changes in project implementation: drawing,
parties material, contract, etc.
Investigating Construction Project Delay …
B3 Lack of control over every job Y Y Y Y Y Y Y Delay in approval, the meeting didn’t been implemented
well, etc.
D1 The occurrence of work accidents N N N N Y Y Y The worker has an accident, the equipment didn’t well,
the material was improper, etc.
D2 Problems with residents around the project Y Y Y Y N N N So many complaints due to project: community property
damage, loading activities, equipment activities, etc.
D3 Late work due to environmental conditions Y Y Y Y N N N Land acquisition and weather problem
C2 Planning and control has no well carried out by Y Y Y Y Y Y Y The construction method can’t be implemented,
contractors incomplete in scope of work, etc.
Note Y = Yes (occurrence); N = No (No occurrence)
885
Table 6 The scenario of project delay probability rate

Code Simulation Scenario 1 Scenario 2 Scenario 3 Scenario 4
A1 0.0001 0.0001 0.0000 0.0000 0.0000
A4 0.0074 0.0074 0.0000 0.0000 0.0000
B3 0.0065 0.0000 0.0065 0.0000 0.0000
D1 0.0200 0.0000 0.0000 0.0200 0.0000
D2 0.0498 0.0000 0.0000 0.0498 0.0000
D3 0.0102 0.0000 0.0000 0.0102 0.0000
C2 0.0138 0.0000 0.0000 0.0000 0.0138
Total 0.1000 0.0074 0.0064 0.0766 0.0138
Scenario Model
This research developed the scenario to determine the impact of each intermediate
event on project delays by designing some basic events that occur together or not. This
scenario can be used as a priority scale determination to control events, especially
those that significantly impact project delays. The author arranges 4 scenarios based
on the main variable that caused project delay: (1) Scenario 1: the project delay due
to owner related-delay; (2) Scenario 2: due to the CM Consultant related-delay; (3)
Scenario 3: because of the contractor related-delay; and (4) Scenario 4: caused by
the environmental and worker condition. The results of these scenarios are shown in
Table 6.
From the scenario above, this study concluded that the project environment and
worker variables significantly affect the construction project delay, namely 0.0766.
The 2nd ranking of the variable that affects delay is the contractor’s performance
with a probability of 0.0138. The 3rd is the owner’s role, which is 0. 0074, and the
least effective is the performance of CM consultant with a contribution of 0.0064.
Based on this scenario, managing the project should maintain the project environ-
ment and worker properly to reduce the possibility of project delays. Furthermore,
the contractor’s performance must also be maintained, especially in conducting
internal planning, scheduling, and monitoring. Meanwhile, the owner and the CM
consultant contributed to the relatively small causes of delay because of their role in
implementing the project as supporting only.
4.5 Mitigate to Minimize Project Delay
Based on the scenario above, this research mitigates reducing the risk of project
delays. This mitigation is carried out based on the priority of the variable due to
project delays that have been identified in the scenario of the FTA model. This
mitigation results from discussions with the project manager and the team whose
concepts have been drafted based on the debate results in this paper. This mitigation
can be an input for owners, contractors, and CM consultants to improve their future
performance. This study resulted in the mitigation in Table 7.
Table 7 The risk mitigation due to project delay

No Variable Risk response
1 Environment and worker variable
1.1 The occurrence of work accidents Carry out risk control by compiling the JSA
and implementing the safety control
hierarchy, namely Elimination, Substitution,
design, and PPE (Protect Personal
Equipment) and WPE (Work Protect
Equipment)
1.2 Problems with residents around the project Outreach and coordinate with the
community around the project
1.3 Late work due to environmental conditions Develop and evaluate work methods to suit
environmental conditions and stop work
when the weather is extreme
2 Contractor variable
2.1 Planning and control of work carried out by • Arranging planning and scheduling using
contractors is not going well historical data/previous project data as a
reference in project work
• Always monitor and evaluate every
activity regularly. If there is a deviation,
immediately take corrective action
3 Owner variable
3.1 Delay in the owner in making payments to • There is a need for mutual agreement and
contractors commitment in carrying out the work
contract, especially in carrying out each
party’s rights and obligations
• Conduct regular meetings with all project
parties, namely the owner, contractors, and
MC consultant
• The owner always maintains a financial
presence so as not to interfere with
physical progress on the project, which has
an impact on project delays
• Contractors are expected to be able to find
alternative solutions through additional
working capital
4 CM consultant variable
4.1 Lack of control over each employee • CM consultant should reference in
conducting supervision based on the work
plan that the contractor has prepared
• Always coordinate/meet so that any
problems are be resolved quickly
5 Conclusion
This study has identified the variables that cause delays in construction projects
based on literature studies fitted in project conditions. The main variables obtained
are quite a lot caused difficulty in in-depth analysis. Therefore, to determine the
dominant variable, use a risk analysis and Pareto diagram.
The FTA model established is proven to assess the probability of project delays
well. In addition, with the framework arranged, this model can also predict the
likelihood of project delays based on the prediction of basic events in the scheme
project time. This model can also prioritize the main variables that cause project
delays and make set mitigation to control and reduce project delays.
While this study has several contributions to predict the probability of project
delays, it is only an estimate of the event that has not determined the behind schedule.
It needs to study further to afford to determine the delay. In addition, this model uses
a one-way hierarchical assumption, whereas the actual event of each variable can
influence each other. Therefore, this model needs to be improved by the System
Dynamics to solve interrelationships and looping between variables to make it as
close as possible to the actual response.
References
1. Gündüz M, Nielsen Y, Özdemir M (2013) Quantification of delay factors using the relative
importance index method for construction projects in Turkey. J Manag Eng 29(2):133–139
2. Silva GASK, Warnakulasuriya BNF, Arachchilge BJH (2016) Criteria for construction project
success: a literature review. In: 13th International conference on business management. Faculty
of Management Studies and Commerce, Sri Lanka, pp 697–717
3. Ramlee N, Tammy NJ, Mohd RNH, Musir A, Karim AKN, Chan HB, Nasir SRM (2016) Critical
success factors for construction project. In: International conference on advanced science.
In: 2015, AIP conference proceeding engineering and technology (ICASET), pp 030011–1–
030011–6
4. Silvianitaa, Mahandekaa DS, Rosyida DM (2014) Fault tree analysis for investigation on the
causes of project problems. In: 2nd International seminar on ocean and coastal engineering,
environment and natural disaster management, ISOCEEN. Science Direct, Indonesia, pp 213–
219
5. Bhatti IA, Abdullah AH, Nagapan S, Sohu S, Lohar MA, Bhatti NB (2018) Delay in high-rise
building construction projects of dubai: a review. Eng Sci Tech Int Res J 2(2):7–15
6. Fashina AA, Omar MA, Shikh AA, Fakunle FF (2021) Exploring the significant factors that
influence delays in construction project in Hargeisa. Heliyon 7(2021):e06826
7. Sweis G, Sweis R, Hammad AA, Shboul A (2008) Delays in construction projects: the case of
Jordan. Int J Proj Manag 26(6):665–674
8. Zidane YJT, Andersen B (2018) The top 10 universal delay factors in construction projects.
Int J Proj Manag Bus 11(3):650–672
9. Hamzah N, Khoiry MA, Arshad I, Tawil NM, Che Ani AI (2011) Cause of construction
delay—theoretical framework. In: The 2nd international building control conference 2011.
Science Direct, Malaysia, pp 490–495
10. Jeffrey BHY, Pei LG, Yoke BW, Martin S (2021) Revisiting critical delay factors for
construction: analysing projects in Malaysia. Alex Eng J 2021(60):1717–1729
11. Al-Humaidi HM, Hadipriono FT (2010) A fuzzy logic approach to model delays in construction
projects using rotational fuzzy fault tree models. Civil Eng Environ Syst 27(4):329–351
12. Soetjipto JW, Adi T, Anwar N (2016) System dynamics approach for bridge deterioration
monitoring system. Int J Eng Technol Innov 6(4)
13. Maurice PO, Charles MR, Paul AO (2015) Effects of delayed payment of contractors on the
completion of infrastructural projects: a case of Sondu-Miriu hydropower project. Chin Bus
Rev 14(7):325–336
Implementing a Relational Database
in Processing Construction Project
Documents
Mik Wanul Khosiin and Ardian Umam
Abstract Construction project documents are extensive and complex, and most of
those are still carried out with traditional systems. Consequently, it takes much time
to process, low accuracy, cannot be stored for long periods, and make it difficult
for future engineers to access historical data project for the next project planning.
The current researchers have responded to this situation by implementing database
management. However, it has not had a significant impact on the industrial field.
Thus, the study of the application of relational databases through SQL with a
different approach, namely integrating project tables such as resource tables, cost
items, work items, and project drawings, which are currently being worked on sepa-
rately. As the output, it would be demonstrated into a supply chain management table
(table_3), construction cash flow (table_4), and the quantity take-off management
table (table_5). The total project price from the result of the SQL program shown in
supply chain and project cash flow tables is IDR 1.42 billion, with a project duration
of 109 days. The SCM table describes the types of resources used in the project,
including the quantity and price values. Then, the table of CSF contains a list of
cost items and the volume and price of work for the budget plan and cash flow plan.
Moreover, the QTO table describes the association between cost items and project
drawings intended to project surveyors for the quantity take-off management process.
Project managers are always required to make decisions for project control appro-
priately, such as cost estimation, resource planning, and project schedule effectively.
Therefore, through those three tables, the authors hope they can help construction
practitioners to improve their construction projects efficiently in the future.
Keywords Relational database · Construction · SQL
M. W. Khosiin
Civil Engineering, East Java Branch of Veterans National Development University, Surabaya,
Indonesia
A. Umam (B)
School of Electrical Engineering and Informatics, Bandung Institute of Technology, Bandung,
Indonesia
892 M. W. Khosiin and A. Umam
1 Introduction
The construction industry has a large market and will be stable for the next few years
because all countries are competing to build monumental buildings as the identity
of the nation’s progress [1]. Therefore, all construction companies also compete
in the use of information technology as part of project management [2]. Most of
the current construction projects have implemented building information modelling
(BIM), artificial intelligence (AI), virtual reality (VR), as well as data science [3].
Scientists and practitioners agree that in the future, the construction sector will be
dominated by the automation system in construction by answering various problems
and challenges currently faced by project managers [4].
During the construction phase, the project will be faced with various kinds of
documents with a large amount of data [5]. These data are interrelated, which are
always reviewed, revised and controlled during the project. However, in practice, in
the construction field today, they still use conventional methods so that the project
database process is very long and the level of accuracy is quite low. Thus, the project
manager is difficult to make decisions in real time, and then it will have an impact
on project implementation performance [6].
This study was carried out to solve problems regarding complexity of project
databases, by totally using SQL programming, all project documents will be inte-
grated and simplified to make it easier for engineers in creating supply chain
management table, cash flow construction, and quantity take-off management table.
2 Database Management System (DBMS)
Today databases are very important for every industry, including the construction
sector. There is always a database behind the scenes on every data or document we
request, because all companies always maintain all important data for future use.
The database management system is a powerful tool for creating and managing large
areas of data efficiently and can maintain it for very long periods in a secure manner.
Basically a database is nothing more than a collection of information that has existed
for a very long time, then users can create a new database with a different scheme,
dig up the desired data, be able to accommodate a very large amount of data, and
can be done with access control for a sufficient period of time long. The DBMS
can be designed through several programming languages such as structured query
language (SQL), R language, python, unified modelling language (UML), object
definition language (ODL), Application Programming Interface (JDBC), hypertext
pre-processor (PHP), and etc. [7, 8].
Several previous studies on the role and application of relational databases in
the construction industry have been carried out by many researchers, starting from
simple management in the form of table management in MS. Excel to involve building
information modelling (BIM) and the internet in data management. In addition, the
Implementing a Relational Database in Processing Construction … 893
programming language that used in the current studies are different according to the
their research objectives.
Database in Construction Project
Project database management in Indonesia has not been well managed. The hundreds
or even thousands of cost items and work items which are the core of all project
documents are still done traditionally and have not been integrated. Through the use
of data analytics (Pivot Table) the correlation between the cost breakdown structure
(CBS) and the work breakdown structure (WBS) can be done better, thus helping
contractors to estimate costs and schedule better and more efficiently [9]. Project
documents that require integration between CBS and WBS are very diverse, including
S-curve, bill of quantity, budget plan, resources plan, project scheduling, progress
reports, payment progress, and others. Research collaborations between countries
such as Indonesia and Taiwan have been carried out to develop the best scheme for
integrating CBS and WBS and demonstrated in the form of a construction progress
curve [14].
Internet-Based Database Management
The advanced database management system can be generated by the internet, all work
packages, data history, and formulas are integrated to develop the project database
stores. Therefore, this system can help the fund contractor to track and control their
construction projects in real time and accurately [10].
Building Information Modelling and DBMS
Information about historical project data is very difficult to trace, because it is devel-
oped separately, so that the engineers worked hard to estimate the project costs. The
application of a relational database that contains tables of project history connected
to BIM models is proven to be able to carry out costs tracking to determine the final
costs required for future projects [11]. The BIM system can also be used to main-
tain the sustainability of project procurement by considering life-cycle cost analysis
(LCCA) through a database management system, this is very helpful for project
managers to manage supply chains in construction projects [12].
DBMS for Current Construction Issues
Construction projects have a high risk of delays due to natural conditions or misman-
agement. Through the use of historical data for certain project types, database
management is carried out which can identify what jobs are often affected by delays,
as well as the costs and duration of work due to delays. By using a relational database,
contractors can anticipate delays or can submit claims supported by accurate data
through the results of studies that have been conducted [13]. In another study, the
application of the DB2 database management system as a database design can also
improve management and information levels. By using UML class diagrams to find
a logical structure in order to accelerate the development of project costs to make it
faster and more accurate [14].
The review that has been carried out from previous studies (Table 1) can be
concluded that DBMS is proven to be able to help the construction sector to manage
Table 1 Comparison model of the previous studies

Year Researchers Model Brief concept Output
2006 Ji Li et al. Internet-based A relational database Developed system to
in storing all project track an control the
data through internet construction project
based in real time
2016 Nannan Zhang UML, DB2 Integrating the Index design on
et al. database construction project
management system cost system
to accelerate the
construction cost
system
2015 Gozde Bilgin et al. MS. Access Relational database Delayed activities and
in processing response party
previous projects data
to predict
construction delay
2020 Hang Thu Thi Le BIM, MS. Access BIM integrated with Building life-cycle
et al. database costs for sustainable
management to procurement
compute building
cycle
2020 Wijatmiko et al. MS. Excel Using pivot table and The association table
unique code to of cost item and work
integrate WBS and item
CBS
2021 Khosiin et al. SQL Assigning entity Supply chain, cash
relationship on flow and take-off
complex project data tables
through SQL script
2021 Khosiin et al. SQL Assigning entity Supply chain, cash
relationship on flow and take-off
complex project data tables
through SQL script
data efficiently. An advanced scheme using the internet and a combination with
project controlling software such as BIM is also able to overcome the latest problems
in the construction industry such as mitigating the risk of delay, cost estimation, data
mining, etc. However, due to the complexity of the problem, and the scope of the
data to be faced, there are still many challenges that must be resolved immediately.
Therefore, this indicates that there is a need for further study on project database
management through one of the programming languages like SQL to correlate data
so that it can be realized in a simpler and more accurate table form to help project
managers make real-time and efficient decisions (Table 1).
3 Case Study
A simple one-story residential building and all its construction data are plotted to
demonstrate the function of a relational database. The project document is formed
in several stages, starting with a detailed engineering design (DED) which contains
a project drawing, then the contractor will list the work items and their work order
and end with the creation of a project schedule. At the same time the engineer will
also compile a cost item for making a project budget plan. Each of those has a
unit item, quantity, and unit price that previously analysed according to the price
of the resources used (labour, material, equipment and subcontractor). All of these
data categories are interrelated, project drawings are used as a reference for making
project schedules and budgets, while cost items are integrated with work items with
one to many or one to one relationship model as in Fig. 1. These project documents
will always be reviewed by engineers during the project.
Step One, Relationship Design
To implement a relational database begins with creating an entity relationship
diagram (ERD) through the MySQL programming language such as Fig. 2, this
diagram consists of several tables that represent the project database that has been
mentioned earlier. Then, the tables are integrated through the ID numbers that have
been assigned to each database category, the details of the table used as an ERD
compiler are,
1. Drawing table, contains drawing_id, description and number of drawing pages
2. The cost_item table has a cost_id, cost breakdown structure (CBS), units, work
volume, and project overhead
3. Table work_item, contains the work_id, work breakdown structure (WBS),
project schedule
4. The resource table, consisting of the resource_id, resource name, unit, and
resource unit price
Project Design Cost Breakdown Work Breakdown

(DED) Structure (CBS) Structure (WBS)
Output of CBS : Output of WBS :

One to Many
1. Cost item 1. Work item
or
2. Quantity take-off 2. Project schedule
One to One
3. Resources
4. Unit Price analysis
5. Budget Plan
Fig. 1 The overview of construction database

Fig. 2 ER diagram for construction database
5. Table workitem_costitem, describes the relationship between work_id and

cost_id
6. Table resource_costitem, is the relation between resource_id (along with
index_resource) and cost id
Step Two, SQL Script

The process of joining between tables is the second step in the project database
management process. SQL scripts are arranged according to the required output table
form. This study will present three project management tables as a demonstration of
the application of a database management system, namely the supply chain manage-
ment table, project financial plans, and quantity take-off management. Table 2 SQL
script to prove that a relational database can help project managers in processing
construction documents better through the three tables previously mentioned.
To obtain the Table_3 with columns of resources name, unit, quantity (qty),
start_date, finish_date, duration, price and cost, we can join the tables of resource,
resource_costitem, costitem, workitem_costitem and workitem using attributes of
Table 2 SQL script for construction database

#Table 1
WITH temp(name, cost_id, unit, qty, start_date, finish_date, duration, price, cost) as
(SELECT r.name, c.cost_id, r.unit, c.qty, MIN(w.start_date), MAX(w.finish_date),
DATEDIFF(MAX(w.finish_date), MIN(w.start_date)) as duration, r.price,
CAST(((rc.index_res*c.qty*r.price)) AS DECIMAL(14,2)) as cost
FROM (((resource r JOIN resource_costitem rc USING(resource_id)) JOIN costitem c
USING(cost_id)) JOIN workitem_costitem wc USING(cost_id)) JOIN workitem w USING(work_id)
GROUP BY r.name, c.cost_id, rc.index_res, r.price, r.unit)
SELECT name, cost_id, unit, qty, start_date, finish_date, duration, price, cost
FROM temp
UNION
SELECT "~TOTAL", "-", "-", "-", MIN(start_date), MAX(finish_date),
DATEDIFF(MAX(finish_date),MIN(start_date)), "-", SUM(cost)
FROM temp
ORDER BY name
#Table 2
WITH temp1(cost_id, cost_item, unit, qty, price, total_price, start_date, finish_date, duration) as
(WITH temp2(cost_id, cost_item, resource, unit, qty, price, start_date, finish_date) as
(SELECT c.cost_id, c.description, r.name, c.unit, c.qty, (rc.index_res*r.price), MIN(w.start_date),
MAX(w.finish_date)
FROM (((costitem c JOIN resource_costitem rc USING (cost_id)) JOIN resource r USING (re-
source_id)) JOIN workitem_costitem wc USING(cost_id)) JOIN workitem w USING(work_id)
GROUP BY c.cost_id, r.name, rc.index_res, r.price
ORDER BY c.cost_id)
SELECT cost_id, cost_item, unit, qty, CAST(SUM(price) AS DECIMAL(14,2)) as price,
CAST((qty*SUM(price)) AS DECIMAL(14,2)) as total_price, MIN(start_date) as start_date,
MAX(finish_date) as finish_date, DATEDIFF(MAX(finish_date), MIN(start_date)) AS duration
FROM temp2
GROUP BY cost_id)
SELECT cost_id, cost_item, unit, qty, price, total_price, CAST((total_price/(SELECT
SUM(total_price) FROM temp1)*100) AS DECIMAL(3,2)) as "weight (%)", start_date, finish_date, dura-
tion
FROM temp1
UNION
SELECT "TOTAL", "-", "-", "-", "-", SUM(total_price), "100", MIN(start_date), MAX(finish_date),
DATEDIFF(MAX(finish_date),MIN(start_date))
FROM temp1
#Tabel 3
SELECT DISTINCT c.cost_id, c.description as cost_item, c.unit, c.qty, w.drawing_id, d.description
FROM (costitem c JOIN workitem_costitem wc ON (c.cost_id = wc.cost_id)) JOIN workitem w ON
(wc.work_id = w.work_id) JOIN drawing d ON(w.drawing_id = d.drawing_id)
ORDER BY c.cost_id, w.drawing_id
resource_id, cost_id, cost_id and work_id, respectively. The price is calculated by

multiplying qty with price in ‘SELECT’ clause, meanwhile for total cost is aggregated
using ‘sum function’ toward the ‘cost column’, and in order to do that, we implement
a temporary table called ‘temp’ using ‘WITH’ clause. Table_4 is composed of several
columns, namely cost_id, cost_item, unit, qty, price, start_date, and finish_date. Thus,
it is necessary to combine the resource_costitem, resource, workitem_costitem, and
workitem tables through the attribute id of each table. To retrieve the unit price
for each cost item, it is necessary to analyse the resources consisting of human
resources, equipment, and material as detailed items for unit price compilers through
SELECT clause, DATEIFF, and WITH clause as the formula in Table_2. In the
last table is quantity take-off management, the column cost_id, cost_item, column
unit, qty, drawing_id and drawing description column are presented. This is to show
Table 3 Supply chain management

Resources Cost ID Unit Qty. Start Finish Dur. Unit price Cost
Bar bender & CS001 Ls 2K 21-01-13 21-02-23 41 5K 10 M
Cutter
Bared steel CS009 Kg 250 21-03-14 21-03-23 9 70 K 14.9 M
Baseplate & CS008 Kg 200 21-03-14 21-03-23 9 14 K 3M
stiffener
Wooden glue CA004 Kg 4.2 21-03-10 21-03-13 3 13 K 55 K
Wooden CS004 M3 500 21-02-02 21-02-23 3 3.6 M 72.5 M
rafters
Etc. .. .. .. .. .. .. .. ..
Total 21-01-01 21-04-20 109 1.42B
the relationship between each volume of work on the cost item having an asso-
ciation with a particular project drawing, so that the quantity surveyors can more
easily calculate the quantity of buildings. Table_5 is obtained from the merging of
the workitem_costitem, workitem, and drawing tables by using ORDER BY and
SELECT clause.
Step Three, SQL Statement Execution

The rendering process of the SQL script for the three demo tables cannot be generated
at same time, but must be executed one by one script, so that there is no failure in the
execution process. Table 3 contains the types of resources used in the project along
with detailed data including the volume of work, resource usage schedules and cost
analysis to determine the total project value. Based on the results of the analysis in
Table 3, it is also obtained 36 types of resources involved or 146 resource items after
the repetition process, this shows that the construction project has a large enough
order for material, equipment, and labour, so it’s no wonder this will be a huge data
when the project is scaled up. The next table is the project cash flow, which explains
the project cost budget plan through a cost breakdown structure along with a time
schedule to estimate the project cost flow during the construction process. Table 4 can
also be transformed into an S-curve, which is the correlation between the accumulated
costs and the project implementation time (weeks or months). Therefore, the project
manager can easily control the amount of costs that must be spent within a certain
time and can find out what kind of jobs and resources involved in it, [9]. Table 4
shows that of the 23 cost items, the total project value is IDR 1,420,522,875.06 with
a project duration of 109 days starting from January 1 to April 20, 2021. Table 5
provides information that those cost items whose work volume is calculated based on
the particular project drawing as its association. This process is part of the quantity
take-off management that is usually managed by an engineer or quantity surveyor in
a construction project.
Table 4 Project cash flow

Cost ID Cost item Unit Qty. Price Total price % Start Finish Dur.
CA001 Brick M2 500 149 K 74 M 5 21-03-02 2103-09 7
laying
CA002 Stucco M2 500 90 K 44 M 3 21-03-10 21-03-09 7
work
CA003 Wooden M2 14.4 1M 14 M 1 21-03-14 21-03-13 3
Door and
windows
CS009 Steel Kg 250 630 K 15 M 1 21-03-14 21-03-23 9
bracing
CS010 Metal M2 750 167 K 125 M 9 21-01-01 21-03-23 9
roof
Etc. .. .. .. .. .. .. .. ..
Total 1.42B 100 21-01-01 21-04-20 109
Table 5 Quantity take-off management

Cost ID Cost item Unit Qty. Drawing ID Description
CA001 Brick laying M2 500 AR001 Walls and stucco
CA002 Stucco work M2 500 AR001 Walls and stucco
CA003 Wooden Door and windows M2 14,4 AR002 Doors and window
CS009 Steel bracing Kg 250 ST005 Roof plan
CS010 Metal roof M2 750 ST005 Roof plan
Etc. .. .. .. .. ..
4 Conclusion
A large market in the construction industry sector must be accompanied by profes-

sional project database management. Project data including project design tables,
work items, cost items, and resource lists are interrelated and generated separately.
The database management system (DBMS) is a tool for managing a broad set of
information effectively so that it can be accessed easily and can last a long time.
Previous researchers have initiated the application of relational databases to address
construction issues but their role is not significant yet, so a new approach is needed
in this case. A case study through a simple residential project database was fully
demonstrated in SQL language in the form of three strategic tables in the construc-
tion industry including supply chain table, the table of construction cash flow, and
quantity take-off management table.
There are three stages that must be made in designing a database in SQL language.
The first step is integration design, by compiling the project tables such as cost_item,
work_item, resource, etc. through entity relationship diagram (ERD) with the unique
ID in each data. The next step is arranging the data analytical script to combine, clas-
sify, and analyze all of tables automatically and obtain the purposed tables. Finally,
the tables in question are obtained by running the script (final step) shows that the total
project price is IDR 1.42 billion with a project duration of 109 days. Furthermore,
resource types and project scheduling can also be identified properly.
Acknowledgements The authors would like to thank the scientific committee and the organizing
committee especially the editor, the paper reviewer for their evaluation and supports to this paper.
The authors thank the Department of Civil Engineering, East Java Branch of Veterans National
Development University for financially supporting this 5th International Conference on Rehabili-
tation and Maintenance in Civil Engineering (ICRMCE). In addition, The authors are also grateful
for the research collaboration between universities (UPNVJT-ITB) in Indonesia, and hopefully this
collaboration can be upgraded to a top research level in the future.
References
1. T. B. Research Comp (2020) Global construction market, the business research company,
Philadelphia, USA
2. Halpin DW, Senior BA (2011) Construction management fourth edition, Colorado. Wiley, USA
3. Haynes ME (2002) Project management practical tools for success. Axzo Press, Arkansas
4. Fryer B (1996) The practice of construction management, 3rd edn. Blackwell Science, London
5. Institute PM (2013) A guide to the project management body of knowledge (PMBOK Guide)-
fifth edition. Project Management Institute Inc., Pennsylvania
6. Oxley R, Poskitt J (1971) Management techniques applied to the construction industry, 2nd
edn. Crosby Lockwood Staples, London
7. Molina HG, Ulman JD, Widom J (2009) Database systems the complete book, 2nd edn. Pearson
Prentice Hall, New Jersey
8. Sumanthi S, Esakkirajan S (2007) Fundamentals of relational database management systems.
Springer, Tamil Nadu
9. Khosiin MW, Wang KC, Wijatmiko I, Wang HH, Wang WC (2019) S-Curve development model
by integrating Taiwanese and Indonesia Practices. In: The 23rd Symposium on construction
engineering and management, Taichung, Taiwan
10. Li J, Moselhi O, Alkass S (2006) Internet-based database management system for project
control. Eng Constr Archit Manag 13(3):242–253
11. He X, Liu R (2020) A relational database scheme of construction cost tracking from BIM
model to final costs. In: American association of civil engineering, Florida, USA
12. Thi Le HT, Likhitruangsilp V, Yabuki N (2020) A BIM-Integrated relational database
management system for evaluating building life-cycle costs. Eng J 75–86
13. Bilgin G, Eken G, Dikmen I, Birgonul MT (2015) A relational database for construction delay.
In: Eighth international conference in the 21st Century, Thessaloniki, Greece
14. Zhang N, Song W (2016) Database design on construction project cost system. In: 3rd
International conference on materials engineering, manufacturing technology and control
(ICMEMTC), Jinzhou, China
Evaluation of the Public Procurement
Principles Implementation in Surabaya
Construction Projects
Patrisius Valdoni Sandi and Mohammad Arif Rohman
Abstract Public procurement is essential to the success of infrastructure devel-

opment as through this process the best service providers can be selected. Within
Indonesian context, there are seven basic principles of good public procurement,
namely: efficient, effective, transparent, open, competitive, fair, and accountable.
This study aims to investigate and assess seven principles implementation based on
the providers’ perceptions in the public construction procurement in Surabaya as
the second biggest city in Indonesia. Importance Performance Analysis (IPA) was
employed to analyze the level of importance and performance of each principle’s
indicator. The result shows that the implementation of the construction procurement
principles in Surabaya obtained score of 87.55%, which can be considered quite
good as the score was relatively high and close to the maximum (100%). In more
details, from each procurement’s principle, the implementation of the ‘efficient’ prin-
ciple had score of 87.91%, ‘effective’ was 86.31%, and ‘transparent’ was 86.18%.
Meanwhile ‘open’ principle was 89.87%, ‘competitive’ was 89.11%, 88.90% for
‘fair’, and 85.61% for ‘accountable’. Meanwhile, there were some gaps between the
importance and the performance levels. This means the Surabaya City Government
still needs to improve or at least to maintain the existing performance to promote a
successful infrastructure development in their region.
Keywords Public procurement principles · Construction procurement ·

Importance performance analysis
1 Introduction
Construction procurement leads to a system of policies that can be used to acquire

new buildings, either through purchasing or leasing, or by designing and constructing
infrastructure facilities [1]. Public procurement, which is a government policy
P. V. Sandi · M. A. Rohman (B)

Construction Management Laboratory, Department of Civil Engineering, Sepuluh Nopember
Institute of Technology, Surabaya, Indonesia
902 P. V. Sandi and M. A. Rohman
system, is closely related to issues of culture, leadership, management, economy,

environment, and even politics [2]. On the other hand, construction must ensure the
integration of the interests of all stakeholders, including the community, in meeting
its objectives. This is also related to the issue of the vulnerability of public trust due
to the use of public funds as a source of system funding.
Surabaya as the second biggest city in Indonesia has implemented e-procurement
system since 2004, which is considered successful in the Indonesian national scale.
The success of Surabaya in implementing this system in the midst of various global
issues related to the application of this system is an interesting thing to research.
This will inform how the Surabaya City Government implements the system and
becomes a national pilot, especially in terms of implementing the principles of
procurement, especially in the construction sector. As part of Indonesia which
also implements national procurement regulation, namely the Presidential Regu-
lation 16/2018, Surabaya have to apply the principles of efficiency, effectiveness,
transparency, openness, competition, fairness, and accountability in its procurement
system.
However, even though the procurement principles have been implemented, there
is still limited study which investigates these principles within Indonesian context.
According to that, this study aims to evaluate the implementation of the principles of
procurement public construction works in Surabaya. The data in this study were based
on a questionnaire which was distributed to construction sector involving contrac-
tors, consultants, and suppliers who have been involved in government construction
procurement in Surabaya.
2 Literature Review
2.1 Public Construction Procurement
Public procurement has been consistently used to advance public policy in a variety
of areas, including industrial policy, reducing unemployment, supporting small
businesses, local development, and equalizing wages [3]. In its implementation,
public procurement must guarantee the effectiveness of the implementation of work
contracts for those involved [4]. Meanwhile, construction procurement can be inter-
preted as a procurement system related to the purchase of construction services
to build new buildings or structures. This includes all construction works such as
repairing, maintaining, expanding, or demolishing a building or structure. Several
previous studies focused on irregularities, problems, or issues hindering the imple-
mentation of procurement, or the impact of the application of electronic technology
on the procurement system. For example, Vaidya and Campbell examine public
procurement in terms of information systems, e-commerce/e-government, supply
chain management, and their impact on procurement efficiency [5]. Detkova et al.
Evaluation of the Public Procurement Principles Implementation … 903
examined the relationship between procurement competition and the problem of

corruption in the form of violations [6].
2.2 Procurement Principles in Indonesia
Indonesia has seven principles of public procurement that should be implemented,

namely: efficient, effective, transparent, open, competitive, fair, and accountable.
Based on the principles, (i) Efficient is a condition that is closely related to efforts
to optimize the use of resources [7]; (ii) Effective can be defined as the use of
available resources to obtain goods and services that have the highest benefits [5];
(iii) Transparent means that this principle requires that information related to the
implementation of public procurement to be obtained through sources that are easily
accessible and freely available [8]; (iv) Open emphasizes providing all competent
providers of goods and services the opportunity to be involved in procurement [8]; (v)
Competitive emphasizes that public procurement should avoid having a single bidder
and prevent restrictions on competition [9]; (vi) Fair is related to equal treatment that
applies to every party or individual involved in the public procurement process [7];
and (vii) Accountable emphasizes the existence of an effective control mechanism
in public procurement accompanied by a clear and structured chain of responsibility
based on their respective authorities at each stage of public procurement [10].
Importance Performance Analysis (IPA) was used to analyze performance by

measuring customer satisfaction with a product or service. IPA is a simple tool,
but it has strong and strategic benefits in identifying the performance of product or
service attributes, so it can be seen that the attributes most need to be improved under
conditions of cost savings but not significant quality loss [11].
Research indicators were obtained from relevant previous studies which were
then validated using preliminary survey to understand their relevancy to Indonesian
context. There are 63 indicators of the 7 principles developed from various literatures
which were validated using preliminary survey involving 4 experts in the field with
academic and practitioner backgrounds and work experience ranges of at least 11–
15 years. From the survey, 48 indicators were considered relevant for this study. Due
to space limitation, the detail of the indicators is inserted in Table 1 together with
their performance assessment.
The 48 indicators were then used as criteria to evaluate the procurement princi-
ples. The main survey was conducted involving 30 respondents through questionnaire
distribution involving respondents from 18 contractors (60%), 10 consultants (33%),
and 2 suppliers (7%) who had been participating in public procurement in Surabaya
Table 1 The result of importance (I) and performance (P) analysis

No Indicators Source (s) Mean Gap score Conformity level (%)
P I
A Efficient
1 Use of electronic based [13] 3.967 4.633 −0.667 85.61
systems
2 Short processing time [5, 7] 3.733 4.500 −0.767 82.96
3 Decreased number of [5] 3.967 4.100 −0.133 96.75
objections or claims
4 Centralized access [13] 3.933 4.400 −0.467 89.39
5 Less processing cost [5] 3.833 4.467 −0.633 85.82
6 Reducing the number of [6] 3.767 4.533 −0.767 83.09
errors due to human
error
7 A classification of the [13] 4.200 4.533 −0.333 92.65
type of work or service
Mean A 3.914 4.452 −0.600 87.91
B Effective
1 The achievement of the [14, 15] 4.100 4.500 −0.400 91.11
objectives of the
procurement
implementation
2 Completion of the [16] 3.933 4.500 −0.567 87.41
procurement process on
time
3 Procurement is in [17] 3.800 4.600 −0.800 82.61
accordance with
community needs
4 Integration into [14] 3.600 4.367 −0.767 82.44
procurement programs
5 There is a control [13] 4.033 4.667 −0.633 86.43
mechanism for the
implementation of
procurement
6 Quality system [15] 3.867 4.467 −0.600 86.57
7 A communication [18] 3.767 4.300 −0.533 87.60
process that minimizes
face-to-face systems
Mean B 3.871 4.486 −0.614 86.31
C Transparent
1 Simple rules and [19] 3.833 4.633 −0.800 82.73
procedures
(continued)
Table 1 (continued)
P I
2 Access to procurement [18, 20] 4.033 4.033 −0.367 91.67
information
3 Availability of complete [4, 9] 4.033 4.633 −0.600 87.05
information
4 Easily accessible system [8] 4.067 4.533 −0.467 89.71
5 Visibility of the [21] 3.867 4.567 −0.700 84.67
procurement budget
6 The existence of a [22] 3.733 4.333 −0.600 86.15
control mechanism and
sanctions
7 Archiving procurement [21] 3.667 4.233 −0.567 86.61
documents
8 Mass media involvement [21] 3.533 4.367 −0.833 80.92
in the procurement
process
Mean C 3.846 4.463 −0.617 86.18
D Open
1 Acceptance of [4] 3.967 4.367 −0.400 90.84
participants based on
qualifications
2 Blacklisted business [23] 4.067 4.500 −0.433 90.37
entity publications
3 Disseminate information [24] 3.867 4.400 −0.533 87.88
on procurement
implementation
4 There are no restrictions [8] 4.067 4.500 −0.433 90.37
on participation due to
the participant’s origin,
location, religion,
ethnicity, or political
orientation
Mean D 3.992 4.442 −0.450 89.87
E Competitive
1 There is no intervention [4] 3.967 4.700 −0.733 84.40
from any party
2 There are clear rules [9, 25] 3.867 4.500 −0.633 84.40
regarding the
procurement
competition cycle
(continued)
Table 1 (continued)
P I
3 Participants have the [25] 4.000 4.567 −0.567 87.59
same opportunity to
compete
4 The procurement agency [26] 4.133 4.367 −0.233 94.66
is not affiliated with or
cooperating with any
particular party
5 Completeness of the [25] 4.200 4.500 −0.300 93.33
required documents
6 Involvement of three or [9, 25] 3.967 4.467 −0.500 88.81
more bidders as
participants
7 Compliance with [8] 3.967 4.433 −0.467 89.47
procurement ethics
Mean E 4.014 4.505 −0.490 89.11
F Fair
1 Winners are determined [4] 4.033 4.733 −0.700 85.21
based on eligibility for
qualifications
2 Avoiding conflicts of [24] 4.100 4.400 −0.300 93.18
interest
3 Equal distribution of [7] 3.967 4.667 −0.700 85.00
rights and obligations
among participants
4 There is a code of ethics [27] 4.067 4.467 −0.400 91.04
that regulates the
integrity of the
procurement
5 Each participant is [7, 20, 25] 4.100 4.433 −0.333 92.48
treated equally
6 The required technical [13] 3.967 4.367 −0.400 90.84
specifications must not
refer to certain brands
7 The number of [28] 3.833 4.400 −0.567 87.12
complaints against
procurement
implementation has
decreased
8 Bid consideration is [20] 3.967 4.567 −0.600 86.86
based on compliance
with bid documents
(continued)
Table 1 (continued)
P I
Mean F 4.004 4.504 −0.500 88.90
G Accountable
1 There is a strict [4, 21] 3.733 4.633 −0.900 80.58
monitoring mechanism
2 There are internal and [18, 21] 3.700 4.500 −0.800 82.22
external audits
3 Reporting System which [13] 3.833 4.400 −0.567 87.12
is open to the public
4 Professionalism standard [29] 3.967 4.600 −0.633 86.23
5 There are rewards for [4, 21, 25] 4.100 4.567 −0.467 89.78
those who have good
performance and
sanctions
6 Clear organizational [21, 29] 3.933 4.667 −0.733 84.29
framework
7 Conflict resolution [21] 4.100 4.600 −0.500 89.13
Mean G 3.910 4.567 −0.657 85.61
Total mean for overall 3.936 4.488 −0.552 87.55
principles
construction projects. The respondents had background of 5–20 years working expe-
rience and involved in at least 5 procurement processes in Surabaya. The question-
naire employed 5-point Likert scale, ranging from 1 (= not at all important) to 5 (=
extremely important) for the importance, and from 1 (= poor) to 5 (= excellent) for
the performance part.
4 Analysis Result
The data acquired from the main survey were analyzed using Importance Perfor-
mance Analysis (IPA). Table 1 shows the respondents’ perceptions regarding the
performance of the overall indicators. According to the results, it was found that
the indicators’ performance is below the expectation (level of importance). This is
indicated by negative gaps on the overall indicators.
The overall mean score of the indicators’ performance (P) was 3.936 while the
mean score of the indicators’ importance (I) was 4.488. As the gap between them
was −0.552 (smaller than 0), it means it still has not met the user’s expectation [12].
This also can be seen from the parameter of suitability level between the performance
and interests (%), which is still below the ideal (<100%).
In addition, the overall implementation of the procurement principles reached

87.35%. In more detailed, the implementation of the principle of ‘efficient’ has a
score of 87.91%, ‘effective’ of 86.31%, ‘transparent’ of 86.18%, ‘open’ of 89.87%,
‘competitive’ of 89.11%, ‘fair’ of 88.90%, and ‘accountable’ at 85.61%. Even though
the result can be considered good as the value of the conformity level is still less than
100%, it can be concluded that the quality of the services provided still has not met
the criteria or expectation [30].
5 Findings and Discussion
The analysis results can be discussed with regard to the indicators which had low
performance. For the principle of efficient, A2 (Short processing time) is the indicator
with the lowest score. The public procurement service system should be conducted
in short period of time [5, 7]. As the indicator had low score, it means this indicator
performance has not met the providers’ expectation, so it should be improved.
With regard to the principle of effective, indicator B3 (Procurement is in accor-
dance with community needs) is the indicator with the lowest score. The procure-
ment must be suitable with the community needs of the users. Non-conformity with
community needs will lead to waste of resources, including the budget. This can
occur due to lack of community involvement or lack of monitoring of the absorption
of priority community needs.
Indicator C8 (Mass media involvement in the procurement process) is the lowest
ranking in the ‘transparent’ principle. This means media has not been effectively
involved in the process. Indeed, the media is an important component for control-
ling the implementation of public procurement to ensure ‘transparent’ principle as
this can be source and channel of information regarding procurement planning and
implementation. The media also has a role in educating public regarding the public
procurement process [21].
Meanwhile, indicator D3 (Disseminate information on procurement implementa-
tion) is the lowest on the ‘open’ principle. Information dissemination is very impor-
tant to increase market responsiveness (potential bidders) to participate with compet-
itive bidding [24]. This low score could happen due to the impact of the relation-
ship with media involvement. As previously explained, media involvement is very
important as a source of information.
Indicator E2 (There are clear rules regarding the procurement competition cycle)
is the lowest on the principle of competition. This indicator relates to the clarity of
regulations, including those related to competition restrictions such as experience,
technical resources, origin, and other documentation requirements [25, 31]. The low
score of this indicator is probably due to the relatively complicated regulations. As
explained in previous studies, procurement procedures should be simple and flexible
[31].
In the ‘fair’ principle, the indicator with the lowest score is F3 (Equal distribution
of rights and obligations among participants). The low score of this indicator is
probably due to some parties who feel they are not being treated fairly in meeting the
qualification requirements. As explained by Khan, this indicator relates to the equal
treatment in meeting the qualification requirements, including the provision of equal
time and the process of obtaining information [7].
Furthermore, on the ‘accountable’ principle, indicator G1 (There is a strict moni-
toring mechanism) is the lowest indicator. This indicator relates to a clear mechanism
related to the responsibilities and roles of each party in each stage of procurement,
so that the implementation of procurement is directed and in accordance with the
planned [4]. The low score of this indicator is probably due to the current mechanism
which has not provided understanding to all parties, especially to the participants who
are involved in the procurement process.
6 Conclusion and Recommendation
This paper presents several attributes or indicators to measure the implementation

of the procurement principles in Surabaya construction project as well as to assess
their performance. Based on the analysis, the overall implementations of the public
procurement principles in Surabaya construction work can be said quite good. This
can be seen from the achievement of the conformity level which reached 87.55%.
However, this achievement is still not satisfactory as the gap between the importance
and performance implementation was −0.552. Based on these results, the imple-
mentation of these principles should be improved to achieve successful procurement
process.
Acknowledgements The authors thank the Indonesian Ministry of Research and Tech-
nology/National Agency for Research and Innovation for providing grants to have conducted this
research in 2021.
References
1. Mohsini R, Davidson CH (1991) Building procurement-key to improve performance. Build

Res Inf 19(2):106–113
2. Noor MA, Khalfan MMA, Maqsood T (2013) The role of procurement practices in effective
implementation of infrastructure projects in Pakistan. Int J Manag Proj Bus 6(4):802–826
3. Hawking P, Stein A, Forsters S (2004) E-Procurement: is the ugly duckling actually a swan
down under. Asia Pac J Mark Logist 16(1):1–26
4. Lynch J (2013) Public procurement: principle, categories and methods. Leanpub
5. Vaidya K, Campbell J (2016) Multidisciplinary approach to defining public e-procurement and
evaluating its impact on procurement efficiency. Inf Syst Front 18(2):333–348
6. Detkova P, Podkolzina E, Tkanchenko A (2018) Corruption, centralization and competition:
evidence from Russian public procurement. Int J Public Adm 41(5–6):414–434
7. Khan N (2018) What is public procurement: introduction. Public Procurement Fundamentals,
Emerald Publishing Limited
8. Hardjowijono B, Muhammad H (2009) Prinsip dasar dan kerangka hukum pengadaan barang
dan jasa pemerintah. Indonesian Procurement Watch, Jakarta
9. Lembaga Kebijakan Pengadaan Barang/Jasa Pemerintah and Millennium Challenge Corpora-
tion United States of America (2016) Modul 1: Menerapkan Prinsip-Prinsip Dasar Pengadaan,
Pelatihan Pengadaan—Tingkat Dasar, Proyek Modernisasi Pengadaan LKPP/MCA-I.
LKPP/MCA-I, Bogor
10. Organisation for Economic Co-operation and Development (2009) OECD principles for
integrity Public Procurement. OECD, Paris
11. Abalo J, Varela J, Manzano V (2007) Importance values for importance performance analysis:
a formula for spreading out values derived from preference rankings. J Bus Res 60(2):115–121
12. Shia BC, Chen M, Ramdansyah AD, Wang S (2016) Measuring customer satisfaction towards
localiation website by WebQual and Importance Performance analysis (case study on Aliex
Press Site in Indonesia. Am J Ind Bus Manag 6:117–128
13. Institute for Development of Freedom of Information (2018) Recommendations for the public
procurement system of the Republic of Indonesia. IDFI, Jakarta
14. Siagian SP (1982) Manajemen modern. Gunung Agung, Jakarta
15. Joesbury P (2016) Improving the effectiveness of procurement identification and improve-
ment of key determinant factors—the PEPPs project. Doctorate Thesis, Aston University,
Birmingham
16. Panayiotou NA, Gayialis SP, Tatsiopoulos IP (2004) An E-Procurement system for govern-
mental purchasing. Int J Prod Econ 79–102
17. Andriani S (2012) Analisa efektivitas hukum dalam penerapan pengadaan barang dan jasa
secara elektronik (e-procurement) serta peranan lembaga pengawas terhadap pengadaan barang
dan jasa pemerintah. Tesis Universitas Indonesia, Jakarta
18. Neupane A (2014) The potential of public e-procurement technology to reduce corruption in
public procurement. Doctorate Dissertation, University of Southern Queensland, Australia
19. Vicente RS (2008) Evaluation of efficiency and effectiveness of the regional procurement
service depots in the Philippines: a case of region 1. Master Thesis Massey University, New
Zealand
20. Krivinsh A, Vilks A (2013) Prevention of corruption in public procurement: Importance of
general legal principles. Jurisprudence 20(1):235–247
21. Organisation for Economic Co-operation and Development (2016) High-level principle for
integrity, transparency, and effective control of major events and related infrastructures. OECD,
Paris
22. Asian Development Bank (1999) International survey of government procurement system.
International Government Solution Ltd
23. Thai KV (2017) Global public procurement theories and practices. Springer International
Publishing, Switzerland
24. Eady R, Parera S, Heany G, Carlisle J (2007) Construction contractor’s perspective on web-
based advertising of public sector construction contracts. In: Boyd D (ed) Proceedings 23rd
annual ARCOM/association of researchers in construction management. Belfast, 767–776
25. Komakech RA (2016) Public procurement in developing countries: objectives, principles and
required professional skills. Public Policy Adm Res 6(8)
26. Racca GM, Perrin RC, Albano GL (2016) Competition in the execution phase of public
procurement. Public Contract Law J 41(1):89–108
27. Byung GS, Chang HL, Byoung CH, Hyunjeonget N (2019) Investigating the fair treatment of
suppliers and its trust fostering role and performance benefits. Int J Prod Econ 216:54–66
28. Jourdain R, Balgobin N (2003) Controlling corruption in Asia and pacific. In: 4th Regional anti-
corruption conference of the ADB/OECD anti-corruption initiative for Asia and the Pacific,
Kuala Lumpur
29. Diggs SN, Roman AV (2012) Understanding and tracing accountability in the public
procurement process. Public Perform Manage Rev 36(2):290–315
30. Supranto J (2006) Pengukuran tingkat kepuasan pelanggan untuk menaikkan pangsa pasar.
Rineka Cipta, Jakarta
31. Khorana S, Bouher KF, Kerr WA (2015) Governance Issues in EU’s e-Procurement framework.
J Common Market Stud 53(2):292–310
Relationship Model Between Conceptual
Cost Estimation Process of Flyover
Development in the Provincial
Government of DKI Jakarta
with the Accuracy Level
Putika Yussi, Yusuf Latief, and Rossy Armyn Machfudiyanto
Abstract The conceptual cost estimate for flyover development is the cost esti-
mate when the flyover construction does not yet have complete planning data. In the
DKI Jakarta Provincial Government, the forecast is used for the proposed flyover
construction budget in the 5 Year Strategic Plans and Multi-Year Activities. In the
estimation process, there is a level of accuracy that still needs to be improved, which
causes the regional development planning to be less effective because other alloca-
tions have shifted in priority. Various factors can cause this in the estimation process
that affects the level of accuracy. This study aims to identify the factors or vari-
ables in flyover construction’s conceptual cost estimation process in the DKI Jakarta
Provincial Government that affect accuracy. And then, to analyze the model of the
relationship between variables of conceptual cost estimation process to the level
of estimation accuracy, using the SEM-PLS method. 8 variables in the conceptual
cost estimation process, which affect the estimation accuracy, have been identified
through literature studies and expert validation. The variables are information quality,
scope quality, estimator performance, estimating procedures, project definition, cost
information, project characteristics, and government regulations. Based on respon-
dents’ results and the relationship model’s analysis with the SEM-PLS method, the
model produces 12 significant relationships between the variables of the conceptual
cost estimation process for flyover development, with an R-square value of 0, 279.
Keywords Conceptual cost estimation process · Flyover development ·

Accuracy · Structural equation modeling
1 Introduction
Conceptual cost is calculated without complete information from design details and
does not have a clear work scope [1]. Conceptual cost is also part of the feasibility
P. Yussi · Y. Latief (B) · R. A. Machfudiyanto (B)

Civil Engineering Department, Faculty of Engineering, University of Indonesia, Depok, Indonesia
P. Yussi
914 P. Yussi et al.
analysis at the start of a project, with limited information on project scope, design, and
technical data [2]. Meanwhile, a flyover is an intersection with the unique construc-
tion of a bridge built over a level intersection to allow two ways free passage on one
of the main roads. Its function is to increase traffic flow capacity and reduce traffic
congestion at intersections. And the traffic under the existing bridge intersection is
still usable [3]. For flyover construction budgeting in the DKI Provincial Government
that has not had complete planning data, the DKI Jakarta Provincial Government uses
the conceptual cost estimate for the proposed Multi-Year Activities proposal. By defi-
nition, accuracy means close to the actual value, minimal error [4]. The conceptual
cost estimation accuracy is also compared with the cost of completion, adjusted for
agreed changes [5]. According to True (1998), accuracy can be expressed as the
percent difference between the estimated value and the actual value [6]. Table 1
gives a range of conceptual cost estimates accuracy based on literature.
Based on Dinas Bina Marga DKI Provincial Government’s data in 2016–2021, 3
flyovers use conceptual cost estimate for the proposed Multi-Year Activities Proposal.
Each of them has 104.21; 23.67 and 12.47% for conceptual cost estimate accuracy.
The flyover with 23.67 and 12.47% conceptual cost estimates accuracy has not yet
reached optimal accuracy based on [5] and [7]. And the flyover with 104.21% accu-
racy, beyond the conceptual cost estimate accuracy range based on [5, 7], and [8].
It describes that the Provincial Government of DKI Jakarta’s problem is the accu-
racy of the conceptual cost estimation of flyover construction, which still needs
improvement.
Based on interviews with conceptual cost estimators at the Dinas Bina Marga of
the DKI Jakarta Provincial Government, there are several causes for the accuracy
of the conceptual cost estimation of flyover construction that needs to be improved.
Figure 1 uses a fishbone diagram to describe those causes.
In actual condition, this problem causes regional development planning to be less
effective or wasteful and less beneficial because other shifting allocations prioritize.
Meanwhile, from a literature review, the impact is that if the conceptual stage cost
estimate is too low, it will experience difficulties in the contract. If the forecast is
too high, it will reduce funds for other projects [9]. Based on this, the Provincial
Government of DKI Jakarta’s accuracy of the conceptual cost estimation of flyover
construction needs improvement.
This research aims to identify the factors or variables in flyover construction’s
conceptual cost estimation process in the DKI Jakarta Provincial Government that
affect the accuracy and analyze the relationship between the conceptual cost estima-
tion process variables with accuracy through a model. Therefore, this research can
Table 1 Conceptual cost

Source Conceptual cost estimate
estimate accuracy range
accuracy range
based on literature
Ballard and Pennanen [5] −10% – +10%
Asal [7] +10% – +20%
AACE [8] +20% – +30%
Relationship Model Between Conceptual Cost Estimation Process … 915
Fig. 1 Fishbone diagram of causes for the accuracy of the conceptual cost estimation of flyover
development needs to be improved
be used as a base to improve the accuracy of the conceptual cost estimation of flyover
development.
2 Literature Study
2.1 Conceptual Cost Estimation Process
The conceptual cost estimation process is a process to estimate project costs before
there are significant design information and clear scope of work. Due to the lack of
knowledge and the element of uncertainty, the conceptual cost estimation process
is carried out based on deep consideration and experience [1]. Bley (1990) states
that conceptual estimation is a process that is based mainly on judgment and knowl-
edge because of the lack of information and uncertainty that always exists at the
conceptual stage [2]. This statement is supported by Serpell [1], who stated that the
conceptual cost estimation process is a process to estimate project costs before there
are significant design information and clear scope of work. Due to the lack of knowl-
edge and uncertainty, the conceptual cost estimation process is carried out based on
deep consideration and experience. According to Honn An et al. [10], the conceptual
cost estimation process is one of the critical stages in the early stages of the project
cycle because it relates to several project uncertainties. It needs an estimator who
must have specific expertise and knowledge to reduce the level of risk caused by
uncertainty to be an acceptable level of risk in estimating the conceptual cost.
Conceptual cost estimates are generated at the conceptual stage where there is
little information, no design details are available, and the project definition is unclear.
916 P. Yussi et al.
The estimation approach looks at the project in outline and is calculated on a top-
down basis. This estimate is the basis for decision-making for funding so that if
this estimate is inaccurate, it will result in an incorrect funding decision [11]. At the
conceptual cost estimation stage, the project definition has not been fully defined.
Most of the estimators will calculate the conceptual cost based on the characteristics
of the project, which is an outline of the project [12]. In preparing the project budget,
the estimated performance is influenced, among other things, by cost data, which
comes from the Government and factors of regulatory conditions [13].
This research concludes that the conceptual cost estimation process is a series of
stages carried out in the early phases of the project when there is no clear project
information, scope definition, and project design. The process uses an approach
to analyze the project in general and calculate with top-down estimates using cost
information related to similar projects and government regulations. The process needs
estimators who have the expertise and knowledge to reduce the level of risk caused
by uncertainty to an acceptable level of risk to produce conceptual cost estimation.
2.2 Factors in Conceptual Cost Estimation Process

that Affect the Level of Accuracy
There is a need to identify the factors and determine what factors may affect the
level of accuracy in the conceptual cost estimation process. Several factors affect
the accuracy of conceptual cost estimates in the estimation process, which these
factors are essential to identify [6]. As mentioned by [10], which states that various
factors influence conceptual costs, it is necessary to identify the influencing factors
to determine its degree [4] also indicated that it is essential to identify critical factors
for the practical estimation of each project.
Based on literature studies, some factors affect accuracy at each stage of the
conceptual cost estimation process. The research performed by Serpell [6] identified
5 factors that affect the level of accuracy, namely scope quality, information quality,
uncertainty level, estimator’s performance, and estimating procedure. Serpell draws
on these factors from expert interviews. Then a qualitative causal model is created
using these factors to assess the accuracy and how this model is used to improve the
scoring system. The journal provides a further explanation of each factor’s details
that helps further understand and clarifies the meaning of each of these factors.
According to [10], the research uses information quality, uncertainty level, esti-
mator’s performance, estimating procedure, and project definition. This study aims
to detect and assess the factors that affect the cost estimation of a project. The author
seeks to provide other solutions to evaluate the factors that influence conceptual
cost estimates and evaluate conceptual cost estimates with a method that is easier to
understand than the method offered by Serpell [6].
The study performed by L Liu and Kai, who stated that output measurability is
likely to be low at the conceptual stage because the estimation is based on inaccurate
data, identified the critical factors for practical estimation at various stages of typical
construction. From 9 categories, there are essential factors that affect conceptual cost
estimates, namely: project information, team experience and cost data [14].
Meanwhile, [15], which conducted research based on the data analysis, produces
11 groups of factors that affect the accuracy of estimation, in the following order:
basic process design, team experience and cost information, the time allowed to
prepare the estimate, site requirement, bidding and labour climate, team alignment,
owner’s cost, contingency and reviews, formal estimating process, money issues,
technology issues. Then the five most dominant are taken, namely basic process
design, team experience and cost information, time allowed to prepare the estimate,
site requirements, and bidding and labour climate.
In [16], the research identifies factors that influence cost estimation in 5 construc-
tion phases, states that conceptual cost estimation is a rough estimate whose accu-
racy is effected by 12 factors. Among them are the clarity and adequacy of the client
brief, the volume of the consultant’s workload during estimation, project size and
complexity and estimation method. This study describes clearly through a detailed
table the factors and their influence on each phase. It can be seen what factors have
an impact and do not have an effect in that phase.
Thomas Pasco [17] explored project characteristics that affect the variances
between conceptual cost estimates and tender estimates to investigate the most effec-
tive solution to increase early-stage forecast accuracy. With the initial hypothesis,
Ballard [5] states that the conceptual cost accuracy rate can reach 10%, saying that
information quality, project definition, and target cost can affect accuracy. The results
of this study answer the initial hypothesis that conceptual cost accuracy can reach
±10% with a 95% confidence level with the Haahtela method.
According to [13], there are factors used to evaluate project cost proposals
and propose quantitative project costing procedures. These factors consist of
project conditions, project duration, environmental conditions, regulation conditions,
planning conditions, estimation conditions.
Eventually, this research uses factors that affect accuracy in the conceptual cost
estimation process based on those existing literature studies. These factors are
Information Quality, Project Definition, Estimator’s Performance, Scope Quality,
Estimating Procedure, Cost Information, Project Characteristics, and Government
Regulations.
2.3 Factor’s Definition in Conceptual Cost Estimation

Process that Affect the Level of Accuracy
After identifying the factors in the conceptual cost estimation process that affect
accuracy, each factor’s definition needs to be described.
X.1 Information Quality. Gustavsson and Wanstrom (2009) explain that informa-
tion quality is the ability of information to meet user needs. Information quality
918 P. Yussi et al.
is the embodiment of workers’ knowledge and customer expectations, which

consists of 2 elements: the level of truth and usefulness of the information [18].
The minor difference between the information required and the information
obtained increases the quality of the information [19].
X.2 Project Definition. Project definition is the result of the process of exploring
all aspects of a proposed project [20]. Project definition can also be understood
as a process by which stakeholder needs are identified [21, 22].
X.3 Estimator’s Performance. An estimator is a person who is responsible for
organising and analysing all information into estimates based on the expertise
and experience he has. With expertise, an estimator will better understand new
projects based on previous projects, collect data by making assumptions and
justifying existing information, and visualize the scope of work [2].
X.4 Scope Quality. Quality of scope is the suitability of investment returns or
results that have been delivered by the performance of an activity or project,
with the requirements set by project stakeholders [20].
X.5 Estimating Procedure. Estimating procedure is an activity stage or step-by-step
method determined to estimate the number and costs of uncertain resources
required for the project’s scope or activity to minimize the level of uncertainty
from the quality of the scope definition and maximize the opportunity for costs
[20].
X.6 Cost information. From a cost management perspective, the cost is an invest-
ment of resources in strategic assets, including time, finance, human resources,
and physical resources [20]. Meanwhile, concerning estimation, information
is a reference estimator for generating costs [2].
X.7 Project Characteristic. According to Skitmore (1991), project characteristics
are project features, which can be physical or non-physical characteristics,
which explain the nature of the target. The factors are objective and measurable
[23].
X.8 Government Regulations. Regulations can take many forms and can be
enforced by governments at various levels. At the local level, regulations
can take permit fees, installation and implementation of construction stan-
dards that directly increase construction costs or may cause delays resulting
in increased costs higher [24].
3 Conceptual Framework
The relationship between variables in the conceptual cost estimation process and the
level of accuracy based on literature consists of:
• Information Quality and Accuracy Relationship. The accuracy of conceptual cost
estimation is primarily determined by the level of information provided by both
current information and historical information [2, 6].
• Information Quality and Estimating Procedures Relationship. According to [25],
current cost data and historical cost data will affect the estimation method.
• Information Quality and Cost Information Relationship. The cost information

consists of historical data related to similar projects and historical data quality
[14]. Similar project data can be a source for conceptual estimation cost data and
can be the basis for calculating the bridge’s average cost per m2 [12].
• Project Definition and Accuracy Relationship. Project definition is one factor that
affects the accuracy of conceptual cost estimates. It increases the level of detail
definition to produce a more credible conceptual cost estimate [12].
• Project Definition and Estimating Procedure Relationship. The estimating proce-
dure used during the initial project design stage depends on its completeness; a
certain design percentage is complete [26].
• Project Definition and Project Characteristic Relationship. According to [12], with
an incomplete project definition, the estimator must see the project as a whole.
Thus, the estimation will focus on “bigger picture” characteristics such as project
location, potential environmental problems, and utility impacts.
• Project Definition and Cost Information Relationship. A project’s definition can
influence cost information in design quality, construction complexity, and the
project components’ condition [27].
• Estimator’s Performance and Accuracy Relationship. According to [2, 6, 10, 25],
estimator performance is one factor that affects the accuracy of conceptual cost
estimates [2, 6, 10, 25].
• Estimator’s Performance and Information Quality Relationship. Oberlender and
Trost (2001) state that quantity surveyor’s competence and experience determine
and affect information quality [25].
• Estimator’s Performance and Estimating Procedure Relationship. According to
Bley (1990), the estimator’s ability to visualize the scope of work from incom-
plete scope definitions is essential in the conceptual cost estimation process. The
estimator is a process coordinator and brings his expertise and experience in esti-
mating costs. Guidance is sometimes needed to assess the inputs that enter the
process [2].
• Scope Quality and Accuracy Relationship. The scope quality is one factor that
affects the accuracy of conceptual cost estimates [6].
• Scope Quality and Estimator’s Performance Relationship. According to Bley, the
availability of a complete scope definition is the essential factor in the conceptual
cost estimation process. An estimator’s first thing is learning and interpreting the
project scope and producing an estimated plan [2].
• Estimating Procedure and Accuracy Relationship. In its effect on the accuracy
level, the estimation procedure’s quality is influenced by the time available to
calculate conceptual costs [6, 10, 25].
• Cost Information and Accuracy Relationship. Cost information is one factor that
affects the accuracy of conceptual cost estimates [14, 15].
• Cost Information and Estimating Procedure Relationship. Cost data taken from
current data and historical data can affect the estimation method [25].
• Characteristic Project and Accuracy Relationship. According to several studies,
project characteristics are one of the factors that affect the accuracy of conceptual
cost estimates [2, 12].
920 P. Yussi et al.
• Characteristic Project and Estimator’s Performance Relationship. The estimator

prepares estimates based on project characteristics related to other factors, and
some are subjective [25].
• Characteristic Project and Cost Information Relationship. The use of previous
actual project cost data following the project’s physical data allows the formation
of cost factors to form conceptual cost estimates, with different types, for example,
the cost per meter or the cost per square meter [12].
• Government Regulation and Accuracy Relationships. According to [13, 28], regu-
lations are a factor that affects the quality of proposed costs and is an external
factor to the accuracy of estimates.
The relationship between variables and accuracy based on the literature study
was then used to compile 19 hypotheses. These hypotheses are shown in Table 2.
The arrows indicate that one variable has a significant effect on the other variable.
These hypotheses also form a relationship model and are considered as a model’s
component of SEM. The relationship model is described in Fig. 2.
Table 2 Hypothesis significance relationship table

No. Hypothesis No. Hypothesis significance No. Hypothesis significance
significance relationship relationship
relationship
1. Information quality 8. Estimator’s performance → 15. Cost information →
→ estimating information quality accuracy
procedure
2. Information quality 9. Estimator’s performance → 16. Project characteristic →
→ cost information estimating procedure estimator’s performance
3. Information quality 10. Estimator’s performance → 17. Project characteristic →
→ accuracy accuracy cost information
4. Definition of project 11. Scope quality → 18. Project characteristic →
→ estimating estimator’s performance accuracy
procedure
5. Definition of project 12. Scope quality → accuracy 19. Government regulation →
→ cost information accuracy
6. Definition of project 13. Estimating procedure →
→ project accuracy
characteristic
7. Definition of project 14. Cost information →
→ accuracy estimating procedure
Fig. 2 Relationship model
4.1 Research Methodology Stages
The research methodology in this study consists of several stages. In the early stage,
this study uses the literature study method to obtain variables and indicators in the
conceptual cost estimation process that affect estimation accuracy at the conceptual
stage. A review of several literature studies resulted in a conceptual definition of
the variables. Then based on the literature study, the conceptual definition is broken
down into indicators that can measure variables.
And then, 5 experts in relevant fields will validate the variables and indicators
with the criteria of having a minimum education of bachelor degree and having work
experience of at least ten years in construction or flyover planning. The result of
expert validation is analysed using the Delphi method. Delphi method is a method
to reach agreement from experts to deal with problems [29]. In this study, the ques-
tionnaire was given to five experts in several rounds. The first round is carried out
to obtain opinions from experts, which will be summarised to serve as the basis
for the second round until we reach a mutual agreement. This expert validation
impacts mutually agreed variables and indicators. Also, variables and indicators that
experience improvements in the form of additions, subtractions, or changes.
The results from expert validation are used for a pilot survey. The pilot survey
was conducted on 7 respondents with a minimum bachelor in engineering education
qualifications in this study. The purpose of this pilot survey was to find out whether the
delivery in the questionnaire was easy to understand or not. If there is a questionnaire
that is not easy to understand, it will be corrected.
922 P. Yussi et al.
After a pilot survey, the survey was conducted with respondents using a question-
naire with corrected questions based on input from the pilot survey. This survey was
conducted on 151 respondents with the criteria they had or are currently working in
the DKI Jakarta Provincial Government and have a background or work relationship
with flyover construction or planning. In this research, the number of respondents
has exceeded the reference. With the maximum number of arrows pointing to the
latent variable in the model, the minimum sample size needed is 84 [30].
Respondent survey results are compiled and applied to the relationship model
between variables in the conceptual cost estimation process to accuracy. Then the
model is analyzed using the SEM-PLS method with Smart-PLS software. The model
in SEM is described in Fig. 3.
Structural Equation Modeling (SEM) is a multivariate analysis technique that
combines factor analysis and regression methods, allowing researchers to perform
simultaneous analyses and examine a series of dependency relationships simulta-
neously [31]. This analysis determines the influence level of each indicator in the
variable on the level of accuracy, based on the hypothesized relationship model that
has been made previously. After the SEM analysis is carried out, then validation will
be carried out by experts.
Fig. 3 Relationship model in SEM

4.2 Research Variables and Indicators
Based on literature studies and validation experts, this research identifies 8 variables
in the conceptual cost estimation process that affect the accuracy. The variables are
information quality consisting of 6 indicators, project definition composed of 4 indi-
cators, estimator’s performance composed of 6 indicators, scope quality composed
of 6 indicators, estimating procedure composed of 7 indicators, cost information
composed of 7 indicators, project characteristics composed of 6 indicators, and
government regulations composed of 6 indicators. For accuracy, the variable has
2 indicators. Table 3 gives a list of variables and their indicators.
Table 3 Research variables and indicators

No Variables and indicator’s of variable References
X.1 Information quality
X.1.1 Quality of historical information [2, 6, 10, 25]
X.1.2 Quality of current information [2, 6, 25]
X.1.3 Accuracy of available information [6, 32]
X.1.4 Information accessibility [6, 32]
X.1.5 Fulfillment of user expectations [19, 32]
X.1.6 Meeting stakeholder needs [33]
X.2 Project definition
X.2.1 Planning definition level [15]
X.2.2 Plan drawing quality level [10]
X.2.3 The level of project complexity [10, 12, 26, 34]
X.2.4 Project communication level [21]
X.3 Estimator’s performance
X.3.1 Estimator’s experience and knowledge level [6, 10, 25, 35, 36]
X.3.2 Estimator’s commitment and perceptions level [6, 10, 25, 34]
X.3.3 Estimator’s ability to communicate [25, 36]
X.3.4 Estimator’s ability to see cost components and analyse project [36, 37]
scope
X.3.5 Estimator’s ability to make assumptions about project details [28, 37, 38]
X.3.6 Estimator’s ability to gain access to cost information [28]
X.4 Scope quality
X.4.1 Completeness of the scope relating to the level of consistency [6]
X.4.2 Project definition that is filled in for execution [2, 39, 40]
X.4.3 Stability or change in project scope related to owner commitment [6]
level
(continued)
924 P. Yussi et al.
Table 3 (continued)
X.4.4 Stability or changes in the project scope related to the level of [2, 6, 16]
complexity of the project
X.4.5 Level of scope control [26, 39]
X.4.6 Compliance with stakeholder understanding or requirements [39–41]
X.5 Estimating procedure
X.5.1 Estimated time [6, 10, 25, 42]
X.5.2 The expected level of accuracy [6, 25, 43]
X.5.3 Standard estimation process procedure [10, 26, 42]
X.5.4 Documentation [26, 42]
X.5.5 Communication in the estimation process [26]
X.5.6 Management support [26]
X.5.7 Estimation review [26]
X.6 Cost information
X.6.1 Accuracy, reliability and completeness of cost information [15, 25, 34]
X.6.2 Formulation of construction component costs, in the form of costs [12, 26]
per m2 based on similar previous projects
X.6.3 Formulation of construction component costs, in the form of costs [12, 26]
per m2 typical project section based on similar previous projects
X.6.4 Formulation of the cost of the non-construction component to [12, 15, 27, 43]
calculate contingency costs
X.6.5 Formulation of the cost of the non-construction component is in [12, 26]
the form of inflation
X.6.6 Price regulation [13, 44]
X.6.7 Economic conditions, market and price level [6]
X.7 Project characteristic
X.7.1 The physical characteristics of the project are the dimensions of [25, 45]
the project
X.7.2 The physical characteristics of the project include the type of [42, 46]
structure used
X.7.3 The non-physical characteristics of the project are the project [47]
location
X.7.4 The non-physical characteristics of the project are the type and [25, 42]
duration of the contract
X.7.5 Non-physical characteristics of the project in the form of project [25]
type
X.7.6 The uniqueness of the project conditions and the availability of [12]
information on the related data references
X.8 Government regulation
X.8.1 Application of applicable regulations regarding licensing [47]
X.8.2 Application of applicable regulations regarding technical road [12, 40, 44]
requirements, general requirements of bridges, construction
standards
(continued)
Table 3 (continued)
X.8.3 Implementation of regulations containing the budget proposal and [48]
budget phasing for strategic planning and multiple years
X.8.4 Establishment of regulations regarding fees [44]
X.8.5 Application of regulations regarding fulfillment of project [47, 49]
requirements relating to environmental policies and public
engagement
X.8.6 Application of building requirements within the DKI Jakarta area [50]
and accessibility technical requirements for buildings and the
environment
Y.1 Accuracy
Y.1.1 The percentage represents the difference between the conceptual [10, 51]
cost estimate and the actual cost of building the flyover
Y.1.2 The Percentage represents the average difference between the [25, 52]
conceptual cost estimate and the actual cost of building flyovers
Table 4 Analysis results on

Variable Construct reliability and validity
construct reliability and
validity based on results from Cronbach’s Composite AVE
SEM-PLS Alpha reliability
X.1 Information 0.704 0.818 0.529
quality
X.2 Project 0.621 0.839 0.723
definition
X.3 Estaimtor’s 0.851 0.889 0.574
performance
X.4 Scope quality 0.854 0.892 0.579
X.5 Estimating 0.785 0.853 0.538
procedure
X.6 Cost 0.837 0.877 0.507
information
X.7 Project 0.825 0.873 0.536
characteristic
8. Government 0.875 0.901 0.606
regulations
Y.1 Accuracy 0.813 0.914 0.842
Based on SEM-PLS analyzed, after testing the validity and reliability, the model
meets the condition that AVE is higher than 0,5 and composite reliability is higher
than 0,7 [53], Cronbach’s alpha is higher than 0,6 [54], as shown in Table 4. AVE for
926 P. Yussi et al.
Table 5 Analysis results on discriminant validity—Fornell-Larcker criterion based on results from

SEM-PLS
Variable Discriminant validity—Fornell Larcker criterion
X1 X2 X3 X4 X5 X6 X7 X8 Y1
X.1. Information 0.727
quality
X.2. Project 0.314 0.850
Definition
X.3. Estaimtor’s 0.517 0.565 0.757
Performance
X.4 Scope 0.519 0.550 0.726 0.761
Quality
X.5 Estimating 0.442 0.516 0.685 0.626 0.734
Procedure
X.6 Cost 0.402 0.417 0.636 0.574 0.666 0.712
Information
X.7 Project 0.329 0.384 0.561 0.520 0.553 0.566 0.732
Characteristic
X8. Government 0.190 0.318 0.376 0.470 0.519 0.522 0.474 0.778
Regulations
Y.1 Accuracy 0.241 0.293 0.311 0.367 0.259 0.436 0.245 0.401 0.918
each latent variable is higher than R2 with all other latent variables [53], as shown in
Table 5. Based on the bootstrapping results of the SEM-PLS analysis, the relationship
model’s output is as shown in Table 6.
Based on the SEM-PLS analysis results, from the 19 relationship hypotheses, there
are 12 hypotheses with a T-statistic value (significance value) above 1.96, which indi-
cates that the relationship between these variables has a significant relationship. The
significant relationships are information quality and cost information relationship,
project definition and estimating procedure relationship, project definition and cost
information relationship, project definition and project characteristics relationship,
estimator’s performance and information quality relationship, estimator’s perfor-
mance and estimating procedure relationship, scope quality and estimator perfor-
mance relationship, information cost and estimation procedures relationship, cost
information and accuracy relationship, project characteristics and estimator perfor-
mance relationship, project characteristics and cost information relationship, and
government regulations and accuracy relationship. Besides that, there is 1 hypoth-
esis (X.5 Estimating procedure and Y.1 Accuracy level relationship) with a T-statistic
value of 1.627, which is considered to have a significant relationship even though
it is less significant. Based on the T-statistic, the value of X.5 → Y.1 can increase
(although not powerful) as the respondents’ number in analyzing increases.
Table 6 Analysis results of relationship between variables based on SEM-PLS
No Hypothesis relationship T Statistics The relationship between No Hypothesis relationship T Statistics The relationship between
variable variable
1 Information quality → 1.042 Insignificant 11 Scope quality → estimator’s 9.728 Significant
estimating procedure performance
2 Information quality → cost 2.904 Significant 12 Scope quality → accuracy 1.029 Insignificant
information
3 Information quality → 0.836 Insignificant 13 Estimating procedure → 1.627 Less significant
accuracy accuracy
4 Definition of project → 2.097 Significant 14 Cost information → 3.680 Significant
estimating procedure estimating procedure
5 Definition of project → cost 2.942 Significant 15 Cost information → 2.972 Significant
information accuracy
6 Definition of project → 4.294 Significant 16 Project characteristic → 3.160 Significant
project characteristic estimator’s performance
7 Definition of project → 1.066 Insignificant 17 Project characteristic → cost 4.870 Significant
accuracy information
Relationship Model Between Conceptual Cost Estimation Process …
8 Estimator’s performance → 7.884 Significant 18 Project characteristic → 0.743 Insignificant

information quality accuracy
9 Estimator’s performance → 3.318 Significant 19 Government regulation → 2.536 Significant
estimating procedure accuracy
10 Estimator’s performance → 0.014 Insignificant
accuracy
927
928 P. Yussi et al.
Based on the outer loading results, the model has an R-square of 0.279, categorized
as weak [53]. The constructed variable, the accuracy of conceptual cost estimation,
is affected by information quality, project definition, estimator performance, scope
quality, estimating procedure, cost information, project characteristics, and govern-
ment regulations amounted to 27.9%. And this model has an NFI value of 0, 545,
which means that this model can represent the field’s actual conditions amounted to
54.5%.
After the SEM-PLS analysis is carried out, 3 experts validate the relationship
model. Expert validations show that they agree with significant variable relation-
ships based on SEM-PLS analysis, and there are 3 additional significant relation-
ships. The additional significant variables relationships are information quality and
accuracy relationship (from insignificant become significant relationship), the esti-
mator’s performance and accuracy relationship (from insignificant become signifi-
cant relationship), and the estimating procedure and accuracy relationship (from less
significant become significant relationship). There are differences in the relation-
ship model based on the results of SEM-PLS analysis and expert validation. These
differences can be illustrated by the relationship model’s result based on SEM-PLS
analysis in Fig. 4 and the relationship model’s result based on expert validation in
Fig. 5.
Fig. 4 The relationship model’s result based on SEM-PLS analysis

Fig. 5 The relationship model’s result based on expert validation
6 Conclusions
Based on literature studies and expert validation, 8 factors affect the accuracy of
conceptual cost estimates: information quality, scope quality, estimator performance,
estimating procedures, project definition, cost information, project characteristics,
and government regulations. And from SEM-PLS analysis, those factors affect the
accuracy as big as 27.9%. Other factors outside of this study still determine the rest.
Based on the SEM-PLS analysis of 19 relationship hypothesis results, 12 relation-
ships between variables in estimating the conceptual cost of flyover construction are
considered significant. In comparison, based on expert validation results to the SEM-
PLS, there are 3 additional significant relationships. So it becomes 15 relationships
between variables in estimating the conceptual cost of flyover construction is consid-
ered significant. These differences between SEM-PLS analysis and expert validation
show different perceptions of respondents, literature, and experts. Therefore, it is
suggested to do more outreach to respondents due to the difference in perceptions.
Based on the results of SEM analysis of the 8 factors that affect the level of
accuracy, it turns out that cost information and government regulations are factors
that have a significant direct effect on the level of accuracy. This result indicates that
these two factors may be the main cause of the accuracy in estimating the conceptual
cost of flyover development in DKI Jakarta Province needs to be improved. Suppose
they need to increase accuracy; first, they need to improve government regulations
and cost information. In addition, it is also required to enhance factors that have a
significant effect, either directly or indirectly, to cost information, such as information
quality factors, project definition, project characteristics, and estimator performance.
930 P. Yussi et al.
References
1. Serpell AF (2005) Improving conceptual cost estimating performance. 2005 AACE Int Trans
2. Phaobunjong K (2002) Parametric cost estimating model for conceptual cost estimating of
building construction projects, 2002
3. Slatoom N, Taneerananon P (2015) A study of a flyover-bridge—improved intersection. Eng
J 19(1):1–12. https://doi.org/10.4186/ej.2015.19.1.1
4. Jumas DY, Zainon N, Rahim FAM (2015) The significance of cost variables and their influences
on the conceptual cost estimation accuracy 2015(December)
5. Ballard G, Pennanen A (2013) Conceptual estimating and target costing. In: 21st Annual
conference of the international group for lean construction, IGLC 2013, January 2013:1–10
6. Serpell AF (2004) Towards a knowledge-based assessment of conceptual cost estimates. Build
Res Inf 32(2):157–164. https://doi.org/10.1080/0961321032000172373
7. Asal EM (2014) Factors affecting building construction projects’ cost estimating
8. AACE International (2020) Cost estimate classification system-as applied in engineering,
procurement, and construction for the building and general construction industries
9. Azman MA, Abdul Samad Z, Ismail S (2013) The accuracy of preliminary cost estimates
in Public Works Department (PWD) of Peninsular Malaysia. Int J Proj Manag 31:994–1005.
https://doi.org/10.1016/j.ijproman.2012.11.008
10. Honn An S, Cho H, Kyun Lee U (2011) Reliability assessment of conceptual cost estimates
for building. Int J Civ Eng, vol 9
11. Gardner BJ (2015) Applying artificial neural networks to top-down construction cost estimating
of highway projects at the conceptual stage
12. AASHTO (2013) Practical guide to cost estimating
13. Lai Y, Wang W, Wang H (2008) AHP- and simulation-based budget determination procedure
for public building construction projects. Autom Constr 17:623–632. https://doi.org/10.1016/
j.autcon.2007.10.007
14. Liu L, Zhu K (2007) Improving cost estimates of construction projects using phased cost
factors. J Constr Eng Manag 133(1):91–95. https://doi.org/10.1061/(ASCE)0733-9364(200
7)133:1(91)
15. Trost SM, Oberlender GD (2001) Predicting accuracy of early cost estimates based on estimate
quality. J Constr Eng Manag 127(June):173–182
16. Dandan TH, Sweis G (2019) Factors affecting the accuracy of cost estimate during various
design stages cost estimate. https://doi.org/10.1108/JEDT-08-2019-0202
17. Pasco T, Aibinu Ayodeji A (2008) Project factors influencing the accuracy of early stage
estimates. In: Proceedings from international conference on building education and research
(BEAR), 2008, pp 102–112
18. Woodall P, Parlikad AK (2013) Approaches to information quality management: state of the
practice of UK asset intensive organizations (January):0–19. https://doi.org/10.1007/978-1-
4471-2924-0
19. Eppler MJ, Christ DW (2000) Conceptualizing information quality: a review of informa-
tion quality frameworks from the last ten years. In: Proceedings of the 2000 conference on
information quality, 2000, January
20. AACE International (2019) Cost engineering terminology
21. Kuchta D, Chojnacka JM (2020) An approach to increase the sustainability of projects and
their outcomes in public sector through improving project definition
22. Cano JL, Lidón I (2011) Guided reflection on project definition. Int J Proj Manag 29:525–536.
https://doi.org/10.1016/j.ijproman.2010.04.008
23. Azman MA (2012) The accuracy of preliminary cost estimates in Public Works Department
(PWD) of Peninsular Malaysia dissertation submitted in fulfilment of the requirements for the
Degree of Master of Science Faculty of Built Environment University of Malaya. University
of Malaya, Kuala Lumpur
24. Emrath P (2011) How government regulation affects the price of a new home. HousingEco-
nomics.com
25. Azman MA (2012) The accuracy of preliminary cost estimates in Public Works Department
(PWD) of Peninsular Malaysia. University of Malaya, Kuala Lumpur
26. Anderson S, Molenaar K, Schexnayder C (2006) Final report for NCHRP Report 574: guidance
for cost estimation and management for highway projects during planning, programming, and
preconstruction
27. GSA Public Building Service (2007) Project estimating requirements. Washington
28. Enshassi A, Mohamed S, Abdel-Hadi M (2013) Factors affecting the accuracy of pre-tender
cost estimates in the Gaza Strip. J Constr Dev Ctries 18(1):73–94
29. Grime MM, Wright G (2016) Delphi method (2016). https://doi.org/10.1002/9781118445112.
stat07879
30. Kay W (2013) Partial least squares structural equation modeling (PLS-SEM) techniques using
SmartPLS. Mark Bull 24(November)
31. Hair JF, Black WC, Babin BJ, Anderson RE (2014) Multivariate data analysis
32. Alshikhi OA, Abdullah BM (2018) Information quality: definitions, measurement, dimensions,
and relationship with decision making. Eur J Bus Innov Res 6(5):36–42
33. Azemi NA, Zaidi H, Hussin N (2018) Information quality in organization for better decision-
making. Int J Acad Res Bus Soc Sci (January). https://doi.org/10.6007/IJARBSS/v7-i12/3624
34. Hatamleh MT, Hiyassat M, Sweis GJ, Sweis RJ (2018) Factors affecting the accuracy of cost
estimate: case of Jordan Engineer. Eng Constr Archit Manag (January). https://doi.org/10.1108/
ECAM-10-2016-0232
35. Lim B, Nepal M, Skitmore M, Xiong B (2015) Drivers of the accuracy of developers’ early
stage cost estimates in residential construction. J Financ Manag Prop Constr 21:4–20. https://
doi.org/10.1108/JFMPC-01-2015-0002
36. Alroomi A, Seok Jeong DH, Oberlender GD (2012) Analysis of cost-estimating competencies
using criticality matrix and factor analysis. J Constr Eng Manag 138(November):1270–1280.
https://doi.org/10.1061/(ASCE)CO.1943-7862.0000351
37. Carr BRI (1990) Cost-estimating principles. 115(4):545–551
38. Hwang S (2016) A Bayesian approach for forecasting errors of budget cost estimates. J Civ
Eng Manag 22(2):178–186. https://doi.org/10.3846/13923730.2014.897981
39. Kraus W, Kenneth R (1992) Project scope definition: a practical approach: a publication of the
.... Cost Eng 34(12):15–19
40. Project Management Institute (2017) A guide to the project management body of knowledge
(PMBOK guide), 6th edn. Project Management Institute, Inc.
41. AACE International (2019) Cost estimate classification system-as applied in engineering,
procurement, and construction for the process industries
42. Arif F, Lodi S, Azhar N (2015) Factors influencing accuracy of construction project cost
estimates in Pakistan: perception and reality. Int J Constr Manag (July). https://doi.org/10.
1080/15623599.2015.1012141
43. Dysert LR (2006) Is ‘estimate accuracy’ an oxymoron? AACE Int Trans, pp 1–5
44. Khemani R, Shapiro DM (1999) Glossary of industrial organisation economics and competition
law. Organization for Economics Co-Operation And Development
45. Elbeltagi E, Hosny O, Abdel-Razek R, El-Fitory A (2014) Conceptual cost estimate of Libyan
highway projects using artificial neural network. J Eng Res Appl 4(8):56–66. www.ijera.com
46. Fayek AR, Ruben J, Flores R (2010) Application of fuzzy logic to quality assessment of infras-
tructure projects at conceptual cost estimating stage. 1147(December):1137–1147. https://doi.
org/10.1139/L10-036
47. Washington State Department of Transportation (WSDOT) (2015) Cost estimating manual for
projects
48. Jakarta PPD (2013) Peraturan Gubernur Provinsi DKI Jakarta Nomor 100 Tahun 2013
49. Singh RR, Chugh G (2016) Various risks involved in highway projects. Int J Innov Res Sci
Technol 3(06):131–135
50. Jakarta PPD (1991) Peraturan Daerah Daerah Khusus Ibukota Jakarta Tentang Bangunan Dalam
Wilayah Daerah Khusus Ibukota Jakarta
932 P. Yussi et al.
51. Aibinu AA, Pasco T (2008) The accuracy of pre-tender building cost estimates in Australia.
Costr Manag Econ 6193. https://doi.org/10.1080/01446190802527514
52. Skitmore M (1991) Early stage construction price forecasting: a review of performance
53. Hair JF, Ringle CM, Sarstedt M (2011) PLS-SEM: indeed a silver bullet. J Mark Theory Pract
19(2):139–151. https://doi.org/10.2753/MTP1069-6679190202
54. Memon AH, Badullah NH, Rahman IA (2013) Using structural equation modelling to assess
effects of construction resource related factors on cost overrun. World Appl Sci J (January).
https://doi.org/10.5829/idosi.wasj.2013.21.mae.9995
The Sustainability Aspect
of the Consulting Firm in Terms of Its
Competitiveness in Indonesia
Nofita Harwin, Mairizal, Zelmi Sriyolja, Abd Rahman Mohd Sam,

and Mohd Zaimi Abd Majid
Abstract Infrastructure development aims to improve services and accessibility.

In terms of realizing the ideals of independence, infrastructure development for
economic growth is a must. The government increased the infrastructure budget in
the 2020 State Revenue and Expenditure Budget or RAPBN by 4.9%. To achieve this
economic growth scenario, we need quality human resources, thus contributing to
Indonesia’s competitiveness in the 4.0 industrial era. Reliable, competent, and quality
resources to compete will also encourage the growth of business entities in all areas,
including consulting firms. The term competitiveness comes from the word power
which means strength, and the word competitiveness which means to achieve more
than others, or differ from others in terms of quality, or have certain advantages. The
determinants of competitiveness are determined by factor conditions, demand condi-
tions, relevant supporting industries, and company strategy & competitive structure.
In this study, a literature review methodology is used to see what competitiveness
factors are needed by a consulting firm to sustain the sustainability of getting a project
every year. The results of this study obtain 13 competitiveness factors and two deter-
minant factors that impact the sustainability of consulting companies validated by
an expert.
Keywords Sustainability aspect · Consulting firm · Competitiveness
N. Harwin · Mairizal · Z. Sriyolja · A. R. M. Sam · M. Z. A. Majid

School of Civil Engineering, Faculty of Engineering, Universiti Teknologi Malaysia,
81310 Skudai, Johor, Malaysia
A. R. M. Sam
e-mail: [email protected]; [email protected]
M. Z. A. Majid
Mairizal (B)
Program of Industrial Engineering, Faculty of Engineering, Universitas Pamulang, 15417
Tangerang Selatan, Banten, Indonesia
934 N. Harwin et al.
1 Introduction
1.1 Infrastructure in Indonesia
Infrastructure development aims to improve services and accessibility. In order to

realize the ideals of Indonesian independence, infrastructure development built for
economic growth is a must. One of the important factors in infrastructure develop-
ment is the quality of human resources [1]. As reported by the McKinsey Global
Institute in September 2012, Indonesia will become the 7th largest economy in the
world by 2050. In a report entitled “Archipelago Economy: Unleashing Indonesia”,
it is stated that Indonesia’s economic growth is very fast and fast. Indonesia will
become the 4th largest country in the world for the largest GDP, which is equivalent
to 7275 USD billion by 2050 [2].
In this regard, the government of Indonesia has increased the infrastruc-
ture budget allocation in the 2020 year State Revenue and Expenditure Budget
(Rancangan Anggaran Pendapatan dan Belanja Negara) posture by 4.9% to IDR
419.2 trillion. The 2020 infrastructure budget allocation consists of Rp. 405.1 tril-
lion for economic infrastructure, Rp. 8.7 trillion for social infrastructure, and Rp. 5.3
trillion for supporting infrastructure (Fig. 1).
Fig. 1 2015–2020 Infrastructure growth and budget. Sources https://databoks.katadata.co.id/dat

apublish
The Sustainability Aspect of the Consulting Firm in Terms … 935
1.2 Infrastructure in Indonesia
To achieve this economic growth scenario, Indonesia needs the support of quality
human resources, thereby contributing to Indonesia’s competitiveness in the indus-
trial era 4.0. Sources that can be trusted, competent and qualified to compete will
also promote the growth of commercial entities in all fields, including negotiating
companies [3].
The need for infrastructure investment in Indonesia amounts to USD 1.7 trillion
until 2045. If on average every year he needs USD 77 billion, together with IDR 1078
trillion (Global Infrastructure 2017). The need for large infrastructure, of course, the
need for a negotiation agreement is also very large. In 2018 the number of negotiators
accompanying Inkindo was 4776 companies, a decrease from the previous year’s
5053 companies.
2 Literature Review
2.1 Sustainability
In research conducted by Mairizal [4] it has been explained that what is meant
by sustainability according to Kotler and Keller [5] is an ability to meet human
needs without harming the future of the current generation which is the top of
many company agendas. Large companies have described in detail how they try
to increase the long-term impact of all their actions on society and the environment.
As one sustainability consulting company has also said, “There are three very basic
things—people, planet and profit—and also which parts of society should come first.
Sustainability. This means more than just being friendly to people, environment, and
that also means that you are there for a long time.
While in other Refs. Kotler [6] and Ermon [7] sustainability is a dynamic concept,
not a static one, in which the concept is built through intensive interactions between
the company and its dynamic business environment. Sustainability is the result of a
series of incremental changes and step-by-step efforts which collectively become a
large cumulative result [4].
2.2 Consulting Firm
Consulting companies are usually companies that have various experts or profes-
sionals who have experience in certain fields and also have extensive knowledge
about these various problems [8]. There is also another opinion that gives meaning
to professional consultants or consultants according to Shenson [9], which states that
individuals or companies have special skills and talents available to those who need
them (clients), by receiving a certain amount of compensation (wages). The job of a

professional consultant is to advise and often assist others in carrying out that advice
both on the same side and for the client.
According to Darmadi’s opinion, written in Diversity of Indonesia [10], stated
that the first consultant was Arthur D. Little who founded his business in 1886
in Cambridge, Massachusetts. He also provides technical (technical) assistance to
his clients. A man named Booz Allen Hamilton later founded a company with a
similar structure in the early twentieth century, after the first company went bankrupt.
After that in 1926, James McKinsey a professor from the University of Chicago
also founded a company with the field of “accounting advisor and engineer” which
introduced the concept of a different approach and framework. Mc Kinsey did not
recruit traditional engineers, but instead employed experienced executives who were
trained in the various collections of contemporary analysis and knowledge concepts
that existed at that time, which included strategy, policies, objectives, organization,
procedures, facilities and personnel. In historical facts, several extraordinary innova-
tions have been made by the Boston Consulting Group (BCG). They used a different
approach, and BCG developed the concept of the growth division matrix which is
a tool for assessing the attractiveness of a company in various industries. Finally,
the framework they created was later adopted by many other consulting firms in
understanding the problems and business opportunities that could be exploited. So
since then, various consulting companies in the world have shown themselves to be
growing and developing rapidly. Some of them have merged with others, and some
have changed their brands and changed their organizational structures. Likewise,
the approach taken, the methodology and framework used and developed are also
increasingly complex and more comprehensive.
2.3 Competitiveness
The history of the scientific field notes that the term competitiveness comes from the
word “power” which means strength, and the word “competitiveness” which means
achieving more than others, or something that is very different from others in terms of
certain qualities, or has certain advantages too. Competitiveness can Competitiveness
can also be interpreted as a strength that is used to try to be superior in certain matters
and is carried out by certain groups or institutions [4, 7, 11–13].
According to the opinion of Porter [14] in research by Mairizal [4], it is said that
the meaning of competitiveness or competition is the essence of the success or failure
of a company, where competition is very similar to an activity in a company that will
be able to contribute to its performance. such as innovation, a cohesive culture, or
good practice. In addition, the meaning of the word competitive strategy is an effort
to find a competitive position in an industry that is the basis of competition, which
aims to form a profitable and sustainable position. Sustainable prosperity is governed
by an underlying economic competitiveness which is largely determined by the level
of productivity achieved by a company.
Table 1 Competitiveness
No. Author’s Competitiveness factors
factors
1 Porter [13], Ermon [7], Human resources
Mairizal [17]
2 Porter [13], Ermon [7], Physical resources
Mairizal [17]
3 Porter [13], Ermon [7], Knowledge resources
Mairizal [17]
4 Porter [13], Ermon [7], Capital resources
Mairizal [17]
5 Porter [13], Ermon [7], Infrastructure
Mairizal [17]
6 Porter [13], Ermon [7], Demand segment
Mairizal [17]
7 Porter [13], Ermon [7], Sophistication
Mairizal [17]
8 Porter [13], Ermon [7], Anticipation
Mairizal [17]
9 Porter [13], Ermon [7], Material supply
Mairizal [17]
10 Porter [13], Ermon [7], Supporting equipment
Mairizal [17]
11 Porter [13], Ermon [7], Product innovation
Mairizal [17]
12 Porter [13], Ermon [7], Supply of experts
Mairizal [17]
13 Porter [13], Ermon [7], Company qualification
Mairizal [17]
14 Porter [13], Ermon [7], Strategy focused
Mairizal [17]
15 Porter [13], Ermon [7], Management structure
Mairizal [17]
16 Porter [13], Ermon [7], Future orientation
Mairizal [17]
Meanwhile, according to a competitiveness expert, Porter [13] in the discussion

of his book The Competitive Advantage of Nations, states that the potential compet-
itiveness of a country or company comes from several sources, namely 4 groups of
factors, namely: Factor Condition, Demand Condition, Related & Supporting and
Firm Strategy. There are consist of 16 sub-factors of competitiveness (Table 1).
Factor Condition
Competitive- People
Demand Condition ness
Planet
Related & Supporting
Sustainability Profit
Firm Strategy
Fig. 2 Relationship model between competitiveness and sustainability. Source Porter [13] and
developed by researcher
2.4 Relationship Model Between Competitiveness

and Sustainability
In this study it will be seen to what extent the relationship between aspects of
sustainability with competitiveness factors as illustrated in Fig. 2.
3.1 Data Collecting
This paper uses two types of statistical approaches that can be used in research,
namely quantitative and qualitative approaches [15, 16]. The two statistical
approaches differ in terms of the methods and techniques used such as objectives,
concepts, research approaches, data collection methods, data analysis and sampling
[4, 15]. In the early stages of this study only collected all forms of relevant liter-
ature on issues of sustainability and competitiveness as mentioned by the various
researchers reviewed. Literature review is not only limited to articles published in
peer-reviewed and prestige journals, but also covers various theses and books. The
literature search dates ranged from 1990 and 2018 to cover a wide range and at
the same time capture new findings. The next stage, all factors or aspects related
to sustainability and competitiveness are collected, and then used as a questionnaire
which is distributed to respondents using the purposive sampling method. All respon-
dents involved had experience in technical consulting activities in various projects
in Indonesia. The questionnaire format consists of some general information and the
respondent’s background, followed by data on the importance of competitiveness
factors identified through the literature as shown in Tables 1 and 2.
Table 2 Sustainability factors

No. Author’s Sustainability factors
1 Kotler and Keller [5], Ermon [7], Mairizal [17] People
2 Kotler and Keller [5], Ermon [7], Mairizal [17] Planet
3 Kotler and Keller [5], Ermon [7], Mairizal [17] Profit
Delivery of questionnaires to experts is also through self-administered question-

naires. A total of 20 responses have been received in full. Respondents consist of
contractors, consultants and project owners. The Likert scale consists of five points,
which is used to measure the importance of competitiveness for the sustainability of
a consulting company. All data obtained were analyzed using descriptive statistical
methods. All data from the questionnaire answers were analyzed using the Statis-
tical Package for Social Sciences (SPSS). The results of the reliability value were
calculated and found to be above 0.70 (Cronbach’s Alpha) while the validity value
must exceed the “r” value in the reference table. The highest factor value will be a
determining factor for competitiveness in the sustainability of consulting companies
in Indonesia.
3.2 Data Analysis
Data processing uses SPSS version 24.0 and according to Supriyadi [18], the testing
techniques that are often used to test validity are Pearson Bivariate Correlation
(Pearson Moment Product) and Item-Total Correlation Correction. In this paper using
Corrected Item-Total with the results for the amount of data (n) = 16 obtained r table
= 0.4259 (Table r for Pearson Product Moment) for a significance level of 0.05. The
results of the SPSS data analysis can be seen in Table 3.
Referring to Table 3 of the Corrected Item-Total Correlation value, the calculated
r value (Pearson Correlation) is smaller than r table 0.4259, namely X1 (Human
Resources), X2 (Physical Resources) and X5 (Infrastructure). In the reliability testing
method is to use the Cronbach’s Alpha method whose value is >0.7 which is used to
determine reliability and its shown the Cronbach’s Alpha of all factors >0.700.
Table 3 Corrected item-total correlation results

Scale mean if item Scale variance if Corrected item-total Cronbach’s Alpha if
deleted item deleted correlation item deleted
X1 60.9000 67.463 0.135 0.904
X2 61.2500 65.987 0.174 0.906
X3 61.5000 59.000 0.728 0.888
X4 61.3500 58.871 0.775 0.887
X5 61.5000 63.421 0.392 0.899
X6 62.8000 56.484 0.777 0.886
X7 62.0500 59.418 0.618 0.892
X8 61.9000 57.358 0.673 0.890
X9 61.9000 57.779 0.643 0.891
X10 61.8000 57.432 0.659 0.891
X11 61.3500 61.397 0.545 0.895
X12 61.7500 61.250 0.525 0.895
X13 61.3500 61.608 0.527 0.895
X14 61.6500 58.976 0.808 0.887
X15 61.7000 60.116 0.505 0.897
X16 61.5000 61.842 0.481 0.897
4.1 Result
The first step has to ratify the Competitiveness Factors of sixteen (16) Porter’s
Competitiveness Factors in 3 Concepts of Competency Theory. The analysis of
thirteen (13) factors what has reliable and valid in this research such as follows:
1. Knowledge Resources
2. Capital Resources
3. Demand Segment
4. Sophistication
5. Anticipation
6. Material Supply
7. Supporting Equipment
8. Product Innovation
9. Supply of Experts
10. Company Qualification
11. Strategy Focused
12. Management Structure
13. Future Orientation.
4.2 Discussion
Seen from the highest means of the 13 competitiveness factors, there are 2 compet-
itiveness factors with the highest value, namely X6 (Demand Segment) with means
= 62.8000 and X7 (Sophistication) with means = 62.0500. This means that these
two factors are determining factors in the competitiveness of a consulting firm in
Indonesia. All 13 factors mentioned above are mentioned to 3 Competitiveness
experts and quite understanding about the competitiveness concept and they agree
that all of these factors apply to the competitiveness of consultants in Indonesia.
It is argued that all marketing strategies involve the search for competitive advan-
tage (Bharadwaj and Varadarajan 1993; Day and Wensley 1988; Varadarajan and
Cunningham 1995). For market segmentation strategy factors are based on the
achievement of competitive advantage and thus superior financial performance is
generated by the company by: (1) identifying demand segments, (2) targeting specific
segments, and (3) developing specific marketing “mix” for each segment. targeted
market (Dibb, Simkin, Pride, and Ferrell 1994; Hunt 2002b) [19, 20].
On the other hand, in research Hunt and Arnett [19] say as confirmed by
Shawney [21], p. 54, that “Customers become very sophisticated and demand prod-
ucts and services that are tailored to individual preferences and tastes.“ Likewise
with Lancaster (1990) who argues that product variation is the result of consumers
looking for variations for their own consumption and/or different consumers will
want different variants of different flavors. From this stated perspective, a company
that uses the concept of a market segmentation strategy will actually benefit both
consumers and society by providing market offerings that better satisfy individual
consumers’ own wants and needs.
5 Conclusion
From all parts of the discussion above, the competitiveness of consulting firms in
Indonesia is determined by 13 competitiveness factors referring to Porter’s Compet-
itiveness and previous research. From the 13 competitiveness factors, the results of
data analysis using SPSS, obtained 2 competitiveness factors with the highest means
value, namely Demand Segment and Sophistication. This means that consulting firms
in Indonesia must take these two competitiveness factors into account in planning a
competitive strategy for projects in increasingly fierce competition.
References
1. Bappenas (2017) Bahan Paparan Rapat Kerja Bappenas - Kementerian Lembaga dan Temu
Konsultasi Triwulanan I Bappenas - Bappeda Provinsi Seluruh Indonesia pada Senin, 22
Februari 2016. Kementerian PPN/Bappenas
2. Price Waterhouse Coopers. Indonesia Akan Menjadi Negara Dengan Perekonomian Terbesar
ke-4 di 2050. https://www.pwc.com/id/en/media-centre/pwc-in-news/2017/indonesian/pwc--
indonesia-akan-menjadi-negara-dengan-perekonomian-terbesar-.html
3. Inkindo (2018) Roadmap Menuju INKINDO Emas 2030: “INKINDO Mandiri Demim
Keunggulan Negeri”. Ikatan Nasional Konsultan Indonesia (INKINDO), Jakarta, Indonesia
4. Mairizal (2020) Kompetensi Pengurus Projek Dalam Penilaian Prestasi Projek Pembinaan
Bersepadu di Indonesia. Universiti Teknologi Malaysia (UTM), Johor, Malaysia
5. Kotler P, Keller KL (2012) Marketing management, 14th edn. Prentice Hall, New Jersey
6. Kotler P (2003) Marketing Management, 11th edn. Prentice Hall. Inc., New Jersey
7. Ermon DHN (2011) Strategi peningkatan daya saing untuk meningkatkan profit perusahaan
jasa konstruksi, UI, Depok, Indonesia
8. Tordoir PP (1995) The Professional Knowledge Economy. The Management and Integration of
Professional Services in Business Organizations. Springer Science+Business Media Dordrecht,
Springer Nature Switzerland AG
9. Shenson HL (1990) The Contract and Fee-Setting Guide for Consultants and Professionals.
Penerbit Wiley, USA
10. Darmadi S (2011) Board diversity and firm performance: the Indonesian evidence. Diversity
of Indonesia. Corporate Ownership Control J 8:1–38
11. Durand M, Giorno C (1987) Indicators of international competitiveness: conceptual aspects
and evaluation. OECD J: Econ Stud, ISSN 1995-2848, ZDB-ID 858174-5. 9:147–182
12. Hudori M (2013) analysis of competitiveness of the agribusiness sector companies using
Porter’s five forces. In: Proceedings: 2nd international conference on adaptive and intelligent
Agroindustry (ICAIA) IPB International Convention Center Bogor—Indonesia, Sept 16–17
(2013)
13. Porter ME (1990) Competitive advantage-creating and sustaining superior performance. The
Free Press, New York
14. Porter ME (2004) Global competitiveness report 2004–2005: findings from the business
competitiveness index, Press Conference, New York
15. Chua YP (2001) Kaedah dan statistik penyelidikan: kaedah penyelidikan. Mcgraw-Hill
Education, Kuala Lumpur, Malaysia
16. Fielding N, Schreier M (2001) Qualitative and quantitative research: conjunctions and diver-
gences. Forum: qualitative social research. Center for Digital Systems, Freie Universität
Berlin
17. Mairizal (2016) Perancangan Strategi Keunggulan Bersaing Perusahaan EPC. Institut Sains &
Teknologi Nasional, Jakarta, Indonesia
18. Supriyadi (2014) Edy.SPSS+AMOS, Perangkat lunak statistic, In Media, Jakarta
19. Hunt SD, Arnett DB (2004) Market Segmentation Strategy, Competitive Advantage, and
Public Policy: Grounding Segmentation Strategy in Resource-Advantage Theory. Australasian
Marketing Journal (AMJ) 12(1):7–25
20. Hunt SD, Arnett DB (2006) The Explanatory Foundations of Relationship Marketing Theory.
Journal of Business & Industrial Marketing 21(2):72–87
21. Shawney MS (1998) Leveraged high-variety strategies: from portfolio thinking to platform
thinking 26(1)54–61
Sustainability Aspects in Water Resources
Management
Sustainability Analysis of Minimization
of Spills from a Reservoir
Syamsul Hidayat, Ery Setiawan, Ida Bagus Giri Putra, M. Bagus Budianto,
and Salehudin
Abstract Reservoir operation in general is planned towards either maximization of

benefit or minimization of loss. The operation plans may take form in optimization,
simulation or combination of the two. Spills from reservoir or system of reservoirs
can be considered as a loss, so that it must be minimized. In order to get a better
perception of a study reservoir’s behavior to which minimization of spills is imple-
mented, sustainability of the plan needs to be assessed. The present work is aimed
to evaluate the sustainability of an optimization-simulation model applied to the
Batujai reservoir in Lombok Island, Central Indonesia. The model is set to minimize
the spills from the study reservoir system, after which the sustainability of spill mini-
mization is appraised in terms of reliability, resilience and vulnerability. The model
was run at three consecutive years (2014–2017) so that sensitivity of the model to
hydrologic variability can be evaluated. Results show that reliability, resilience and
Sustainability Index (SI) clearly have same trends while vulnerability is in inversely
trend with these three measures. In contrast, no conclusion can be made about the
trade-offs between SI and irrigable area.
Keywords Sustainability · Minimization · Spills
1 Introduction
Reservoir management and operations deal with minimizing the loss caused by water
shortages, flooding, adverse environmental impacts; or optimizing the benefit of
S. Hidayat (B) · E. Setiawan · I. B. G. Putra · M. B. Budianto · Salehudin

Civil Engineering Department, Faculty of Engineering, University of Mataram, Mataram,
Indonesia
E. Setiawan
I. B. G. Putra
946 S. Hidayat et al.
using natural resources such as land, water and energy, allocating water between
multiple users and uses, and developing reservoir regulation strategies and operating
rules [1]. Systems Approach is a concept underlying associated models and analysis
methods in reservoir management and operations, which is aimed to improve the
decision making process by applying optimization, simulation, multi-objective or
combination of these previous three techniques to a whole picture of the reservoir
problems [2].
Existing operating rules in multi-reservoir systems which have been formulated
based on simulation or experience can be found, for instance, in [3, 4]. One of the
purposes of these rules summarized in [3] is to gain a better understanding of the
study system’s behavior. In spite of their usefulness for quite a long time, alterations
to these rules may be required due to changes in objectives and demands. Meanwhile,
optimization refers to mathematical formulation which aimed to seek the maximum
or minimum values of objective functions subjected to constraints [5]. While it is
often said that simulation is descriptive and optimization is prescriptive, sometimes
simulation cannot be distinguished clearly from optimization as some optimization
models also simulate the study water system [1].
Objective functions in optimization may vary from irrigation, flood control,
hydropower production, and environmental preservation [6], recreation and navi-
gation [7]. A huge number of studies have been accomplished in optimizing these
objectives, employing various techniques such as linear programming e.g. [8], non-
linear programming e.g. [9], dynamic programming e.g. [10], network flow program-
ming e.g. [11], neural network e.g. [12], fuzzy logic e.g. [13] and heuristic techniques
e.g. [11]. Little to no attention, however, has been paid to sustainability analysis of
the optimization results in spite of the importance of such an analysis as a means to
gain a better understanding of a study water system’s behavior.
The present work is aimed to a better understanding of the behavior of a reservoir
to which an optimization plan is implemented. The objectives of the present work are
twofold: firstly, to develop an optimization model to minimize spills from the study
reservoir, and secondly to analyze the sustainability of the optimization results. The
sustainability is measured through the most often used terms which are reliability,
resilience and vulnerability.
2 Methods
2.1 Study Reservoir System and Data
Batujai reservoir, built in 1982, is located in Central Lombok Regency which is about
25 km from Mataram, the capital of West Nusa Tenggara Province. The specification
and the location of the reservoir system is summarized and depicted in Table 1 and
Fig. 1, respectively.
Sustainability Analysis of Minimization of Spills … 947
Table 1 Specification of the

Specification Batujai reservoir
study reservoir system [14]
Dead storage (MCM) 23.00
Maximum storage (MCM) 1.40
Purposes Irrigation, flood control, raw water
supply
Irrigation command area 3350 ha
Fig. 1 Study water system [14]
The optimization was conducted for three years from November 2014 to October
2017, which bi-monthly was taken as optimization time step. One year consists of
three cropping seasons lasting for four months each with the first season commences
on Nov and cropping patterns paddy-paddy-corn. The evaporation rate is expressed in
mm/day and is obtained from local evaporation pan observations. Trade-offs between
reservoir’s water elevation and reservoir’s storage is represented by a quadratic form
(Eq. 1) with a coefficient of regression (R2 ) of 0.9957, whereas that of reservoir’s
storage and reservoir’s free water surface area is interpolated based on Table 2.
Vt = 10−12 ∗ A2t + 0.0004 ∗ At + 616.02 (1)

Table 2 Study reservoir’s

Storage Surface area Storage Surface area
storage and corresponding
(MCM) (ha) (MCM) (ha)
surface area
23.000 8900 12.575 5112
21.517 7869 9.567 4276
18.434 6908 4.650 2222
14.941 5989 1.400 1042
2.2 Optimization Model
The objective function is to minimize total spills from the study reservoir (Eq. 2),
subject to constraints expressed in water balance (Eq. 3), storage bounds (Eq. 4) and
steady state condition (Eq. 5)
12
Minimize Z = St , t = 1, 2, . . . , 12 (2)
t=1
Subject to:
Vt+1 = Vt + It − IRt − RWt − Et − St (3)
Vmin ≤ Vt ≤ Vmax (4)
V1 = V24 (5)
where:
Vt : volume at time t (MCM),
It : inflow at time t (MCM),
IRt : irrigation releases at time t (MCM),
RWt : raw water supply at time t (MCM),
Et : evaporation at time t (MCM),
St : spill at time t (MCM),
Vmin : minimum volume (1.400 MCM),
Vmax : maximum volume (23.00 MCM),
V1 : volume at the beginning of Nov-1 (MCM),
V24 : volume at the end of Oct-2 (MCM).
2.3 Sustainability Analysis
The terms reliability, resiliency and vulnerability were first introduced by [15] and
were later used in evaluating water management sustainability by, among others
[16–20]. The definition of each term and the formulas here are taken from [17] as
follows and are modified accordingly to spill minimization instead of hydropower
maximization. Reliability (α) is defined as the probability of the system being in
satisfactory state, which in the present work is represented by the number of occasions
whenever no spill occurs at any time period.
24
St = 0
α= t=1
(6)
T
The transition of a system from being unsatisfactory to satisfactory is denoted by
Wt , whose value is either 1 if the transition exists or 0 elsewhere [15]. The speed of
recovery from an unsatisfactory condition is defined as resiliency (β); where in the
present work is represented by a time period when spill occurs followed by another
period without spill and can be expressed by Eq. 7. Both α and β values are ranging
from 0 to 1.
24
t=1 Wt
β= 24 (7)
T − t=1 St = 0
The extent or duration of a failure can be measured by expected extent-

vulnerability expressed by Eq. 8. In the present work, deficit at period t (Dt ) is
defined as an event when the reservoir volume is below the dead storage. Further, a
dimensionless vulnerability measure is proposed by [21] as expressed in Eq. 9.
24
Dt
V ulavg = 24
t=1
(8)
T− t=1 (Vmin ≤ Vt ≤ Vmax)
V ulavg
ϕ= (9)
Max Dt
Finally, Sustainability Index (SI) is a geometric mean of reliability, resiliency and

vulnerability and can be computed as follows:

SI = 3
αβ(1 − ϕ) (10)
The optimization model was run three times for one agricultural year each run
commencing on November at any year and concluding on October the following
year. Irrigable area, reliability, resiliency, vulnerability and SI for each period are
summarized in Table 3 and depicted in Figs. 2 and 3.
Figure 2 shows that there is no clear trend between SI and irrigable area. While the
irrigable area increases from 1551 hectares at 2014–2015 to 2290 hectares at 2015–
2016, the values of SI decrease from 0.368 to 0.329. Meanwhile, at 2015–2016 period
both irrigable area and SI increase.
In contrast to trade-offs between SI and irrigable area, Fig. 3 shows that there
exists clear trend between reliability, resilience and SI. The more reliable the study
reservoir, the more resilience and the more sustainable it is and so forth. Conversely,
the more reliable the study reservoir, the less vulnerable it is and so forth.
Table 3 Results of sustainability analysis at each period

Period Irrigable area (ha) A B SI
Nov 2014–Oct 2015 1551.10 0.708 0.143 0.490 0.368
Nov 2015–Oct 2016 2289.82 0.625 0.111 0.510 0.329
Nov 2016–Oct 2017 2388.05 0.792 0.400 0.153 0.535
3,000 2,388.00 0.600

2,290.00 0.535 Area
2,500 0.329 0.500 SI

Sustainability Index (SI)
1,551.00
Irrigable area (ha)
2,000 0.368 0.400
1,500 0.300
1,000 0.200
500 0.100
0 0.000
2014-2015 2015-2016 2016-2017
Year of opƟmizaƟon
Fig. 2 Irrigable area (ha) and Sustainability index (SI)

0.900
0.792
0.800
0.708
0.700
0.625
0.600 0.535
0.490 α
0.500 0.510
β
0.400 0.400 ϕ
0.300 0.368 SI
0.329
0.200 0.143
0.111 0.153
0.100
0.000
2014-2015 2015-2016 2016-2017
Fig. 3 Reliability (α), resilience (β), vulnerability (ϕ) and Sustainability index (SI) of the study
reservoir during 3-year optimization horizon
4 Suggestions and Recommendation
By comparing the irrigable area to the designated irrigation area of 3350 ha (see Table
1), it can be seen that the former is much lower to the latter; this requires further
research to investigate the possible causes which may include sediment deposition
in the reservoir, or significant decreasing of inflows. Another result obtained in the
present work is that there exist similar trends between reliability, resilience and SI
whereas these measures’ trends are inversely compared to that of vulnerability.
The formulas of reliability, resilience, vulnerability and Sustainability Index
(SI) in the present work were adopted from a study whose research objective is
hydropower maximization. While the mathematical formulas remain exactly the
same, their physical meaning are modified accordingly to the objective of the present
work i.e. spill minimization. This is believed to be true for future studies with
different research objectives such as, for example, mitigation of droughts and reser-
voir hedging. The measures of sustainability mentioned above may need to be defined
appropriately in relation to the future studies’ objective functions.
Acknowledgements The authors would like to express their thankfulness to the Rector of Univer-
sity of Mataram as well as the Dean of the Faculty of Engineering for providing international
publication research grant. The authors also wish to extend their gratitude to the Head of Civil
Engineering Department for his encouragement to the authors to conduct and to publish research
in international conferences and journals.
References
1. Wurbs RA (1993) Reservoir-system simulation and optimization models. J Water Resour Plan
Manag 119(4):455–472
2. Simonovic SP (1992) Reservoir systems analysis: closing gap between theory and practice. J
Water Resour Plan Manag 118(3):262–280
3. Oliveira R, Loucks DP (1997) Operating rules for multireservoir systems. Water Resour Res
33(4):839–852
4. Smith J (1993) Conditional-yield operating rules for low-flow management. J Water Resour
Plan Manag 119(3):324–338
5. Fayaed SS, El-Shafie A, Jaafar O (2013) Reservoir-system simulation and optimization
techniques. Stoch Env Res Risk Assess 27(7):1751–1772
6. Rani D, Moreira MM (2010) Simulation-optimization modeling: a survey and potential
application in reservoir systems operation. Water Resour Manage 24(6):1107–1138
7. Ahmad A, El-Shafie A, Razali SM, ZS Mohamad (2014) Reservoir optimization in water
resources: a review. Water Resour Manag 28:3391–3405
8. Yoo J-H (2009) Maximization of hydropower generation through the application of a linear
programming model. J Hydrol 376(1–2):182–187
9. Abdulrahman MKA, Ismail AAMd, Kek SL (2013) Determination of normal operating heights
of reservoirs in a network using non-linear reservoir method. Br J Appl Sci Technol 3(1):34
10. Zhang Z et al (2013) Use of parallel deterministic dynamic programming and hierarchical
adaptive genetic algorithm for reservoir operation optimization. Comput Ind Eng 65(2):310–
321
11. Lerma N et al (2013) Development of operating rules for a complex multi-reservoir system by
coupling genetic algorithms and network optimization. Hydrol Sci J 58(4):797–812
12. Cancelliere A et al (2002) A neural networks approach for deriving irrigation reservoir operating
rules. Water Resour Manage 16(1):71–88
13. Kumar ARS et al (2013) Application of ANN, fuzzy logic and decision tree algorithms for the
development of reservoir operating rules. Water Resour Manag 27(3):911–925
14. Perdana C (2015) Map of catchments within Lombok Greater Basin (in Indonesia language)
15. Hashimoto T, Stedinger JR, Loucks DP (1982) Reliability, resiliency and vulnerability criteria
for water resources system performance evaluation. Water Resour Res 18(1):14–20
16. Moy WS, Cohon JL, ReVelle CS (1986) A programming model for analysis of the reliability,
resilience, and vulnerability of a water supply reservoir. Water Resour Res 22(4):489–498
17. Ghimire BNS, Reddy MJ (2013) Optimal reservoir operation for hydropower production using
particle swarm optimization and sustainability analysis of hydropower. ISH J Hydraul Eng
19(3):196–210
18. Shin S et al (2018) A systematic review of quantitative resilience measures for water
infrastructure systems. Water 10:1–25
19. Sule BF, Surajudeen M (2017) Evaluation of the reliability, resilience and vulnerability of
Jebba hydropower reservoir operation, Nigeria. Saudi J Eng Technol 2(9):315–323
20. Jain SK, Bhunya PK (2008) Reliability, resilience and vulnerability of a multipurpose storage
reservoir. Hydrol Sci J 53(2):434–447
21. Kjeldsen TR, Rosbjerg D (2001) A frame work for assessing the sustainability of a water
resources system. In: Regional Management of Water Resources, 2001. Maastricht, the
Netherlands
Infiltration Wells as an Alternative Eco
Drainage System a Case Study
in Mangkubumen Surakarta
Retnayu Molya, R. R. Rintis Hadiani, and Adi Yusuf Muttaqien
Abstract The city of Surakarta is a lowland area located in the slopes of Mount
Lawu and Mount Merapi. Therefore, the city of Surakarta is prone to inundation and
flooding. Mangkubumen Village is one of the urban villages that is quite dense and
located in the middle of the city, often inundation and flooding occurs during the
rainy season. This is due to the possibility of runoff flooding from the Pepe River
that crosses the edge of the Mangkubumen Village. Besides that, this problem occurs
because the conventional drainage system cannot accommodate runoff water. Urban
drainage planning must take into account the environmental drainage function (Eco
drainage). One of the eco drainage methods is to use infiltration wells. This research
is descriptive quantitative. The main objective of this research is to determine the
reduction of runoff by implementing infiltration wells in Mangkubumen Village so
that the channel can accommodate water according to its normal capacity. Compara-
tive analysis shows the return peak discharge of 2 and 5 years of Kali Pepe is 15.4848
and 19.5255 m3 /s. The peak discharge of 5 years resulted in inundation of 224,322.08
m3 while the drainage channel capacity of 223,415.7 m3 was not sufficient to accom-
modate the discharge. There are two alternatives in designing infiltration wells. The
first alternative can re-duce the burden of drainage channels existing by 5.4% and
the second alternative can reduce the burden of drainage channels existing by 2.8%
so that the channel can accommodate runoff discharge by its capacity.
Keywords Eco drainage · Infiltration well · Urban drainage · Conventional

drainage
R. Molya (B) · R. R. Rintis Hadiani · A. Y. Muttaqien

Sebelas Maret University, Ir. Sutami 36A Street, Surakarta 57126, Indonesia
A. Y. Muttaqien
954 R. Molya et al.
1 Introduction
Population growth increases every year and has an impact on increasing the level of
population density in an area. This has resulted in a change in land use from open
land to developed land [1]. The development of land is being developed rapidly and
uncontrollably resulting in the built-up land becoming a dense residential area. This
results in an increase in runoff flow and a decrease in the quantity of water that
is absorbed into the soil out of balance with the ability of the drainage channels,
and this is what causes inundation and even flooding during the rainy season [2].
Mangkubumen Village which is a village in the middle of the city, also experiencing
the same symptoms as indicated by population growth and urban development. In
the end, there was very little open land in this village. Every rainy season there are
always inundation that interfere with the activities of local residents.
Drainage is the most feasible and economical solution for managing surface runoff
by draining it as quickly as possible [3]. However, the conventional drainage concept
is considered ineffective for long-term use because it has to be expanded gradually
over time and requires a large amount of money and a design that does not pay atten-
tion to water quality [4]. Eco drainage is an option to replace conventional drainage
that has been used in urban areas. Environmentally friendly drainage has the principle
of controlling excess water by absorbing it into the ground rather than discharging
it directly into drains. One of the eco drainage methods is to use infiltration wells to
support drainage channels [5]. Starting from this background the authors conducted
a study “Infiltration Wells as an Alternative to Eco Drainage Systems, a Case Study
in Mangkubumen Surakarta”. The main objective of this research is to determine the
reduction of runoff by implementing infiltration wells in Mangkubumen Village.
2 Research Method
The location of this research is located in Mangkubumen Village, Banjarsari District,

Surakarta City which is geographically located at 7° 33 34,83 South Latitude
and 110° 48 43,48 East Longitude and is administratively located surrounded by
Manahan Village, Kestalan Village, Gilingan Village and Punggawan Village. The
border between Mangkubumen Village and Kestalan Village is bordered by Kali
Pepe.
The research stages are first, primary data collection, namely measuring the dimen-
sions of the existing drainage channel. Then, rainfall data processing at Pabelan
station to determine the rainfall intensity. After that, calculate the runoff discharge.
The next is calculating the need for infiltration wells. The last, is calculate runoff
reduction and compare with existing drainage capacity.
Infiltration Wells as an Alternative Eco Drainage System a Case … 955
2.1 Data
The data is divided into two, namely primary data and secondary data. Primary data
is data taken directly in the field by researchers. In this study, the primary data is
the dimension of existing drainage channels. Secondary data is data obtained from
various existing sources. This study uses secondary data from related agencies. The
data are daily rainfall data, climatological data, topographic maps, dan DEM map.
2.2 Data Durability Test
The data durability test was conducted to determine whether the data used was
consistent or not. If after being tested the results are inconsistent, it is necessary to
replace or fill in the missing data. The method in testing the robustness of the data
in this study is the RAPS (Rescaled Adjusted Partial Sums) method.
2.3 Eco Drainage
A drainage is a form of effort to drain or divert water from rainwater, seepage, or

excess irrigation water from an area that does not interfere with the function of
the area [6]. Conventional drainage is a drainage system that has the principle of
draining excess water into drains as quickly as possible so as to allow the channel
to be overloaded. Meanwhile, the environmentally drainage or ecodrainage has the
principle of controlling excess water by absorbing it into the soil so that it has a role
in contributing to groundwater reserves during the dry season [7].
2.4 Mononobe Method
The analysis of rainfall intensity using the Mononobe method is carried out to esti-
mate the peak discharge in small catchments such as urban drainage system planning,
culverts, and bridges [8]. The rainfall intensity is determined as
2
R24 24 3
I = (1)
24 tc
where I = the rainfall intensity in mm per hour; R24 = the maximum annual daily
rainfall for Return Period in mm; and t = concentration time in hours.
956 R. Molya et al.
2.5 Rational Method
The Rational Method is the development of a rational method where the concentration
of rainfall occurs over time. The rational method is intended to consider the effect
of discharge capacity in estimating peak runoff [9]. The modified formula for the
rational method in determining the peak discharge is as follows
Q = 0., 278 x C x I x A (2)
where Q = the peak discharge with a certain period of time in m3 per second;
0.278 = the conversion factor; C = runoff coefficient; I = the rainfall intensity in
mm per hours; and A = catchment area in km2 .
2.6 Soil Permeability
According to the SNI 8256:2017 about ditch and infiltration wells, soil structures
that can be used for infiltration wells must have a permeability coefficient of more
than 2 cm per hour [10].
2.7 Number of Infiltration Wells
The number of infiltration wells that need to be calculated using the following
formula:
L +s
n= (3)
D+s
where L = the length of drainage channel; s = the space between wells and D =
diameter of wells.
2.8 Dimension of Infiltration Well
The depth of the infiltration wells is generally designed not exceeding the ground-
water depth in the area. The formula that can be used to calculate the depth of
infiltration wells is Sunjoto formula [11]. Calculations using the Sunjoto formula get
a shallower well depth (H) than other formulas [12].
Q − F×K ×T

H= 1 − e π×R2 (4)
ωπr K
where H = infiltration well depth in m; Q = runoff discharge m3 per hours; r =

the width of the infiltration well in m; K = soil permeability coefficient in m per
hours; ω = 2, for empty infiltration wells with impermeable walls or infiltration wells
without walls with fill stones; and ω = 5, for empty infiltration wells with porous
walls.
3 Analysis and Discussion
3.1 Type of Soil
Most of Surakarta has sandy clay types, including alluvial and gray regosol soil types
[13]. In the western region leading to the south of the Surakarta area, including in
Mangkubumen Village, is a gray regosol soil type. This gray regosol soil is a type
of soil with quick rather-fast infiltration rate. This permeability coefficient is about
at 36 cm per hours.
3.2 Soil Permeability
The permeability value that I use comes from previous research in the area around
the study with the same soil type, gray regosol soil type, namely “Analisis Hidrologi
dan Kapasitas Sistem Drainase Kota Surakarta”. The results of the laboratory test in
this study showed that the permeability coefficient was k = 1.1 × 10–4 m per second
[14], so according to the permeability coefficient value, it could be categorized as soil
with fast absorption. This fast absorption soil type is qualified to make infiltration
wells.
3.3 Ground Water Level
The existence of the Surakarta free aquifer is at a depth of 4–10 m [15] which means
that it meets the requirements for making infiltration wells according to SNI 8456:
2017 the depth of the groundwater level is more than 2 m. According to SNI 8456:
2017, the depth of infiltration wells should not exceed the depth of free aquifer
(groundwater) so the dimension depth of infiltration wells is about less than 4 m. In
this research, infiltration wells are designed with a depth of 1 m.
958 R. Molya et al.
Table 1 Rainfall intensity

Repeat period Kt Concentration time Maximum rainfall Rainfall intensity
(year) (h) (R24) (mm) (mm/h)
2 −0.1484 3.46 102.08 15.4848
5 0.9113 3.46 128.718 19.5255
Source Calculation
3.4 Catchment Area
The area of the rainwater catchment is the area of Kali Pepe watershed. Based on
the Surakarta City Profile book, the area of watershed is 5.65 km2 . According to the
area, the peak discharge of the Pepe River is calculated using the Rational Method.
Before that, we should calculate the rainfall intensity.
3.5 Rain Data Processing
The rainfall intensity is determined by rainfall data at Pabelan Station in 21 years

(2000–2020). The rainfall data obtained need to be calculated for the probability
of recurrence in the future, namely at a return period. This is necessary to plan the
peak discharge at a certain return period. The type of distribution used in this study
is the Gumbel distribution. The distribution method needs to be tested to determine
whether the distribution is carried out to determine whether the distribution method
for determining design rain can represent the statistical distribution of the analyzed
data. Distribution checking using the Chi Square method. Then, the rainfall intensity
is calculated using the Mononobe method as shown in Table 1.
Based on Fig. 1. IDF curves above, it can be seen by the time the rainfall intensity
equation. The equation for various 2-year and 5-year is I2 return = 542.38x−0.667 and
I5 = 683.92x−0.667 .
3.6 Peak Discharge
Furthermore, the return period of peak discharge in this study was calculated using
the rational method. In this study, the return period used is a return period of 2 years
and 5 years because the research area is an urban residential area with the most
channels being tertiary channels. Here is Table 2 the result of peak discharge of Kali
Pepe and Fig. 2 is the hydrograph of peak discharge of Kali Pepe.
IDF Curve Mononobe Method

250
Rainfall Intensity (mm/hour)
200
150
2 yrs
100
5 yrs
y = 683,92x-0,667
50 y = 542,38x-0,667
0
0 50 100 150 200 250
Duration (minute)
Fig. 1 IDF curve of rainfall intensity using Mononobe method. Source Calculation
Table 2 Peak discharge of

Repeat period (year) Rain intensity (mm/h) Q (m3 /s)
Kali Pepe
2 15.4848 18.2089
3 19.5255 22.9937
Source Calculation
Hydrograph of Peak Discharge Rational

Method
25
Peak Discharge (m3/s)
20
15
10 2 yrs
5 yrs
5
0
0 2 4 6 8
Time Concentration (hour)
Fig. 2 Hydrograph of peak discharge of Kali Pepe using rational method Source Calculation
960 R. Molya et al.
3.7 Inundation Resulting to Peak Discharge
After obtaining the peak discharge of Kali Pepe with a return period of 2 and 5 years,
it was then modeled using Hecras to determine the inundation that occurred [16]. The
results of the analysis using Hecras produce an inundation area of 0.25 km2 when
the peak discharge is applied for a 2-year return period and 0.28 km2 for a 5-year
return period as shown in Fig. 3. Existing channel’s capacity is able to accommodate
runoff due to return period discharge Q2 years but it can’t accommodate the runoff
due to period discharge Q5 years at 906.42 m3 .
Then to calculate the volume of inundation that occurs, the assumption is that the
inundation is in the shape of a rectangular pyramid so that the volume of inundation
due to Q5 can be calculated by the following formula:
1
V = × area × elevation (5)
3
1
V = × 280402.6 × (2.7 − 0.3)
3
V = 224,322,08 m3 and so Q = 224,322,08 m3 /3600 = 62.3889 m3 /s.
Fig. 3 Inundation area of Q5 years from running Hecras. Source Author document
3.8 Existing Drainage Capacity
Based on the drainage channel tracing survey, the length of drainage channel in
Mangkubumen is 10.127 km and the capacity of the channel can accommodate the
discharge of 62.0599 m3 /s.
3.9 Number and Dimension of Infiltration Wells
The steps in designing an infiltration well include determining the geometric factor,
the depth of the infiltration well, and the number and distance of the infiltration well.
The infiltration well is planned to be located at the bottom of the channel existing
with the same diameter as the smallest tertiary channel width, which is 0.4 m. In
this research, there are two alternative of infiltration wells as shown in Fig. 4. The
different between those alternatives is about the construction and cost.
The designed infiltration wells are cylindrical with a flat bottom. Here is the
example of calculating infiltration wells alternative 1. The first step is to determine
the geometric factor of the infiltration wells [17].
F = 5 × 3.14 × 0.2
F = 3.14 m
where ω = 2, for empty holes with waterproof walls or holes without walls with
fill stone and ω = 5, for an empty hole with a porous wall. Then, look for the required
total depth of infiltration hole with the following Sunjoto method:
Fig. 4 Alternatives of infiltration wells design. Source Author document

962 R. Molya et al.
Q − F×K ×T

H= 1 − e π×R2
ωπr K
−4 ×207.3
62.3889 − 3.14×1.1×10
H= 1−e π×0.2 2
3.14 × 1.1 × 10−4

H = 77956 m
where H = total depth of infiltration well (m), Q = runoff discharge (m3 /s), F =
geometric factor (m), K = coefficient of permeability (m/s), T = flow time (min),
and R = radius of infiltration well (m).
After obtaining the required total depth, then look for the number of infiltration
wells with a designed depth of 1 m for each hole.
H
n=
1m
n = 77, 956 wells.
From the calculation, the number of wells is 77956 while the length of the existing
channel is 10.127 km or 10,127 m. Of course, this cannot be done because there is
no space for the infiltration wells. Therefore, it is necessary to recalculate so that the
space in the channel can accommodate the infiltration hole. If the depth of each hole
is designed at 1 m and the distance between holes is planned 2 m, then the number
of holes that can be made are as follows:
L +s
n=
D+s
10127 + 2
n=
0.4 + 2
n = 4221 wells.
The same calculation above to calculate the second alternative. Table 3 is a

recapitulation of the dimensions of the two alternatives.
Table 3 The dimension of infiltration wells Alternative 1 and 2

Alternative D (m) H (m) n S (m) Q (m3 /s)
Alternative 1 0.4 1 4421 2 3.3739
Alternative 2 0.3 1 4404 2 1.0560
Source Calculation
Table 4 Construction and cost of Alternative 1 and 2

Alternative D (m) Q (m3 /s) Reduction (%) Construction Cost (IDR)
Alternative 1 0.4 3.3739 5.4 PVC pipe 800.000
Alternative 2 0.3 1.0560 2.8 Concrete 89.000
Source Calculation
3.10 Runoff Reduction
After obtaining the dimensions of the infiltration wells, the following is the discharge
that is absorbed into the well alternative 1.
F × H×
Q= ×T

− F×K
1−e π×R 2
Q = 3.3739 m3 /s.
While the discharge capacity of channels is 62.0599 m3 /s, the drainage channel
load reductions are:
Qcapacity = Qexisting − Qabsorp

Qcapacity = 62.0599 − 3.3739
= 58.6860 m3 /s
% = 100 − (58.6860/62.0599)
= 5.4%
From the calculations above, a total of 4221 pieces infiltration wells placed in the
Village Mangkubumen able to reduce the burden of existing drainage channels by
5.4%. The same calculations above to calculate the second alternative. Table 4 is a
recapitulation of the construction and cost of the two alternatives.
4 Conclusion
The conclusions of this study are as follows:

There are two alternatives on infiltration wells. The first one is by applying 4221
infiltration wells that was designed with a radius of 0.4 m, a depth of 1 m and the
space between each hole is 2 m can reduce the burden of drainage channels existing
by 5.4%. And the second is by applying 4401 infiltration wells that was designed
with a radius of 0.3 m, a depth of 1 m and the space between each hole is 2 m
can reduce the burden of drainage channels existing by 2.8%. The different between
964 R. Molya et al.
those alternatives is the construction. The first one using PVC pipe construction and
the second using concrete cylinder construction.
References
1. Karya J, Sipil T (2014) Perencanaan Eko-Drainase Kawasan Perumahan Tembalang Pesona

Asri-Semarang. In: ejournal3.undip.ac.id 3(2). Halaman. http://ejournal-s1.undip.ac.id/index.
php/jkts
2. Kusumastuti C, Chandra H, Wibisono K, Hartono A (2019) Eco drainage system for surface
runoff reduction in Indonesia. http://repository.petra.ac.id/18281/
3. Idfi G, Yulistyorini A, T A-M W, (2018), undefined. (n.d.). Injection well as an eco-drainage
solution to reduce surface run-off at the State University of Malang. Matec-Conferences.Org.
Retrieved March 17, 2020, from https://www.matec-conferences.org/articles/matecconf/abs/
2018/63/matecconf_imiec2018_03017/matecconf_imiec2018_03017.html
4. Mguni P, Herslund L, Marina Jensen B (2016) Sustainable urban drainage systems: examining
the potential for green infrastructure-based stormwater management for Sub-Saharan cities.
Springer 82:241–257
5. Wahyuningtyas A, Hariyani S, Sutikno FR (2011) Strategi Penerapan Sumur Resapan sebagai
Teknologi Ekkodrainase di Kota Malang (Studi Kasus : Sub DAS Metro). Jurnal Tata Kota Dan
Daerah 3(1):25–31
6. Suripin (2004) Sistem Drainase Perkotaan yang Berkelanjutan. ANDI, Yogyakarta
7. Hamzah Syahruddin M, Halide H (2019) To cite this article: Muhammad Hamzah Syahruddin
et al. IOP Conf Ser: Earth Environ Sci 279:12021. https://doi.org/10.1088/1755-1315/279/1/
012021
8. Suleman AR, Yusuf H, Nabi A, Erdiansa A, Aulia N (2020) Determining biopore infiltration
hole as catchment flood in inundation area Eastern of Makassar City Based on Geographic
Information System (GIS). J Eng Appl Sci 15(2)
9. Prayitno G, Hariyani S (2013) The effectiveness of bio-pore as an alternative eco drainage
technology to control flooding in Malang City (case study: metro subwatershed). In
researchgate.net
10. Indonesia BSN (2017) Sumur dan Parit Resapan Air Hujan. Sni 8256(2017):12–13
11. Sunjoto (2011) Teknik Drainase Pro-Air. Universitas Gadjah Mada, Yogyakarta
12. Fachrurazie C, Firman Arifin Y, Sri Susanti D (2002) Analisa Drainase Sumur Resapan Pada
Kampus UNLAM Banjarbaru. Info-Teknik 3(1):24–34
13. Pemerintah Kota Surakarta (2016). Badan Perencanaan, Penelitian, dan Pengembangan Daerah.
Rencana Program Investasi Jangka Menengah (RPIJM) Bidang Cipta Karya Tahun. Bab ii profil
wilayah kota surakarta 2.1:1–28
14. Sataloff RT, Johns MM, Kost KM (n.d.) Buku II Informasi Kinerja Pengelolaan Lingkungan
Hidup Daerah Kota Surakarta. Pemerintah Kota Surakarta Dinas Lingkungan Hidup, Surakarta
(2017)
15. Sipil JT, Teknik F (2016) Analisis hidrologi dan kapasitas sistem drainase kota surakarta
16. Amdan Y, Studi P, Sipil T, Teknik F, Maret US (2019) Drainage channel capacity analysis
for flood controlling solution in Sewu Surakarta drainage channel capacity analysis for flood
Yuliarini Amdan
17. Kamila N, Wardhana IW, Sutrisno E (2016) Perencanaan Sistem Drainase Berwawasan
Lingkungan (Ecodrainage) Di Kelurahan Jatisari, Kecamatan Mijen, Kota Semarang. Jurnal
Tehnik Lingkungan 22(2):63–72. https://doi.org/10.5614/j.tl.2016.22.2.7
Analysis of Leb Irrigation Patterns
of Pipe System in Sorgum Plants in Sand
Dry Lands Akar Akar Village
I Dewa Gede Jaya Negara, Sasmito Soekarno, Suwardji, Humairo Saidah,

and Atha Adi
Abstract To take advantage of the dry land potential in North Lombok district which
reaches 38,000 ha, the government has facilitated groundwater irrigation networks
(JIAT). The most visible use of groundwater irrigation networks so far has been in
the community for maize farming with wasteful water use, and it is necessary to
develop crops such as sorghum in order to obtain more feed availability that supports
the provision of animal feed. An adequate irrigation system is necessary so that it can
provide economic improvement or reduce costs used. This study aims to examine
the pattern of leb pipe irrigation water use in sorghum to until harvest on a land area
500 m2 . Leb irrigation network using 2 inches pvc pipe on 6 blocks of sorghum crop
land, with irrigation duration of 20 and 30 min. The irrigation test is carried out on
the uniformity, depth irrigation and soil water content. The test results show that
the irrigation uniformity produced is 94% very good, reaching a soil depth of about
20–30 cm with a water content of about 35–40% so that this irrigation system has
the potential to support irrigation of sorghum plants.
Keywords Uniformity · Pipe width · Irrigation pattern
I. D. G. J. Negara (B) · S. Soekarno · H. Saidah · A. Adi

Department of Civil Engineering, Mataram University, Mataram, Indonesia
S. Soekarno
H. Saidah
Suwardji
Department of Soil Science, Faculty of Agriculture, University of Mataram, Mataram, Indonesia
966 I. D. G. J. Negara et al.
1 Introduction
The Akar-Akar area of Bayan District, North Lombok Regency is one of the dry land
potentials in the province of West Nusa Tenggara with an area of about 38,000 ha.
The obstacle to providing plant irrigation water that is often faced in the field is the
provision of irrigation water on land using open boxes, which causes a lot of water
loss due to evaporation of the land surface during the trip and uneven infiltration into
the soil resulting in uneven plant growth and crop yields. not good. The results of the
optimization survey on the use of deep groundwater wells show that only 10–15%
of the total deep groundwater wells built by the government are utilized by farmers
while the rest is optimal due to high operational costs [1]. The practice of inundation
irrigation systems has been carried out so far. 1 ha of land requires irrigation for
about 14 h, the results of the study [2] show that the inundation irrigation system that
has been carried out is a system that is very wasteful of water and is not profitable
for farming. irrigation on dry land so it is necessary to develop an irrigation system
to support agricultural development on dry land which is more water efficient.
To increase the productivity of agricultural land in this area, it has been tried to
apply water-efficient irrigation techniques such as JIAT, sprinkler and drip systems
to increase agricultural yields on dry land so that it can work well, especially in maize
and horticultural crops. Based on field experience, it is not enough to improve the
welfare of people in dry land, but other innovations are still needed, especially in the
selection of plant types so that the community can benefit from multiple economic
effects. Besides, the harvest can produce fruit multiple times, but it is also able to
provide a continuous reserve of animal feed, so that livestock production and sales
can contribute together in helping the economy of dry land communities.
To increase the productivity of agricultural land in this area, water-efficient irri-
gation techniques such as the JIAT system, big gun and drip sprinklers have been
tried to increase agricultural yields on dry land. Based on field experience, in order
to improve the welfare of the community on dry land, the irrigation system requires
new investment which is expensive, so that it is not necessarily possible for the
general public in JIAT to implement it. For this reason, it is necessary to develop an
existing irrigation system so that the irrigation method becomes more efficient and
agricultural crops can provide a higher economic effect. Sorghum plants are consid-
ered suitable, so it needs to be tested on land with over-pipe irrigation so that water
use is known according to its growth phase. With this effort, it is hoped that later
sorghum can be cultivated so that it can help livestock businesses towards improving
the economy of the community on dry land. According [3], an irrigation system over
a pipe with a duration of 10 min in dry land with finely graded soil can provide
irrigation up to 30 cm with an average humidity of 25% and at 30 min it can provide
about 32% moisture. After four days of testing, the soil moisture is obtained, the
remaining moisture is above 20%. So the irrigation system may be more suitable to
be applied to sandy soil such as in this study location, to support future development
of sorghum crops.
Analysis of Leb Irrigation Patterns of Pipe System … 967
Test of over-irrigation pipe system at the Salut location, North Lombok, shows
that the over-irrigation test system on an area of about 1 acre requires a shorter
irrigation time than the inundation system, with a uniformity of more than 91%. At
irrigation duration of 10 min, 20 min and 30 min, the irrigation duration efficiency
was 0.42 h, 0.83 h and 0.63 h, respectively, with cost efficiency of 25%, 33% and
25%, respectively [4]. Based on this test, irrigation provides water more efficiently
than the inundation system, so that the puddle irrigation method which is also applied
to JIAT needs to be developed with a pipeline system.
Soil moisture is water in the soil in the root area/root zone that fills part of the soil
pore space expressed in units of weight percent or volume percent. According to [5],
the most accurate measure of water content is the gravimetric method, namely by
weighing the soil sample, drying it in an oven at 100–110 °C for 24 h and weighing
it again.
High output uniformity is very important in the development of irrigation systems.
The goal is to achieve a uniform irrigation water output level at each pipe outlet that
can meet the needs of plants, especially in the root zone. Irrigation uniformity can
be calculated by the formulation of Christianes [6].
To overcome this, it is necessary to innovate dry land plants in order to obtain a
double economic effect in future farming. Tests need to be carried out on types of
sorghum plants that have the ability to grow long with a long harvest period. The
stover from the sorghum harvest can be used as animal feed, and can grow more and
more so that animal feed reserves will be obtained due to repeated harvesting. Plant
testing needs to be applied with a drip irrigation system and a wide pipe so that the
distribution of water to plants can be evenly distributed to support optimal sorghum
planting results.
To overcome the above problems, it is necessary to innovate dry land plants in
order to obtain a double economic effect in future farming. Plant tests need to be
carried out on types of sorghum plants that have the ability to grow longer than corn
plants with a harvest period of more than once. The stover from the sorghum harvest
can be used as animal feed, and can grow more and more so that animal feed reserves
will be obtained due to repeated harvests. Plant testing needs to be applied with an
over-pipe irrigation system, so that the water distribution pattern to the land is known
and the discharge size and duration required for the depth of sorghum roots can be
known. With this effort, sorghum plants will later be developed in a certain area with
more water efficiency, and provide better economic benefits to farming communities.
This study aims to determine the amount of irrigation water demand for sorghum
farming tests, pipeline and drip irrigation systems including irrigation time and irri-
gation duration. It also includes how the water services are carried out by the two irri-
gation systems until the sorghum plants can be harvested and what are the advantages
and disadvantages of the irrigation system used during the growing season.
The equipment used during the study was a water meter, sample pipe, hoe, oven,
gravimetric, hammer and stop time. For the research material consists of 2 “pvc pipe,
2” pvc dop, L 2 inches pvc, T 2 inches pvc and pvc 2 inches timer. This research was
carried out by field test research where drip irrigation and over-irrigation systems
with sorghum plants were tested directly in the dry land of Akar Akar, North Lombok
district Fig. 1. The stages of conducting field research were carried out as follows.
The preparation stage consists of preparing references, tools and materials for
the location of sorghum plants including clearing 4 acres of land for two irrigation
systems, making 6 blocks of more irrigation test fields. In 6 irrigation blocks, block
1 consists of 9 outlet holes, land 2 has 6 outlet holes, test area 3 has 9 outlet holes,
test area 4 with 4 outlet holes, test area 5 has 6 outlet holes and test area 6 there are
also 6 irrigation outlet holes.
Testing the uniformity of irrigation over pipes, taking soil samples before and
after irrigation at 3 (three) different location points in one plot of land and testing
the initial soil moisture content. Provision of irrigation water for over-pipe irrigation
systems is carried out in 20 and 30 min. The recording of the discharge of water used
is carried out at each different irrigation time. Take a soil sample before and after
irrigation, and test soil moisture.
Data analysis and discussion. Data analysis of test results was carried out on used
discharge data, soil moisture data and irrigation uniformity. The need for irrigation
water at each stage of plant growth is calculated as the volume of used water at each
provision of irrigation water and the amount of water used until the initial harvest is
used as data on the use of early harvest irrigation water. The results of the analysis
were presented in tables and graphs and concluded descriptively.
Akar Akar
Fig. 1 Akar Akar, North Lombok district

3 Discussion
Based on the results of the analysis of the test data using the Chrietiansen’s formula,
the irrigation uniformity value is 94% so that it is considered good, so that this
irrigation system is thought to be applicable for farming such as sorghum. The results
of the analysis are shown in Table 1.
This indicates that over-piped irrigation will be able to provide irrigation services
as well as drip irrigation, but the application of over-irrigation on a wider plot of land
and on a large availability of discharge.
The amount of water flowing through the cross section for each unit of time is
called the flow rate and is given the notation Q. for the calculation of the discharge,
the following equation can be used [7] Table 2.
For irrigation, the discharge given is around 0.00017 m3 /s.
Soil moisture is water in the soil of the root area/root zone that fills part of the soil
pore space expressed in units of weight percent or volume percent. According [5],
the most accurate measure of water content is the gravimetric method, namely by
weighing the soil sample, drying it in an oven at 100–110 °C for 24 h and weighing
it again. The results of soil moisture analysis from wide irrigation can be seen in
Figs. 1 and 2.
The result of providing irrigation to land is in the form of soil moisture, which is
the amount of water stored in the soil layer which will later be used by plant roots for
plant growth. Based on the existing soil conditions, soil moisture that can be given
from wide irrigation is shown in Fig. 1. Based on this figure, the irrigation is given
with a duration of 20 min, at three sample points a depth of 10–15 cm where soil
Table 1 Uniformity of pipe

Outlet pipe CU (%)
Leb irrigation system
1 94.02
2 93.89
3 93.44
4 97.24
5 93.80
6 93.38
Table 2 Irrigation hole

Outlet pipe leb (m3 /dt)
discharge over
1 0,00,016
2 0,00,017
3 0,00,016
4 0,00,018
5 0,00,017
6 0,00,017
50.000
point 1
Water Conten (%)
40.000 point 2
30.000 point 3
20.000
10.000
0.000
0 24 48 72 96 120 144 168
Time ( hour)
Fig. 2 Soil moisture and daily changes at a depth of 10–15 cm
moisture after irrigation is obtained at 38.88%, and after 7 days of irrigation, the
remaining moisture is 21.77% approaching the initial soil moisture.
The result of providing irrigation to land is in the form of soil moisture which
will later be used by plant roots at a depth of 30 cm for plant growth shown in Fig. 2.
Based on this figure, irrigation is given with a duration of 30 min, at three sample
points a depth of 20–30 cm where Soil moisture after irrigation was obtained by
40.66% and after 7 days of irrigation, the available moisture in the soil depth was
25.15%. This moisture is still higher than the initial soil moisture.
Based on the moisture condition of the soil obtained from excessive irrigation, the
provision of irrigation in its application can still follow the JIAT irrigation pattern in
general, where irrigation is given in 7 turns so that it does not change the irrigation
schedule of the existing pumper. However, the provision of irrigation will be shorter
than the existing JIAT irrigation system. With more irrigation it is still possible to
adjust the uniformity of irrigation into the soil and this can shorten the irrigation time,
reduce irregularity in soil and plant growth critical down that will be detrimental to
plant growth and irrigation can be done sooner than 7 days.
The graph of the results of pipe irrigation for sorghum plants on sandy dry land
can be seen in Fig. 4. So the amount of irrigation water used is irregular and follows
the conditions of the plant so that it will follow the growth phase of the sorghum
plant.
Based on Fig. 3, it is known that the provision of irrigation water on average in one
irrigation in the initial growth phase is 1.242 m3 and at the planting age of 20 days
to 40 days is given irrigation water of 1.242 m3 , at the age of the plant 41 days to
67 days it is given irrigation water. an average of 1.446 m3 and after the plants are
70 days old until harvesting, the average amount of irrigation water given by sorghum
plants is 1.86 m3 . Increasing the provision of irrigation water from the initial phase
to harvest to 1.242–1.86 m3 when the sorghum plants are ready to be harvested at
88 days.
Water Content (%) 50.000

40.000
30.000
point 1
20.000
point 2
10.000
point 3
0.000
0 24 48 72 96 120 144 168
Time (hour)
Fig. 3 Soil moisture and daily changes at a depth of 20–30 cm
6
Volum, m 3
0
0 5 10 15
Week
Fig. 4 Provision of low irrigation water for each planting phase
The provision of irrigation water during the growing season for each growth phase
of sorghum is certainly not the same, because it depends on plant development. When
compared with the provision of more irrigation water, it can be seen in Fig. 5.
Leb irrigation water is given every 1 week with the amount of water used for
initial irrigation of 1.242 m3 , plant age up to 20 days, plant age 21 days to 43 days
given irrigation water of 1.242 m3 , at 44 days to 67 days of 1.46 m3 and at the age
of 70 days of sorghum until harvest 1 is given 1.86 m3 of water in one irrigation.
When compared to the irrigation water distribution pattern of the two irrigation
systems above, it is very different, in drip irrigation the variation of the discharge
from the irrigation system is still very high, whereas in irrigation the pattern is very
regular.
Based on Fig. 6, it is known that, for the over-pipe irrigation system in the 5-
acre sorghum area, it is irrigated once every 7 days with the amount of irrigation 13
2.5
Volume , m3 2
1.5
1
0.5
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Week
Fig. 5 Distribution of Leb irrigation water supply to harvest 1
25
20
Volum, m3
15
10
5
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Week
Fig. 6 Accumulation of irrigation water Leb
times, and the water used is 21.8 m3 . So the use of water by irrigation is lower than
by pipeline irrigation, by a difference of about 4.75 m3 .
In irrigation, plants are given water with irrigation duration, namely the duration
of irrigation is 20 and 30 min with the aim of achieving water at the roots with a
depth of 10–15 cm and 20–30 cm. Irrigation is given once every 6 days with 8 times
the duration of the irrigation is 20 min and 5 times the duration of the irrigation is
30 min. The total duration of irrigation over pipes needed to harvest crops is 26 h,
with a used water volume of 21.80 m3 .
4 Conclusions
The test results show that the irrigation uniformity produced is 94% very good,
reaching a soil depth of about 20–30 cm with a water content of about 35–40% so
that this irrigation system has the potential to support irrigation of sorghum plants.
References
1. Eterna and Suwardji (2003) Survey of deep groundwater pump utilization throughout Lombok
Islan, Collaborative research report on the groundwater development project section of the NTB
Provincial Public Works Office And The Unram Dryland Research and Development Center.
76h
2. Suwardji (2005) Assessment of irrigation water needs for inundation systems and sprinkle big
gun systems. Research Report of the Mataram University Research Institute
3. Zulvarisandy I (2017) Analysis of providing combined water for irrigation LEB pipes and mini
sprinklers against changes in daily soil moisture in Pringgabaya District, Mataram
4. Muta’al Atha Hadi (2018) Analysis of time and costs of irrigation in the Leb pipe irrigation
system in the dry land of Arungan Bali Hamlet, Akar-Akar Village, Bayan District, North
Lombok, FT Unram Thesis, Mataram
5. Soemarto CD (1986) Engineering hydrology. National Business, Surabaya
6. Triatmodjo B (2003) Hidraulika II, Beta Offset, Yogyakarta
7. Negara JDG, Supriyadi A (2016) Analysis of the design of integrated water-saving irrigation
systems based on groundwater irrigation network (JIAT) on dry land, smoothly degraded soils
in Pringgabaya, East Lombok Regency, Mataram
Hourly Rainfall Simulation Using Daily
Data
Suroso, Fatimatus Sholihah Marush, Purwanto Bekti Santoso, Irfan Sudono,

Edvin Aldrian, and Nelly Florida Riama
Abstract Rainfall is one of the main inputs in the analysis of the hydrological
system. However, rainfall data are often found to be inadequate in length and
completeness. The purpose of this study is to overcome the problem of the limited
availability of high resolution of rainfall data. Rainfall data on a daily scale generally
have a better quantity and quality of data available to be used as a basis or reference
for deriving hourly rainfall data. The research location is limited to Java Island, with
25 rain gauge stations used. Hourly rainfall has a diurnal cycle, where some rain-
fall is concentrated in the afternoon or evening. The result of the evaluation using
Root Mean Square Error, Mean Absolute Error, Pearson Correlation, and Spearman
Correlation shows that the hourly rain simulation model using daily rainfall data can
be used for design, especially in the field of water structures due to producing similar
statistical characteristics.
Suroso (B) · F. S. Marush · P. B. Santoso

Department of Civil Engineering, Faculty of Engineering, Jenderal Soedirman University,
Purbalingga 53371, Indonesia
P. B. Santoso
I. Sudono
Research Centre for Water Resources, Agency of Research and Development, Ministry of Public
Works and Housing, Jakarta, Indonesia
E. Aldrian
Agency for the Assessment and Application of Technology (BPPT), Jakarta, Indonesia
N. F. Riama
Meteorology, Climatology, and Geophysical Agency (BMKG), Jakarta, Indonesia
976 Suroso et al.
1 Introduction
A flood is a pool of water that causes a detrimental impact on human life. A flood is a
flow of water whose height exceeds the normal water level. Among all the disasters,
floods are the most destructive because they can cause serious damage to human life,
infrastructure, agriculture, and social-economic systems [1]. In Indonesia, almost all
areas have a high potential for flooding such as West Java [2], Central Java [3], and
East Java [4]. Flooding can take in many forms, including slow river flooding, rapid
flash floods, accumulation of rainwater in poorly drained environments, and coastal
flooding caused by tides and extreme waves.
Rainfall is one of the main inputs in the analysis of the hydrological system.
However, rainfall data are often inadequate in length and completeness due to the
small number of observations or instrument errors. Insufficient rainfall data series
can cause severe problems in hydrological analysis. Accessibility of high-resolution
rainfall is one of the main problems for hydrologists due to various reasons such
as cost and government policies [5]. Therefore, scientists have attempted to produce
accurate high-resolution rainfall data with many computer programs. The simulation
of hourly rainfall is an important area of hydrological research, for example for
designing hydraulic structures in urban areas or urban drainage systems, and for
evaluating the frequency of flooding in small natural catchments characterized by a
rapid hydrological response [6]. Hourly rainfall data or even smaller by temporal scale
are critical for applications in many sectors, such as urban hydrology, infrastructure
design, and risk assessment [7].
Daily rainfall is generally a difficult thing in rainfall modelling because daily
rainfall has unique characteristics and is difficult to make a model. Unlike the annual
and monthly rainfall, which can be modelled by the simple autoregressive moving
average (ARMA) process, the incidence and amount of daily rainfall are challenging
to model with the ARMA model [8]. There are many problems in hydrology, agricul-
tural science and engineering that require extensive rainfall records [9], especially
for flood risk assessment, water resources management or hydraulic structure design
[10]. However, hourly or higher resolution rainfall data tend to have lower temporal
and spatial coverage due to high instrument costs and technical limitations.
The purpose of this study is to overcome the problem of the limited availability
of high-resolution rain data. High-resolution rainfall data in the hourly or higher
(minute) is necessary for the hydrological aspect. Rainfall data on a daily scale
generally have a better quantity and quality of data availability. Hence, daily rainfall
can be used as a basis or reference for deriving hourly rainfall data.
Hourly Rainfall Simulation Using Daily Data 977
2.1 Data
The data used were obtained from the Meteorology, Climatology, and Geophysical
Agency (BMKG). The data required is hourly and daily rainfall data in Java. The
location of selected rainfall stations is presented in Fig. 1, and the name of that
can be seen in Table 1. The rainfall data used in this study are selected with having
relatively complete data from all stations in Java. By knowing the number of NA (Not
Fig. 1 Map of 25 rain station in Java Island
Table 1 The name of rain station in Java Island

Code Rain station ID rain station Code Rain station ID rain station
0 Tangerang 96,735 13 Semarang_Klimatologi 96,835
1 Serang 96,737 14 Semarang_Maritim 96,837
2 Bawean 96,925 15 Ahmad Yani_Semarang 96,839
3 Curug 96,739 16 Surabaya Tg. Perak 96,937
4 Mataram 96,742 17 Surabaya_Perak1 96,933
5 Jakarta BMKG 96,745 18 Juanda 96,935
6 Cengkareng 96,749 19 Karang Ploso 96,943
7 Citeko 96,751 20 Tretes 96,945
8 Dermaga 96,753 21 KarangKates 96,949
9 Bandung 96,783 22 Kalianget_Madura 96,973
10 Jatiwangi_Cirebon 96,791 23 Sawahan 96,975
11 Tegal 96,797 24 Banyuwangi 96,987
12 Cilacap 96,805
978 Suroso et al.
Available) or the countless data, there are three rain recording stations are obtained.
The relatively complete rainfall recording stations are Citeko (96,751) and Dermaga
(96,753).
2.2 Research Steps
Data Preparation
The following steps in preparing the data are the hourly rainfall recording stations that
have a relatively long measurement period, complete and continuous are identified
and selected for further analysis. Then at each hourly rainfall recording station that
has been selected, the cumulative rainfall calculation is carried out into a daily period.
Two data sets of rainfall are collected, so that hourly and daily rainfall data will be
obtained at the same time.
Model Development
There are two types of modelling approaches were taken, namely the approach that
the rainfall data series for both zero and non-zero rainfall was calculated and another
approach that only non-zero rainfall data series were carried out in the modelling
stage.
Model Evaluation
After obtaining the hourly rain simulation modelling using daily rainfall data, it is
necessary to evaluate it to obtain the performance of the model. There are several
measuring tools to find out how well the model developed consists of RMSE, MAE,
Pearson Correlation, Spearman Correlation.
RMSE (Root Mean Square Error)

RMSE is a way to evaluate linear regression models by measuring the accuracy of the
estimated results of a model. RMSE is calculated by squaring the error (prediction
- observation) divided by the amount of data, then rooted. The RMSE formula is
shown in Eq. (1). This prediction model is said to be the best if the value is 0 (zero).
1/2
yi − ŷi
RMSE = (1)
n
where y = the value of the observation result, ŷ = predicted value, i = sequence

of data in the database, N = amount of data.
MAE (Mean Absolute Error)

MAE is one of the methods used to measure the accuracy of the forecasting model.
The MAE value shows the absolute average error between the prediction results
with the real value. In general, the MAE formula is shown in Eq. (2). This prediction
model is said to be the best if the MAE value is 0 (zero).
Table 2 Class of correlation

Class Correlation coefficient value Information
coefficient value
1 0–0.25 Very weak
2 0.25–0.50 Weak
3 0.50–0.75 Strong
4 0.75–1 Very strong
1
n
MAE = yi − ŷi (2)
n i=1

where n = amount of data, yi − ŷi = absolute error, yi = predicted value, ŷi =
real value.
Pearson and Spearman Correlation

Pearson correlation is used to determine whether there is a relationship between two
variables on an interval or ratio scale. The Pearson correlation denoted by r is shown
in Eq. (3).
(

x)( y)
xy −
r =

n
(3)
(

x)2 (

y)2
x2 − n
y2 − n
where r = correlation value, x = variable x, y = variable y. The strength of the

correlation relationship classified into 4 classes that shown in Table 2.
The Spearman-rho rank correlation coefficient was used for the measurement of
correlation in nonparametric statistics.
3.1 Analysis of the Quality of Hourly and Daily Rainfall Data
Completeness of rain data is an important aspect in knowing the quality of hourly

and daily rainfall data. Rainfall data on the hourly and daily time scales used in this
study were taken from 1981 to 2019. However, the available rain data, especially on
the hourly time scale, is inadequate so that a lot of rain data is not recorded or is called
NA (Not Available). From the 25 stations, the percentage of missing values (NA)
that is less than 31% is the Dermaga station (96,753) and Citeko (96,751), that are
24.57 and 30.44%. Meanwhile, there are 4 stations that have rainfall data between
30–50%. The remaining 19 stations with more than 70% missing values. For the
daily rainfall scale, the percentage of missing values (NA) for the 25 stations is not
980 Suroso et al.
Fig. 2 Percentage of number of non-NA and NA at each station
much different from the hourly rain scale. The following graph shows the percentage
of NA numbers at each station as in Fig. 2.
In another study, it was stated that in Malaysia, for the daily rainfall scale, the
number of missing values was less than 10% for the period 1975–2004. The number
of missing values is estimated using various types of weighting methods. Meanwhile,
in Singapore, the daily time scale analyzed from 30 stations for the period 1980 to
2010 shows very few missing values data, namely less than 1%. However, the area
used in Singapore is relatively small, around 716 km2 compared to Java, which is
128.297 km2 .
3.2 Statistical Analysis of Hourly and Daily Rainfall Data
From 25 rain stations for the hourly rain scale, the largest average value of rainfall
occurred at Semarang Maritim Rain Station (96,837) of 1.17 mm. Meanwhile, for the
daily rain scale, there is only one rain station that has an average value below 1 mm,
namely Banyuwangi Rain Station (96,987) of 0.06 mm. Then for the hourly scale
of rain, the largest maximum amount of rain occurs at the Jatiwangi Cirebon Rain
Station (96,791) of 147 mm which occurs precisely at 17.00 on 25 December 2013.
For the daily rain scale, the Cilacap rain station (96,805) has the largest maximum
amount of rain, which is 302.7 mm. For the minimum amount of rain in the hourly
and daily rain scale at each rain station, it is 0 mm, which means that each rain station
has dry hours. Furthermore, the standard deviation value for the largest hourly rain
scale occurred at Semarang Maritim Rain Station (96,837) of 4.08. Meanwhile, the
largest standard deviation value for the daily rain scale occurred at the Cilacap Rain
Station (96,805), which was 19.56. Statistical analysis for 25 rain stations can be
seen in Tables 3 and 4.
Diurnal cycles and regional variations in rainfall are essential in the tropics.
Because tropical rainfall involves the uptake and release of large amounts of latent
heat from evaporation, in the study of energy and water cycles on a large scale such
as the Asian monsoon, an understanding of the diurnal rainfall cycle is needed. Most
studies have shown that diurnal variation in precipitation in the Tropics peaks in the
afternoon overland areas and the morning over adjacent ocean areas (e.g. [11–14]).
Table 3 Statistical analysis of hourly rainfall data

Rain station Average (mm) Maximum (mm) Minimum (mm) Standard
deviation
Tangerang 0.69 55 0 3.16
Serang 0.18 85 0 1.61
Curug 0.31 80 0 2.57
Jakarta BMKG 0.24 82 0 2.15
Cengkareng 0.15 73.9 0 1.69
Citeko 0.37 100 0 2.21
Dermaga 0.41 107 0 2.96
Bandung 0.3 67 0 2.09
Jatiwangi_Cirebon 0.27 147 0 2.46
Tegal 0.18 99.7 0 1.86
Cilacap 0.35 111.2 0 2.56
Semarang_Klimatologi 0.37 94 0 2.71
Semarang_Maritim 1.17 63 0 4.08
Ahmad Yani_Semarang 0.64 85 0 3.37
Surabaya_Perak1 0.23 99 0 2.3
Juanda 0.26 80 0 2.31
Karang Ploso 0.21 77 0 1.89
Tretes 0.8 99 0 4.03
KarangKates 0.62 61.5 0 3.17
Kalianget_Madura 0.18 100 0 1.84
Sawahan 0.53 77.2 0 3.21
Banyuwangi 0.36 74 0 2.5
Mataram 0.18 70 0 1.86
Bawean 0.32 99 0 2.51
Surabaya Tg. Perak 0.21 91.1 0 2.01
982 Suroso et al.
Table 4 Statistical analysis of daily rainfall data

Rain station Average (mm) Maximum (mm) Minimum (mm) Standard
deviation
Tangerang 2.13 67.8 0 8.64
Serang 4.33 101.4 0 10.41
Curug 7.55 187.8 0 17.07
Jakarta BMKG 5.68 277.5 0 15.47
Cengkareng 3.58 125.2 0 11.72
Citeko 8.91 245.3 0 15.98
Dermaga 9.7 240 0 18.58
Bandung 7.28 108.4 0 13.39
Jatiwangi_Cirebon 6.6 157.5 0 15.44
Tegal 4.38 140 0 12.15
Cilacap 8.27 302.7 0 19.56
Semarang_Klimatologi 8.78 245.2 0 17.86
Semarang_Maritim 7.14 100 0 16.26
Ahmad Yani_Semarang 1.99 54.5 0 6.78
Surabaya_Perak1 4.31 187 0 14.78
Juanda 5.73 118.2 0 14.15
Karang Ploso 5.11 152 0 12.63
Tretes 7.27 187.4 0 18.35
KarangKates 5.26 91.9 0 11.6
Kalianget_Madura 4.31 157.9 0 11.79
Sawahan 9.96 176 0 18.67
Banyuwangi 0.06 13.3 0 0.71
Mataram 4.4 164 0 12.78
Bawean 7.09 221.2 0 16.53
Surabaya Tg. Perak 4.72 130 0 12.36
3.3 Daily Cycle Pattern of Hourly Rainfall
In this study, two seasons were used to determine the daily rain pattern at selected
stations, namely the wet and dry seasons. Based on the rain event in Java, it is chosen
for the wet season to occur from October to March and the dry season occurs from
April to September. Furthermore, a rain station that has the least amount of NA (Non-
Available) is selected to determine the daily rainfall pattern based on hourly rainfall
data. The selected rain stations are the Dermaga rain station (96,753) and Citeko
rain station (96,751). The results of the daily rainfall patterns for the two stations in
the two predetermined seasons can be seen in Fig. 3. The boxplot for the daily rain
pattern at the Citeko rain station and Dermaga in the wet season and dry season have
Fig. 3 Boxplot for the daily rain pattern at Citeko and Dermaga rain station
many outlier values, which means that there is a lot of hourly rain data which has
very high values differ greatly from other data.
3.4 Simulation Model of Hourly Rainfall
In this research, to develop a simulation model, the assumption is that for rain below
0.1 mm in an hourly period, it is considered 0 mm. The results of the daily rain
pattern in Fig. 4 shows that the daily rain has the same pattern for both the Citeko
and Dermaga rain stations in two seasons, the wet and dry seasons. Furthermore, a
daily rainfall model is made by considering the mean hourly rainfall as shown in
Fig. 4. The results of the development of a simulation model from known patterns
show that intermittent rainfall has a daily cycle pattern; that is, some concentrated rain
occurs in the afternoon or at night. This is because morning to noon has evaporated
and formed clouds. So that in the afternoon or evening, the clouds begin to condense
so that it rains.
984 Suroso et al.
Fig. 4 Daily rain modeling for Citeko and Dermaga rain station
3.5 Model Evaluation
After obtaining the hourly rain simulation modelling using the daily rainfall data, it is
necessary to evaluate it to obtain the performance of the model. Evaluation based on
the mean value for simulated and observed rain can be seen in Fig. 5. The mean value
for the simulated and observed rain has quite a small difference. For the Citeko rain
station, the average difference in the mean value for simulated rain and observations
in the wet season is 0.193 mm and for the dry season 0.185 mm. Then for Dermaga
rain station, in the wet season the average difference in the mean value is 0.214 mm
and for the dry season is 0.317 mm.
The next model evaluation is MAE and RMSE. The two prediction models are
said to be the best if the value is 0 (zero). For the Citeko rain station in the wet season,
the average of MAE value is 1.05 mm, and for the dry season 0.85 mm. Furthermore,
the average of RMSE value in the wet season is 2.56 mm and for the dry season is
2.12 mm. For the Dermaga rain station in the wet season, the average of MAE value
is 1.17 mm and for the dry is 1.25 mm. Furthermore, the average of RMSE value in
the wet season is 2.96 mm and for the dry season is 2.98 mm. From Fig. 6 shows that
the pattern of MAE and RMSE values is close to the rain pattern simulations that
have been made. The average error value is both above 0 but not too big. The average
value for the Citeko rain station tends to be smaller than Dermaga rain station.
The last model evaluation is seen from the value of the Pearson and Spearman
correlation coefficient. The average value of the Pearson correlation coefficient for the
Citeko rain station in the wet season is 0.29 which indicates that entering into class 2
is sufficiently correlated. Meanwhile, the dry season falls into class 1 with an average
of 0.21 which indicates that the correlation is very weak. The Spearman correlation
coefficient for the Citeko rain station is smaller than the Pearson correlation. In the
wet season, the average coefficient value is 0.24 and 0.15 in the dry season indicates
that it is classified as class 1. For the Dermaga rain station, the value of the Pearson
and Spearman correlation coefficient in the wet and dry season falls into class 1. For
the Pearson correlation in the wet season, the average value is 0.21 and 0.17 in the
dry season. Furthermore, the average Spearman correlation value in the wet season
is 0.14 and 0.12 in the dry season. Figure 7 shows that the pattern of the Pearson and
Spearman correlation coefficient values is close to the rain pattern simulations that
have been made. However, the correlation coefficient for both of them is quite small,
so it can be concluded that the evaluation of rain simulations and observations has a
very weak correlation. For this reason, further modelling is needed if it is to be used
for forecasting rain data.
Fig. 5 Comparison of mean value of simulation and observation rain

986 Suroso et al.
Fig. 6 Comparison of MAE

and RMSE values
Fig. 7 Comparison of Pearson and spearman correlation coefficient value

4 Conclusion
Based on the quality of hourly and daily rainfall data, it shows that the quality of
rain at 25 rain stations in Java Island is not good so that only two rain stations
were selected with the least amount of NA (Not Available) data, that are Citeko
and Dermaga. The results of the development of a simulation model show that the
hourly rainfall has a daily cycle pattern, that is some concentrated rain occurs in
the afternoon or at night. The mean values for simulated and observed rain at the
Citeko and Dermaga rain stations in both the wet and dry seasons have quite small
differences. The average error (MAE and RMSE) value of the two rain stations is
above 0 but not too big. The average error value for the Citeko rain station smaller
than Dermaga. The value of the Pearson and Spearman correlation coefficient for
the two rain stations is quite small. The evaluation of simulated and observed rain
has a very weak correlation. Hourly rain simulation modelling using daily rainfall
data can be used for design, especially in the field of water construction. However,
for forecasting purposes, further modelling is required.
Acknowledgements This research is part of a strategic research collaboration between the Indone-
sian government and the UK government with the title “Extreme Rainfall and Its Impact on Flood
Risk in Indonesia”.
References
1. Mosavi A, Ozturk P, Chau KW (2018) Flood prediction using machine learning models:
literature review. Water 10(11):1–40
2. Jati MIH, Suroso S, Santoso PB (2019) Prediction of flood areas using the logistic regression
method (a case study of the provinces Banten, DKI Jakarta, and West Java). J Phys Conf Ser
1367(1):012087 (IOP Publishing)
3. Pratidina G, Suroso S, Santoso PB (2019) Detection of satellite data-based flood-prone areas
using logistic regression in the central part of Java Island. J Phys Conf Ser 1367(1):012086
(IOP Publishing)
4. Suroso S, Ghifari IA, Santoso PB (2021) Detection of trend behaviour of extreme rainfall over
Java using Mann-Kendall. In: IOP conference series: earth and environmental science, vol 698,
issue No. 1. IOP Publishing, p 012010
5. Bakhshi M, Al Janabi F (2019) Disaggregation the Daily rainfall dataset into sub-daily
resolution in the temperate oceanic climate region. Int J Mar Environ Sci 13(1):11–16
6. Brigandì G, Aronica GT (2019) Generation of sub-hourly rainfall events through a point
stochastic rainfall model. Geosciences 9(5):11–14
7. Diez-Sierra J, del Jesus M (2019) Subdaily rainfall estimation through daily rainfall down-
scaling using random forests in Spain. Water 11(1).https://doi.org/10.3390/w11010125
8. Suroso S, Bárdossy A (2018) Multisite daily precipitation simulation in Singapore. MATEC
Web of Conferences 195:05007 (EDP Sciences)
9. Mehrotra R, Li J, Westra S, Sharma A (2015) A programming tool to generate multi-site daily
rainfall using a two-stage semi-parametric model. Environ Model Softw 63:230–239
10. Vandenberghe S, Verhoest NEC, Onof C, De Baets B (2011) A comparative copula-based
bivariate frequency analysis of observed and simulated storm events: a case study on Bartlett-
Lewis modeled rainfall. Water Resour Res 47(7)
988 Suroso et al.
11. Gray WM, Jacobson RW Jr (1977) Diurnal variation of deep cumulus convection. Mon Weather
Rev 105(9):1171–1188
12. Murakami M (1983) Analysis of the deep convective activity over the western Pacific and
Southeast Asia Part I: Diurnal variation. J Meteorol Soc Jpn Ser II 61(1):60–76
13. Mori S, Jun-Ichi H, Tauhid YI, Yamanaka MD, Okamoto N, Murata F, Sribimawati T (2004)
Diurnal land-sea rainfall peak migration over Sumatera Island, Indonesian Maritime Conti-
nent, Observed by TRMM satellite and intensive rawinsonde soundings. Mon Weather Rev
132(8):2021–2039
14. Sakurai N, Murata F, Yamanaka MD, Mori S, Hamada JI, Hashiguchi H, Suhardi B (2005)
Diurnal cycle of cloud system migration over Sumatera Island. J Meteorol Soc Jpn Ser II
83(5):835–850
Water Quality Mapping on the Coast
of Bangkalan Madura Based
on the Acidity Value from Aqua MODIS
Satellite Imagery
Hendrata Wibisana , Zetta Rasullia Kamandang , and Kartini
Abstract This study aims to present a map of the acidity values distribution along
the Bangkalan, Madura coastline to illustrate the water quality based on its acidity
value by implementing remote sensing technology that used the reflectance value
at wavelengths of visible light, which is correlated with acidity value (pH). This
information could be essential to people who live on the coast, so they could be
more aware of the water quality changing because most of them depending on their
livelihood from fishery activities. Based on the calculation, the mathematical models
representing the pH value are exponential and linear algorithms due to their number
of R2 values compared to other models, with 0.779 and 0.773 values, respectively.
Furthermore, the study result indicates that acidity in the Bangkalan coast is in the
range of 6.4 to 6.98, obtained by an exponential algorithm, and 6.57 to 6.98, obtained
by a linear algorithm, respectively. This study concludes that the distribution of
acidity in the Bangkalan coast is still within the safe pH range for aquatic ecosystems
on the coast.
Keywords Acidity value · Aqua MODIS satellite imagery · Madura
1 Introduction
Good water quality affects the healthiness of the ecosystem. It has a significant influ-
ence on the growing process of living creatures because it is contained of compounds,
minerals, and ions. In terms of monitoring the quality of water, various existing tech-
nologies and also parameters are used such as total suspended solids [1, 2], sea
H. Wibisana (B) · Z. R. Kamandang · Kartini

Universitas Pembangunan Nasional “Veteran” Jawa Timur, Jl.Raya Rungkut Madya,
Gunung Anyar, East Java 60294 Surabaya, Indonesia
Z. R. Kamandang
Kartini
990 H. Wibisana et al.
surface temperature [3, 4], salinity, chlorophyll-a, and acidity of waters. One of the
considered technologies that able to map parameters of water quality determining
is remote sensing. Remote sensing utilizes satellite imagery in a monitoring process
that is currently widely used by researchers in various parts of the world. The satel-
lite imagery has a variety of characteristics in its use, some use low resolution such
as Meris satellite imagery [5], Aqua MODIS satellite imagery [6–8], Terra MODIS
[9–11], and Suomi VIIRS [12, 13]. Some experts also prefer intermediate resolution
such as ASTER satellite imagery [14], Landsat imagery [9, 15, 16], and imagery
SPOT [17]. Overall, this satellite image relies on sunlight to record data in the form
of objects on the earth’s surface, so that experts who have examined the parameters
of water quality many use wavelengths of visible light. In this case, the wavelength is
represented by red light, green light, and blue light or ordinary given RGB notation
(red, green, blue).
Bangkalan, Madura is a coastal area that is surrounded by fish ponds, shrimp
farms, and mangrove areas. The coastal area is a home for people who depends
on the water creatures such as fish and shrimp for their livelihood. Therefore, the
ecosystem problems that might be caused by water quality issues, such as water
pollution could affect their fishery products.
The purpose of this study is to analyze water quality in the Bangkalan coast by
mapping the acidity distribution of the coastal area using the Aqua MODIS satellite
imagery. The expected results could be essential to people who live on the coast, so
they could be more aware of the water quality changing, regarding its acidity.
2.1 Satellite Image Data
To collect Aqua MODIS satellite imagery data, a process including downloading the
data is conducted by accessing http://oceancolor.glfc.nasa.com [18] for May 2020
with name A2020128061500.L2_LAC_OC.nc, where A is an in initial for Aqua
MODIS and the next four-digit means year 2020, three-digit following it is 128
indicates the 128th of days in a year according to the ‘Julien Days’ rule, the rest of
the next digit 061,500 is the satellite recording time wherein recorded at 6 a.m. past
15 min. This satellite image is a Level-2 image that has been corrected geometrically.
The Aqua MODIS imagery in full view can be seen in Fig. 1.
Figure 2 above illustrates the scene of the Aqua MODIS satellite image of Madura
island. The figure was obtained by cropping the wider image (Fig. 1) using SeaDAS
7.5 software.
Water Quality Mapping on the Coast of Bangkalan Madura Based … 991
Fig. 1 A full scene of

Indonesia Aqua MODIS
satellite images on 8 May
2020
Fig. 2 A cropped scene from Fig. 1 of research location, Madura Island

Table 1 Acidity data (pH)

Point Longitude Latitude pH
with the field coordinates
1 112.8911 −7.2262 7.5
2 112.9138 −7.2261 7.8
3 112.9365 −7.2260 7.3
4 112.9478 −7.2259 7.2
5 112.9705 −7.2372 7.1
6 112.9819 −7.2371 6.7
7 113.0045 −7.2370 6.8
8 113.0159 −7.2370 6.5
9 113.0385 −7.2369 6.6
10 113.0386 −7.2482 6.5
11 113.0159 −7.2483 6.7
12 113.0046 −7.2483 6.8
13 112.9819 −7.2484 6.8
14 112.9706 −7.2485 6.7
15 112.9592 −7.2485 6.9
16 112.9478 −7.2373 7.1
17 112.9366 −7.2486 7.3
18 112.9139 −7.2487 7.6
19 112.8912 −7.2488 7.5
20 112.8798 −7.2375 7.6
2.2 Sea Water Acidity Data (pH)
Acidity value data denoted by the pH value is obtained directly in the field by
measuring sea level at a depth of 10 to 20 cm using a digital pH meter and directly
reading the pH value generated at the corresponding coordinates. The coordinates
that accompany the pH measurements were performed using GPS Navigation and at
each stop, the measurements of coordinates Latitude and longitude were conducted
before measuring the pH value. The results of the field measurements can be seen in
Table 1, where the total number of measurements is 20 points.
2.3 Algorithm of Mathematical Models
After resulting the Aqua Modis satellite image reflected data generated by extracting
these values in the SeaDAS 7.5 software by inserting the coordinate values, the next
step was to analyze the mathematical model with trend analysis to obtain mathemat-
ical equations and the existing R2 correlation values. From the obtained R2 -value, it
can be seen which the most optimal model that can describe the acidity distribution
conditions in the Bangkalan coast.
2.4 Thematic Map of pH Distribution on the Bangkalan,

Madura Coast
To illustrate the distribution of pH value on the coast, this study implemented the two
most optimal mathematical model algorithms as indicated by the accepted R2 -value
of other models using SeaDAS 7.5 software. It concluded that acidity distribution
which has processed using a wavelength of 667 nm has the best results compared to
412 nm and 531 nm.
The most optimum mathematical model results using the trend analysis method
presented in the scatter diagram form are shown in Table 2. It presents the resume
of all analyzed mathematical models with the existing R2 correlation value and the
initial of variable x that represents a reflectance of satellite images.
Table 2 shows that the linear model and the exponent model have the highest R2
correlation value with 0.773 for the linear and 0.779 for the exponent, respectively.
The corresponding wavelength is 667 nm which is in the red spectrum, so it can
be concluded that the red wavelength can provide a strong enough signal to detect
changes in the acidity value in coastal waters.
Furthermore, to see how far the chosen mathematical models could provide the
same response to others coordinate points, a validation test was conducted using 5
Table 2 Resume the

No Algorithm Mathematical models R2
algorithm of a mathematical
model from 667 nm 1 Linear pH = 83.531x + 6.3879 0.773
wavelength 2 Exponent pH = 6.4089e11,834x 0.779
3 Logarithmic pH = 0.4845ln(x) + 9.4612 0.692
4 Power pH = 9.9189x0,0689 0.703
Source Calculation results
remaining field data that has not been used in the previous calculation. The result of
the validation test is presented in Fig. 3.
As seen in Fig. 3, the pH value of each mathematical model has a slight difference
compared to each other. After the pH value of the field (in-situ) data has processed
using algorithm calculation of satellite image data, the mathematical models that can
represent the situation of Bangkalan, Madura coast water were obtained, however,
an advanced correlation test is still required to identify their correlation to each other.
The results are seen in Table 3.
7.7
7.6
7.5
7.4
7.3
7.2
pH
Series1
7.1
7 Series2
6.9 Series3
6.8 Series4
6.7 Series5
Exponent Logarithmic
In-situ (pH) Linear (pH) Power (pH)
(pH) (pH)
Series1 7.1 7.112 7.101 7.161 7.151
Series2 7.2 7.015 7.004 7.091 7.08
Series3 7.4 7.307 7.3 7.276 7.27
Series4 7.5 7.556 7.562 7.392 7.391
Series5 7.4 7.44 7.439 7.342 7.338
Fig. 3 Validation test result for each mathematical model
Table 3 Comparison of the correlation coefficient of mathematical model

In-situ Linear Exponent Logarithmic Power
In-situ 1
Linear 0.9056 1
Exponent 0.9059 0.99995 1
Logarithmic 0.8970 0.99676 0.99590 1
Power 0.8979 0.99719 0.99638 0.99998 1
Source Calculation results
Table 4 shows that the in-situ pH value has a large correlation with the linear
model and the exponent model with a 0.9056 correlation value or around 90.56% of
the available data. It can be represented as a description of the actual field conditions
in comparison with the logarithmic model and the power model that only shows
0.8979 of value, but indeed these values are also able to describe the situation well.
The final result is shown by depicting acidity mapping in Bangkalan, Madura
coast, where Fig. 4 shows the results of the thematic map for linear algorithms at
667 nm wavelength, and Fig. 5 shows the results of thematic maps for the exponent
algorithm at the same wavelength. Those figures show that the distribution of acidity
ranges is in the value of 6.4–6.98.
Table 4 Comparison of the correlation of mathematical model

In-situ Linear Exponent Logarithmic Power
In-situ 1
Linear 0.9056 1
Exponent 0.9059 0.99995 1
Logarithmic 0.8970 0.99676 0.99590 1
Power 0.8979 0.99719 0.99638 0.99998 1
Fig. 4 Map of acidity (pH) using the linear algorithm at 667 nm wavelength
Fig. 5 Map of acidity (pH) using the exponent algorithm at 667 nm wavelength
4 Conclusion
Aqua MODIS image data can be used as a reference to identify the mathematical
model algorithms that able to calculate any parameters such as the acidity or pH
value of the coastal water environment. This study concludes that the distribution of
acidity in the Bangkalan coast is still within the safe pH range for aquatic ecosystems
on the coast.
References
1. Wibisana H, Soekotjo BM, Lasminto U (2019) Preliminary study of total suspended solid distri-
bution in coastal ujung Pangkah Gresik based reflectance value of Landsat satellite imagery.
Indones J Geogr 51(1):. https://doi.org/10.22146/ijg.38967
2. Sirjacobs D et al (2011) Cloud filling of ocean colour and sea surface temperature remote
sensing products over the Southern North Sea by the data interpolating empirical orthogonal
functions methodology. J Sea Res. https://doi.org/10.1016/j.seares.2010.08.002
3. Wang Y, Huang F, Fan T (2017) Spatio-temporal variations of Arctic amplification and their
linkage with the Arctic oscillation. Acta Oceanol Sin 36(8):42–51. https://doi.org/10.1007/s13
131-017-1025-z
4. Koner PK , Harris A (2016) Sea surface temperature retrieval from MODIS radiances using
truncated total least squares with multiple channels and parameters. Remote Sens 8(9). https://
doi.org/10.3390/rs8090725
5. Maritorena S, Fanton D’andon OH, Mangin A, Siegel DA (2010) Merged satellite ocean color
data products using a bio-optical model: characteristics, benefits and issues. Remote Sens
Environ 114:1791–1804. https://doi.org/10.1016/j.rse.2010.04.002
6. Marpaung S et al (2020) Analysis of water productivity in the Banda Sea based on remote
sensing satellite data. Int J Remote Sens Earth Sci 17(1):25. https://doi.org/10.30536/j.ijreses.
2020.v17.a3280
7. Wibisana H, Kamandang ZR, Zaenab S (2020) Analysis of mathematical models in mapping
total suspended solids (TSS) on the coast of Banyuwangi beach based on satellite image data.
Int. J Sci Eng Investig 9. Accessed: 23 Nov 2020 [Online]. Available https://oceancolor.gsfc.
nasa.gov/
8. Brando V, Dekker A, Marks A, Qin Y, Oubelkheir K (2006) Chlorophyll and suspended sedi-
ment assessment in a macrotidal tropical estuary adjacent to the Great Barrier Reef: spatial and
temporal assessment using remote sensing. Coop Res Cent Coast Zo Estuary Waterw Manag
Tech Rep 74:1–128
9. Fan X, Liu Y, Wu G, Zhao X (2020) Compositing the minimum NDVI for daily water surface
mapping. Remote Sens 12(4):700. https://doi.org/10.3390/rs12040700
10. Ody A et al (2016)“Potential of high spatial and temporal ocean color satellite data to study
the dynamics of suspended particles in a micro-tidal river plume. Remote Sens 8(3). https://
doi.org/10.3390/rs8030245
11. Moreno-Madrinan MJ, Al-Hamdan MZ, Rickman DL, Muller-Karger FE (2010) Using the
surface reflectance MODIS Terra product to estimate turbidity in Tampa Bay, Florida. Remote
Sens 2(12):2713–2728. https://doi.org/10.3390/rs2122713
12. Shi W, Zhang Y, Wang M (2018) Deriving total suspended matter concentration from the
near-infrared-based inherent optical properties over turbid waters: a case study in Lake Taihu.
Remote Sens 10(2):1–17. https://doi.org/10.3390/rs10020333
13. Platnick S et al (2020) The NASA MODIS-VIIRS continuity cloud optical properties products.
Remote Sens 13(1):2. https://doi.org/10.3390/rs13010002
14. Wang C, Chen S, Li D, Wang D, Liu W, Yang J (2017) Engineering technology center for remote
sensing big data application. Geosci Model Dev 10:4347–4365. https://doi.org/10.5194/gmd-
10-4347-2017
15. Miller RL, Liu CC, Buonassissi CJ, Wu AM (2011) A multi-sensor approach to examining the
distribution of total suspended matter (TSM) in the Albemarle-Pamlico estuarine system, NC,
USA. Remote Sens 3(5):962–974. https://doi.org/10.3390/rs3050962
16. Nazirova K et al (2021) Comparison of in situ and remote-sensing methods to determine
turbidity and concentration of suspended matter in the estuary zone of the Mzymta River,
Black Sea. Remote Sens 13(1):143. https://doi.org/10.3390/rs13010143
17. Rocchini D (2013) Seeing the unseen by remote sensing: satellite imagery applied to species
distribution modelling. J Veg Sci 24(2):209–210
18. Feldman GC (2020) NASA Goddard Space Flight Center, Ocean Ecology Laboratory, Ocean
Biology Processing Group. Moderate-resolution Imaging Spectroradiometer (MODIS) Aqua
A2020128061500.L2_LAC_OC.nc Data; NASA OB.DAAC, Greenbelt, D, USA, 2021. https://
oceancolor.gsfc.nasa.gov/data/aqua/
Multi-attribute Analysis of Raw Water
Treatment from Deep Wells at PDAM
Tirta Mahottama, Klungkung Regency,
Bali
Ni Kadek Dian Utami Kartini, Nurulbaiti Listyendah Zahra,

Ariyanti Sarwono, Intan Rahmalia, Almira Davina Nastiti,
Iva Yenis Septiariva, and I. Wayan Koko Suryawan
Abstract PDAM Tirta Mahottama is a regional drinking water company located

in Klungkung Regency, Bali Province. The community uses 21 deep wells to meet
their water needs for daily activities. This study aims to determine the reliability
characteristics of raw water from deep wells and the quality of raw water used in the
Klungkung Regency. This study was conducted to determine what processing can be
done to improve raw water quality using the Multi-Attribute Utility Theory (MAUT).
The deep well used has a production capacity of 1–9 L/s. In terms of raw water quality,
the parameter that does not meet the quality standard is aluminum, followed by color
and TDS. The selected processing is adsorption with consideration of its high ability
to remove color and aluminum. Based on MAUT analysis, the best adsorbs are using
GAC (granular activated carbon).
Keywords Sludge treatment · Water treatment · Decision analysis
1 Background
PDAM Tirta Mahottama is a PDAM located in Klungkung Regency, Bali Province.

PDAM Tirta Mahottama serves four sub-districts and uses water sources in ground-
water, springs, and seawater. PDAM Tirta Mahottama has 21 deep wells, seven
N. K. D. U. Kartini · N. L. Zahra · A. Sarwono · I. Rahmalia · A. D. Nastiti ·

I. W. K. Suryawan (B)
Department of Environmental Engineering, Faculty of Infrastructure Planning, Universitas
Pertamina, Jakarta 12220, Indonesia
N. L. Zahra
A. Sarwono
I. Y. Septiariva
Department of Civil Engineering, Faculty of Engineering, Universitas Sebelas Maret, Surakarta
57126, Indonesia
1000 N. K. D. U. Kartini et al.
springs, and one reverse seawater osmosis (SWRO) treatment unit. According to the
2019 PDAM Performance Book, PDAM Tirta Mahottama had received a healthy
category from 2017 to 2019 with customers’ water quality aspect value of 88% in
2019. This value shows the quality of water in customers has not been maximal or
has not fully met quality standards. Because it has not reached 100%, this value is
still included in the three lowest scores compared to other PDAMs in Bali Province.
Improving the quality of the environment in Klungkung Regency will be in line
with water resources management. Water resources in Klungkung Regency have not
been utilized until 100%, with most deep wells [1]. Deep wells pollution depends on
groundwater infiltration, which carries wastewater [2].
The quality and quantity of deep well water in the Klungkung Regency need
to be studied further to determine community consumption feasibility. Besides, the
technology required to improve quality needs to be studied further. This study aims
to develop the right technology selection based on the quantity and quality of deep
wells located in Klungkung Regency.
2 Method
2.1 Data Collection
The data used in this research is from monitoring the deep well quality of PDAM Tirta
Mahottama. Brucine measured inorganic chemical parameters Nitrite was measured
by colorimetric method and Nitrate. The physical parameters measured were odor,
color, TDS, taste, and temperature tested by organoleptic, calorimetric, organoleptic,
and thermometer methods, respectively. The chemical parameters measured were
aluminum with titrimetry, iron with phenanthroline, hardness with titrimetry, chlo-
ride with titrimetry, manganese with trichloride, pH with pH meter, sulfate with
turbidimetry, and ammonia with Nessler.
2.2 Selection of Alternative Processing
The selection of alternatives is carried out using the Multi-Attribute Utility Theory
(MAUT) method. This method is a method of decision-making by considering more
than one criterion. According to Kiker et al., in applying multicriteria decision anal-
ysis in environmental decision making, this method has often been used in various
fields to assist decision-making, including management, environmental management,
economics, and engineering [3].
Multi-attribute Analysis of Raw Water Treatment from Deep Wells … 1001
3.1 Reliability of Raw Water Sources
PDAM Tirta Mahottama is one of the PDAMs that uses deep wells as a source of raw
water. There are 21 deep wells utilized, with a total production discharge of 164.5
L/s (Table 1). In the archipelago area, namely Nusa Penida Island, the water sources
in Kutampi I, II, and III deep wells are only installed half of the mainstay discharge.
Development of groundwater sources has been carried out in Nusa Penida by drilling
wells on Nusa Penida’s island. The availability of deep well water drops dramatically
during the dry season, so it is necessary to think about other ways to meet the water
needs in the Klungkung Region archipelago [4].
The production process in the deep well of PDAM Tirta Mahottama only consists
of taking water from the well through a submersible pump and disinfection using
chlorine gas. Then the water is directly distributed to customers. The disinfec-
tion process is carried out by injecting chlorine gas directly into the distribution
pipe. However, some deep wells still have some parameters that exceed the quality
standard, so that these parameters are not excluded.
3.2 Water Quality
Table 2 showed the predominant parameter is aluminum, followed by color and TDS.
The excess of TDS occurs in deep wells located in the Nusa Penida area. The Nusa
Penida area is composed of limestone, which causes seawater intrusion to occur
quickly and affects the quality of deep well water to become brackish. It is planned
that the advantages of TDS in the area will be processed using a distillation process.
Water treatment in the Nusa Penida area is well planned [1], so that in this study, no
further discussion was carried out. Also, nutrient parameters are seen under normal
conditions, so they will not negatively impact streams [5, 6].
Aluminum metal is the metal most found in nature. Aluminum can be in the
soil, water, and air and makes up about 8% of the earth’s crust. Aluminum is a
very reactive element that is easily oxidized. Next source of intake of aluminum
in drinking water. According to Permenkes No. 492 of 2010 concerning Drinking
Water Quality Requirements, aluminum’s full drinking water content is 0.2 mg/L.
The source of aluminum in water can come from chemical residues used during the
drinking water treatment process, such as aluminum, commonly used as a coag-
ulant. Besides, aluminum can naturally be contained in groundwater. The color
in water indicates organic content and ions metals such as iron, manganese, and
turbidity [5]. Color is a component of physical water quality. The presence of color
will interfere with the water’s aesthetics and indicate polluted water [6, 7]. True
color was measured in water samples filtered using a filter with a size of 0.45 µm.
In simple terms, the true color is the color that is measured when the turbidity
Table 1 Characteristics utilization of raw water from deep wells in Klungkung Regency
Location The reliability Installed (L/s) Production Not yet Be Sub-district
of clean water (L/s) used (L/s)
(L/s)
Sema Agung 5 5 4 1 Banjarangkan
Desa Pikat 4 2.5 2 2 Dawan
Paksebali 10 10 7 3 Dawan
RSU 10 10 4 6 Klungkung
Klungkung
Tegalinggah 5 5 5 0 Klungkung
Kutampi (I) 5 2.5 2 3 Nusa Penida
Kutampi (II) 5 2.5 1 4 Nusa Penida
Kutampi (III) 5 2.5 2 3 Nusa Penida
Dawan Kaler 10 10 5 5 Dawan
Kusamba 8 4 2,5 5,5 Dawan
Buayang 18 15 8 10 Dawan
Mincidan 18 15 8 10 Dawan
Swecapura 10 5 3 7 Klungkung
Sedap Malam 7 5 5 2 Klungkung
Akah 20 12 12 8 Klungkung
Koripan 5 5 5 0 Banjarangkan
Pura 2,5 2,5 0 2,5 Banjarangkan
Kentelgumi
Balai Budaya 12 12 9 3 Klungkung
Cucukan 0 0 0 0 Klungkung
Besan 0 0 0 0 Dawan
Pesinggahan 5 5 5 0 Dawan
in the water has been removed. True color comes from the breakdown of organic
compounds, such as hummus, lignin’s, tannins, and other organic acids. The pres-
ence of this compound makes the color in water challenging to remove. Water that
contains true color is generally characteristic of yellow to reddish-brown or relatively
clear.
3.3 Water Treatment Decision Analysis
The design unit’s determination is done through a literature study by looking at

previous research on aluminum removal and color. Several removal methods can be
used to remove aluminum and color, including adsorption, precipitation, membranes,
Table 2 The water quality of deep wells in Klungkung Regency
Location NO2 NO3 Color TDS Al Fe Hardness Chloride Mn pH Sulfate NH3
mg/L mg/L TCU mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L
Sema Agung nd 1.78 3 217 0.26 0.0004 194 16.08 nd 7.7 11.64 0.10
Desa Pikat nd 0.99 0 414 0.29 0.047 388 38.32 0.0008 8 5.27 0.19
Paksebali nd 1.01 7 291 0.28 nd 291 16.66 nd 7.9 11.88 0.11
RSU Klungkung nd 2.79 1 264 0.26 0.012 232.8 22.49 nd 8.2 19.84 0.08
Tegalinggah nd 2.7 40 368 0.25 0.247 291 20.9 0.0012 7.4 17.63 0.12
Kutampi (I) nd 4.55 1 2740 0.23 nd 853.6 1037.16 0.0187 7 58.35 1.95
Kutampi (II) nd 4.23 4 1979 0.28 0.026 659.6 755.76 0.021 7.3 45.72 0.98
Kutampi (III) nd 3.41 7 2650 0.33 nd 756.6 1157.76 0.0184 7.1 52.43 1.27
Dawan Kaler nd 1.43 7 339 0.27 0.007 426.8 17.49 nd 8.1 0.69 0.14
Kusamba nd 1.32 7 558 0.26 nd 388 141.62 nd 8.3 24.12 0.13
Buayang nd 4.42 20 300 0.33 0.018 388 27.44 0.0021 7.5 7.71 0.26
Mincidan nd 1.89 11 327 0.25 nd 291 16.08 0.0005 7.4 12.90 0.14
Swecapura 0.0004 3.15 9 335 0.25 0.002 291 27.49 nd 8.4 27.23 0.12
Akah 0.0015 1.85 2 274 0.28 0.039 252.2 18.32 nd 7.9 15.06 0.11
Koripan nd 3.46 4 224 0.26 0.018 213.4 17.68 0.0005 7.1 12.49 0.10
Multi-attribute Analysis of Raw Water Treatment from Deep Wells …
Pura Kentelgumi nd 1.33 9 246 0.22 nd 194 15.27 0.0009 7.9 11.32 0.10
Balai Budaya nd 3.01 2 309 0.28 0.66 213.4 27.49 nd 8.1 25.68 0.11
Cucukan 0.0715 0.68 20 194 0.27 0.334 174.6 17.68 0.0012 7.8 20.56 0.18
Besan 0.0269 0.88 40 251 0.24 nd 194 38.59 0.0022 10.1 19.45 0.29
Pesinggahan 0.442 0.61 40 621 0.2 nd 465.6 60.3 0.0011 7.4 16.45 0.59
Standard 3 50 15 500 0.2 0.3 500 250 0.4 6.5–8.5 250 1.5
1003
* nd: not detected

granular filtration, and ion exchange. However, the PDAM wants processing that is as
simple as possible and does not require high costs. Therefore, several considerations
were made to choose a design alternative. The choice of alternatives at this stage is
early selection. At this stage, the selection unit will be varied in processing methods,
materials, or configurations for further evaluation using the Multi-Attribute Utility
Theory method. Table 3 shows the considerations used in the selection process.
Table 3 Various considerations for suitable treatment methods to be applied in water treatment
Method Can be remove Installation fee less than Ease of operation and
aluminum and color all Rp 250,000,000 treatment
at once
Adsorption Adsorption can be used 40.000.000–65.000.000 The adsorption
to set aside the color [8] operation is not
and aluminum metal [5] required materials or
specific treatment. The
only backwash is
necessary and
desorption periodic [9]
for treatment
Granular Filtration The pure filtration 5.000.000–10.000.000 Operational filtration is
process can only [10] not required materials
remove a specific size or specific treatment,
of material, but and only backwash is
aluminum with a necessary periodically
minimal size can only [5, 11, 12]
be removed through
membrane filtration [9]
Precipitation Precipitation can be 50.000.000–65.000.000 Operations and care
used to remove [10] relatively easy
aluminum by forming a precipitation, but
precipitate Al (OH) 3+ precipitation need tub
[9] but cannot be used sediment or filter tub to
for color removal put aside residue [9]
In this method, six steps must be taken are describing the alternatives and criteria,
determining the criteria’s importance, determining the qualitative criteria weights,
calculating the utility value, calculating the relative weight, and calculating the final
evaluation value. The description of alternatives has been discussed in the previous
discussion that will be proposed three alternatives processing, namely adsorption
using GAC (granular activated carbon) adsorbent with the adsorption column (fixed-
bed adsorption column) processing method, adsorption with silica adsorbent with the
adsorption column processing method (fixed-bed adsorption column), and adsorp-
tion with PAC (powdered activated carbon). After the weight value for the qualita-
tive criteria is determined, a summary is made of all the measures values for each
alternative as in Table 4.
Table 4 Considerations in choosing alternative media in the adsorption process

Alternatives Criteria
Aluminum Color Cost Area Operational Maintenacne
removal removal required
(m2 )
GAC 99.2% 99.67% 63,792,200.00 3.82 Very easy Kinda easy
Silica 94.6% 63% 63,792,200.00 5.73 Very easy Kinda easy
PAC 99% 80.1% 65,262,304.40 25,22 Kinda easy Very easy
The utility value is calculated on each criterion for each alternative. This utility
requires quantified criterion value data. Relative weight calculation is carried out
on each measure based on the value of the level of importance. The calculation
results can be seen in Table 5. In calculating the final evaluation, v(x), the utility
Table 5 Utility value is calculated on each criterion for each alternative

Alternatives Criteria
Aluminum Color Cost Area Operational Maintenacne
removal removal required
(m2 )
GAC 1 1 1 1 1 0
Silica 0 0 1 0,91 1 0
PAC 0.956 0.466 0 0 0 1
Relative 0.2083 0.2083 0.1667 0.1667 0.125 0.125
Weights
value data and the weight normalization of each criterion are needed. Based on the
analysis of the multi-attribute utility theory (MAUT) method, an alternative ranking
was obtained for the Deep Well as listed in Table 6. Based on the analysis using the
Table 6 Alternate ranking

No Alternative Final evaluation results Ranking
order
1 GAC 0.875 1
2 Silica 0.443 2
3 PAC 0.421 3
The bold indicate the selected water treatment method through the
MAUT analysis.
multi-attribute utility theory (MAUT) method, adsorption with GAC adsorbent was
chosen as the best alternative in the well.
The adsorption process using granular adsorbents usually uses the batch method
column or fluidized-bed operation. The contact system usually applied is a fixed bed
Fig. 1 Diagram of deep well processing in Klungkung Regency with fluidized-bed granular
activated carbon
contacting system or countercurrent moving bed. This application is because these

two methods have low operational costs and streamline the adsorbent’s absorption
ability [13]. Fixed bed columns usually have an up-flow or downflow flow. Figure 1
shows an example of a circuit used from a fixed bed adsorption column in a deep well
in Klungkung Regency. The water will be entered through the influent distributor
and will undergo adsorption in the adsorption zone (carbon bed) and flowed through
the underdrain system as an effluent. The spent water drain tank tube helps store
carbon during backwash to be cleaned, and then after cleaning the carbon, it will
be regenerated using a furnace. GAC can be used in water or air treatment. GAC
can be divided into three types based on the pores’ size: micropores, mesopores,
and macropores [14]. According to IUPAC, micropores have a pore size of <2 nm,
mesopores have a pore size of 2 nm-50 nm, while macropores have a pore size of
>50 nm. The choice of pore size depends on the type of adsorbate to be adsorbed.
Using GAC as an adsorbent, pore size and pore structure are the most important
things that need to be considered. These pores’ size and design depend on the level
of carbon activation and the relative carbon area. The activation rate is determined
by the raw material’s ratio and dehydrating agent and heating time and temperature.
After selecting the best alternative, detailed planning should be carried out
regarding the GAC processing implemented in Klungkung Regency. Good environ-
mental infrastructure planning is supported by considering the needs of the commu-
nity. Currently, the COVID19 pandemic outbreak has caused the need for drinking
water to increase [15]. In Bali Province, the environmental impact of the COVID19
pandemic affects the amount of waste generation [16, 17]. If it is linearly predicted
that a pandemic, COVID19 will also impact the need for clean water needed, which is
immensely increasing. Therefore, future water treatment planning to improve public
health in Klungkung Regency.
4 Conclusion
The measurement of water discharge produced is 1–9 L/s for each deep well in
Klungkung Regency. The parameters that need to be considered in water treatment
are aluminum and color. Based on evaluation using MAUT on several considera-
tions, namely the efficiency of aluminum removal and color, area required, cost,
ease of operation and maintenance, processing of environmental parameters can use
fluidized-bed granular activated carbon.
References
1. Septiariva IY, Suryawan IWK, Suhardono S, Sofiyah ES, Zahra NL (2021) Strategy and master
plan for water supply and domestic wastewater in Klungkung Regency, Bali. In: IOP conference
series: materials science and engineering, vol 1096, issue No. 1. IOP Publishing, p 012058
2. Septiariva IY, Suryawan IWK (2021) Development of water quality index (WQI) and hydrogen
sulfide (H2s) for assessment around Suwung Landfill, Bali Island. J Sustain Sci Manag
16(4):137–148
3. Kiker GA, Bridges TS, Varghese A, Seager TP, Linkov I (2005) Application of multicri-
teria decision analysis in environmental decision making. Integr Environ Assess Manag Int J
1(2):95–108
4. Sudipa N, Mahendra MS, Adnyana WS, Pujaastawa IB (2020) Daya Dukung Air di Kawasan
Pariwisata Nusa Penida, Bali, Jurnal Sumberdaya Alam dan Lingkungan 7(3):117–123
5. Apritama MR, Suryawan I, Afifah AS, Septiariva IY (2020) Phytoremediation of effluent textile
wwtp for NH3 -N and Cu reduction using pistia stratiotes. Plant Arch 20(1):2384–2388
6. Afifah AS, Suryawan IWK, Sarwono A (2020) Microalgae production using photo-bioreactor
with intermittent aeration for municipal wastewater substrate and nutrient removal. Commun
Sci Technol 5(2):107–111
7. Davis ML (2010) Water and wastewater treatment engineering: design principle and practice.
McGraw-Hill. Companies, Inc.
8. Suryawan IWK, Prajati G, Afifah AS, Apritama MR (2021) NH3 -N and COD reduction in
Endek (Balinese textile) wastewater by activated sludge under different DO condition with
ozone pretreatment. Walailak J Sci Technol 18(6):1–11
9. Suryawan IWK, Septiariva IY, Helmy Q, Notodarmojo S, Wulandari M, Sari NK, Sarwono
A, Pratiwi R, Jun-Wei L (2021) Comparison of ozone pre-treatment and post-treatment hybrid
with moving bed biofilm reactor in removal of remazol black 5. Int J Technol 12(2)
10. Rini ADAM, Ali F, Suwartha N (2013) Desain Kolom Adsorpsi granular activated carbon.
Jurusan Teknik Lingkungan Fakultas Teknik Universitas Indonesia, Jakarta
11. Crittenden JC, Trussell RR, Hand DW, Howe KJ, Tchobanoglous G (2012) Water treatment
principle and design, 3rd edn. Wiley, USA
12. Andini DRSDJ, Ainun S (2017) Perencanaan Unit Pengolahan Air Bersih di Kecamatan
Sumedang Selatan. Jurnal Reka Lingkungan 5(2):503–514
13. ReynoldsT D, Richards PA (1996) Unit operation and processes in environmental engineering.
PWS Publishing Company, California
14. Water Quality Association (2016) Granular activated carbon (GAC) fact sheet. Water Granular
Activated Carbon (GAC) Fact Sheet, Illinois
15. Abu-Bakar H, Williams L, Hallett SH (2021) Quantifying the impact of the COVID-19
lockdown on household water consumption patterns in England. NPJ Clean Water 4(1):1–9
16. Suryawan IWK, Rahman A, Septiariva IY, Suhardono S, Wijaya IMW (2021) Life cycle assess-
ment of solid waste generation during and before pandemic of Covid-19 in Bali Province. J
Sustain Sci Manag 16(1)
17. Sari MM et al (2022) Identification of face mask waste generation and processing in tourist
areas with thermo-chemical process. Arch Environ Prot (48)
Determination of Produced Wastewater
Treatment Systems for Reclaim Water
in the Oil and Gas Industry
Novena Lany Pangestu, Nurulbaiti Listyendah Zahra, Ariyanti Sarwono,

Intan Rahmalia, Iva Yenis Septiariva, and I. Wayan Koko Suryawan
Abstract Reclamation of water from the oil and gas industry activities from the
produced water can reduce the environmental impact and increase its proactive
efforts. This study was conducted to assess the appropriate treatment system for
produced water. This study was conducted by taking a case study at the Lindai
Gathering Station (GS). The highest quality parameters that do not meet the quality
standards for class III water reclamation are COD and oil and grease. Some of the
recommended processing units are the American Petroleum Institute (API) separa-
tors, Parallel Plate Interceptors (PPI), and Corrugated Plate Interceptors (CPI). Each
alternative is given the addition of ultra-filtration technology to improve the quality
of the effluent produced water. The results of selecting the best technology using the
Analytical Hierarchy Process (AHP) method show the use of the most suitable CPI
to be applied for water reclamation.
Keywords Produced water · CPI · Reclamation water
1 Background
The activities of the oil and gas industry in the upstream sector are one of the activities
that can cause pollution and environmental disturbance. In the oil and gas industry,
N. L. Pangestu · N. L. Zahra · A. Sarwono · I. Rahmalia · I. W. K. Suryawan (B)

N. L. Zahra
A. Sarwono
I. Y. Septiariva
Department of Civil Engineering, Faculty of Engineering, Universitas Sebelas Maret, Surakarta
57126, Indonesia
1010 N. L. Pangestu et al.
the exploration activities are considered as the primary key of upstream activities.
Exploration activities carried out by the oil–gas industry can produce fluid containing
oil and produce gas, mud, and water. Based on the Ministry of Environments and
Forestry Regulation (PermenLH) Number 19 of 2010, the produced water is defined
as the water carried out from the strata that contain hydrocarbon to the surface during
the oil and gas extraction activities. The quantity of produced water is often more
significant than the quantity of oil produced, especially in old fields. According to
Hayes, 80% of the total volume of wastewater produced by the oil and gas industry
is produced water [1].
Produced water, often referred to as brine, is groundwater present with hydro-
carbons in oil and gas rock formations and becomes waste produced by drilling [2].
Tiana also states that produced water is a by product of the processing of petroleum,
natural gas, or Coal Bed Methane (CBM) [3]. Produced water is always found with
petroleum. This water is slightly acidic and resides under hydrocarbons in formations
or reservoirs. According to Ivory, produced water can be used as clean water to meet
water needs [4]. The use of produced water as clean water requires paying atten-
tion to its quality because the government has regulated the quality standard of raw
water according to its designation in Government Regulation (PP) of the Republic
of Indonesia Number 82 of 2001. Produced water can be used for irrigation, wildlife
consumption, industrial use, or processed into drinking water [5].
The primary treatment of produced water is used to remove oil and grease. Gravity-
based separation technology is a technology that companies generally choose in
removing oil and grease in produced water. According to Forero et al., gravity-based
separation technologies that are usually used in industry are American Petroleum
Institute (API) separators, Parallel Plate Interceptors (PPI), and Corrugated Plate
Interceptors (CPI) [6]. This type of separator generally consists of an initial sepa-
rator that functions to separate large particles and the primary separator used to
separate oil from water. The produced water’s primary treatment does not separate
the hydrocarbons contained in produced water 100% [7]. Secondary treatment is
also often required if the desired effluent has not reached the target. The technolo-
gies commonly used for secondary processing of produced water are membrane
technology, coagulation, and biological processing. Some companies choose to use
biological processing because the nutrient concentration is still very high. Biolog-
ical processing that several companies commonly use is the Biological Aerated Filter
(BAF) and activated sludge [8]. However, the physic-chemical processing method has
been preferred by companies because of the need for a smaller area and operational
simplicity [9].
Various alternatives can be used in water reclamation from oil and gas-produced
water; it is necessary to select the best alternative according to the company’s priority
scale. This research aims to determine the most appropriate alternative in water
reclamation from oil and gas industry activities.
Determination of Produced Wastewater Treatment Systems … 1011
2 Method
2.1 Location
The design of produced water treatment is carried out in the onshore facility oil
and gas industry, namely PT Pertamina Hulu Energi Kampar-Siak, to be precise, at
the Lindai Gathering Station (GS). Administratively, Lindai GS is in Lintas Suram,
Senama Nenek Village, Tapung Hulu District, Kampar Regency, Riau.
2.2 Produced Wastewater Quality
Secondary data includes the produced water characteristics data in Lindai GS, which
can be found in the Environmental Management Plan and Environmental Monitoring
Plan (RKL-RPL) document belonging to PHE Kampar-Siak. The main parameters
that are seen, are the parameters regulated in PermenLH No. 19 of 2010, specifically
for produced water. These parameters include Chemical Oxygen Demand (COD), oil
and grease, temperature, pH, and Total Dissolved Solid (TDS) measured according
to the standard method. The data taken as a design reference is data from 2016 to
2019.
2.3 Decision Method
Treatment alternatives were selected based on literature studies from journals,

proceedings, and books on produced wastewater treatment. The selection of alter-
native processing units is carried out using the Analytical Hierarchy Process (AHP)
method. The AHP method is quite commonly used to determine the best wastewater
treatment unit, which similar as in research conducted by Ramadan et al. [10]. There
are five criteria for selecting alternative production water treatment units.
The assessment criteria include the removal of oil and grease, the effectiveness of
eliminates the COD, the operation and maintenance (O&M), the land requirements,
and the initial cost of processing unit materials. These criteria will then be assessed
and compared in pairs to determine which standards are prioritized. The rating scale
1–9 was chosen because it is the best scale in formulating opinions for various
problems [11].
3.1 Existing Produced Wastewater Treatment System
The production water treatment unit is carried out in the oil and gas industry, namely
at the Lindai Gathering Station on PT Pertamina Hulu Energi (PHE) Kampar-Siak.
In treating produced water from Gathering Station (GS) Lindai used a storage pool
called a pit. There are four pits in Lindai GS, which function to reduce the tempera-
ture and oil levels in the produced water before being channeled into the canal and
forwarded to the Lindai River. The industry’s proactive efforts to reduce environ-
mental impact are water reclamation [12, 13]. For this reason, an additional treatment
unit is required in the existing wastewater treatment system.
3.2 Produced Water Quality
Produced water meets the wastewater quality standards of PermenLH No. 19 of 2010
regarding wastewater quality standards for oil, gas and thermal power businesses
and activities. The processing load of producing water into class III raw water is
more easier and more possible to do. Table 1 shows the parameters of produced
water that compared to the effluent quality standard. Produced water at Lindai GS
can be reused to watering plant. However, there is a need for further processing of
produced water so that its quality can be accepted following applicable regulations.
The quality standard reference used is PP. 82 of 2001, specifically on third-class
clean water quality. Third-class water quality can be used as water for freshwater
fish cultivation, animal husbandry, watering plants, and similar uses [4]. It can be
seen in Table 1 that the parameters of concern are COD and oil and grease. Nutrient
parameters in the form of NH3 -N are not a problem in water production. Usually,
nutrients parameters are the main factors in the pollution flow in streams [14, 15].
Table 1 Comparison of produced water quality in effluent quality standards

Parameters Example Maximum produce Average produce Standard PP No. 82
water quality water quality 2001 Class III
COD mg/L 287.8(*) 61.29(*) 50
Oil and Grease mg/L 16.28(*) 5.9(*) 1
TDS mg/L 989 725.87 1000
pH - 8.8 8.04 6–9
NH3 -N mg/L 2.09 0.78 Not required
Phenol total mg/L <0.001 <0.001 0.001
H2 S mg/L <0.002 <0.002 0.002
(*) The parameter being concern because the value that exceeding the quality standard
3.3 Decision Analysis
The produced water treatment unit’s design aims to obtain water quality following
class III water quality which then to be used as a new water source for watering
plants. Parameters that exceed the quality standard are COD and oil and grease, so
the processing process focuses on setting these parameters aside. In the process of
designing a produced water treatment unit, there are three alternative treatment series.
According to Kundu and Mishra, the stages of produced water treatment consist of
primary processing and secondary processing [12]. In water processing production
at PHE Kampar Siak, primary processing is used to remove oil & grease. Secondary
processing is used to remove COD content to comply with quality standards. The
alternative manufacture for the series of produced water treatment units at Lindai GS
has differences in primary processing, namely oil and grease removal technology.
This is because removing oil and grease is crucial process and can affect the subse-
quent processing unit. The technology to be designed to remove oil and grease must
be considered carefully. The most effective and efficient processing unit is obtained
so that the oil and grease content will not interfere with the following unit’s perfor-
mance. Three alternative series of produced water treatment units with variations in
primary treatment are shown in Fig. 1.
The selection of alternative processing units is carried out using the Analytical
Hierarchical Process (AHP) method. The AHP method is quite commonly used to
determine the best wastewater treatment unit. There are five criteria for selecting alter-
native production water treatment units. The evaluation criteria include oil and grease
removal efficiency, COD removal efficiency, operation and maintenance (O&M),
land requirements, and the initial cost of processing unit materials. According to PP,
if the produced water is will be used as plant water, the produced water must meet
the Class III water quality standard. 82 of 2001, which regulates the parameters of
oil, grease, and COD. Therefore, the efficiency of removing oil, grease, and COD
is used as a criterion for consideration because it has not met the specified quality
standards. It is hoped that the processing unit will set aside these two parameters.
Operation and maintenance (O&M) is directly proportional to the human resources
required. The frequency of cleaning and maintaining the unit and the costs to pay
Fig. 1 Schematic diagram of produced wastewater treatment

for staff and experts can be reduced. Land requirements are considered so that the
unit wastewater treatment plant can be built in a predetermined vacant area and mini-
mize construction costs. Cost criteria also need to be considered so that costs can be
reduced. The smaller the unit material cost, the more profitable the company will be.
The efficiency criteria for removing oil and grease were compared with the
removal efficiency of COD. The efficiency of removing oil and grease is consid-
ered very important compared to removing COD so that it is given a value of 7.
COD concentration can be removed in the subsequent processing unit, namely ultra-
filtration membrane technology. Membrane technology can only remove emulsified
oil, because free-oil type can damage membranes and interfere with the processing
process. Therefore, the concentration of oil and grease needs to be released with as
best as possible to not interfere with the membrane process and cause clogging. Based
on these reasons, the comparison value of COD efficiency criteria to oil and grease
efficiency is the opposite of 7, 1/7. Oil and grease removal efficiency criteria were
compared with O&M criteria. The efficiency of oil and grease removal is more impor-
tant than operation and maintenance. It has a score of 5. the efficiency of removing
oil and grease affects the unit’s effluent quality. If the removal’s efficiency is small,
the oil and grease concentrations cannot meet the class III water quality standards,
so that the produced water cannot be used as plant flush. On the other hand, O&M
can still be overcome by assigning experts to handle it if maintenance and operations
are difficult or quite complicated. The value for O&M criteria compared to oil and
grease efficiency is the opposite of 5, which is 1/5. The criteria for the efficiency of
oil and grease removal compared to land requirements criteria. Compared to land
requirements, the removal efficiency of oil and grease is more critical so that the
value is 5. The removal efficiency of oil and grease is mandatory because the main
function of the unit is intended to set aside oil and grease. The available vacant land
at the Lindai GS location is still sufficient to build a processing unit with a large land
area. Based on this, the land requirement criteria’ comparison value against the oil
and grease removal efficiency criteria is 1/5.
The criteria for oil and grease removal efficiency are compared with unit material
costs criteria. The costs incurred to build a treatment unit will be replaced by the cost
of using raw water from produced water. Thus, compared with the cost criteria, oil
and grease removal efficiency is critical, because the concentration of oil and grease
have to meet class III water quality standard, which limited to a maximum of 1 mg/L.
The treatment unit selected must ensure that it can produce effluent less than 1 mg/L.
Therefore, the comparison value of the oil and grease efficiency criteria against the
unit material cost is given a value of 7. The ratio of the cost criteria to oil and grease
efficiency is the opposite, which is 1/7. The COD removal efficiency criteria were
compared with the O&M criteria. COD efficiency does affect the quality of produced
water effluent to be used as plant flush. Still, COD concentration is not a priority in
selecting the primary processing unit because it can be removed in the subsequent
processing unit. The comparison value of COD removal efficiency criteria against
O&M criteria is given a value, 1/5. The O&M criterion is crucial compared to COD
efficiency so that the comparison value of O&M to COD efficiency is 5.
Efficiency criteria for COD removal compared to the requirements for land
requirements the criteria for the land requirement are moderate or slightly more crit-
ical than COD removal efficiency requirements. COD removal will be focused on
the subsequent processing unit, namely the ultrafiltration membrane. Ultrafiltration
membrane technology is well-known as a technology with high removal efficiency,
so the value of the COD removal efficiency criterion against the land requirement
criteria is 1/3. The COD removal efficiency criteria were compared with the unit
material cost criteria. Compared to the cost criteria, the efficiency of COD removal
is assigned a score of 3 because it is slightly more critical. This is because the company
prioritizes the quality of produced water to be used later than the processing unit’s
cost. Besides, the water used for watering the plants will be replaced with water from
this processing to make savings in the future. Therefore, the comparison value of cost
criteria to COD removal efficiency is 1/3. O&M criteria compared to land require-
ments. Operation and maintenance are slightly more important than land because
operation and maintenance activities are carried out continuously as long as the
processing unit is used. The comparison value of the O&M criteria to land require-
ments is 3. The comparison value of the land requirement criteria against O&M is
1/3. O&M criteria compared to unit material costs. Cost is not a problem for the
company. Still, the availability of human resources is quite limited. The company
needs to increase workers’ time or recruit operators or experts to operate and main-
tain the processing unit. Therefore, the O&M criterion was slightly more critical than
the cost criterion and got a score of Criteria for land requirements compared to unit
material costs. The criteria for land needs are considered to be slightly more critical
than unit costs and get a value of 3. This is because the large land area will also affect
costs, such as clearing the area. Therefore, a unit that requires an area that is not
too large will save on company costs. The comparison value between the material
cost criteria and the land requirement criteria is 1/3. After conducting a comparative
assessment, calculations are carried out according to the AHP method to determine
the eigenvector value. The eigenvector value is directly proportional to the priority
level of the criteria. Eigenvector values for each measure are listed in Table 2. The
comparison assessment of each alternative with each criterion is based on the criteria
data for each alternative listed in Table 3.
The assessment of the comparison of each alternative with each criterion is based
on the criteria data for each alternative listed in Table 3. The selected processing
alternative is the processing unit with the highest weight value than other alternatives,
Table 2 Eigen vector each

No Criteria Eigenvector
criteria
1 Efficiency of removal of oil and grease 0.54
2 COD removal efficiency 0.04
3 Operations and maintenance (O&M) 0.21
4 Land requirements 0.13
5 Material cost unit 0.08
Table 3 Alternative calculation results against criteria

Alternative Criteria Total
Efficiency of COD Operations Land Material cost
removal of oil removal and requirements unit
and grease efficiency maintenance
(O&M)
API 0.07 0.09 0.65 0.07 0.08 0.19
CPI 0.24 0.22 0.23 0.77 0.19 0.51
PPI 0.69 0.69 0.12 0.16 0.72 0.30
Fig. 2 Schematic diagram of the mass balance of selected produced wastewater treatment
namely CPI, with a weight of 0.51 or 51%. Thus, based on the AHP analysis results,
a CPI unit will be built at the Lindai GS location to treat produced water into plant
water. After processing in the CPI unit, the produced water will be reprocessed using
ultrafiltration membrane technology (Fig. 2).
4 Conclusion
Parameters that must be eliminated in water reclamation are COD along with high oil
and grease. A process hierarchy analysis for a suitable treatment system is prescribed
consisting of a CPI unit, UF, and reservoirs. Reclamation of produced water from
upstream activities in the oil and gas industry is carried out using a CPI unit and
Ultrafiltration that selected based on the Analytical Hierarchy Process (AHP) method
to eliminate high COD and oil and grease content from the water.
References
1. Hayes (2010) Produced water management: challenges and solutions. E&P Center Gas
Technology Institute.
2. Astono W, Suswantoro E (2016) The analysis of oil and grease distribution and the evaluation
of produced water process unit on Kf star energy platform (Kakap) LTD. Indonesian J Urban
Environ Technol 6(5):123–131
3. Tiana AN (2015) Air Terproduksi: Karakteristik dan Dampaknya Terhadap Lingkungan.
Fakultas Teknologi Industri, Institut Teknologi Bandung 1–11
4. Ivory D (2015) The prospect of utilization produced water. Teknik Kimia Institut Teknologi
Bandung, Bandung
5. Arthur JD, Langhus BG, Patel C (2005) Technical summary of oil and gas produced water
treatment technologies. Gulf Professional Publishing, Tulsa
6. Forero JE, Ortíz OP, Nariño FA, Díaz J, Peña H (2008) Design and development of a high
efficiency tank for crude oil dehydration (I), CT&F-Ciencia. Tecnología y Futuro 3(4):185–199
7. Arnold K, Stewart M (2011) Produced water treatment field manual (1st edn.) Gulf Professional
Publishing
8. Tellez GT, Nirmalakhandan N, Gardea-Torresdey JL (2005) Kinetic evaluation of a field-scale
activated sludge system for removing petroleum hydrocarbons from oilfield-produced water.
Environ Prog 24(1):96–104
9. Camarillo MK, Stringfellow WT (2018) Biological treatment of oil and gas produced water: a
review and meta-analysis. Clean Technol Environ Policy 20(6):1127–1146
10. Ramadan HS, Moersidik SS, Priadi CR (2014) Selection of wastewater treatment plant effluent
recycling unit at Muara Karang steam power plant with analytical hierarchy process method.
Fakultas Teknik, Universitas Indonesia, Jakarta
11. Saaty TL (2008) Decision making with the analytic hierarchy process. Int J Serv Sci 1(1):83–98
12. Suryawan IWK, Helmy Q, Notodarmojo S (2020) Laboratory scale ozone-based post-
treatment from textile wastewater treatment plant effluent for water reuse. J Phys Conf Ser
1456(1):012002 (IOP Publishing)
with moving bed biofilm reactor in removal of Remazol Black 5. Int J Technol 12(2)
14. Apritama MR, Suryawan I, Afifah AS, Septiariva IY (2020) Phytoremediation of effluent textile
wwtp for NH3-N and Cu reduction using pistia stratiotes. Plant Arch 20(1):2384–2388
ozone pretreatment. Walailak J Sci Technol 18(6):1–11
Preference of Sludge Treatment Plan
in IPA II Pejompongan Water Treatment
Plant
Nailatul Fadhilah, Nurulbaiti Listyendah Zahra, Fatimah Dinan Qonitan,

Imroatus Sholikhah, Intan Rahmalia, Iva Yenis Septiariva,
and I. Wayan Koko Suryawan
Abstract Water treatment is inseparable from the processing of sludge resulting

from the deposition process. IPA II Pejompongan currently has no sludge treatment,
where the treated sludge is discharged directly into the surrounding water bodies. This
research was conducted to determine the suitable sludge treatment system for existing
conditions. This study was conducted using data on the quantity of sludge and the
quality of the resulting sludge, which was then discussed descriptively. This study also
uses a literature review to find various appropriate treatments for the Pejompongan II
IPA. The alternative is selected with the Technique for Order Preference by Similarity
to Ideal Solution (TOPSIS) method. The highest mud generation in 2019 is during the
rainy season, namely January-June. In terms of quality, the parameters of suspended
solids, iron, nickel, lead, manganese, organic (KMnO4 ), BOD, and COD do not meet
Jakarta City’s quality standards government. There are two alternatives: the belt filter
press and the screw press, where the decision analysis results show a better screw
press.
Keywords Sludge treatment · Water treatment · Decision analysis
1 Background
The sludge from the IPA II Pejompongan, the water treatment plant, is currently
directly discharged into surface water, namely the Krukut River, without any treat-
ment process. Discharge of mud into rivers can damage the river’s aesthetics and
endanger the aquatic biota in it. The release of sludge into rivers can increase water
bodies’ turbidity, accumulate sludge, and ultimately lead to rivers’ silting. Increased
turbidity in water bodies will interfere with sunlight’s penetration, which will then
N. Fadhilah · N. L. Zahra · F. D. Qonitan · I. Sholikhah · I. Rahmalia · I. W. K. Suryawan (B)

I. Y. Septiariva
Department of Civil Engineering, Faculty of Engineering, Universitas Sebelas Maret,
Surakarta 57126, Indonesia
1020 N. Fadhilah et al.
block the photosynthesis process and disrupt aquatic biota’s life [1]. Research
conducted by Pradel et al., states that direct disposal of sludge into water bodies
has the potential to cause eutrophication due to the nitrogen content in the sludge
[2]. The Krukut River is a river of concern to the DKI Jakarta Provincial Govern-
ment because the current depth of the Krukut River is less than 3 m, while the ideal
depth for the river is around 8 m. From this explanation, it can be stated that IPA
II Pejompongan has an urgency in building a sludge treatment plant to comply with
applicable regulations and maintain environmental quality so as not to be disturbed.
Preventing environmental pollution can be done by adequately treating waste by
considering various existing technologies even to waste to energy [3]. The benefit of
the study of alternative design is to provide input to IPA II Pejompongan regarding
the sludge treatment plant, superannuation management, and solid cake that can be
carried out in the future. Suppose the design of the sludge treatment plant is applied.
In that case, it can comply with Government Regulation Number 16 of 2005, which
states that the water treatment plant’s residue must be treated first before being
discharged into the environment. The purpose of this study was to determine the
existing conditions of sludge treatment at IPA II Pejompongan and to determine the
appropriate alternative for sludge treatment based on the characteristics of the sludge
in IPA II Pejompongan. By carrying out a study on selecting the best alternative for
sludge treatment in IPA II Pejompongan, it is hoped that it will become the basis
of reference in determining the sludge classification unit to then design the sludge
treatment plant.
2 Method
2.1 Location and Sampling Period
This study was conducted at IPA II Pejompongan, located in Central Jakarta. IPA II
Pejompongan can implement the drinking water supply system in the western part of
DKI Jakarta. The data used in this study is secondary data during January–December
2019.
2.2 Data Collection
In the design of the sludge treatment plant, data on the quantity of sludge is needed
to determine the sludge treatment plant’s design capacity. The sludge quantity data
comes from the sludge discharge calculation at the sludge discharge valve. The sludge
discharge value will depend on the interval and duration of the sludge discharge valve
opening. Determination of the density value of mud and sludge solids content was
Preference of Sludge Treatment Plan in IPA II Pejompongan Water … 1021
carried out by sampling in the field and continued testing at the IPA II Pejompongan
Science Laboratory.
2.3 Selection of Alternative Processing
Determination of design alternatives is carried out using the Technique for Order
Preference by Similarity to Ideal Solution (TOPSIS) method. The TOPSIS method
can determine the best alternative from several alternative solutions with various
criteria [4]. The selection of alternatives to the TOPSIS method is carried out using
the Euclidean distance associated with the Pythagorean Theorem. Determining the
ideal solution matrix is carried out to determine each alternative’s distance with a
positive ideal solution (Si+) and a negative ideal solution (Si−). The following are
the equations used to determine positive and negative ideal solutions:

Si+ = j = n1(vij − vj+)2 1/2 (1)

Si− = j = n1(vij−vj−)2 1/2 (2)
With: Si+ = distance to positive ideal solution; Si− = distance to negative ideal
solution; vj+ = best weighted normalized data for each criterion; vj− = worst
weighted normalized data for each criterion.
The preference value states the alternative distance to the positive and negative
ideal solution. The greater the preference value, the smaller the distance to the positive
ideal solution and the further away distance to the negative ideal solution. The value
of greatest preference will be the chosen alternative. The following is the equation
in determining the preference value:
P = (Si− /(Si− + Si+) (3)
With P = preference value; Si+ = positive ideal solution; Si− = negative ideal
solution.
3.1 Sludge Generation
The sludge produced from IPA II Pejompongan comes from the pre-sedimentation
unit, pulsator, and backwash filter processes. The sludge that will be treated in the
sludge treatment plant is only the sludge from the pre-sedimentation unit and the
Sludge Flow Rate (m3/day) 5.E+06 Primary Settling Sludge flow rate (m³/day)
Pulsator Sludge flow rate (m³/day)
4.E+06 Total Sludge flow rate (m³/day)
3.E+06
2.E+06
1.E+06
0.E+00
1 3 5 7 9 11
Time (Month)
Fig. 1 Sludge flowrate produced at IPA II Pejompongan from January (1)–December (12) 2019
pulsator, while the sludge from the backwash filter process will be directly flowed to
the pre-sedimentation unit to be reprocessed. The sludge from the pre-sedimentation
unit and pulsator will flow to the sludge drain through the sludge discharge valve.
The longer the duration of sludge disposal, the greater the daily sludge discharge
produced. Interval, period, and monthly average sludge discharge data through the
sludge discharge valve are shown in Fig. 1.
From Fig. 1, the most significant sludge discharge comes from the pulsator unit
because the most significant allowance for suspended solids is in the pulsator unit.
In January–June, the resulting mudflow is more important than in July–December.
This was due to an increase in suspended solids in January–June. The increase in
suspended solids was caused by the rainy season during January–June. Rainwater
carries a high content of suspended solids due to soil erosion. Thus, there is an
increase in turbidity during the rainy season [5].
3.2 Sludge Quality
The sludge sampling method is area-integrated, that is, by combining the samples
taken from the pre-sedimentation valve and the pulsator. This sampling is carried
out to determine the samples’ physical quality that enter the processing plant and
calculate the dimensions of the sludge treatment plant. The quality of the combined
sludge of the pre-sedimentation unit and the pulsator, testing the sludge’s quality is
also carried out in each unit. The sludge sampling method is also carried out by using
the area-integrated sampling method, which is by combining the samples taken from
each MPA pre-sedimentation unit or pulsator. Sampling was carried out to calculate
the mass balance of the resulting sludge. Sludge density data, sludge solids content,
and contribution to total sludge are shown in Table 1.
The sludge tested came from the sludge drainage channel, a combined channel
between the sludge from the pre-sedimentation unit and the pulsator. Routine sludge
quality testing is carried out to compare the quality of the IPA II Pejompongan
sludge with the quality standards stipulated in the Regulation of the Governor of
Table 1 Characteristics of sludge produced in the primary settling unit and pulsator at IPA II
Pejompongan
Parameters Primary settling Pulsator Combined
Density of sludge (kg/m3 ) 1004 1013 1011
Solid level 0.30% 2.42% 1.92%
Contribution to total sludge at IPA II Pejompongan 24% 76% 100%
the Special Capital Region of Jakarta Number 69 of 2013. The average value of the
results of testing the IPA II Pejompongan sludge parameters is shown in Table 2.
Several parameters exceed the quality standards of the Regulation of the Governor
of the Special Capital Region of Jakarta Province Number 69 of 2013 as effluent
standards, namely parameters of suspended solids, iron, nickel, lead, manganese,
organic (KMnO4), BOD, and COD. The value of suspended solids (TSS) in the
sludge exceeds the quality standard because the pre-sedimentation and sedimentation
processes aim to remove TSS from raw water. Thus, the resulting sludge will contain
high TSS. The high TSS content contributes to increasing the BOD and COD content
in waste [6]. The high levels of iron, nickel, lead, and manganese in the sludge are
caused by the raw water’s quality containing these metals. The metal contained in
sludge is generally in the form of metal hydroxide (MOH), insoluble in water. In
the process of clarification, these metals will experience precipitation and residue
together with the floc. Thus, the resulting sludge contains high metal content. Based
on Table 1, the value of sludge-specific gravity produced by IPA II Pejompongan is
greater than the value of water-specific gravity. Thus, the separation of solids and
water in the sludge by gravity can be carried out. Based on Table 2. The comparison
of BOD and COD values in the sludge is very low, namely 0.135. The value of the
BOD/COD ratio for processing with a biological treatment unit more than 0.31 [7–
10]. Thus, the sludge produced by IPA II Pejompongan cannot be treated biologically
and cannot be discharged to a wastewater treatment plant that uses a biological
treatment unit.
3.3 Existing Mass Balance
The sludge volume balance describes the volume flow of sludge into and out of the
treatment unit in units of m3 /day. Solid mass balance states the amount of solid that
enters and leaves a processing unit in units of kg/day. The mass flow in the pre-
sedimentation unit and pulsator represents the number of solids suspended in the
sludge flow. The sludge mass and volume flow diagrams are prepared to determine
which units produce the most significant sludge volume. The flow of the average
volume of existing sludge at IPA II Pejompongan is shown in Fig. 2.
In general, the sludge treatment plant consists of collecting, thickening, condi-
tioning, dewatering, and final disposal units. The thickening unit for all alternatives
Table 2 Comparison of sludge quality with standards in Jakarta city

No Parameters Unit Standard* Concentration (Average)
1 Temperature °C 38 25.933
2 TDS mg/L 1000 186
3 TSS mg/L 100 8337.5*
4 Hg mg/L 0.02 0.003
5 NH3 mg/L 5 0.242
6 As mg/L 0.1 0
7 Fe mg/L 5 154.71*
8 F− mg/L 2 0.315
9 Cd mg/L 0.05 <0.01
10 Free chlorine mg/L 1 0.142
11 Chrome (total) mg/L 0.5 0.075
12 Hexavalent chrome mg/L 0.1 <0.003
13 Ni mg/L 0.1 0.158*
14 Nitrate mg/L 10 8.072
15 Nitrite mg/L 1 0.0425
16 pH – 6–9 7.033
17 Zn mg/L 2 6.475*
18 Sulfide mg/L 0.05 0.0097
19 Copper mg/L 1 0.35
20 Pb mg/L 0.1 0.457*
21 Mn mg/L 2 25.857*
22 Phenol mg/L 0.5 0.0132
23 Oil and Grace mg/L 5 1.58
24 Methylene blue active compound mg/L 1 0.13
26 Organic (KMnO4 ) mg/L 85 241.77*
27 BOD5 mg/L 75 92.417*
28 COD mg/L 100 682.5*
* Regulation of the Governor of the special capital region of Jakarta province Number 69 of 2013
Fig. 2 Mass balance of

sludge production from
existing water treatment at
IPA II Pejompongan
Fig. 3 Schematic diagram of alternative sludge processing with belt filter press and screw press
is gravity thickening because the sludge’s quality is still possible to use this unit.
Thus, alum sludge is easier to settle with a gravity thickening unit than to float
with a dissolved air flotation thickening unit. Besides, mechanical thickening, such
as gravity belt thickening, is rarely used in water treatment plant sludge because it
requires much energy and chemicals to treat sludge in water treatment plants with
large discharges. Conditioning is necessary when using a mechanical dewatering
treatment unit to increase the compressibility of the sludge. The type of conditioning
commonly used in water treatment plants is chemical conditioning with polyelec-
trolytes or polymers because it does not produce additional sludge. The use of inor-
ganic chemical conditioning reagents produces additional sludge, namely metal hi-
hydroxides and sludge from lime [11]. Dewatering is the final process before the
waste is discharged into the environment. The dewatering unit consists of mechan-
ical and non-mechanical units. The mechanical units commonly used are the belt
filter press, centrifuge, screw press, and pressure filter press. In selection alternative
sludge treatment plants, the parameters that must be met are land area and elec-
trical energy requirements. The available land area is 1300 m2 and the maximum
electrical energy that can be used for a sludge treatment plant is 100 kW. Based
on these four mechanical units, the units that require more than 100 kW of energy
are the centrifugal and pressure filter press unit, so that the alternative options to be
considered are the belt filter press and screw press (Fig. 3).
The land area criterion has a weight value of 1 because the land area will not
be related to future operations and costs. In the land area criteria, the smaller land
area required by an alternative, the closer the alternative is to a positive ideal solu-
tion. The alternative screw press requires a smaller space than the alternative belt
filter press. The TSS value will affect compliance with quality standards. Based on
the Regulation of the Governor of the Special Capital Region of Jakarta Number
69 of 2013 concerning Wastewater Quality Standards for Activities and/or Busi-
nesses, the maximum total suspended solids that can be discharged into water bodies
is 100 mg/L. Supernatant TSS values obtained for alternatives 1 and 3 have met
the quality standard. The smaller the supernatant TSS value of an alternative, the
closer an option is to the positive ideal solution. Alternative belt filter press produces
smaller supernatant TSS than screw press. The belt filter press unit requires regular
maintenance, especially for belt cleaning. The belt filter press unit has an automatic
belt cleaning system using water with a minimum pressure of 6 kg/cm2 [12]. The
belt on the belt filter press unit needs to be replaced periodically [13]. In the screw
press unit, the fixed and moving rings’ gap has an automatic cleaning system using a
nozzle located above the ring. Things that need to be considered in cleaning are the
interval and duration of cleaning [14]. The easier the treatment of an alternative, the
closer the alternative is to a positive ideal solution.
The alternative screw press is easier to maintain than alternative belt filter press.
The potential odor and noise criteria have a weighted value of 1 because they will
not be related to future operations and costs. The belt filter press unit’s noise level
without a cover is ±80 dB [15]. Meanwhile, the screw press unit’s noise level without
a cover is ±70 dB [16]. The belt filter press can have high odor and noise because it is
designed to be open [12]. Alternative screw press has a lower odor and noise potential
than alternative belt filter press. From the criteria data for each alternative and the
weighted value of each criterion, the normalized and weighted data are determined
using Eqs. 1 and 2. Normalized and weighted data are shown in Table 3. The value of
V+ represents the best value of normalized and weighted data, while V− represents
the worst value of normalized and weighted data. From the weighted normalization
data, then a positive ideal solution (Si+) and a negative ideal solution (Si−) are
determined using Eqs. 1 and 2. A preference value (Pi) is selected, which states the
proximity of the alternative to ideal conditions using Eq. 3.
The alternative with the highest preference value (Pi) will be the best design
alternative applied in IPA II Pejompongan is screw press. Screw press was chosen
because it is closer to the positive ideal solution than the belt filter press. In addition,
screw press is also further away from negative ideal solution than belt filter press.
With the sludge treatment at IPA II Pejempongan, a continuous process is needed
to achieve zero waste. Although the screw press requires higher energy, the energy
recovery of the mud can help supply energy to the IPA II Pejempongan. Determination
Table 3 Comparison of sludge quality with standards in Jakarta city

Criteria Belt filter press Screw press (V+) (V−)
Land requirements 0.0322 0.0320 0.0320 0.0322
Dryness cake solid 0.0650 0.0636 0.0650 0.0636
Volume cake solid (m3 /day) 0.0955 0.0973 0.0955 0.0973
Massa cake solid (kg/day) 0.0645 0.0641 0.0641 0.0645
TSS supernatant 0.0169 0.0422 0.0169 0.0422
Energy requirements 0.1237 0.0575 0.0575 0.1237
Chemical requirements 0.1169 0.0702 0.0702 0.1169
Operation and maintenance 0.0407 0.0813 0.0813 0.0407
Order impact 0.0407 0.0203 0.0203 0.0407
Noise 0.0407 0.0203 0.0203 0.0407
Budget 0.0410 0.1301 0.0410 0.1301
Si+ 0.0951 0.0926
Si− 0.0926 0.0951
Pi 0.4935 0.5065
Ranking 2 1
of the energy potential of solid waste can be done by knowing the application that
will be applied in the field [16–18].
4 Conclusion
The best alternative selection analysis results show that the screw press can be used as
sludge treatment in IPA II Pejompongan. Before using the screw press, pretreatment
in the form of gravity thickening and chemical conditioning is required.
References
1. Bilotta GS, Brazier RE (2008) Understanding the influence of suspended solids on water quality
and aquatic biota. Water Res 42(12):2849–2861
2. Pradel M, Reverdy AL, Richard M, Chabat L (2014) Environmental impacts of sewage sludge
treatment and disposal routes: a life cycle assessment perspective. In: 4th European conference
on sludge management, vol 8
3. Afifah AS, Suryawan IWK, Sarwono A (2020) Microalgae production using photo-bioreactor
with intermittent aeration for municipal wastewater substrate and nutrient removal. Commun
Sci Technol 5(2):107–111
4. Roszkowska E (2011) Multi-criteria decision-making models by applying the TOPSIS method
to crisp and interval data. Multiple Criteria Decis Mak/Univ Econ Katowice 6(1):200–230
5. Saputra DGTB, Arthana IW, Pratiwi MA (2016) Analisis Kualitas Fisika Perairan Berdasarkan
Nilai Padatan Tersuspensi dan Kekeruhan Perairan di Bendungan Telaga Tunjung Desa Timpag,
Kecamatan Kerambitan, Kabupaten Tabanan-Bali. ECOTROPHIC 10(2):130–136
6. Aygun A, Nas B, Berktay A (2008) Influence of high organic loading rates on COD removal
and sludge production in moving bed biofilm reactor. Environ Eng Sci 25(9):1311–1316
7. Wang XJ, Xia SQ, Chen L, Zhao JF, Renault NJ, Chovelon JM (2006) Nutrients removal
from municipal wastewater by chemical precipitation in a moving bed biofilm reactor. Process
Biochem 41(4):824–828
8. Suryawan I, Koko W, Rahman A, Lim JW, Helmy Q (2021) Environmental impact of municipal
wastewater management based on analysis of life cycle assessment in Denpasar City. Desalin.
Water Treat. 244:55–62
with moving bed biofilm reactor in removal of Remazol Black 5. Int J Technol 12(2)
ozone pretreatment. Walailak J Sci Technol (WJST) 18(6)
11. Turovskiy IS, Mathai PK (2006) Wastewater sludge processing. Wiley
12. Andreoli CV, von Sperling M, Fernandes F (2007) Sludge treatment and disposal. IWA
Publishing
13. EPA U (2000) Biosolids technology fact sheet–belt filter press. EPA, Washington D.C.
14. Mumbi AW, Fengting L, Mwarania F, Uuganchimeg B (2017) An assessment of multiplate
screw press in Dewatering process of sludge treatment. Int J Adv Res 740–747
15. Aventura (2011) Drum thickener belt-type filter press. Aventura, New Delhi
16. Macomber S (2016) Huber’s New Q-Press. Huber, Huntersville
17. Suryawan IWK, Sarwono A, Septiariva IY, Lee CH (2021) Evaluating marine debris trends
and the potential of incineration in the context of the COVID-19 Pandemic in Southern Bali,
Indonesia. Jurnal Ilmiah Perikanan dan Kelautan 13(2)
18. Qonitan FD, Suryawan IWK, Rahman A (2021) Overview of municipal solid waste generation
and energy utilization potential in major cities of Indonesia. J Phys Conf Ser 1858
19. Sarwono A, Septiariva IY, Qonitan FD, Zahra NL, Sari NK, Fauziah EN, Ummatin KK, Amoa
Q, Faria N, Lim JW, Suryawan IWK (2021) Municipal solid waste treatment for energy recovery
through thermal waste-to-energy in Depok City, Indonesia. J Adv Res Fluid Mech Thermal
Sci 85
Design of Typical Rainwater Harvesting
Storage Tanks Based on Housing Type
(Case Study in Indonesia)
Lina Indawati, Setyo Budi Kurniawan , Siti Rozaimah Sheikh Abdullah ,

Abstract Indonesia, a tropical country, has high rainfall intensity and has an oppor-
tunity to harvest rainwater as a daily water supply. However, most Indonesians still
rely on groundwater or tap water as their primary water source. Rainwater harvesting
is a feasible technology to be applied to decrease the dependency on the current water
source. A simple rainwater harvesting using rooftop is considered to be applicable
in a residential area. This study calculated the typical storage tank for rainwater
harvesting based on the house type in Indonesia. Two scenarios were used during
the calculation. In the first scenario, storage was designed to collect water during
rainfall to be used during the rainy season only. In the second scenario, the tank
was designed to collect rainwater to be used for the whole year. Calculation results
showed the number of typical storage tank (∅ 1.8 m and height of 1.9 m) required for
the tiny house type was 1 tank, the medium house type was 1–2 tanks, and the large
house type was 3–6 tanks. In the second scenario, the number of storage tanks for
each type of house was 25 tanks. The second scenario is suggested to be used in areas
where water scarcity is the main issue during the dry season, while the first scenario
is suggested to be applied for a user with no land availability and construction cost
issues.
Keywords Design · Rainwater harvesting · Storage tank
1 Introduction
Some areas in Indonesia still utilize groundwater as their primary water supply [1].
The use of groundwater for a long time can result in the depletion of groundwater,
L. Indawati (B) · R. H. Dananjaya

Department of Civil Engineering, Faculty of Engineering, Universitas Sebelas Maret,
57126 Surakarta, Indonesia
S. B. Kurniawan · S. R. S. Abdullah
Department of Chemical and Process Engineering, Faculty of Engineering and Built Environment,
Universiti Kebangsaan Malaysia, UKM, 43600 Bangi, Selangor, Malaysia
1030 L. Indawati et al.
while it can also cause land subsidence due to the empty aquifer [2]. Meanwhile,
a catchment of rainwater is also reduced due to the development of housing and
industry. The reduced catchment of rainwater may cause the increase of runoff and
the decrease of water infiltration, which contribute to the groundwater recharge [3].
The increase of rainwater runoff is the leading cause of flood in many areas, especially
in the densely populated area in Indonesia.
Reducing the rainwater runoff can become a solution to prevent a flood. Flood
reduction in urban area by rainwater harvesting can reach up to 28.66% [4]. One tech-
nology to utilize rainwater further, while also reducing runoff is rainwater harvesting
using a storage tank [5]. Rainwater harvesting is a technology to catch rainwater,
store it, and utilize it for daily needs [6]. A simple rainwater harvesting technology
involves house roofs, soil surface, or rock as catchment unit, while storage might be
carried out in reservoirs or ponds [7]. The utilization of harvested rainwater might go
to the households need and watering/irrigational purposes to reduce the dependence
on groundwater.
Most of areas in Indonesia have medium to high rainfall intensity, which become
very suitable for applying rainwater harvesting technology. The harvested rainwater
may be used during the dry season, which usually caused drought in several areas of
Indonesia. Rainwater harvesting can be applied in the rainy and dry season. Moreover,
the difference in rainfall intensity can have an impact on the harvested rainwater
[8–10]. Harvesting can be considered effective if the rainfall is over the minimum
threshold to hold reliable water consumption [11–13]. However, rainwater harvesting
is currently not familiarized in Indonesia. Most housing in the residential areas still
rely on groundwater and or tap water supply, while the application of simple rainwater
harvesting using the rooftop is currently very limited. This phenomenon is occurred
due to the limitation of the knowledge, especially in terms of the technical part for
applying rainwater harvesting technology, including the minimum requirement for a
storage tank. Based on the problems as mentioned above, the aim of this paper is to
design a typical rainwater harvesting tank, which focused based on the house type
commonly built in Indonesia. The result of this study is expected to contribute as
a basic understanding to provide a typical tank design for the application of simple
rainwater harvesting in a residential area. The obtained result may also be adapted
outside Indonesia, especially in tropical countries with the similar season and rain
intensity characteristics.
In this study, an approach was carried out using 2 scenarios. The first scenario assumes
that rainwater harvesting was used to fulfil the water supply in the rainy season only.
The second scenario assumes that rainwater harvesting was used to fulfil the water
supply in the rainy season and dry seasons. All data used in this study were collected
from 131 rain station of 33 provinces in Indonesia. This research method started from
Design of Typical Rainwater Harvesting Storage … 1031
literature study, data collection, data processing, and calculation of typical rainwater
storage tanks.
2.1 Literature Study
Literature studies focused on scientific journals, which also sourced from websites
that discuss rainwater harvesting design criteria, the application of rainwater
harvesting technology in tropical and developing countries, rainwater intensity in
Indonesia, and the typical water consumption in Indonesia.
2.2 Data Collection and Analysis
The data used in this research were secondary data such as statistical data on housing
types in Indonesia, rainfall data, and water consumption data. The rainfall data in
Indonesia was obtained from the Meteorology, Climatology, and Geophysics Council
(BMKG). Housing data was obtained from Central Bureau of Statistics (BPS). The
water consumption data was obtained from Local drinking water company (PDAM)
and Minister for Public Works and Human Settlements (PUPR). This study observed
rainfall data in February 2021 and rainfall data started from 1981 to 2010 (the average,
minimum, and maximum). Then, those rainfall data were visualized and mapped with
GIS.
The volume of the storage tank depends on runoff and water consumption. Deter-
mination of the storage tank volume was carried out based on the rainwater precipita-
tion and rainwater as a water supply. Moreover, the period of collection was calculated
based on the ordinarily rainy season in Indonesia. The following equations were used
to calculate the required water balance.
Calculation of the roof area as harvesting area
Floor ar ea
A= (1)
cos(the degr ee o f r oo f )
T otal volume o f water consumption (V )
A= × 1000 (2)
R
Calculation of water consumption
T he total volume o f water consumption (V ) = V a + V b (3)
n × hh × h
Va = (4)
1000
n × hk × k
Vb = (5)
1000
where, n is the number of water users (person), hk is the number of days during
the dry season (day), hh is the number of days during the wet season (day), k is the
water consumption in the dry season (L/person/day), h is the water consumption in the
rainy season (L/person/day), A is the roof area (m2 ), Va is the volume of accumulated
water consumed during the wet season (m3 ), Vb is the volume of accumulated water
consumed during the dry season (m3 ), and R is the precipitation during the rainy
season (mm).
Calculation of runoff flowrate, rainwater harvesting quantity, and spill on tth day
[14].
Runoff flow rate
Qt = Ie f f x Cr x A (6)
Rainwater harvesting quantity
St = St+1 + Qt − Dt, 0 ≤ St+1 ≤ C (7)
Spill on t th day
Spill on t th day = St+1 − C (8)
where, Ieff is the effective daily rainfall intensity–first flush; with the first flush value
of 0.33 mm, St + 1 is the storage volume at the end of the rainy period, St is the storage
value at the beginning of the rainy period, Qt is a runoff from the roof, Dt is the total
demand for water supply in the rainy period, C is the trawling capacity that can be
assumed based on the available size in the market, and Cr is a runoff coefficient [15].
3.1 Typical Houses in Indonesia Based on Floor Area
A typical house in Indonesia can be seen from the floor area of the building. According
to BPS 2010 statistical data, 56.98% of Indonesia’s population owns a house with
a floor area of more than 13 m2 [16]. Most areas in Indonesia, such as Yogyakarta,
classify the types of houses based on the middle to upper economic class. Upper-class
house type has an area of 169–216 m2 . Middle-class house types have a floor area of
73–120 m2 [17]. Other city such as Palembang City has 4 types of houses, namely
type A with a floor area of 72 m2 , type B with a floor area of 65 m2 , type C with a floor
area of 85 m2 and type D with a floor area of 49 m2 [18]. In addition, houses on stilts
in urban areas have 3 types, including small, medium and large types. The tiny type
has a floor size of 30 m2 , the medium type has a medium size of 56 m2 , and the large
type has a size of 359 m2 [7]. In general, minimalist home designs in Indonesia have
a floor area of 30–125 m2 [19]. Based on the type of the floor area, the approximate
roof area will be known. Based on typical house data, a typical rainwater harvesting
storage tank will be classified based on its size. Based on the several obtained data,
the house type was classified into three types, namely tiny, medium, and large types.
A tiny rainwater harvesting storage tank was estimated from a floor area of 30–72
m2 , the medium size rainwater tank was determined from the floor area of 73–120
m2 , while the large size was calculated from the floor area of 169–359 m2 .
3.2 Rainwater Quality in Indonesia
Rainwater quality in 2015 in several areas of Indonesia such as Java, Sulawesi and
Sumatra still met the quality standards (Table 1). However, based on BMKG data
in March 2021, the rainwater quality in most parts of Indonesia tended to be acidic,
with a pH range of 4.11–5.55 [20]. Only in some areas such as Beto Ambari, Bone,
Branti, Gawpalu, Jalaludin, Lasiana, Baai Island, Siantan, Simpang Tiga, Sorong,
Table 1 Rainwater quality in Indonesia

Parameter Unit Rainwater quality (source: 22–25)
Malang Residential Industrial Bagan Siapi Sulawesi Standard
area area (Riau) (minister
(Lampung) (Lampung) of health
regulation
no. 32 of
2017)
pH – 7.4 7.31 6,72 8 5.3 6.5–9.0
Iron mg/L <0.26 0.392 0.157 0.04 – 0.3
Fluoride mg/L 0.34 0.494 0.332 0.003 – 1.5
CaCO3 mg/L 39.6 26 32 15.29 – 500
Mangan mg/L 0.01 11 0.308 0.02 – 0.5
Nitrate mg/L 0.53 0.095 0.359 0.376 0.41 10
Nitrite mg/L <0.001 0.005 0.011 0.0074 – 1
Cyanide mg/L – – – – – 0.1
Zinc mg/L – 4 1.174 – – 15
Organic mg/L 0.9 – – 7.79 – 10
Sulphate mg/L – – – – 1.34 400
Supadio had an ideal pH (5.6 ≤ pH ≤ 6.21). According to the Ministry of Health

No. 32 of 2017, the water pH that can be used for sanitation purposes was around
6.5–8.5 [21]. Therefore, the operation of rainwater harvesting must be considered
to maintain the quality of rainwater. Rainwater for the first 15 min is channeled as
a daily flush so that the rainwater only harvested after the first 15 min of rain to
maintain the rainwater quality.
3.3 Rainfall in Indonesia
Most area in Indonesia has a high peak rainfall, especially in Java, Southeast Nusa,
Sulawesi, Papua, which the maximum rainfall exceed 500 mm/year (Fig. 1a). The
minimum rainfall was around 50–100 mm/year in most area in Indonesia (Fig. 1b).
Moreover, North Sumatra, North Sulawesi, Northern Papua, North Kalimantan, and
a small part of East Kalimantan had a minimum rainfall of less than 50 mm/year.
However, most area in Indonesia was still in the range of moderate category, and
rainwater harvesting technology can still be used in minimum rainfall condition.
The average rainfall in some parts of Indonesia was classified as moderate to high.
(Fig. 2a). In addition to data for a period of 10 years from 1987 to 2010 which was
used to determine rainfall trends in Indonesia, the latest rainfall data was used as a
comparison. In February 2021, most areas in Indonesia had moderate to high rainfall
intensity except Central to North Sumatra which had low rainfall intensity (Fig. 2b).
3.4 Estimated Roof Area
There are three types of roofs based on the angle of inclination. First, the flat roof
has a slope of <10°, second, sloping roofs have an angle of inclination between 10°
and 30°, and the sloping roof has a slope above 30°. Most housings in Indonesia used
the high inclination; thus a 30° was used to calculate the roof area in this study. The
estimated roof area based on determined house type is summarized in Table 2.
3.5 Estimated Rainwater Flowrate for Harvesting
The flow rate of rainwater was estimated from the runoff that flows on the roof, so
that, it can be obtained from processing rainfall data, Cr and roof area. The average
monthly rainfall data in Indonesia is 246 mm/month, with the daily rainfall data is
8.2 mm/day [26]. Effective daily rainfall data was obtained from daily rainfall data
minus 0.33 mm, where this value was assumed as rainwater mixed with pollutants
(a)
(b)
Fig. 1 Rainfall intensity in Indonesia mapped using GIS a maximum rainfall; b minimum rainfall
[26]
(first rainfall till 15 min’ rainfall) which will not be collected [27]. The runoff coef-
ficient for the roof of the house is 75–95% [28]. Therefore, 80% was defined as the
average runoff coefficient. Table 3 shows the calculation of rainwater flow rate.
(a)
(b)
Fig. 2 Rainfall intensity in Indonesia mapped using GIS a average rainfall from 1987 to 2010; b
average rainfall in 2021 [26]
3.6 Analysis of the Volume and Requirements of Rainwater

Storage Tanks
The volume and requirement of rainwater storage tanks were calculated from the
runoff flow rate, the water consumption in the rainy and the dry seasons, and
Table 2 Roof area each type

Type Floor area (m2 ) Roof area (m2 )
of house
Tiny 30 194.49
73 473.25
Medium 73 473.25
120 777.95
Large 169 1095.61
359 2327.37
Table 3 Rainwater flowrate for harvesting

Type Floor area (m2 ) Roof area (m2 ) Ieff (mm/day) Cr Q (m3 /day)
Tiny 30 194.49 7.88 0.8 1.23
72 466.77 7.88 0.8 2.94
Medium 73 473.25 7.88 0.8 2.98
120 777.95 7.88 0.8 4.91
Large 169 1095.61 7.88 0.8 6.91
359 2327.37 7.88 0.8 14.68
the storage volume at the end of the rainy period (St+1 ). The final volume of the
storage period was obtained from the difference between runoff discharge and water
consumption in the rainy season. The results of the calculation (Table 3) show that
the rainwater harvesting flowrate still meets water consumption for each unit of the
house type. The water consumption for each housing unit was estimated from the
number of residents (assumed to be occupied by 5 people), and the water consump-
tion per person per day. The water consumption per person per day was set to 126.9
L/person/day [29–32]. In the first scenario, it was assumed that continuous use and
continuous disposal of excess runoff or meeting the needs of water supply with
rainwater harvesting during rainy season only.
The quantity of rainwater storage tanks was obtained from the volume of the tanks
divided by the size of the tanks with a height of 1.9 m and a diameter of 1.8 m. The
determination of the storage tank size considered with the limited availability of land
in Indonesia. The number of tank requirements for the tiny house type was set to
1 tank, the medium house type was 1–2 tanks, and the large house type was 3–6
tanks (Table 4). The difference in the number of tanks for each type of house that
has different roof sizes was considered to be not much different, but the difference
in the size of the roof of the house as a rainwater catchment area provides a quite
visible difference in the spilled rainwater.
Table 5 shows the number of the storage tank in the second scenario, in which there
was no rainfall for the next 6 months in the dry season. Therefore, the fulfillment of
clean water needs in the dry season was obtained from rainwater harvesting carried
out in 6 months of the rainy season. As a result, the number of rainwater harvesting
Table 4 The number of rainwater harvesting tank for the first scenario
Type Floor area Roof area Q Va St+1 (m3 ) Tank Number
(m2 ) (m2 ) (m3 /day) (m3 /day) volume of tanks
(St) (m3 ) (unit)
Tiny 30 194.49 1.23 0.635 0.59 1.18 1
72 466.77 2.94 0.635 2.31 4.62 1
Medium 73 473.25 2.98 0.635 2.35 4.70 1
120 777.95 4.91 0.635 4.27 8.54 2
Large 169 1095.61 6.91 0.635 6.28 12.55 3
359 2327.37 14.68 0.635 14.04 28.09 6
Table 5 The number of rainwater harvesting tank for the second scenario
Type Floor Roof area Number of Vb Number of Additional
area (m2 ) tanks for the (m3 /6 month) tanks in both tank (unit)
(m2 ) first scenario season (unit)
(unit)
Tiny 30 194.49 1 115.80 25 24
72 466.77 1 115.80 25 24
Medium 73 473.25 1 115.80 25 24
120 777.95 2 115.80 25 23
Large 169 1095.61 3 115.80 25 22
359 2327.37 6 115.80 25 19
storage tanks at the beginning of the rainy season was quite a lot compared to the
first scenario.
3.7 Analysis of Spilled Rainwater
In rainwater harvesting with the provided storage tank capacity, it was impossible
for all runoff to be harvested. Some runoffs need to be discharged due to the limited
storage tank capacity. In the first scenario, the discharged runoff was greater for each
day than in the second scenario because the number of tanks in the second scenario
was larger. The first scenario was recommended if there is no issue related to the
land availability and the cost of constructing a rainwater storage tank. Table 6 shows
the amount of daily spilled rainwater.
In the second scenario, rainwater runoff spilled was less than the first scenario
(Table 7) during the rainy season. It is also able to meet the water consumption
for both seasons. The type of tiny house with a roof area of 194.49 in scenario 2
even has no spillage due to the available volume of the tank. However, it would be
Table 6 Spilled rainwater in the first scenario

Type Floor area Roof area V (m3 ) The number of Spilled Spilled
(m2 ) (m2 ) tank (unit) rainwater rainwater
(m3 /month) (m3 /day)
Tiny 30 194.49 1.18 1 30.95 1.03
72 466.77 4.62 1 133.98 4.47
Medium 73 473.25 4.70 1 136.43 4.55
120 777.95 8.54 2 247.15 8.24
Large 169 1095.61 12.55 3 362.78 12.09
359 2327.37 28.09 6 815.14 27.17
Table 7 Spilled rainwater in the second scenario

Type Floor area Roof area V (m3 ) The number of Spilled Spilled
(m2 ) (m2 ) tank (unit) rainwater rainwater
(m3 /month) (m3 /day)
Tiny 30 194.49 4.99 25 −1.27 −0.04
72 466.77 8.43 25 50.3 1.67
Medium 73 473.25 8.51 25 51.5 1.72
120 777.95 12.35 25 109 3.64
Large 169 1095.61 16.36 25 110.3 3.68
359 2327.37 31.89 25 402.3 13.41
difficult to apply this second scenario if there are issues related to land availability
and construction costs.
3.8 Operational Procedure and Maintenance
Operational procedures and maintenance include preparation of rainwater harvesting,

collecting rainwater and maintenance including the cleaning process, rainwater
quality monitoring and checking the installation. Following are the operation and
maintenance procedures for rainwater storage tanks and rainwater quality:
1. Check if there is a blockage in the inlet or outlet pipe, if any, clean it immediately.
2. Check the roof area every week, if there is dirt, it must be cleaned.
3. Rainwater for the first 15 min is channeled as a daily flush so that the rainwater
only harvested after the first 15 min of rain because rainwater quality after 15
min of rain is less pollutant.
4. Close the faucet when the rain stops, or the tank is full
5. Rainwater storage tanks, inlet, and outlet pipes should be cleaned if there is
sediment, moss or dirt.
6. Check the pH, turbidity, TDS, temperature, taste and odor. If those do not meet
the standard (Minister of Health Regulation No. 32 of 2017), further treatment
will be needed.
4 Conclusion
Rainwater harvesting is an applicable technology with many benefits for a tropical

country like Indonesia. This study shows the number of typical storage tank (∅ 1.8 m
and height of 1.9 m) requirements in the first scenario (rainfall occurred for the whole
year) for the tiny house type was 1 tank, the medium house type was 1–2 tanks, and
the large house type was 3–6 tanks. In the first scenario, rainwater harvesting only
meets water consumption during the rainy season and at the beginning of the dry
season unless the area has rain throughout the year. In the second scenario, the
number of storage tanks for each type of house is quite a lot, namely 25 tanks. The
first scenario is considered to be very suitable for a user with no land availability
and construction costs issues, while the second scenario is suggested to be used in
areas where water scarcity in the dry season is the main issue. Based on the results
of this study, further research is needed regarding the design of rainwater harvesting
systems that pay attention to land use, geography, rainwater quality, and innovation
in rainwater harvesting infrastructure that considers the limited availability of land.
References
1. Yudhistira MH, Sastiono P, Meliyawati M (2020) Sustainable water resources through

harvesting rainwater and the effectiveness of a low-cost water treatment. Water Resour Econ
32(100161):1–15
2. Patunru AA (2015) Access to safe drinking water and sanitation in Indonesia. Asia Pac Pol
Stud 2(2):234–244
3. Krishnaswamy J, Bonell M, Venkatesh B, Purandara BK, Rakesh KN, Lele S, Kiran MC,
Reddy V, Badiger S (2013) The groundwater recharge response and hydrologic services
of tropical humid forest ecosystems to use and reforestation: support for the ‘“infiltration-
evapotranspiration trade-off hypothesis.”’ J Hydrol 49(8):191–209
4. Akter A, Tanim AH, Islam MK (2020) Possibilities of urban flood reduction through distributed-
scale rainwater harvesting. Water Sci Eng 13(2):95–120
5. Helmreich B, Horn H (2009) Opportunities in rainwater harvesting. Desalination J 248(1–
3):118–124
6. Richards S, Rao R, Connelly S, Raj A, Raveendran L, Shirin S, Jamwal P, Helliwell R (2021)
Sustainable water resources through harvesting rainwater and the effectiveness of a low-cost
water treatment. J Environ Manage 286(112223):1–8
7. Angkasa Z (2017) Application of architecture concepts stage house in urban area. Arsir J
1(2):175–183
8. Ali S, Zhang S, Chandio FA (2021) Impacts of rainfall change on stormwater control and water
saving performance of rainwater harvesting systems. J Environ Manage 280:111850
9. Wang R, Dijk SV, Lounsbury AW, Hoekstra AY (2020) Strategic design and finance of rainwater
harvesting to cost-effectively meet large-scale urban water infrastructure needs. Water Res
184:116063
10. Zhang S, Zhang J, Jing X, Wang Y, Wang Y, Yue T (2018) Water saving efficiency and reliability
of rainwater harvesting systems in the context of climate change. J Clean Prod 196:1341–1355
11. Che-Ain AI, Shaari N, Sairi A, Zain MFM, Tahir MM (2009) Rainwater harvesting as an
alternative water supply in the future. Eur J Sci Res 34:132–140
12. Lee KE, Mokhtar M, Hanafiah MM, Halim AA, Badusah J (2016) Impacts of rainfall change
on stormwater control and water saving performance of rainwater harvesting systems. J Clean
Prod 126:218–222
13. Shaaban AJ (2009) Research and development of rainwater harvesting. IRCS 14:1–56
14. Khastagir A, Jayasuriya N (2007) Parameters influencing the selection of an optimal rainwater
tank size: a case study for Melbourne. In: Proceedings of the rain water and urban design
conference 2007, Sydney 21–23 Aug 2007
15. Khastagir A, Jayasuriya N (2010) Optimal sizing of rain water tanks for domestic water
conservation. J Hydrol 381(3–4):181–188
16. Central Bureau of Statistics (BPS) (2010) Indonesian housing statistics. Central Bureau of
Statistics, Jakarta
17. Leo KSD (2020) Preferences of selecting house type and housing locations of middle and upper
communities near Yogyakarta city. Reka Ruang J 2(2):75–85
18. Triyuly W (2013) Residential development identification of house type in the Palembang city.
Sriwijaya Eng J 22(1):1–7
19. Anditya, Afriansyah H (2016) Minimalist home design creative inspiration. Elex Media
Komputindo, Jakarta
20. BMKG (2021) Rainwater quality information. BMKG, Jakarta
21. Minister of Health (2017) Regulation of The Minister of Health of the Republic of Indonesia No
32 of 2017: environmental health quality standards and water health requirements for hygiene
sanitation, swimming pools, solus per aqua, and public baths. The Minister of Health of the
Republic of Indonesia, Jakarta
22. Joleha J, Mulyadi A, Wawan W, Suprayogi I (2019) Assessment of rainwater quality in coastal
area for domestic water supply. Civ Eng Nat Conf 13
23. Environmental Agency of South Sulawesi: Regional Environmental Status. Environ-mental
Agency of South Sulawesi, South Sulawesi (2015)
24. Rahmayanti AE, Soewondo P (2015) Provision of drinking water in the coastal areas of Bandar
Lampung city through rainwater harvesting. Environ Eng J 21(2):115–126
25. Untari T, Kusnadi J (2015) Utilization of rainwater as water consumption in Malang city using
simple filtration modification method. J Food Agroindustry 3(4p):1492–1502
26. BMKG (2021) Rainfall analysis for February 2021 and rainfall forecast for April, May and
June 2021. BMKG, Jakarta
27. Li Z, Boyle F, Reynold A (2010) Rainwater harvesting and greywater treatment systems for
domestic application in Ireland. Desalination 260(1–3):1–8
28. Agustianto DA (2014) Rainfall and runoff relationship model. Civ Environ Eng J 2(2):214–224
29. PDAM (2014) Rispam of PDAM of Gorontalo city. PDAM of Gorontalo City, Gorontalo
30. PUPR Ministry (2010) Urban drinking water technical instructions and manuals. PUPR
Ministry
31. Rustan FR, Sriyani R, Talanipa R (2019) Water consumption analysis of Bumi Graha Asri
residents, Kendari city. Stab J 7(2):151–159
32. Salilama J (2018) Analysis of water consumption (PDAM) in Gorontalo city. J Sci Civ Eng
Technol 6(2):102–114
Risk Analysis of Shared Marine Space
in the View of Traditional Fishermen
Perceptions in the National Tourism
Strategic Area of Lombok, Indonesia
Ida Ayu Oka Suwati Sideman , R. M. Nyoman Budiartha,

Ida Bagus Putu Adnyana, and Ngakan Ketut Acwin Dwijendra
Abstract This research is in Gili Air, North Lombok Regency. This island is one of
the three islands, which is the National Tourism Strategic Area. The dominant liveli-
hoods of residents are traditional fishermen, self-employment and tourism support
activities. It is a natural thing if the port in the region is designed as a tourist port.
Likewise, various risks are reviewed when submitting a plan for operating a tourist
port, but they often do not involve the economic value of fishermen’s failure to catch
fish. This is a gap that is considered important in this study. This study is the result
of preliminary research from the Doctor Program thesis; the purpose of this study is
a risk analysis that places fishermen in the common use of marine space as part of
the risk forming components of a tourist port. The research method used “Cause and
Effect Analysis”, which is a structured method to identify possible causes of risk.
This method classifies the causes of risk into categories to facilitate identification.
This study proves that the causes of risk formation from the low income of fishermen
are water traffic, equipment and catch marketing, respectively. The detailed analysis
shows that the synergy of the organization, schedules and services of the wharf and
anchor grooves are the sub-cause of the risk.
Keywords Risk analysis · Traditional fisherman · Cause and effect analysis ·

Perceptions
I. A. O. S. Sideman (B)
Engineering Science, Faculty of Engineering, Udayana University, Bali, Indonesia
Department of Civil Engineering, University of Mataram, Mataram, NTB, Indonesia
R. M. N. Budiartha · I. B. P. Adnyana
Deparment of Civil Engineering, Faculty of Engineering, Udayana University, Bali, Indonesia
I. B. P. Adnyana
N. K. A. Dwijendra
Engineering Science, Faculty of Engineering, Udayana University, Bali, Indonesia
1044 I. A. O. S. Sideman et al.
1 Introduction
1.1 Research Background
Several countries, including Indonesia, This tourism industry is used by businesses

providing services to visitors as a source of taxes and revenue [1, 2]. Tourism is
defined as a temporary trip, carried out from one place to another to enjoy sightseeing
and recreational trips [3] Unfortunately, there are several weaknesses of tourism
factors such as weak human resources, long distances, lack of transportation access
and less of community creativity [4]. Gili Indah area is one of the Strategic National
Tourism Areas (KSPN) that located in West Nusa Tenggara Province. This area has
3 small islands that use sea transportation as the only access to the island of Lombok.
Sea lanes are used as shared spaces in tourism transportation, fishing, and domestic
movements. Shared space projects have been shown to greatly improve the public’s
view of an intersection, corridor, or locale through the use of enhanced architectural
elements [5]. So that various ways must be done to fix this problem, one of this is
by analyzing sea transportation through the perceptions of local communities such
as fishermen. That is the reason why this research is important to be held.
All ship movements require port services in various classes and completeness. A
port is a place that is deliberately built to be a place for ships to dock. This area is
used as a stopover for ships before finally anchoring or continuing their journey [6].
Through this research, it is hoped that the risks borne by fishermen in the movement
in the common room can be identified. The risk referred to is the risk based on the
fishermen’s perception as one of the actors in the movement. Risk is defined as the
possibility of a potentially detrimental event or an uncertain situation faced which
can have an adverse impact [7] in case of traffic, traffic congestion is a global trend
that is expected to intensify in the future [8]. So that, traffic risks and the solving
problems can be work as a universal recommendation in future.
The characteristic of traditional fishermen is to use a navigation system that
combines nature, feeling, and instinct [9] with the characteristic of moving at a slow
speed and stopping for fishing, while the characteristic of local logistics movement
is fast moving because commodities are easily damaged [10]. On the other hand, the
two movements are moving in an area designated as a tourist movement, where the
characteristic is fast movement with higher speed [11]. Due to the condition above,
it is predicted that there will be problems due to differences in the characteristics of
the movement in the same roar from the beginning, namely from the port.
Many ports in tourism areas are built based on the perception of policy makers
(the government) but these ports are not well utilized, the reason is that they are not
in accordance with the needs of their users so they do not participate in supporting
the port [12]. The most common cause of fear over the operation of an activity is
risk [9]. The views or perceptions of different interests on the risk of an activity are
often different [13].
Risk Analysis of Shared Marine Space … 1045
Based on this, it is necessary to carry out research that examines from the point
of view of fishermen who also receive the impact of the operation of the port in the
tourism area.
The area used by the three movements with different characteristics certainly
creates various risks [14]. The most common study to assess risk is the assessment
of incidence [15]. The data ignores the feeling of one of the important components
involved, namely fishermen. Research on risk is rare by looking at fishermen as a
subject who has a sensitivity to risk assessment. Meanwhile, the sensitivity of the
assessment is expressed in perceptions by the fishermen. A risk assessment relying
solely on the number of occurrences eliminates the opportunity for fishermen to
provide an opinion on future potential.
By involving fishermen’s perceptions, neglect of traditional fishermen in poten-
tial risks can be eliminated. Discrimination in assessment in the formation of risk
components can be more fair and humane, considering that fishermen in this area
have existed before the stipulation of tourism areas is determined.
1.2 Literature Review and Novelty
One of the internal factors causing fishermen’s poverty is the high dependence on
fishing occupations, while one of the external factors is the geographical isolation of
fishing villages which does not support the mobility of goods, services, capital and
people [16]. When it is compared with research in Sambas district, where fishermen
have a risk of unpredictable fishing season which is 41%. This number is relative high
for fishermen’s profit perception of 70% [7]. So, it can be said that fishermen have
high optimism about the life expectancy of the traditional fishing profession. Both
studies can be interpreted as strengthening the need for a research on risk factors, as
fishermen still have a high dependence on their work and the sea. So that a research
that based on fishermen’s economic risk is absolutely necessary to provide the justice
for fishermen. The new of this research is that the study area is designed as a tourism
area, not a fishing area. So that, fishermen’s perceptions will be found towards other
movements that have different characteristics from the movements of fishermen, but
work in the same sea space.
The three main characteristics of standard procedures from a typical port project
evaluation are: diversity of port size and operations, diversity of port destinations
and missions, and private sector participation in the tourism industry. A program is
formulated in the framework of the importance of broad community participation
in the success of the program [17]. Thus, the perception of fishermen as an affected
community is important to research. This research took the point of fishermen’s
perception as the new of this research.
1.3 Preliminary Hypothesis
Economic problems are the main risk for fishermen in the Gili Indah region in the
case of sharing marine space in the tourism area. This economic problem is supported
by ship problems, especially the age of the ship’s engine. Engine condition has the
potential to affect safety and comfort, especially in terms of vehicle speed. Similar to
land vehicles, water vehicles also have speed, where the journey speed is the effective
speed of a vehicle traveling between two places [18, 19].
2 Research Method
2.1 Research Method
Various problems that occurred during the pandemic period caused many things to
be done to continue to carry out research but in the health novel. This stimulates this
research to be held by using the ‘cause and effect’ analysis method.
Cause and Effect Analysis is a structured method to identify possible causes of
a risk [20]. This method classifies the causes of risk into categories to facilitate
identification. This cause information is displayed in a fishbone diagram or Ishikawa
diagram or in a tree form. The basic steps for Cause and Effect Analysis are as follows:
(1) determining the risk or effect to be analyzed, this risk can have a positive or
negative impact depending on the circumstances; (2) Determine the main categories
of risk causes to be the main branches in the diagram; (3) Identify each possible
cause covered in each category to the extent possible. The causes are described as
twigs from each branch; (4) check each category branch to ensure completeness and
consistency. 5. Identify the most likely cause based on opinion and evidence [21].
2.2 Research Design
Samples are the share of the population taken in a certain way and in number smaller
than the population [22]. Then the sample is built from the accident sampling method.
Accidental sampling is sampling accidental by taking the respondent who happened
to be there is somewhere in accordance with the place of research [23].
2.3 Characteristic of Sample and Its Validation
Even though the method used was accident sampling, for data accuracy, samples
were grouped based on characteristics related to their work.
Law of Republic of Indonesia No. 13 of 2003 about manpower, appealed that

children can do some light works as long as they do not interfere with their phys-
ical development, are supervised by adults and work a maximum of 3 h a day [24].
Furthermore, in the same law it is stated also if the limit of children is 15 years.
In Indonesia, the age limit for having various official personal documents such as
identity cards and driving licenses is 17 years, so this research places a minimum of
17 years of workforce as fishermen. Furthermore, this limitation will apply in deter-
mining the eligibility requirements for the sample. Further characteristics analyzed
from the sample were gender, work experience (in years), skills (training, courses,
and other supporting skill) and side jobs. This is to obtain the feasibility of using
samples to represent the fishing population in the Gili Indah area.
As the work experience is not clearly stated, a detail was made on the minister
of manpower regulation No. 12 of 2016 [25] which follows the Indonesian National
Work Competency Standards (SKKNI) where the workforce is determined to be
skilled with a minimum work experience of 2 years.
The reason for including the competency component of fishermen is because the
competency standard is expected to be a benchmark for fishermen’s income, where
fishermen with more competence will get more income [26]. Table 1 shows the
characteristics of the respondent’s data in this study.
Table 1 is a requirement to choose if the respondent’s research methods are to
determine the sample using the accidental sampling method, which is to capture
a sample from a number of populations while doing activities [27]. In accident
sampling which is also referred to as captive sample, the population must have high
uniformity so that the sampling does not produce data bias [28].
From these data, the characteristic of respondents that stands out is the dominance
of the productive age with an almost even distribution from the ages of 22–51 years.
The gender is entirely male, very few have received training and it is concentrated
on training on disasters. Although there are 12.5% of respondents who have a side
job, they have a dependency on marine life. Work experience is dominated by a
work period of 6–10 years. The characteristics of these respondents are valid to
be continued as a research sample because they approach the results of previous
research in selecting the sample model. The conclusion of the sample test, the research
can be continued with the sample.
3.1 Risk Analysis of Shared Marine Space
The aim of any risk assessment tool is to ensure a consistent, knowledge-based

decision-making process that represents the shared understanding of stakeholders.
When the implementation of risk matrices is checked, it becomes evident that there
are some notable weaknesses in this simple tool [29]. Risk evaluation at sea is a critical
Table 1 Characteristics of respondents

Number Description % of 16 respondents Sample test standards
1 Age group 0 Meet the requirements in
17–21 6.25 accordance with law of
22–26 6.25 Republic of Indonesia No. 13
27–31 18.75 of 2003
32–36 31.25
36–41 18.75
42–46 18.75
46–51 0
51–56 0
More than 56
2 Gender 100 Meet the requirements in
Male accordance with Law of
Female Republic of Indonesia No. 13
of 2003
3 Residence 80 Meet the requirements in
Gili Indah 20 research problem boundaries
Pemenang made by researchers based on
references
4 Side job 12.5 (in sea tourism field Meet the requirements in
Yes before pandemic) research problem boundaries
No 87.50 made by researchers based on
references
5 Work experiences 12.5 Meet the requirements in the
0–5 years 62.5 minister of manpower
6–10 years 25 regulation No. 12 of 2016
More than 10 years
and time-consuming process. In order to determine the rules and strategies by which
the team can conduct the risk assessment, the selection of suitable risk assessment
tools is critical [30, 31]. Protection is characterized as the state of safety from the risk
of pain, injury, threat, loss, injury prevention or hazard prevention devices. Security
may also be described as the role of ensuring that accidents are prevented. Protection,
however, cannot be strengthened by looking at the past and taking precautions against
injuries [32]. Three key attributes are required by a resilient system: (1) anticipation:
avoidance of disturbance; (2) perception: avoidance of aggravation of the effects of
disturbance; (3) response: recovery and survival after disturbance [33]. The accident
ratio will be an important issue if the respondent’s perception points to sea accidents
as a risk. The risk ratio is the potential for accidents per incident [34].
Of the three keys, the first key cannot be used because fishermen cannot avoid
the risk of using the shared space. Meanwhile, the third key is the key that can be
used by the government as the policy maker. Thus, it is increasingly important to
conduct research on perceptions because it is a key that can be used by fishermen to
overcome risks.
3.2 The Traditional Fishermen Perceptions
Perception in general is a process of obtaining, interpreting, selecting and organizing

sensory information [33]. Fisherman is a community who lives in a coastal area
with the main livelihood utilizes the natural resources contained in it sea, whether
in the form of fish, shrimp, seaweed, shellfish, coral reefs and other products other
marine wealth [35]. The life of coastal communities is different from aspects of life
in agrarian societies (residents living in rural areas on generally).
This is due to natural environmental factors, due to coastal communities more
related to the dominant sea, whereas agrarian society by natural environment in the
form of rice fields, moor or fields. With that condition different, allowing them to
have a culture and knowledge system different in meeting their daily needs [16].
Small island challenges can involve numerous creative approaches, as they all turn
out to be wicked issues because of size, isolation, geography, tradition and culture,
governance and education [36].
Based on the Table 2, the priority sub case handling is on traffic cooperation
between related components in the movement of the sea space together. The first
handling is on the movement schedule information, especially on tourist movements
that have different movement characteristics from fishermen movements. The move-
ments of these different characteristics can influence each other, such as when a
fishing boat is at the anchor stage, the movement of a tourist boat can be a nuisance,
as well as for tourist boats, the existence of a fishermen’s boat can become an obstacle
to movement.
The next risk treatment is information, especially about sea and weather condi-
tions. Fishermen need up-to-date information as a reference to go to sea. Port services
and synergy between organizations can be corrected by implementing cooperation
between the organizations involved in it. The rating for this matter is that it can be
Table 2 Causes and effect

Impact Categories of Causes Sub causes Fishermen
causes perception
(%)
Equipment Machine Machine life 7
Organizational 13
synergy
Low income Water traffic Traffic Time schedule 19
management Services of the wharf 14
and anchor grooves
Information 25
Market Types of fish Not the kind of tourist 10
taste, so it must be
marketed locally and
to Lombok
Table 3 Risk respond

Level of risk Respond
1–3 Acceptable Sufficient controlled
4–6 Monitored Sufficient controlled
6–9 Control of management Sufficient controlled
10–14 Urgent Acceptable only with
excellent controlled
15–25 Unacceptable Acceptable only with
excellent controlled
Source [19, 37]
done in a category that is not very urgent but requires handling. Meanwhile, the
risk of vessel conditions and cooperation in selling fish is cooperation with outside
parties which can be carried out in a non-urgent status and if it is to be implemented
it requires more detailed research because the score is 7 that due to Table 3 it is on
category of control management. For this category the respond of the risk must be
sufficient controlled.
Verification of the data obtained from fishermen’s perceptions is to use source
triangulation. Triangulation is a validation and verification method that can be relied
on to obtain valid data [38]. One of them is by triangulating data sources; this method
uses other sources to test the data that has been obtained [39]. In this study, data
from fishermen’s perceptions were tested by cross check on the port operator. The
data that is cross checked is schedule data and other information, which according to
fishermen’s perception is the cause of risk. From the cross check results, it was found
that the port operator did not provide services to fishermen, because the schedule was
informed at the port for the benefit of passenger movement. Other related information
such as weather and height, waves are also carried out at the port for the benefit
of passenger transportation operators. Thus it can be denied that the fishermen’s
perception of traffic management risk is correct and acceptable.
Furthermore, to support the results of the research, it resulted in two sub causes
as things that must be managed with excellence; the following table of interests was
created by conducting detailed interviews with fishermen.
Based on the results of in-depth interviews of the schedule and information needed
by fishermen to reduce risk, the data obtained are as shown in Table 4.
Table 4 Forms of risk

Needs Form
handling
Time schedule Cooperation agency/organization
Information Regularly sharing information
Weather
Wave
Wind
Accident
Table 5 Fishermen’s
Fishermen’s perceptions of SWOT Level
perceptions of digital
information systems Weakness: ability 25
Strengths; relationship 18
Opportunity: information 10
Threat: decreasing local wisdom 3
Based on Table 4, it can be seen that fishermen representatives expect the involve-
ment of fishermen in compiling the schedule of tourist movements in the Gili Indah
national tourism strategic area. Fishermen also expect to share information regularly
and always be updated as part of reducing risks. This will be risk mitigation for
fishermen in the fishermen’s perception.
To bridge the need for space to receive up-to-date movement schedules and ship-
ping information, it is again to explore opinions or perceptions from fishermen about
digital-based information systems. To comply with the causes and effects method
that has been used with the results in Table 2, point 25 is set as the highest value to
determine fishermen’s perceptions of digital-based information. Perception-forming
factors were extracted from the Strength Weakness Opportunity and Threat (SWOT).
The data obtained are as shown in Table 5.
Furthermore, to facilitate the reading of the SWOT, the mapping was carried out
into a spider web diagram. Thus, the focus of the various organizations involved can
lead to dominant SWOT formers.
The results obtained are illustrated in Fig. 1.
weaknes:
ability
25
20
15
10
threat : 5
Strengths ;
decreasing 0
relaonship
local wisdom
opportunity :
informaon
Fig. 1 SWOT analysis

Based on this figure, the direction of the use of digital systems will clash with
media ownership (gadgets) but if the problem of gadget ownership can be resolved,
the digital information system program will be successful because it has the support
of “good relations” between port organizations in the region.
4 Conclusion
Thus, the existing recommendations, the preliminary hypothesis are refuted in case
of risk causes. Even though the main risk is about economic problem, but the problem
of ship’s engine is not a priority problem in the perception of Gili Indah fishermen,
because the priority issues are capacity institutions and networking port manage-
ment. The recommendation for further research is the development of an institutional
system based on fishermen’s culture and local wisdom.
Funding Acknowledgement This research is a preliminary study for a dissertation of a Ph.D.

student, so it uses the full cost of the student as the main author.
References
1. Pariwisata S (2009) Pitana dan Gayatri, Sosiologi Pariwisata, 2005, hal 45, pp 1–15
2. Ketut N, Dwijendra A (2020) Design model innovations for tourism villages in Bangli, Bali
Indonesia: debate between environmental and cultural protection versus community economic
development, vol 29, no 7, pp 1428–1443
3. Primadany S (2013) Analisis Strategi Pengembangan Pariwisata Daerah (Studi Pada Dinas
Kebudayaan Dan Pariwisata Daerah Kabupaten Nganjuk). J Adm Publik Mhs Univ Brawijaya
1(4):135–143
4. Nugraha YE, Siti YE (2020) Pengembangan Ekowisata Berbasis Masyarakat Di Desa Pemo
Taman Nasional Kelimutu Kabupaten Ende. J Destin Pariwisata 8(2):169. https://doi.org/10.
24843/jdepar.2020.v08.i02.p01
5. Frosch C, Martinelli D, Unnikrishnan A (2019) Evaluation of shared space to reduce traffic
congestion. J Adv Transp 2019. https://doi.org/10.1155/2019/6510396
6. Adnyana I, Dwijendra (2011) Arsitektur dan tata ruang pelabuhan di Bali, January 2012, 2011
[online]. Available http://erepo.unud.ac.id/id/eprint/26973/1/a01638a63e692dd76bbe4d4c1ca
42df0.pdf
7. Lindawati L, Rahadian R (2016) Identifikasi Faktor Dan Penilaian Risiko Pada Usaha Perikanan
Tangkap Di Kabupaten Sambas. J Sos Ekon Kelaut dan Perikan 11(1):99. https://doi.org/10.
15578/jsekp.v11i1.3175
8. Kesuma PA, Rohman MA, Prastyanto CA (2019) Risk analysis of traffic congestion due to
problem in heavy vehicles: a concept. IOP Conf Ser Mater Sci Eng 650(1). https://doi.org/10.
1088/1757-899X/650/1/012011
9. Profesi AK, Navigasi S, Pratomo AH, Rustamaji HC (2020) Analisis Karakteristik Nelayan
Tradisional Berdasar Jenis dan Klasifikasi Nelayan , Kelompok Kerja , Jenis Perairan ,
Teknologi 2(1):29–34
10. Chandra A (2013) Analisis kinerja distribusi logistik pada pasokan barang dari pusat distribusi
ke gerai indomaret di kota semarang
11. Ketut N, Dwijendra A, Dani AA, Sjuchro DW, Klochko EN, Patriani Y (2020) Political and
social infrastructure towards high quality tourism sites 7(11):1764–1773
12. Situbondo K (2013) pelabuhan panarukan, pp 1–8
13. Bilondatu MR, Kepercayaan DAN, Terhadap P (2016) motivasi, persepsi, kepercayaan, dan
keputusan pembelian 1(3):710–720
14. Sideman I (2021) Proc ICST 2:1–9
15. Sideman I (2021) Traffic management of Gunung Sari intersection base on problem solving
hierarchy traffic management of Gunung Sari intersection base on problem solving hierarchy.
https://doi.org/10.1088/1742-6596/1779/1/012082
16. Wijaya AB, Fauzie A (2020) Pemaknaan Hidup Nelayan (Analisis Makro dan Mikro pada
Kemiskinan Nelayan). Indonesia Psychol Res 2(2):96–108. https://doi.org/10.29080/ipr.v2i
2.259
17. Arnatha M, Budiartha NRM (2011) Prioritas Program Terminal Kapal Pesiar Tanah Ampo:
Suatu Penerapan Benefit-Cost Analysis (Bca) Untuk Penilaian Proyek. J Ilm Tek Sipil 15(1)
18. Karyawan IDMA, Widianty D (2014) Analisis Jarak Pandangan Henti sebagai Elemen
Geometrik pada Beberapa Tikungan Ruas Jalan Mataram-Lembar 18(2):40–48
19. Ketut N, Dwijendra A (2020) Expanding the oil and gas sector of Indonesia through better
marketing performance: analysing the role of online engagement, interactivity and brand
attachment, Aug 2020
20. Kurniawan Z, Aknuranda I, Setiawan NY (2018) Analisis dan Penentuan Prioritas Pence-
gahan Masalah Menggunakan failure mode and effect analysis dan cause and effect analysis
Pada Proses Bisnis Pengajuan proposal program Kreativitas Mahasiswa (PKM) (Studi Kasus:
Universitas Brawijaya Malang). J Pengemb Teknol Inf dan Ilmu Komput Univ Brawijaya
2(12):7105–7112
21. Sebab-dan-akibat A. Cause-and-effect analysis
22. Iii BAB (1992) Jenis Penelitian Lokasi Penelitian, pp 55–77
23. Sunarta (2018) Metode Penelitian 53(9):1689–1699
24. Undang-Undang Republik Indonesia No. 13 Tahun (2003) Undang-Undang Republik Indonesia
No. 13 Tahun 2003 tentang Ketenagakerjaan. Undang-Undang 1:1–34 [online]. Available http://
www.kemenperin.go.id/kompetensi/UU_13_2003.pdf
25. Ketenagakerjaan K (2016) Peraturan Menteri Ketenagakerjaan Republik Indonesia Nomor 2
tentang Sistem Standarisasi Kompetensi Kerja Nasional. Lembaran Negara RI
26. Anugerah Y, Nurani TW, Sondita MFA (2016) Kompetensi Nelayan Rawai Tuna Ditinjau
Dari Standar Kompetensi Kerja Nasional Indonesia (Skkni) Di Ppn Palabuhanratu. J Sos Ekon
Kelaut dan Perikan 11(2):251. https://doi.org/10.15578/jsekp.v11i2.3693
27. Ekonomi DR, Perbankan DAN (2020) Menentukan Populasi dan Sampel Dalam Riset-Riset
Ekonomi dan Perbankan Islam, Apr 2020
28. Iii BAB (2006) Vinna Indahtianti, 2013 Hubungan Pelayanan Sirkulasi Dengan Pemben-
tukan Citra Perpustakaan (Studi Deskriptif di Perpustakaan Universitas Pendidikan Indonesia).
Universitas Pendidikan Indonesia|repository.upi.edu, pp 25–41
29. Duijm NJ (2015) Recommendations on the use and design of risk matrices. Saf Sci 76:21–31.
https://doi.org/10.1016/j.ssci.2015.02.014
30. Parra NM, Nagi A, Kersten W (2018) Risk assessment methods in seaports—a literature review,
May 2018
31. Ketut N, Dwijendra A (2018) Quality of affordable housing projects by public and private
developers in Indonesia: the case of Sarbagita Metropolitan Bali, May 2013. https://doi.org/
10.5897/JGRP12.050
32. Raduan S, Dayang RAJ, Assis K, Diana H (2017) Passengers perception on safety level of ferry
transport: a case study in Labuan Island, Malaysia. J Adv Res Soc Behav Sci 1(2017):7–17
33. Langard B, Morel G, Chauvin C (2015) Collision risk management in passenger transportation:
a study of the conditions for success in a safe shipping company. Psychol Fr 60(2):111–127.
https://doi.org/10.1016/j.psfr.2014.11.001
34. Zeisel J (2015) Publication preview source observing environmental behavior—Chapter 8 in
inquiry by design: tools for environment-behavior, Jan 1981
35. Sasmita S, Martasuganda A, Purbayanto S, Hestirianoto T (2013) Keselamatan Kerja pada

Operasi Penangkapan Ikan Cantrang Nelayan Tanjung Sari, Kabupaten Rembang (Safety
Assessment of Cantrang Operation in Tanjung Sari, Regency Rembang Central Java). Bul
PSP 21(1):11–17
36. Mijts E, Arens P (2019) Capacity building for sustainable development in small island states
through science and technology research and education. https://ic-sd.org/wp-content/uploads/
2019/11/eric_mijts, Nov 2019
37. Moeller J (2015) A word on standardization in longitudinal studies, Dec 2015. https://doi.org/
10.3389/fpsyg.2015.01389
38. Ikasari AC (2017) Konsep Kebijakan Alih Kelola Pelabuhan Perikanan di Jawa Barat. J. Ilm.
Magister Ilmu Adm. No. 2 Tahun XI Juni, no 2, pp 44–55
39. Hadi S (2010) Pemeriksaan Keabsahan, pp 21–22
Construction Materials for Sustainable
Infrastructures
Residual Stress Evaluation
on Cold-Formed Steel C-Section
by X-Ray Diffraction
T. Widya Swastika , Heru Purnomo , M. R. Muslih, R. Apriansyah,

Henki W. Ashadi, and Mulia Orientilize
Abstract It is known that the existence of residual stress in cold-formed steel profile
plays an important role in determining the nominal strength of the profile in resisting
the external forces. Ignoring the residual stress will be very risky due to the service-
ability and ultimate strength requirements of the structure constructed from these
profiles. This study focuses on the determination of residual stress from local manu-
facture cold-formed steel lip channel C-sections which are widely used today for
the construction of residential roof truss. Two types of cold-formed steel from two
different manufactures have been investigated from the residual stresses in the cold-
formed lip channel C-sections and the flat steel sheets before being formed into the lip
channel C-section by X-Ray Diffraction apparatus. The results show that the tension
stress is measured on the flanges surface, while the compression stress is found in
the web surface. The residual stress in the bent area of lip channel C-section changes
significantly, while the residual stress in flat steel sheet experiences compression. In
addition, the residual stress in PC-065-1 specimen is around 200 MPa, while it is
about 150 MPa in LB-065-1, whereas for the flat steel sheet specimen of PC-065-1
it is around 100 MPa, while in LB-065-1 it is as high as 240 MPa.
Keywords Residual stress · Cold-formed steel · X-Ray diffraction
T. Widya Swastika · H. Purnomo (B) · H. W. Ashadi · M. Orientilize

Civil Engineering Department, Universitas Indonesia, Depok, Indonesia
T. Widya Swastika
H. W. Ashadi
M. Orientilize
M. R. Muslih · R. Apriansyah
Neutron Scattering Laboratory, PSTBM-BATAN Serpong, South Tangerang, Indonesia
1058 T. Widya Swastika et al.
1 Introduction
Residual stress is the stress that exists within a steel profile after it has been formed.
With the increasing scarcity of wood and the vulnerability of wood to termites as
a roof construction material in Indonesia, cold-formed steel has been commonly
used for roof constructions in the last fifteen years. Cold-formed steel lip channel
C-section is widely used for the roof construction in Indonesia because of the ease of
fabrication and installation. However, it is often the case that builders or contractors
perceive cold-formed steel to behave in the same way as hot rolled steel, where in the
calculation of its strength uses hot rolled steel. On one hand, the literature regarding
the residual stress of cold-formed steels is rare; on the other hand, the builders often
ignore the presence of residual stress in the cold-formed steel strength as there are no
data related to this, although efforts to investigate residual stresses in cold-formed and
hot rolled steels have been carried out overseas [1–7]. In general, the residual stress
measurement method in steel can be selected by using one of the three method groups,
namely non-destructive test, semi-destructive test, and destructive test [8–15]. The
non-destructive test consists of tests based on Ultrasonic, Diffraction, Magnetic,
Raman Spectroscopy, and Nano Indentation technique. Meanwhile, semi-destructive
tests are based on Hole Drilling, Ring Core, and Deep Hole. The destructive test
is based on Sectioning or Contour. The choice of the method will depend on the
availability of equipment in the laboratory and the level of accuracy desired.
This study presents an experimental investigation on cold-formed steel lip channel
C-section to evaluate the residual stress from two local manufacture products having
the same thickness but with a different flat width of the largest sub-element on the
web stiffener. Subsequently, the flat steel sheets before being formed into the lip
channel C-section will also be evaluated to have insight on the history of the residual
stress. Later, the coating of these specimens will be removed by water cooled sanding
on the center of the section and then the residual stress will have been re-evaluated.
2 Method
In the X-Ray Diffraction method, the strain is measured in the crystal lattice. This
method is applied to find grainy crystalline materials that produce a diffraction pattern
and appropriate intensity peaks, and free of interference in high back reflection area
for any orientation sample surface. Surface residual stress measurements are non-
destructive. Good macroscopic residual stress and line widening caused by micro
stresses and crystal damage can be determined independently. The X-Ray Diffraction
method is claimed to have high accuracy and simple operation [15–17].
This study investigated the cold-formed steel lip channel C-sections from two
different local manufactures which have a similar dimension in terms of height and
thickness; however, each will have its associated peculiarities of web stiffeners. In the
initial stage, two types of lip channel C-section were measured for the residual stress.
Residual Stress Evaluation on Cold-Formed … 1059
The next stage was removing the surface coating using water cooled sanding on the
center of the section. The residual stress in the two flat steel sheets has also been
evaluated by the same procedure as the lip channel C-section specimens. Evaluating
the residual stress in flat steel sheets is necessary to see the difference of the residual
stress state due to the fabrication process. The residual stress from all the specimens
was measured by the X-Ray Diffraction cos-α method. The residual stress from the
same coordinates located at the specimens will be compared.
2.1 Residual Stress Evaluation by X-Ray Diffraction
The relationship between normal stress components σij and the strain εfѰ along the
scattering vectors f and Ѱ in the case of elastically isotropic specimens is shown in
Eq. 1 and Fig. 1, as follows [18]:

εφψ = (1/2)S2 σ11 cos2 ∅ + σ12 sin2∅ + σ22 sin2 ∅ sin2

+(σ13 cos∅ + σ23 sin∅)sin2 + σ33 cos2
+ S1 (σ11 + σ22 + σ33 ) (1)
where:
σ11 , σ22 , σ33 = Normal stress in the sample coordinate axes in X, Y, and Z
directions.
σ23 = Shear stress in the Z axis direction normal to the Y axis.
σ12 = Shear stress in the Y axis direction normal to the X axis.
σ13 = Shear stress in the Z axis direction normal to the X axis.
S1 and S2 = The X-Ray elastic compliance constants.
They can be related to Young’s modulus E and Poisson’s ratio v
of the body. S1 = − v/E and S2 = (1 − v)/E.
The strain εfѰ is defined in the following [18]:
Fig. 1 A coordinate system

for stress analysis using
X-Ray diffraction [18, 19]
d∅ − d0
εφψ = (2)
d0
where:
d∅ = Lattice spacing along a scattering vector (the directions will be f and Ѱ).
d0 = Strain-free lattice spacing.
The strain-free lattice spacing varies easily depending on the crystalline state, and
thus it is difficult to know the accurate value for d0. However, d0 can be determined
by an optimization procedure.
2.2 Experimental Investigation
The X-Ray Diffraction method was carried out at the Neutron Scattering Laboratory
of BATAN, located in Serpong, South Tangerang. This method is to determine the
residual stress in the work area on the surface (Plane Stress) with a very thin specimen
thickness (100 μm – 17 mm). The accuracy produced by this technique is very good
as it can reach ±20 MPa. The X-Ray Diffraction measurement is useful for residual
stress analysis if the following conditions are met in the material specimen to be
measured, namely the material must have a crystal structure, the material must have
small grains, and the elastic constant of the material is known.
Figure 2 shows the X-Ray Diffraction residual stress method equipment. The
operation of the tools is simple. In its operation, every time it is used, the tool must
be set in advance with the conditions for using the tool, where the sample distance
is set at 45 mm with ψ angle (in Fig. 1). Then it is calibrated by firing an X-Ray
beam for the stress-free steel specimen. Seeing that in this diffractometer there are
no moving parts (rotating), the center point of its axis, therefore, must be determined
using a standard sample. The standard sample used is a stress-free sample in the form
of iron powder moulded with resin. With the correct settings, the residual stress in
this sample is in the range of 10 MPa (close to zero), and the tool is ready to use for
other specimens.
Fig. 2 The X-Ray

diffraction residual stress
method equipment
The data measured by the X-Ray Diffraction method are classified into 3 classes
with a diffraction pattern of Debye-Scherer circle: very good, good, and bad. Only
the very good and good data were taken into consideration.
The measurement results show the residual stress in the direction 11 (σ11 ) and
direction 12 (σ12 ), where σ11 is the perpendicular direction of the cross section, while
σ12 is the parallel lengthwise direction of the cross section.
2.3 Test Specimens
The test specimens were made from grade G550 AZ100. The specimens used are
cold-formed steel lip channel C-section with dimensions of 75 × 35 of 0.65 mm thick
and with a length of 2 m but with different rib positions from two local manufactures:
LB-065 and PC-065, illustrated on Fig. 3. Meanwhile, the flat steel sheet (SS-LB-065
and SS-PC-065) is 0.65 mm thick with 151 mm of width (3U and 6U for C-section and
SS for flat steel sheet are the label of measurement location configuration). Figure 4
shows the specimens of cold-formed steel lip channel C-section and flat steel sheet,
with coating and uncoating.
The results of the residual stress measurement are shown in Figs. 5, 6 and 7.
Fig. 3 Dimension detail and position of measurement nodes

Fig. 4 Cold-formed steel specimens with coating and uncoating; lip channel C-section (left) and
flat steel sheet (right)
Fig. 5 Residual stress comparison for lip channel C-section specimens LB-065 and PC-065
3.1 Specimens for Cold-Formed Lip Channel C-section
Figure 5 shows the residual stress σ11 of the 2 different cold-formed lip channel
C-section specimens which are taken before and after removing the coating. Lower
residual stresses are given after removing the coating. We can see that the sample LB-
065 gives a better classification than specimen PC-065. Moreover, all specimens show
Fig. 6 Residual stress comparison for steel sheet specimens LB-065 and PC-065
Fig. 7 Residual stress comparison between steel sheet and lip channel C-section specimens (LB-
065 and PC-065)
that the residual stress represent the forming method of the specimens, compression
in web, and tension in flanges.
3.2 Specimens for Steel Sheet Cold-Formed
We can see in Fig. 6 that the uncoated sample gives a lower value of residual stress
than the coated sample. After uncoating the surface, the value of residual stress is all
in compression. In addition, the uncoated surface also shows a better classification,
indicated by the number of nodes read in the X-Ray Diffraction measurement result.
We also see that the sample LB-065 in general gives a better result than the PC-065.
3.3 Comparison of Cold-Formed Steel Sheet and Lip

Channel C-section
Figure 7 indicates that the two specimens of cold-formed lip channel C-sections at
several location have almost similar pick values with the residual stress σ11 for flat
steel sheet. After the coating removal, the residual stress σ11 decreased, and the result
showed that the residual stresses represent the forming method of the specimens; in
other words, the sample lip channel C-section experienced compression in the web
and tension in the flanges, while the steel sheet was in compression. We can see that
the sample LB-065 gives a better classification than specimen PC-065. Furthermore,
the sample LB-065 gives a better result than the sample PC-065.
3.4 Residual Stress Contour
Figures 8 and 9 show the contours of the residual stress σ11 for cold-formed
lip channel C-sections after coating removal and without considering the bad
classification.
From Figs. 8 and 9, all specimens show that the residual stress represents the
forming method of the specimens, compression in web, and tension in flanges.
Fig. 8 Residual stress σ11 of 3U (left) and 6U (right) for specimen LB-065
Fig. 9 Residual stress σ11 of 3U (left) and 6U (right) for specimen PC-065
4 Conclusions
Based on the experimental test results of residual stress measurement with X-Ray
Diffraction cos-α method, conclusions could be taken:
Removing the coating from the specimen’s surface is the key point to have a
success in the X-Ray Diffraction method applied at a cold formed section. After
experiencing forming, the compressive residual stress at the flat steel sheet has been
transformed at both mid flanges to residual tensile stress and at the web; the values
of compressive residual stress have been reduced by almost 50% for the LB-065
specimens. However, for the PC-065 specimens, the same phenomenon happens at
both mid flanges, but at the web, the values of compressive residual stress do not
change too much from its flat steel sheet residual stress state. It might come from the
different forming process from both manufactures. It is interesting to note that ribs at
the web from both specimens LB-065 and PC-65 have caused the residual stress to
shift from compression to tensile stress at some nodes near the rib’s location. In the
near future, another investigation method to obtain residual stress will be conducted
to confirm the average compressive and tensile residual stress resulted from this
study.
Acknowledgements This research was funded by a research grant: “Publikasi Terindeks Inter-
nasional (PUTI) Universitas Indonesia Tahun 2020”, NKB-2015/UN2.RST/HKP.05.00/2020 on
November 6th, 2020. The experiment was conducted at Neutron Scattering Laboratory of
PSTBM-BATAN at Serpong, Indonesia. The authors gratefully acknowledge the support.
References
1. Somodi B, Kövesdi B (2017) Residual stress measurements on cold-formed HSS hollow section
columns. J Constr Steel Res 128:706–720
2. Mehner T et al (2018) Residual-stress evolution of cold-rolled DC04 steel sheets for different
initial stress states. Finite Elem Anal Des 144:76–83
3. Moen CD, Igusa T, Schafer BW (2008) Prediction of residual stresses and strains in cold-formed
steel members. Thin-Wall Struct 46(11):1274–1289
4. Weng CC, Pekoz T (1990) Residual stresses in cold-formed steel members. J Struct Eng
116(6):1611–1625
5. Kabir MZ, Parvizi M, Askari A. determination of residual stresses and strains in cold-formed
steel members due to roll-forming process
6. Chinnaraj K, Prasad MS, Lakshmana RC (2014) Investigation of manufacturing residual
stresses in cold formed truck frame rail sections. Int J Eng Res Appl 4:162–169
7. Akinlabi SA et al (2018) Investigating resulting residual stresses during mechanical forming
process. In: IOP conference series: materials science and engineering. IOP Publishing
8. Queen H et al. Novel electromagnetic sensors for stress measurement and materials testing
9. Civín A et al (2011) Ring-core residual stress measurement: analysis of depth increment
distribution for integral equation method. In: International conference engineering mechanics
2011
10. Epp J, Zoch H-W, Bobrov I (2018) Micromagnetic analysis of residual stress distribution in
42CrMo4 steel after thermal and mechanical surface treatment. Mater Res Proc 6
11. Petrucci G, Scafidi M (2008) A new procedure for the evaluation of residual stresses by the
hole drilling method based on Newton-Raphson technique. In: Proceedings of ICRS-8 the 8th
international conference on residual stress—Denver, CO (USA). Citeseer
12. Lestari S (2004) Residual stress measurements of unblasted and sandblasted mild steel
specimens using X-ray diffraction, strain-gage hole drilling, and electronic speckle pattern
interferometry (ESPI) hole drilling methods
13. Yerman J, Kroenke W, Long W (1996) Accuracy evaluation of residual stress measurements.
Westinghouse Electric Corp., West Mifflin, PA (United States)
14. Schajer GS (2013) Practical residual stress measurement methods. Wiley
15. Haiyang GJAF (2018) Methods of residual stress measurement
16. Deveci M. Types of residual stresses
17. Engstler M (2021) Nondestructive micromagnetic materials characterization at high measuring
speed [cited 2021]. Available from https://www.pm-forum.info/nondestructive-micromagn
etic-materials-characterization-at-high-measuring-speed
18. Rigaku (2020) Smart site RS portable stress analyzer: instruction manual. Rigaku
19. Yasukawa S (2016) X-ray stress analysis technique using the optimization of d0 with error
term—direct refinement solution (DRS) method. Rigaku J 32(2):6–12
Mechanical Properties of Fine-Grained
Concrete Using Fine-Red Sand and Fly
Ash for Road Construction: A Case
Study in Vietnam
Nguyen Thanh Sang, Thai Minh Quan, May Huu Nguyen, and Lanh Si Ho
Abstract Nowadays, in order to decrease environmental impacts such as exhaustion

of natural resources and CO2 discharge, the partial replacement of cement by fly
ash and coarse aggregate by crushed fine aggregate in concrete have been attracted
many researchers. This study concentrated on an experimental examination of the
mechanical properties of fine-grained concrete using fine-red sand placement for
crushed sand. Experimental tests were performed to investigate compressive, splitting
tensile, and flexural strength. The results reveal that the compressive, splitting tensile,
and flexural strength increased with curing time for all mixtures. The increase in the
amount of fine-red sand replacement caused a decrease in all mixtures. It was also
found that there is a strong correlation between compressive and splitting tensile
strength with a high coefficient of determination. Besides, the results of compressive,
splitting tensile, and flexural strength satisfied the requirement of concrete used for
pavement.
Keywords Compressive strength · Fine-grained concrete · Splitting tensile

strength · Flexural strength
N. T. Sang · T. M. Quan
University of Transport and Communications, Hanoi, Vietnam
T. M. Quan
M. H. Nguyen · L. S. Ho (B)
University of Transport Technology, 54 Trieu Khuc, Thanh Xuan, Hanoi 100000, Vietnam
M. H. Nguyen
Civil and Environmental Engineering Program, Graduate School of Advanced Science and
Engineering, 1-4-1, Kagamiyama, Higashi-Hiroshima, Hiroshima 739-527, Japan
1068 N. T. Sang et al.
1 Introduction
In the construction field, the utilization of Portland cement and coarse aggregate
as the main ingredient is a common manner to manufacture conventional concrete.
However, overexploitation has led to the depletion of these resources and caused
numerous environmental problems [1, 2]. Thus, the need for friendly replacement
material is crucial in adapting to environmental requests and sustainable development
[3, 4].
In line with the need for friendly construction material, the utilization of fine
materials as a substitute replacement solution for coarse aggregate is becoming a
typical topic [5, 6]. Theoretically, a fine-grained concrete either does not consist
of any coarse aggregate at all or only a small proportion ensuring the fine/coarse
ratio keeps higher than 1.0 [7, 8]. In this kind of concrete, the local materials (e.g.,
seawater, saline sand) and the various industrial wastes (e.g., ground blast furnace slag
(GGBFS), fly ash) can be effectively used and also further brings many environmental
and economic benefits [9, 10].
The reuse of local sand to utilize fine-grained concrete has been broadly investi-
gated [11, 12]. Bédérina et al. [11] studied the applicability of local sand available in
Algeria as a partial effort to substitute coarse aggregate in concrete. Three types of
local sand, including dune, river, and a mixture of dune and river sands were employed
to produce three corresponding sand concretes that were used to examine the work-
ability and mechanical strength. The obtained results indicated that the local sands
could be used to manufacture a more workable, more compact, and more resistant
sand concrete while still meeting the requirement in compressive strength compared
to conventional concrete. Li et al. [13] presented an investigation to investigate the
application of 100% replacement of river and manufactured sand to produce san
concrete. A series of low and medium strengths (25–60 MPa) was then achieved,
which indicated the application of these sands in producing sand concretes.
The utilization of fly ash (FA) in a concrete mixture can lead to changes in the
hydration process and hence influencing concrete properties such as workability,
mechanical, and durability characteristics [14, 15]. For example, Agrawal et al. [14]
investigated the influence of replacement of natural sand with FA on hardened prop-
erties of concrete including mechanical and durability characteristics. The results
indicated that the FA fine-grained concrete could obtain a comparable mechanical
strength compared to the conventional concrete. The influence of fly ash on the
mechanical properties and other properties such as shrinkage, chloride permeability,
and carbonation of coral sand concretes was also carried out by Cheng et al. [15].
They indicated that the compressive strength development of fine-grained concrete
was higher than conventional concrete at 3 and 7 days, but then lower at 28 days.
In Vietnam, as a developing country, the development of fine-grained concrete
utilizing fine sand and fly ash is a promising approach [8, 16]. Le et al. [8] experimen-
tally studied the performance of fine-grained concrete utilizing sea sand and fly ash.
Several properties of the concrete including workability, compressive strength under
influence of curing conditions were examined. The obtained results indicated that it
Mechanical Properties of Fine-Grained Concrete … 1069
is feasible to manufacture fine-grained concrete with high strength using sea sand
and fly ash. Nevertheless, an investigation of mechanical properties of fine-grained
concrete using red-fine sand and fly ash for road construction has not been carried
out so far in a condition of Vietnam. Thus, the purpose of this study is, therefore, to
investigate mechanical properties of fine-grained concrete using local fine-red sand
as a partial replacement for road-pavement construction in a case study in Vietnam.
2.1 Materials and Specimen Preparation
Materials
To produce fine-grained concrete, ordinary Portland cement (PC 40), crushed sand,
and local fine-red sand (dune sand) were used. To reduce the amount of cement
used, cement was partially replaced with FA, and the content of fly ash was fixed by
90 kg/m3 . The crushed sand (hereafter, CS) was taken from Binh Thuan province,
which can satisfy ASTM C33 standard on grain size distribution. The local fine-red
sand (hereafter, RS) was obtained from a red fine sand mine in Binh Thuan province.
The RS used in this study partially replaces CS to utilize local materials. The chemical
compositions of cement, CS, RS, and FA are introduced in Table 1. The grain size
distribution of CS, RS, and four combined mixtures of two sands is shown in Fig. 1.
The specific density of cement, crushed sand, red sand, and fly ash is 3.10, 2.69,
2.57, and 2.50 g/cm3 , respectively.
Mixture Proportions
The mixtures of fine-grained concrete were calculated based on the absolute volume
of materials. This study used an admixture to reduce the quantity of free water; as a
result, the water/binder (w/b) of 0.37 was prepared for entire mixtures to reduce the
Table 1 Chemical compositions of materials used

Chemical composition (%) PC40 Fly ash (FA) Red sand Crushed sand
SiO2 21.49 49.60 92,69 62.48
Al2 O3 5.4 21.97 1.61 16.82
Fe2 O3 3.49 4.90 2.1 1.04
CaO 63.56 0.66 0.50 1.17
MgO 1.40 0.78 1.08 0.45
Na2 O 0.12 0.16 0.27 4.33
K2 O 0.3 3.52 0.46 2.19
LOI 0.19 12.63 – 1.46
Fig. 1 The grain size

distribution of crushed sand,
red sand, and different
mixtures
Table 2 Mixing proportion of fine-grained concrete

Mixture w/b Water Cement Fly ash Crushed Red sand Admixture
(kg/m3 ) (kg/m3 ) (kg/m3 ) sand (kg/m3 ) (kg/m3 )
(kg/m3 )
70CS30RS 0.37 165 360 90 1259 540 4.50
60CS40RS 0.37 165 360 90 1075 717 4.50
50CS50RS 0.37 165 360 90 892 892 4.50
40CS60RS 0.37 165 360 90 711 1066 4.50
Note mixture name 70CS30RS indicates that this mixture containing 70% crushed sand and 30%
red sand. This is also applied to all remaining mixtures
porosity of this concrete. Four different mixtures were prepared, in which the amount
of crushed sand was partially replaced with red sand at 30, 40, 50, and 60%. The
amount of fly ash (FA) was kept constant for all mixtures (90 kg/m3 ). The amount
of admixture (superplasticizer—a water reduction agent with the name RoadCon-
SPR3000) was designed as 1.25% by mass of cement. The designed mix proportions
of different mixtures are shown in Table 2.
Preparation and Casting of Specimens
All concrete mixtures were produced using a mechanical mixer and the total mixing
time was 8 min. First, aggregates (crushed sand and red sand) including fly ash, and
cement were mixed together in 2 min under dry conditions. Then, approximately
80% of the water was supplied and mixed for 2 min. Finally, the remaining water
and admixture were added, and the mixtures were mixed again for 4 min., The cylin-
drical specimens having a size of 100 × 200 mm were used to determine compressive,
tensile strength and elastic modulus. Regarding the flexural strength, the beam spec-
imens with a size of 400 × 100 × 100 mm were adopted. The cylindrical specimens
were prepared with three layers, each layer was compacted 25 times using a steel rod.
For beam specimens, they were cast with two layers, and they were compacted using
a vibration table. After completing compaction, the top surfaces of the mold were
sealed with polyethylene sheets and stored in room condition at 20 °C. After one day,
all specimens were demolded and cured in water at 20 ± 2 °C until designated ages.
All tests in this study were conducted in triplicate and the mean value was used.
2.2 Testing of the Specimens
Compressive Strength and Elastic Modulus Tests

The compressive strength of this concrete was determined via the compression test
according to ASTM C39. The compression tests were conducted for the specimens
at 3, 7, and 28 days. The value of compressive strength is the mean value of three
specimens. The elastic modulus of concrete was calculated based on ASTM C469
standard for the specimen at 28 days. The elastic modulus of each specimen was
computed using the following Eq. (1).
σ2 − σ1
E0 = (1)
ε2 − 0.000050
where:
E: elastic modulus, Mpa,
σ2 : stress corresponding to 40% of ultimate load, Mpa,
σ1 : stress with a longitudinal strain, ε1 of 50 millions, MPa,
ε2 : longitudinal strain caused by stress σ2 .
Splitting Tensile Strength, and Flexural Strength Tests

It is known that the splitting tensile strength is an important property that influ-
ences many characteristics of concrete such as controlling cracks, stiffness, bonding
capacity to reinforcement, and durability. The splitting tensile strength was carried
out based on ASTM C496 on specimens at 3, 7, and 28 days of curing. To deter-
mine the flexural behavior of this fine-grained concrete, the flexural strength test
was then conducted for the beam specimens at 28 days in accordance with ASTM
C293/C293M–16 standard.
3.1 Compressive Strength and Elastic Modulus
Figure 2 shows the compressive strength development of different mixtures. It can be

observed that the compressive strength increased with curing time for all mixtures as
Fig. 2 Compressive strength

of different mixtures
Fig. 3 Elastic modulus of

different mixtures at 28 days
found in previous studies [8, 16]. The increase in compressive strength, in this case,
is mainly contributed to the hydration of cement. For all curing ages, the compressive
strength decreased with an increase in the amount of RS. This decrease can be caused
by the low strength and frictionl of red sand compared to that of crushed sand. The
rate of decrease in the compressive strength of all curing ages is approximately the
same for all mixtures, this can also be explained by the decrease in the amount of
crushed sand.
The elastic modulus of four mixtures at 28 days is presented in Fig. 3. Similar to
the case of the compressive strength, the elastic modulus decreases with increase in
the amount of red sand (i.e. from 30 to 60%). This was also caused by the increase
in the amount of RS that causes the reduction of compressive strength and modulus.
The highest elastic modulus was observed in mixture 70CS30RS (the lowest amount
of RS), while the lowest value was found in mixture 40CS60RS.
3.2 Splitting Tensile and Flexural Strength
The splitting tensile strength of different mixtures is presented in Fig. 4. It can be

seen that the splitting tensile strength increased with curing time for all mixtures,
Fig. 4 The splitting tensile

strength of different mixtures
as found in previous studies [8, 16]. Comparable to the compressive strength, the
increase in the splitting tensile with time is mainly attributed to the cement hydration.
Comparing with different mixtures, in general, it can be observed that the splitting
tensile strength also decreased with increasing in the amount of RS, it is confirmed
with the result of compressive strength. At the age of 3 days, the splitting tensile
of mixture 70CS30RS was also almost equal to that of mixture 60CS40RS, while it
decreased rapidly from mixture 60CS40RS to mixture 50CS50RS. For other curing
ages, including 7, 14, and 28 days, the splitting tensile strength decreased gradually
from mixture 70CS30RS.
Figure 5 shows the flexural strength of four mixtures at the age of 28 days. It
was found that the flexural strength generally decreased with decreasing amount
of crushed sand (CS) (i.e. increasing amount of RS) and it also agreed well with
the results of the compression test. The flexural strength gradually decreased from
mixture 70CS30RS to mixture 50CS50RS, and then it decreased rapidly from mixture
50CS50RS to mixture 40CS60RS. The highest value (7.70 MPa) of flexural strength
was observed in mixture 70CS30RS, and the lowest value (5.18 MPa) was found in
mixture 40CS60RS. It indicates that the increase in the amount of red sand leads to
a decrease in flexural strength of this type of concrete.
Fig. 5 The flexural strength

of different mixtures at
28 days
Fig. 6 Relation between

splitting tensile strength and
compressive strength
3.3 Correlation Between Compressive Strength and Splitting

Tensile Strength
The correlation between compressive and splitting tensile strength is described in

Fig. 6. Generally, this relationship is presented in a power-type equation as being seen
in many previous studies of conventional concrete [17, 18] and fine-grained concrete
[16]. According to the figure, it can be seen that the splitting tensile strength grew
up with an increment of compressive strength as found in normal concrete and fine-
grained concrete in previous publications [16, 19]. In addition, it is also found that this
relationship has a high coefficient of determination, as found in the previous study
of fine-grain concrete [16]. It indicates that this fine-grained concrete using red sand
has similar behavior compared to normal concreted and fine-grained concrete using
saline sand in the previous study.
4 Conclusions
In this study, the mechanical properties such as compressive, tensile, and flexural
strength of fine-grained concrete using local fine red sand as a partial crushed sand
replacement and fly ash. Some main concluding remarks can be derived from the
experiment results as follows.
• The compressive, tensile, and flexural strength increased with curing time. When
the amount of fine red sand increased, those strengths decreased.
• The increase in the amount of fine-red sand replacement caused a decrease in all
mixtures.
• There is a strong correlation between compressive strength and splitting tensile
strength with a high coefficient of determination, as found in previous studies.
• Compressive, tensile, and flexural strength results satisfied the requirement of

concrete used for the requirement of concrete for pavement.
Acknowledgements This research is funded by the Project of Vietnam Ministry of Science and
Technology‚ “Study on using saline sand to build transport work” under grant number DTDL.CN-
23/19.”
References
1. El-Fadel M, Zeinati M, El-Jisr K, Jamali D (2001) Industrial-waste management in developing

countries: the case of Lebanon. J Environ Manage 61:281–300
2. Henry RK, Yongsheng Z, Jun D (2006) Municipal solid waste management challenges in
developing countries–Kenyan case study. Waste Manage 26:92–100
3. Troschinetz AM, Mihelcic JR (2009) Sustainable recycling of municipal solid waste in
developing countries. Waste Manage 29:915–923
4. Dash MK, Patro SK, Rath AK (2016) Sustainable use of industrial-waste as partial replacement
of fine aggregate for preparation of concrete—a review. Int J Sustain Built Environ 5:484–516
5. Akono A-T, Zhan M, Chen J, Shah SP (2021) Nanostructure of calcium-silicate-hydrates in
fine recycled aggregate concrete. Cem Concr Compos 115:103827
6. Li J, Yang E-H (2017) Macroscopic and microstructural properties of engineered cementitious
composites incorporating recycled concrete fines. Cem Concr Compos 78:33–42
7. Bederina M, Marmoret L, Mezreb K, Khenfer MM, Bali A, Quéneudec M (2007) Effect of the
addition of wood shavings on thermal conductivity of sand concretes: experimental study and
modelling. Constr Build Mater 21:662–668
8. Le HT, Nguyen ST, Ludwig H-M (2014) A study on high performance fine-grained concrete
containing rice husk ash. Int J Concr Struct Mater 8:301–307
9. Bai Y, Darcy F, Basheer PAM (2005) Strength and drying shrinkage properties of concrete
containing furnace bottom ash as fine aggregate. Constr Build Mater 19:691–697
10. Xiao J, Qiang C, Nanni A, Zhang K (2017) Use of sea-sand and seawater in concrete
construction: current status and future opportunities. Constr Build Mater 155:1101–1111
11. Bédérina M, Khenfer MM, Dheilly RM, Quéneudec M (2005) Reuse of local sand: effect
of limestone filler proportion on the rheological and mechanical properties of different sand
concretes. Cem Concr Res 35:1172–1179
12. Gurumoorthy N, Arunachalam K (2016) Micro and mechanical behaviour of treated used
foundry sand concrete. Constr Build Mater 123:184–190
13. Nanthagopalan P, Santhanam M (2011) Fresh and hardened properties of self-compacting
concrete produced with manufactured sand. Cem Concr Compos 33:353–358
14. Agrawal US, Wanjari SP, Naresh DN (2019) Impact of replacement of natural river sand with
geopolymer fly ash sand on hardened properties of concrete. Constr Build Mater 209:499–507
15. Cheng S, Shui Z, Sun T, Yu R, Zhang G, Ding S (2017) Effects of fly ash, blast furnace slag
and metakaolin on mechanical properties and durability of coral sand concrete. Appl Clay Sci
141:111–117
16. Sang NT, Quan TM, Nguyen MH, Ho LS (2021) Performances of eco-fine-grained concrete
containing saline sand as partial fine aggregate replacement. J Appl Sci Eng 24:527–539
17. Behnood A, Verian KP, Gharehveran MM (2015) Evaluation of the splitting tensile strength
in plain and steel fiber-reinforced concrete based on the compressive strength. Constr Build
Mater 98:519–529
18. Legeron F, Paultre P (2000) Prediction of modulus of rupture of concrete. Materials Journal
97:193–200
19. Chhorn C, Hong SJ, Lee SW (2018) Relationship between compressive and tensile strengths
of roller-compacted concrete. J Traff Transp Eng (English Edn) 5:215–223
A Systematic Review of Concrete
Material for Noise Reduction
of Transportation Sectors
Ecky Ferry Ferdyan, Dewi Handayani, Sholihin As’ad, and Ubaidillah
Abstract Noise is a sound whose presence is considered to interfere with hearing

which can cause health problems and is caused by noise sources. The most common
source of moving noise is transportation. Therefore it is necessary to have a noise
absorber to reduce noise levels. The most used noise-reducing material is concrete.
This paper using a systematic review method and aims to investigate the development
of concrete materials as a noise barrier wall. Concrete for noise absorbers is divided
into two, conventional and non-conventional. Non-conventional concrete is a mixture
of concrete with added material and is generally a waste material. Non-conventional
concrete with different added materials produces other characteristics and different
noise attenuation results. Some examples of added materials such as waste materials
from burning coal, fishery waste, waste ceramics, and polymeric materials provide
significant results. The absorption coefficient was able to reach 0.9, and the NRC
value earned 0.61. However, a higher ability to absorb noise, smaller compressive
strength of the concrete produced. It is necessary to understand the characteristics of
the material for use in applications in transportation.
Keywords Reducing noise · Transportation · Material concrete
1 Introduction
Noise is a sound whose presence is considered to interfere with hearing [1]. Noise
is unwanted sound from a business or activity at a certain level and time that can
E. F. Ferdyan (B) · D. Handayani · S. As’ad

Department of Civil Engineering, Faculty of Engineering, Sebelas Maret University,
D. Handayani
Ubaidillah
Department of Mechanical Engineering, Faculty of Engineering, Sebelas Maret University,
1078 E. F. Ferdyan et al.
cause disturbances to human health and environmental comfort. Based on the Decree
of the State Minister for the Environment No. Kep-48/MENLH/11/1996 concerning
noise level quality standards, there are noise thresholds determined and depend on
the area’s designation. The area with the lowest noise limit is green open space, while
the site with the highest noise limit is the trade, industry, recreation, and seaport area,
which is 70 dB [2].
A noise source causes noise. Noise sources are sound sources whose presence is
considered to interfere with hearing both from moving and immovable sources [1]. A
movable source is a moving or non-permanent emission source in a place originating
from a motorized vehicle. Meanwhile, an immovable source is a source of emission
which is fixed in one place [3].
Highway traffic noise is a source of environmental noise and comes from moving
sources, namely motor vehicles. Road traffic noise is influenced by the type of vehicle
and vehicle speed [4]. Apart from the vehicle engine factor, road noise occurs due to
the interaction between tires and pavement. Noise that results from the interaction
between tires and the dominant pavement at speeds above 40 km/h and influenced
by four main factors, namely:
• Pavement surface characteristics (surface texture, friction, aggregate properties,
porosity, pavement materials, and their properties, pavement life, and durability).
• Tire characteristics (tire dimensions, rubber hardness, trend pattern, tire age and
wear, poles, and tire number).
• Environmental factors (temperature, humidity, wind, water layer thickness, and
dust).
• Driver-controlled factors (driving speed, tire load, and pressure, tangential force,
and acceleration) [5].
There are two methods for reducing noise: the noise isolation method and the noise
absorption method. Noise isolation methods reduce noise by interfering with and
blocking the propagation of sound. In contrast, the noise absorption method reduces
noise by absorbing and weakening sound energy [6]. The most frequently used noise
isolation method is a reflective noise barrier. Reflective noise barriers are designed
to reflect most of the traffic noise, which can be problematic when minimizing sound
reflection to noise-sensitive areas adjacent to roadways is required. Simultaneously,
noise reduction can reduce the reflected sound’s intensity on the wall surface [7].
This paper aims to determine how the development of concrete material as a noise
barrier wall in the transportation sector can be helpful for research or use of noise
barrier walls in the future as a reference in determining the material to be selected.
So that later, it can be seen how the properties, characteristics, and effectiveness of
various materials for noise barrier walls.
A Systematic Review of Concrete Material … 1079
2 Method
The method used is a literature review. Literature Review is a description of the theory,
findings, and other research materials obtained from reference materials to be used as
a basis for research activities to formulate a clear frame of mind from the formulation
of the problem to be investigated. This journal uses 32 references consisting of
government regulations, books, journals, and proceedings. There are four references
to government regulations, two books, five proceedings, and 21 Scopus indexed
journals. From 21 journals, there are 14 journals rated Q1 and seven journals rated
Q2.
3 Concrete
3.1 Conventional
In Indonesia, the required noise barrier walls are used ALWA material. ALWA (Arti-
ficial Light Weight Aggregate) is a kind of artificial aggregate processed at high
temperatures and relatively lighter than ordinary aggregate or other artificial aggre-
gates. ALWA can be a by-product of the iron mill as a result/process in a blast furnace.
Composition = Cement:Sand:ALWA = 1:4:4 [8]. Recent studies have shown that
an effective noise barrier wall is used as porous concrete. Porous concrete is a rigid
acoustic absorbent material containing large cavities (pores) intentionally made for
acoustic absorption or permeation. Although ordinary concrete is also an acoustic
absorbent material, Porous concrete is a material that is more widely used because of
its excellent mechanical properties and durability compared to ordinary concrete [9].
Porous concrete is made using coarse aggregate with little or no fine aggregate. So
that porous concrete has a continuous cavity and has high permeability. As a result,
the pores in the porous concrete absorb sound energy through internal friction. The
type of aggregate also affects the properties of the porous concrete itself [10].
3.2 Non-conventional
Coal Bottom Ash. Fly ash, coal bottom ash (CBA), boiler slag, and flue gas desul-
phurization (FGD) sludge are waste generated during the combustion process at
coal-fired power plants [11]. Coal waste is hazardous to health and the environment,
so there is a need for innovation to use this waste not to endanger the environment
but can be helpful, one of which is the use of CBA. CBA is gradually introduced into
the construction industry as a base layer for roads, pavements, structural stockpiles,
drainage media, aggregates for concrete, and prefabricated soil products to reduce
environmental pollution [12]. CBA has been used as a partial substitute for natural
aggregate (fine or coarse aggregate) [13]. CBA as a substitute for fine aggregate with
a percentage of 80, 90, 100% can absorb sound by 35% [11]. CBA as a substitute for
aggregate coarse with a ratio of 100% in concrete can absorb noise up to a coefficient
of 0.43; however, CBA concrete has poor resistance to acids and sulfuric [10]. Apart
from using laboratory testing, testing has also been carried out using the software.
The use of software to determine the absorption coefficient of sound is adequate. The
highest sound absorption coefficient that the material can absorb is more than 0.9.
For semi-applicative use, concrete with coal bottom ash and made semi-industrial
with multilayer can produce a sound reduction of 4 dB [14].
Steel Slag. Steel slag is a by-product of steel production, accounting for about 15% of
steel output [15]. The porosity and NRC of the steel slag’s sound-absorbing material
can reach more than 60.0 and 0.50%, respectively, and the compressive strength of
the sintered specimen is above 6.5 MPa [16]. One product that is similar to steel
slag is Air-Cooled Blast Furnace Slag. Air-Cooled Blast Furnace Slag (ACBFS) is
a by-product of steel and iron manufacturing. ACBFS is produced by compacting
the liquid slag formed on top of the cast iron in a blast furnace under environmental
conditions. It is a slow process that allows the liquid slag molecules to be oriented
to a crystalline state [17]. The larger the ACBFS aggregate size, the greater the
absorption coefficient value, and it can even reach 0.9 at a frequency of 1000 Hz.
However, the higher the absorption value, the smaller the concrete’s compressive
strength [18]. The proportion of steel slag that reaches 50% can produce NRC 0.47,
and the compressive strength is 3.0 MPa [19].
Mollusk Shell. The amount of waste produced from industry increases every year
and can cause environmental and public health problems. In the water sector, about
6–8 million tonnes of waste a year are produced worldwide, and only a quarter of the
waste generated is reused [20]. One of the water wastes is shellfish. Shell waste with
a proportion of 80% as a substitute for aggregate, an aggregate size of 2–7 mm, can
produce NRC reaching 0.45 while ordinary porous concrete only reaches 0.24 [21].
Noise barriers made with shell waste have a higher impact between 32 and 267%
than usual porous concrete noise barriers [22].
Ceramic Waste. It is estimated that about 40% of the daily production in the ceramic
industry is stockpiled. Therefore, the need to find new applications in other sectors
where ceramic waste can be recycled is significant [23]. The proportion of ceramic
waste is 80% as a substitute for coarse aggregate, and the particle sizes are divided
into fine, medium, and coarse. Thickness 40, 80, 120 mm. Ceramic waste with a
percentage of 80% and a large particle size, 120 mm thick, resulted in NRC 0.61
[24].
Fly Ash. Fly ash is a material that pollutes the environment and has the potential to
damage human health. Research on the application of the waste is being intensified,
a mixture for making concrete. Fly Ash has a proportion of 20, 30, 40, 50% used
as a substitute for cement. Fly Ash’s result with a percentage of 50% produces an
NRC of 0.5 [25]. The proportion of fly ash and construction waste is 20, 50, 80%
as a substitute for coarse aggregate. 80% construction waste and 20% fly ash yield
a maximum absorption coefficient of 0.9 [16]. With an average volume density of
0.93 g/cm3 .
Polymeric Material. Polymer foam includes several cells in the polymer matrix.
Foam can have an open, partially open, or closed-cell structure. Polymer foams
require less material than solid polymers of the same volume, so the material costs
are reduced in many applications [26]. The use of polyuteran type polymer foam
with a proportion of 60% and textile waste of 40% produced an NRC of 0.593 [27].
Mixing 70% metakaolin and 30% blast furnace slag powders can produce inorganic
polymer foam with densities ranging from 0.4 to 1.0 g/cm3 . The most significant
sound absorption coefficient is at a density of 0.4 g/cm3 , with a maximum value
reaching 0.9 [28].
Table 1 has summarized the studies using the materials described above. We can
see that each material has a different sound absorption coefficient and NRC value.
There are also materials here that can replace coarse aggregate, fine aggregate, or
cement. It depends on the type of material itself. The maximum absorption coefficient
reaches 0.9, and the highest NRC can earn 0.61.
Table 1 NRC values of the use of unconventional concrete from several studies
References Year Type of As a replacement Percentage (%) Absorption NRC
material coefficient
(Max)
[7] 2017 Coal Fine aggregates 100 0.9 –
bottom ash and coarse
aggregates
[10] 2017 Coal Aggregate coarse 100 – 0.43
bottom ash
[11] 2020 Coal Fine aggregates 80, 90, 100 – 0.35
bottom ash
[16] 2015 Steel slag Fine aggregates 60 – 0.5
[16] 2015 Fly ash Aggregate coarse 80 0.9 –
[19] 2014 Steel slag Cement 50 – 0.5
[21] 2019 Shell waste Fine aggregates 80 – 0.45
and coarse
aggregates
[24] 2016 Ceramic Aggregate coarse 80 – 0.61
waste
[25] 2016 Fly ash Cement 50 – 0.5
[27] 2016 Polymeric Aggregate 60 – 0.593
material
[28] 2014 Polymeric Aggregate 70 0.9 –
material
4 Use for Transportation
A conventional concrete barrier to reduce noise indicates that the noise-absorbing

material is generally uneven and has a porous cavity to reduce transportation noise
[29]. Panel Material Paragon 23-T produces a noise suppression effect on the highway
due to the most effective transportation, namely 13.58 dB, then precast concrete is
second, and the last is steel. It is probably because the 23 T paragon panel is the
thickest material than the others [30]. The tree material, Concrete Hollow, Concrete
Panel, shows that those that meet the requirements of an effective noise barrier
that can reduce noise at least 5 dB are hollow concrete and panel concrete, while
vegetation is not. Concrete Hollow shows the results of noise reduction above 5 dB
but at specific frequencies below 5 dB. It is inconsistent in reducing the impact of
noise compared to concrete panels. It may be because hollow concrete requires many
joints, so leaking sound from these joints [31].
5 Conclusion
One of the efforts to prevent noise in transportation is to use a noise barrier wall.
Research on concrete materials for noise barrier walls has been widely carried out,
and the type of material significantly affects the effectiveness in reducing noise.
Various studies have shown that non-conventional materials such as waste materials
from burning coal, fishery waste, ceramic waste, and polymeric materials can absorb
more than 35% noise. However, because the structure of noise-reducing concrete is
concrete that has many cavities, the strength of the concrete is also tiny.
References
1. Construction and Building Guidelines (Pd T-16-2005-B). Mitigation of noise impacts due to
road traffic. Public Works Department (2005)
2. Minister of Environment Decree no. Kep-48/MENLH/11/1996 (1996)
3. Government Regulation of the Republic of Indonesia Number 41. Concerning air pollution
control (1999)
4. Dubey AK, Shukla AK, Kumar S (2019) Assessment of noise variation on different pavements
with four wheeler vehicle. Int J Sci Res Rev 7
5. Kleiziene R, Šernas O, Vaitkus A, Simanavi R (2019) Asphalt pavement acoustic performance
model. Sustainability 11
6. Ling S, Yu F, Sun D, Sun G, Xu L (2021) A comprehensive review of tire-pavement noise:
generation mechanism, measurement methods, and quiet asphalt pavement. J Clean Prod 287
7. Arenas C, Leiva C, Vilches LF, Ganso J (2017) Approaching a methodology for the development
of a multilayer sound absorbing device recycling coal bottom ash. Appl Acoust 115:81–87
8. Keputusan Direktur Jenderal Bina Marga No. 076/Kpts/Db/1999. Pedoman Perencanaan
Teknik Bangunan Peredam Bising (1999)
9. Kim HK, Lee HK (2010) Acoustic absorption modeling of porous concrete considering the
gradation and shape of aggregates and void ratio. J Sound Vibr 329:866–879
10. Carlos L, Celia A, Luis F, Fatima A, Yolanda L (2017) Assessing durability properties of noise
barriers made of concrete incorporating bottom ash as aggregates. Euro J Environ Civ Eng 10
11. Hannan R, Shahidan S, Ali N, Bunnori N, Zuki S, Ibrahim M (2020) Acoustic and non-acoustic
performance of coal bottom ash concrete as sound absorber for wall concrete. Case Stud Constr
Mater 399
12. Kurama H, Kaya M (2008) Usage of coal combustion bottom ash in a concrete mixture. Constr
Build Mater 9:1922–1928
13. Singh M, Siddique R (2015) Properties of concrete containing high volumes of coal bottom
ash as fine aggregate. J Clean Prod 91:269–278
14. Arenas C, Leiva C, Vilches LF, Cifuentes H, Galán M (2015) Technical specifications for
highway noise barriers made of coal bottom ash-based sound absorbing concrete. Constr Build
Mater 95:585–591
15. Guo WB, Cang DQ, Yang ZJ, Li Y, Wei CZ (2011) Study on preparation of glass-ceramics
from reduced slag after iron melt-reduction. Bull Chin Silica Soc 30(5):1189–1192
16. Sun P, Guo Z (2015) Preparation of steel slag porous sound-absorbing material using coal
powder as pore former. J Environ Sci 536:67–75
17. WS Association: Steel Statistical Yearbook (2014) World steel committee on economic studies,
Brussels
18. Rios JD, Arenas C, Cifuentes H, Peceño B, Leiva C (2019) Porous structure by X-Ray computed
tomography and sound absorption in pervious concretes with air-cooled blast furnace slag as
coarse aggregate. Acoustics, Australia
19. Sun P, Zhan-cheng G (2014) Sintering preparation of porous sound-absorbing materials from
steel slag. Trans Nonferrous Metals Soc China 25:2230–2240
20. Yan N, Chen X (2015) Sustainability: don’t waste seafood waste. Nature 524(7564):155–157
21. Peceno B, Arenas C, Farin B, Carlos (2019) Substitution of coarse aggregates with mollusk-
shell waste in acoustic-absorbing concrete. J Mater Civ Eng 31:6
22. Peceño B, Leiva C, Fariñas B, Schmid A (2020) Is recycling always the best option? Envi-
ronmental assessment of recycling of seashell as aggregates in noise barriers. Processes
8:776
23. Torgal F, Jalalib S (2010) Reusing ceramic wastes in concrete. Constr Build Mater 24(5):832–
838
24. Arenas C, Vilches LF, Leiva C, Fariñas B, Galán M (2016) Recycling ceramic industry wastes
in sound-absorbing materials. Mater Constr 66:324
25. Arenas C, Luna-Galiano Y, Leiva C, Vilches LF, Arroyo F, Villegas R, Fernández-Pereira C
(2017) Development of a fly ash-based geopolymer concrete with construction and demolition
wastes as aggregates in acoustic barriers. Constr Build Mater 134:433–442
26. Niaounakis M (2015) Biopolymers: processing and products. Chapter 9—foaming and foamed
products. Elsevier
27. Tiuca A, Vermeşan H, Gabora T, Vasile O (2016) Improved sound absorption properties of
polyurethane foam mixed with textile waste. Energy Proc 85:559–565
28. Hung T, Huang J, Wang Y, Lin K (2014) Inorganic polymeric foam as a sound-absorbing and
insulating material. Constr Build Mater 50:328–334
29. Christopher W, Menge BD (2011) Reflections from highway noise barriers and use of absorptive
materials in the United States. J Transp Res Board:161–166
30. Clavijo P, Yang JJ (2020) A comparative study of insertion loss of traffic noise barriers in
Georgia. In: International conference on transportation and development
31. Halima H, Abdullah R, Abdullah A, Ali A, Nor M (2015) Effectiveness of existing noise
barriers: comparison between vegetation, hollow block concrete, and concrete panel. Proc
Environ Sci 30:217–221
Mapping Literature of Reclaimed
Asphalt Pavement Using Bibliometric
Analysis by VOSviewer
Mochammad Qomaruddin , Han Ay Lie , Widayat ,

Bagus Hario Setiadji , and Mochamad Agung Wibowo
Abstract The purpose of this research is to map the literature on reclaimed asphalt
pavement (RAP) and to determine future related potential topics. Data were collected
from the Scopus meta-data and analyzed using the bibliometric analysis technique
with the VOSviewer tool. The result showed that there are several relationships
between RAP and the topics related to aggregate, mixing, hot-mix asphalt, and
compressive strength. However, some topics such as extraction, microstructural prop-
erties, furnaces, and interfacial transition zone in the RAP have not been widely
researched. Numerous studies have been carried out on RAP by authors such as
Xiao F, Amirkhanian S N, Daniel J S, Canestrari F, Huang B, Zaumanis M, and
Arulrajah. Some of the countries that have contributed significantly to RAP research
are the United States of America, China, Italy, and India.
Keywords Reclaimed asphalt pavement · Bibliometric · VOSviewer
M. Qomaruddin (B) · H. A. Lie · B. H. Setiadji · M. A. Wibowo

Department of Civil Engineering, Universitas Diponegoro, Semarang, Indonesia
H. A. Lie
B. H. Setiadji
M. A. Wibowo
M. Qomaruddin
Department of Civil Engineering, Universitas Islam Nahdlatul Ulama, Jepara, Indonesia
Widayat
Department of Chemical Engineering, Universitas Diponegoro, Semarang, Indonesia
Advanced Material Laboratory Central of Research and Service Unit, Universitas Diponegoro,
Semarang, Indonesia
1086 M. Qomaruddin et al.
1 Introduction
Reclaimed Asphalt Pavement (RAP) is a waste material obtained from dredging

flexural pavement using a Cold Milling Machine [1]. Numerous studies have been
conducted [2] to determine the more efficient and economical use of RAP [3]. Topic
distribution is very broad in mapping the RAP because it has various purposes.
During data collection, the word “Reclaimed Asphalt Pavement” was entered into a
search engine, and the majority of the information that appeared was on materials,
hot mix, and concrete. However, multidisciplinary RAP research was found from
the Scopus database. The problem associated with RAP is the difficulty in mapping
the scientific development cluster, based on the relationship between the authors and
countries in this research’s citation and correspondence section. Generally, studies on
the mapping of publications related to asphalt and bitumen [3] have been carried out,
although it is not specific to RAP waste. Therefore, this research aims to explore the
topic of RAP using the bibliometric VOSviewer method for authors to identify and
acquire research gaps from its phenomena globally easily. This quantitative research
comprises systematic articles, transparently and synthetically collected, to assess
study findings on topics related to RAP [4, 5]. Bibliometric, which is a statistical
analysis method that connects text or keywords between papers [6], including the
citation relationship between authors [7], the co-occurrence of keywords with author
institutions [8, 9], and analysis was used to carry out this research [10]. An author
from Indonesia also wrote on RAP.
2 Methodology
This research analyzes the SCOPUS indexed literature on RAP using the bibliometric
method with the help of VOSviewer. This software has the ability to map keyword,
author, and citation relationships. Scopus has a large database and complete meta-
data analysis, which was downloaded on February 7, 2021. This method is used to
determine the information that contributes to answers to research questions on RAP.
Systematic Review is carried out through various stages of data collection [11, 12],
Filtering, Identifying and Interpreting on VOSviewer. The process flow is shown in
Fig. 1. The data collection stage is carried out by entering the keyword “Reclaimed
Asphalt Pavement” to obtain the associated articles. The data collection process using
the keywords “Reclaimed Asphalt Pavement” is shown in Fig. 2.
The second process in filtering needs to focus on the subject area by increasing the
relevance between the articles. This filtering is limited to journals published from
Collection Data Filtering Identifying Interpreting
Fig. 1 Scopus bibliometric data interpretation flow

Mapping Literature of Reclaimed Asphalt … 1087
Fig. 2 Scopus document results (scopus.com)
Fig. 3 Export Scopus document to excel CSV (scopus.com)
1999 to 2021. Subject areas selected are Engineering, Material science, Environ-
mental science, Earth and Planetary Sciences, Energy, Chemical Engineering, and
Chemistry. Types of documents include articles, conference papers, book chapters,
reviews, and conference reviews, while the choice of sources is taken from journals,
conference proceedings, book series, trade journals, and books. The publication stage
was selected for the final document and article in press with the filtering process
shown in Fig. 3. A total of 1994 articles were obtained from the search results using
the relevant keywords in accordance with the citation, bibliographical, and abstract.
The selected results are saved in a CSV file format, which is then used to further
analyze VOSviewer.
The identification process using Co-occurrence was summarized in CSV and
created on VOSviewer with the following steps: Create a map based on bibliographic
data → Read data from bibliographic database files → Choose the type of analysis
and counting method (Co-occurrence or Co-authorship) → choose threshold Co-

occurrence with a minimum of 5 keywords → choose the number of keywords to be
selected (500), then run analysis Visualization VOSviewer and interpretation.
3.1 RAP Publication Development
The Scopus data that is mined in relation to the publication on the topic of RAP
increases every year, as shown in Fig. 4. However, from 1999 to 2003, there was no
increase in publications on RAP by looking for asphalt content using the centrifugal
method [13] and mixed research variants [14]. RAP was explored in concrete [15,
16] after 2011 and continued to increase irrespective of the slight decrease from 2015
to 2018.
The authors that conducted the reclaimed asphalt pavement were recorded in a
total of 44 countries divided into 10 clusters, and 17 citations and corresponding
links are shown in Fig. 5. Research on RAP has been carried out in various countries,
which indicates that this topic is quite interesting and of a global nature. However,
this study is dominated in America, China, Italy, and India.
The United States of America has the largest citation and contributors on RAP
research. In addition, a total of 8 countries reinforce each other in citation between
the authors, as shown in Table 1.
The citation analysis in Fig. 6 visualizes the network between one author and
another through digitalization. The total number of authors on the topic of “reclaimed
asphalt pavement” is 239, grouped into 19 clusters and interconnected with citation
and relationship. They are further marked by differences in color and interconnected
350
Number of Publications
300
250
200
150
100
50
0
1999 2001 2003 2005 2007 2009 2011 2013 2015 2017 2019 2021
Publication Year
Fig. 4 Graph of the number of RAP publications each year

Fig. 5 Network visualization country
Table 1 Number of country clusters

Cluster country Number of countries Country name
Cluster 1 8 Egypt, Iran, New Zealand, Oman, Qatar, Saudi Arabia,
United State, Vietnam
Cluster 2 7 Czech Republic, Germany, Greece, India, Italy, San
Marino, Serbia
Cluster 3 6 Belgium, Brazil, France, Portugal, South Korea,
Tunisia
Cluster 4 4 Australia, Canada, Pakistan, Thailand
Cluster 5 4 Denmark, Japan, Malaysia, Nigeria
Cluster 6 4 Netherlands, South Africa, Sweden, Taiwan
Cluster 7 3 Chile, Poland, Spain
Cluster 8 3 Latvia, Switzerland, Turkey
Cluster 9 3 Colombo, Iraq, United Kingdom
Cluster 10 2 China, Hong Kong
Fig. 6 Overlay visualization author
You, Z.
Singh, S.
Mohammad , L.N.
Huang, B.
Author
Daniel, J.S.
Zaumanis, M.
Canestrari, F.
Arulrajah, A.
Amirkhanian, S.N.
Xiao, F
0 5 10 15 20 25 30 35
Number of Publications
Fig. 7 Graph paper per Scopus data author
by 848 networks. Cluster 1 is light blue and consists of an author named Xiao F with
a total of 32 papers.
Author contributions with research topics on RAP are identified in the network and
overlay of Figs. 6 and 7. The author with the most citations is Huang B, with a total
of 1325 citations from 21 articles. The year 2021 is expected to have a significant
increase in the number of new authors interested in the topic of RAP around the
world.
Fig. 8 Network visualization RAP
3.2 Analysis of Research Topics Related to RAP
Figure 8 is a visualization of the article network related to RAP which consists of 396
relevant keywords divided into 10 clusters (color differences) interconnected with
18,569 links. Based on the observations, related studies on RAP have been widely
conducted. However, few topics intersect with RAP, such as extraction, chemical
properties, and microstructure. The blue cluster was more significant in this research
with 123 related keywords on aggregate, mixing, hot-mix asphalt, and compressive
strength relatively high.
The special network on the topic of “reclaimed asphalt pavement” has 3 clusters,
with 395 links in several 1629 papers, as shown in Fig. 8. Meanwhile, the number of
related papers accumulated in the Scopus database is 1994. This shows that there is
a relationship close to the word “asphalt” and “pavement,” which is associated with
“reclaimed asphalt pavement” or “recycled asphalt pavement” with a total of 1629
papers.
The authors were able to find research gaps in linking networks that had few
links between keywords such as extraction, microstructural properties, furnaces,
and interfacial transition zone. For example, the word “extraction” in Fig. 9 used
to visualize the relationship between the words “reclaimed asphalt pavement” and
“extraction” is found in 16 papers out of 1994. The next step is to download these 16
Fig. 9 Network visualization RAP with extraction
papers by breaking and technically describing their state of the art to fill the search
gaps easily.
4 Conclusion
In conclusion, Scopus meta data and VOS Viewer can be used to map the bibliometric
network on RAP. These tools are also used to describe the relationship between
authors’ distribution, countries, and related keywords.
Acknowledgements This research was financially supported by The Faculty of Engineering,

Diponegoro University, Indonesia Through Strategic Research Grant 2021.
References
1. Widayanti A, Soemitro Ria AA, Ekaputri JJ, Suprayitno H (2018) Characterization of reclaimed
asphalt pavement (RAP) as a road pavement material (National Road Waru, Sidoarjo). MATEC
Web Conf 181. https://doi.org/10.1051/matecconf/201818105001
2. Al-Qadi IL, Elseifi M, Carpenter SH (2014) Reclaimed asphalt pavement—a literature review
3. Chang X, Zhang R, Xiao Y, Chen X, Zhang X, Liu G (2020) Mapping of publications on asphalt
pavement and bitumen materials: a bibliometric review. Constr Build Mater 234:117370.
https://doi.org/10.1016/j.conbuildmat.2019.117370
4. Callahan JL (2014) Writing literature reviews: a reprise and update. Hum Resour Dev Rev
5. Sweet M, Moynihan R (2007) Improving population health: the uses of systematic reviews,
Dec 2007
6. Tsay M, Shu Z (2011) Journal bibliometric analysis: a case study on the journal of documenta-
tion. J Doc 67(5):806–822 [online]. Available http://dblp.uni-trier.de/db/journals/jd/jd67.html#
TsayS11
7. Franceschini F, Maisano D, Mastrogiacomo L (2015) Influence of omitted citations on the
bibliometric statistics of the major Manufacturing journals. Scientometrics 103(3):1083–1122.
8. Ravikumar S, Agrahari A, Singh SN (2015) Mapping the intellectual structure of scientomet-
rics: a co-word analysis of the journal scientometrics (2005–2010). Scientometrics 102(1):929–
955 [online]. Available https://econpapers.repec.org/RePEc:spr:scient:v:102:y:2015:i:1:d:10.
1007_s11192-014-1402-8
9. Cainelli G, Maggioni MA, Uberti TE, de Felice A (2015) The strength of strong ties: how
co-authorship affect productivity of academic economists? Scientometrics 102(1):673–699.
10. Zhai L, Yan X, Shibchurn J, Song X (2014) Evolutionary analysis of international collaboration
network of Chinese scholars in management research. Scientometrics 98(2):1435–1454. https://
doi.org/10.1007/s11192-013-1040-6
11. Petticrew M, Roberts H (2006) Starting the review: refining the question and defining the
boundaries. In: Systematic reviews in the social sciences, pp 27–56, Jan 01 2006. https://doi.
org/10.1002/9780470754887.ch2
12. Huang C, Yang C, Wang S, Wu W, Su J, Liang C (2019) Evolution of topics in education
research : a systematic review using bibliometric analysis. Educ Rev:1–17.https://doi.org/10.
1080/00131911.2019.1566212
13. SNI-03-6894 (2002) Metode pengujian kadar aspal dari campuran beraspal dengan cara
sentrifus 1, pp 1–6
14. Taha R, Al-Harthy A, Al-Shamsi K, Al-Zubeidi M (2002) Cement stabilization of reclaimed
asphalt pavement aggregate for road base and subbases. J Mater Civ Eng 14(3):239–245
15. Qiang W, Peiyu Y, Ruhan A, Jinbo Y, Xiangming K (2011) Strength mechanism of cement-
asphalt mortar. J Mater Civ Eng 23(9):1353–1359. https://doi.org/10.1061/(ASCE)MT.1943-
5533.0000301
16. Miller SR, Hartmann T, Dorée AG (2011) Measuring and visualizing hot mix asphalt
concrete paving operations. Autom Constr 20(4):474–481. https://doi.org/10.1016/j.autcon.
2010.11.015
Literature Study: Alternative Materials
for Hot Rolled Sheet-Wearing Course
(HRS-WC) Pavement
Elsa Eka Putri, Purnawan, Bayu Martanto Adji, and Bobby Herman
Abstract HRS mixture is a concrete asphalt mixture using gaps that have a high
bitumen content so that it has high flexibility and is resistant to deformation. More-
over, this pavement is suitable for application in Indonesia because Indonesia is a trop-
ical country with high enough heat. HRS has two types of pavement layers, namely
HRS-Base and HRS-WC. As a non-structural layer containing more fine aggregate
and bitumen, the HRS-WC Mixture’s structural strength is prone to plastic deforma-
tion with rutting on the asphalt surface. Therefore, the quality of HRS-WC needs to
be improved through new specifications and the type of material used must have good
physical and mechanical properties. Efforts have been made to improve the quality
of the existing asphalt. Modifying the physical and chemical properties of asphalt
with various added materials was to obtain cheap and good quality asphalt. This
literature study aims to identify the alternative materials that can improve the HRS-
WC Mixture performance based on previous studies in terms of the Marshall test
results of related research. The collection and analysis of research data from books,
documents, journals, and research results have obtained eight alternative materials
to improve the HRS-WC Mixture’s performance, namely used tire powder, concrete
waste, tras, bagasse ash, rice husk ash, coconut shells, palm oil, natural rubber (latex)
and coal ash. On the other hand, alternative use has advantages from economic and
environmental aspects. They are replacing/reducing the use of new aggregates and
reducing the amount of waste that pollutes the environment.
Keywords HRS-WC · Alternative materials · Marshall parameter · Waste
E. E. Putri (B) · Purnawan · B. M. Adji · B. Herman

Department of Civil Engineering, Faculty of Engineering, University of Andalas, Padang,
Indonesia
Purnawan
B. M. Adji
1096 E. E. Putri et al.
1 Introductions
Hot Rolled Sheet (HRS) is a hot asphalt mixture developed by Bina Marga around
1980, which has durability and flexibility [1]. HRS or Lataston is a concrete asphalt
mixture that uses gaps-graded aggregates and has a high bitumen content [2]. The
HRS-WC layer is non-structural which contains fine aggregate and high bitumen
content. Therefore, the HRS-WC structure is prone to plastic deformation, which is
indicated by the appearance of grooves on the pavement surface [3].
Modification of the physical and chemical properties of the asphalt mixture is
one of the efforts that have been implemented to improve the quality of an asphalt
mixture. Modifications include adding additives or replacing ordinary materials with
better materials to obtain excellent and cheap quality and performance of the asphalt
mixture. As a result, the government’s funds for the manufacture and repair of the
asphalt mixture are becoming more economical [4].
This literature study will discuss alternative materials for the HRS-WC Mixture
based on previous related research to improve the asphalt mixture’s performance.
Such as a substitute for coarse aggregate, fine aggregate and a substitute for asphalt
as a binder.
2 Literature Review
2.1 Hot Rolled Sheet
Hot Rolled Sheet (HRS) or a thin layer of asphalt concrete (Lataston) is a mixture
of road pavements, forming material consisting of asphalt, gaps graded aggregate,
and filler spread and compacted at a specific temperature. The thickness of the layer
ranges from 2.5 to 3 cm [5]. According to the Transportation Infrastructure Research
and Development Center, 2005 there are two HRS mixtures, namely HRS-Base and
HRS-WC.
HRS-WC is a type of HRS pavement layer that is at the top or on the surface. HRS-
WC has water-resistant coating properties, is resistant to the formation of rutting, has
a good surface smoothness, can transmit loads, and is resistant to slipping. The HRS-
WC layer is prone to wear because it directly contacts the vehicle’s wheels and the
weather’s effects. The appearance of cracks on the surface is caused by evaporation
and ageing of asphalt due to temperature and climate. Therefore, the HRS-WC layer
must be designed to have a good level of stability, flexibility, durability and resistance
to fatigue [6].
There are some characteristics after modification of the HRS-WC asphalt mixture.
Sufficient levels of asphalt to produce a durable pavement, high stability to withstand
traffic loads without changing shape, sufficient void in total pavement mixture after
compacting [1]. It can provide space for extra compaction and the development of
the asphalt due to increasing the temperature. Moreover, maximum percentage of
Literature Study: Alternative Materials … 1097
air voids to limit permeability because water can cause damage if it gets into the
pavement, as well as having high workability so that becomes more efficient without
shifting between pavement layers and without reducing the stability of the Mixture
during spreading.
2.2 Marshall Parameters
Asphalt concrete performance is determined through Marshall testing, should fulfill

the following result, stability is the strength of the Mixture to withstand deforma-
tion due to traffic loads. Stability can be obtained through inter-aggregate friction
resistance, interlocking aggregates, and asphalt’s cohesion power [7].
Flow is the magnitude of the change in the plastic shape of an asphalt mixture test
object that occurs due to a load to the limit of collapse, expressed in units of length
(Department of Public Works 2010) [8].
Marshall Quotient is the quotient of stability to flow, which is used to approach
the level of stiffness or flexibility of the Mixture. A high Marshall Quotient value
indicates a high hard layer stiffness value. Hard layers with too high a Marshall
Quotient value will easily crack due to repeated traffic loads. On the other hand,
the Marshall Quotient value, which is too low, indicates the Mixture is too flexible
(plastic), which results in the hard layer being easily deformed if it withstands traffic
loads [9].
Void in the Mineral Aggregate (VMA) is the voids space between the aggre-
gate particles on pavement, including the air voids and the effective asphalt volume
(excluding the volume of asphalt absorbed by the aggregate). Void Filled with Asphalt
(VFA) is the per cent of the voids between the aggregate particles (VMA) filled with
asphalt, excluding asphalt, which is absorbed by the aggregate. Void in Mix (VIM)
is the air voids in the asphalt pavement mixture consisting of air spaces between
aggregate particles covered with asphalt.
3 Methodology
This literature study focuses on finding various ideas, theories, laws, propositions,
principles, or ideas used to analyze, complete and formulate questions in according
to the research object. Through literature research or literature review, knowledge,
ideas and findings contained in academic-oriented literature are summarized and
theoretical and methodological contributions to the topics are formulated [10].
As well as in a descriptive analysis, where the data obtained is presented and
described regularly and given explanations and insights, the research can be easily
understood by readers.
The data collection method used in this research is the documentation method. The
documentation method is a data collection method by finding or extracting data from
literature related to the formulated problem [10]. The data obtained from various
literatures are collected as a single document and used to answer the problems that
have been formulated.
The four procedures used in this research are;
1. Organize, namely organizing the literature to be reviewed. The literature must
be relevant and in accordance with the problem statement. It was conducted
by looking for ideas, general goals, and conclusions from the literature by
reading the abstract, introductory paragraphs and conclusions, and classifying
the literature based on certain categories.
2. Synthesize, which is to combine the results of the literature organization into
a summary so that they become a whole, by looking for linkages between the
contents.
3. Identify, namely identifying controversial issues in the literature. The controver-
sial issue in question is an issue that is considered very important to be explored
or analyzed, in order to get an interesting article to read.
4. Formulate, namely formulating questions that require further research.
4.1 Used Tire Powder
Used tire powder is one of the added materials (additives) that is expected to improve
the asphalt mixture’s quality to fulfil the characteristics of asphalt as a binder and
increase the stability, flexibility, and durability of the asphalt concrete mixture [11].
The use of used tire powder as an added material in improving the quality of asphalt is
motivated by the abundant availability of used tire waste in Indonesia. So, in addition
to improving the quality of the asphalt mixture, the use of used tire waste can also
minimize environmental pollution due to waste [12].
The stability value of the HRS-WC Mixture without used tire powder is higher
than the stability value with the addition of used tire powder [13]. HRS-WC stability
value without using used tire powder is higher than HRS WC with the addition of
used tire rubber but still meets specifications [14].
The flow value of HRS-WC with the addition of used tire powder is higher than
the typical HRS-WC Mixture. In contrast to the research, it stated that the flow value
at HRS-WC without the addition of used tire powder is higher when compared to
HRS-WC with the addition of used tire powder [14]. The variation in the size used
in Hardwiyono’s study was 20, 21 and 22%. The flow value will increase to the
maximum flow value limit, and the flow value will decrease with increasing levels
of used tire powder [11]. The MQ value tends to decrease along with the increase in
the content of the used powder used. It is confirmed that the HRS-WC Mixture with
the addition of used tire powder shows a lower MQ value compared to the HRS-WC
Mixture without the addition of used tire rubber powder waste [15].
The VIM value tends to decrease with increasing levels of used tire powder. The
Voids in the Mixture (VMA) will decrease along with the increasing levels of used
tire powder. The VMA value in all used tire powder mixtures has decreased [16].
The VFA value tends to increase with increasing levels of biofuels powder used. The
research of Faisal (2014) supported that there was an increase in the VFA value and
the increase in the number of biofuels powder, although the increase was not too
significant [17].
It can be concluded that used tire powder can be used as additional alterna-
tive material to improve the performance of the HRS-WC Mixture [13]. It is also
supported by Jatmiko’s research (2019) that the used tire powder is suitable for use
in the HRS-WC asphalt mixture. The asphalt mixture is more elastic, flexible and
the stability will decrease with increasing levels of used tire powder [18].
4.2 Concrete Waste
Concrete waste is a waste product of unused construction materials, which is very

easy to find in Indonesia. Concrete waste originates from the ruins of buildings due to
earthquakes, demolition of buildings and concrete originating from precast concrete
factories, which are no longer used and, if left without handling, can cause problems
for the environment [19].
The physical properties of Concrete Waste are almost as good as, in essence, that
of new aggregates (stones). Based on this equation, the rest of the buildings with
concrete waste can be used as an alternative to coarse aggregates in the construction
of road pavement mixtures [20].
In doing so, the use of new aggregates might be reduced by using aggregates from
recycled concrete waste [21].
In the research of Suwastika et al. (2019), the variations of the desired level used
to obtain the Optimum Asphalt Content (OAC) were 6, 6.5, 7, 7.5, 8%, from the
OAC test results obtained at 7.13% in Mixture using natural aggregate and aggregate
7.08% in the Mixture with concrete waste as coarse aggregate. Then a mixture is
made with variations of 25% natural aggregate and 75% concrete waste, 50% natural
aggregate, 50% concrete waste and the last variation is 75% natural aggregate 25%
concrete waste [19]. The stability value tends to decrease in the use of high levels of
concrete waste. Moreover, the stabilization value is decreasing with each addition of
variations in the bio-concrete waste, which is currently large. The worse ability than
new aggregate is the cause of the decrease in the value of stability in using concrete
waste in the asphalt mixture [19, 21].
The test results show that the flow value has increased from 3.3 to 3.75 mm in
the variation of Concrete waste content from 50 to 100%. Then there is a decrease in
the variation of the concrete waste content of 0–25%, namely 3.8–3.5 mm, that the
value of Flow tends to increase in higher levels of waste. The reason is that concrete
waste has the greatest absorption properties, so mixing with more waste concrete
will easily absorb more asphalt than the mixture with less concrete waste content
[20].
The VIM value has increased as the use of concrete waste content in the mixture
increases. Utomo et al. (2017) confirmed that the higher the level of concrete waste
used, the higher the VIM value obtained will be as well [21].
The test results show that the VMA value tends to increase when the waste concrete
is higher and with the higher variation in the amount of concrete waste, the VMA
value tends to increase due to the high abrasion value of concrete waste [21].
The MQ value decreases along with the increase in the waste content used. As
well as at the higher level of concrete waste, the results of the Marshall test show
the MQ value tends to decrease. The higher the level of concrete waste used, the
VFA value will decrease. The greater the variation of concrete waste, the VFA value
tended to decrease. Asphalt, which is supposed to fill the voids in the mixture, is
absorbed by waste which has a higher absorption rate than new aggregate, which
causes a decrease in the VFA value [21].
Based on the Marshall parameter value obtained from the resolved test results, it
can be concluded that the use of concrete waste aggregates in the HRS-WC Mixture
can still be done because it still meets the General Specifications of Bina Marga 2018
[19].
4.3 Tras
Volcanic rocks that have undergone changes in chemical composition as a result of

the weathering process and affect the water conditions of the soil are called tras [22].
Tras or commonly referred to as natural polypozzolans, is a natural material that
occurs due to the weathering of volcanic rocks, which have a high content of silica
(SiO2 ) and aluminium and alkaline iron compounds chalk, and others that are not in
the inconsiderable amount [23].
It was investigated by varying the coarse, medium and fine aggregates including
sand and tras with a composition according to specifications. The mixture composi-
tion is the proportion of 10% coarse aggregate, 20% of dry aggregate, 50% of fine
aggregate, 10% of sand and 10% of tras and 6% of asphalt percentage [24].
It concluded that with a composition of 10% coarse aggregate, 20% aggregate,
50% fine aggregate, 10% sand and 10% Tras, they are in ideal limits because they are
still within limits required in the specification [22]. Then, at the maximum asphalt
level of 6%, the stabilization value and VIM are far from the minimum limit, the
values of the Marshall test results can still be optimized by trying to re-vary the
use of Tras and asphalt content that are still sufficient so that they can support the
performance of the HRS-WC Mixture. Moreover, the use of raw materials affects
the amount of use of aggregates where the percentage of use of medium and fine
aggregates is reduced by 10–20%. So, there are material and cost savings for the
manufacture of the HRS-WC Mixture.
Hence, Tras can be used as an additive material in the smooth aggregate diffrac-
tion in the hot HRS-WC semi-heated mixture because it can produce a good and
economical mixture while still meeting the specifications [22].
4.4 Bagasse Ash
The sugar factory produces waste in the form of sugarcane waste. In the sugar factory,
sugar cane is used for burning material in the factory’s steam boiler. The results of
burning sugarcane waste will produce ash which can be used as a filler in a mixture
of asphalt concrete [24].
The main content of bagasse ash is Silica (SiO2 ) 70.97% and Na2 O of 22.27%, and
a few other elements such as K2 O of 4.82%, C5 H8 O4 of 0.9%, Magnesium (MgO)
of 0, 82%, Al2 O3 and Al2 O3 each of 0.36%. The content contained in bagasse ash,
when mixed with asphalt, will form a compound reaction that makes the structure of
the Mixture hard and stiff [25].
In the field of road engineering, the use of bagasse ash continues to be optimized in
order to reduce the impact of environmental pollution, improve mixed performance
and at the same time increase the economic activities of the community in an area
[26].
A study on the use of bagasse ash on HRS-WC mixtures’ performance by using
sugar cane ash as a filler and compared it to rock dust filler. From comparing the
Marshall HRS WC Mixture Test with Rock Ash Filler and Sugarcane Ash Filler,
conclusions can be drawn as follows, the optimum asphalt content in the rock ash
filler is 6.75%, while bagasse ash is 6.95%. It indicates that the filler composition
affects the optimum asphalt content. This difference indicates that the absorption
with bagasse ash filler is greater than that with rock ash filler. The effect of using
bagasse ash filler on the VMA value at the optimum asphalt content, the VMA value
in the Mixture with rock ash filler was 20.64%, smaller than the bagasse ash filler of
21.36% [24].
Using bagasse ash filler on the VIM value on the optimum asphalt content shows
that the VIM value of rock ash is 5958% and bagasse ash is 5960%, resulting in
slightly large voids so that the durability of the pavement layer is reduced. The effect
of using bagasse ash filler on the VFB value at the optimum asphalt content shows
that the VFB value of rock ash filler is 71.68%, and bagasse ash is 72.12%, resulting
in small air voids, the greater stiffness of pavement the higher the durability.
Using bagasse ash filler on the stability value of the optimum asphalt content shows
that the stability value of rock ash filler is 1280.98 kg and bagasse ash is 1231.07 kg,
resulting in the asphalt mixture able to withstand traffic loads. The effect of using
bagasse ash filler on the flow value at the optimum asphalt content shows that the
flow value in rock ash filler is 4.76 mm and bagasse ash is 4.72 mm, resulting in the
pavement layer not easily cracked.
The effect of using bagasse ash filler on the Marshall Quotient value on the
optimum asphalt content shows that the MQ value of rock ash filler is 268.97 kg/mm
and bagasse ash is 260.58 kg/mm.
The results of the analysis indicate that the use of bagasse ash filler with HRS-WC
mixture meets the requirements as per General Specifications of Bina Marga 2010
as an alternative material to replace rock ash filler. Hence, bagasse ash waste as well
rock ash filler can be used for HRS-WC asphalt concrete mixture [24].
4.5 Rice Husk Ash
Rice husk ash is a local material that is available in large quantities and relatively
cheap. The high silica content makes rice ashes suitable for use as a substitute filler
in asphalt pavement mixtures [26].
Rice husk ash contains more than 70% silica thus it categorized as the pozzolanic
material. It is a popular material used as an additive in the concrete making, especially
in increasing concrete strength. Silica will react with cement and water to form
calcium silicate hydrate as an adhesive [27]. It is possible to use rice husk ash waste
to replace rock ash for the asphalt mixture on the surface layer. The materials used
are 60/70 penetration bitumen, coarse aggregate, fine aggregate and filler. The use of
husk ash filler can increase the density of the mixture so that the contact area between
the aggregates has increased and the voids in the Mixture are reduced. The increase
in interlocking between aggregates causes the stability value tends to increase. At
8.6% asphalt content, the stability value reaches a maximum value of 2355.7 kg [28].
From the Marshall test results, it can be seen that at the asphalt content of 9.6%,
the flow value reaches a maximum value of 7.87 mm. It proves that the flow value
continues to increase with the increase in the asphalt content used.
The test results show that the VMA value has decreased as the use of asphalt
content increases. In soaked conditions, the minimum VMA value occurs at the
asphalt content of 9.1 at 22.56% and asphalt content of 9.6%. The VMA value has
increased again to 23.02%. Then, the higher the asphalt content used in the mixture,
the higher the VFA value. At 9.6% asphalt content, the maximum VFA value was
90.68 for dry conditions and 86.06 for soaked conditions. The VIM value decreases
as the amount of asphalt used in the mixture increases. Increasing the level of asphalt
used, the MQ value will increase; at the asphalt content of 8.6%, the MQ value
reaches a maximum of 330.10 kg/mm.
The utilization of rice husk ash in the HRS-WC mixture has advantages, such
as increased stability, durability, and flexibility. However, it also has disadvantages,
namely the use of asphalt, which tends to be higher due to the nature of rice husk
ash, which easily absorbs asphalt [27]. Other advantages of the use of raw water as a
filler material include the availability of raw milk as a raw material, which gives the
prospect of the availability of filler material which is more economical compared to
other materials that are relatively expensive and usually difficult to obtain [29].
4.6 Oil Palm Shells
Another alternative material for coarse aggregate substitution in the asphalt mixture
is the waste of oil palm shells. Oil palm shell waste contains a high amount of grit
(SiO2 ), increasing the strength of asphalt concrete [30]. Oil palm shells have a very
thick and hard skin structure and contain high levels of grit (SiO2 ) [31]. Palm shell
waste is available in large quantities at a more economical price [32].
The utilization of oil palm shell waste is generated from oil palm processing. One
way to utilize palm oil waste is to use palm kernel shells to substitute some of the
primary aggregates in an asphalt mixture [31].
The aims of their investigation was to determine the strength and durability of palm
oil waste as a substitute for coarse aggregate in the HRS-WC Mixture. The param-
eters reviewed in this study are Marshall parameters and Wheel Tracking testing.
Variations in the levels of waste oil palm shells used were 0, 30, and 50% [30].
In the variation of 0 and 30% oil palm shell waste, stability values are above the
minimum required Marshall specifications, namely 988.48 and 869.67 kg. Mean-
while, for the variation of oil palm shell waste, 50% of the stability value does not
meet the minimum requirements set, which is 800 kg. So, it can be concluded that
the recommended use of oil palm shell is only at 30% oil palm shell content.
All the oil palm shell content variation in the HRS-WC Mixture still meets the
minimum specifications of flow value of 3 mm. The HRS-WC Mixture with oil
palm shell content of 0 and 30% meet the minimum MQ specifications required,
while the Mixture with an oil palm shell content of 50% does not meet the minimum
requirements for the MQ value of a minimum of 250 kg/mm.
At 0% oil palm shell waste, the VIM value obtained was 5.76%. At the 30%
variation in oil palm shell content, the VIM value is still within the required minimum
specification limit, which is 5.88%. The HRS-WC Mixture at 0 and 30% oil palm
shell content has good durability properties. Whereas in the variation of oil palm shell
content, 50%, the VIM value obtained was 6.13%. This value exceeds the required
maximum value limit of 6%.
From the above research results, it can be concluded that for the use of oil palm
shells with a content of 30% in the HRS-WC pavement mixture, the values of
stability, Flow, Marshall Quotient and Void in Mix are within the required range
of SNI 03-1737–1989. In comparison, the use of oil palm shells at a level of 50% as
coarse aggregate is not recommended because the test results show that the Marshall
parameter values are outside the predetermined specification range.
4.7 Latex
Latex or a natural rubber comes from the sap of the Hevea Brasiliensis tree, which
comes from the Amazon region. It is formed by clumping and drying [33]. Natural
rubber (polysoprene) is classified as an elastomer, a material that can be stretched

and can return to its original form [34].
Natural rubber functions to increase bonds between aggregates, provide elasticity,
and reduce vibrations that arise due to loads [35].
The use of natural rubber in road pavement, especially in the HRS-WC mixture,
has been widely studied by researchers and experts.
The lumpy rubber (lumb), namely rubber or latex, which is directly formed into
crepe (solid rubber) without being given any treatment. Using the same bitumen
content at 7.3% for a standard mixture of HRS-WC specimens was made without the
addition of natural rubber. The comparison between the standard HRS-WC Mixture
with the HRS-WC Mixture plus raw rubber (natural rubber) was obtained with a
variation of 5, 10, and 15% [33].
Moreover, the HRS-WC Mixture with the addition of raw natural rubber has
higher Stability, Flow, VFA, and Marshall Quotient values than the HRS-WC mixture
without the addition of raw natural rubber. However, the HRS-WC Mixture’s VIM
and VMA values with the addition of raw natural rubber are lower than the standard
HRS-WC mixture without the addition of raw natural rubber.
The addition of natural rubber to oil asphalt in the HRS-WC mixture shows a
better Marshall stability value, indicating that the interlocking between aggregates is
getting better, the flow value is getting lower, and the Marshall Quotient is increasing
[2].
The asphalt mixture with the addition of natural rubber has a higher strength to
withstand loads than conventional asphalt mixtures and proved that using rubber
asphalt in the mixture can increase the stability [36].
4.8 Coal Ash
In general, coal ash can be interpreted as the residual or contaminant material of coal
which is not burnt and involve in the combustion process of carbon, hydrogen, sulfur,
oxygen and evaporation of water contained in coal. There are two types of coal ash,
fly ash and bottom ash [3].
Coal ash is very easy to find in Indonesia. It is available in large quantities, which
is the smallest part of a mixture of solids produced from processing coal in mining,
which is no longer used. For this reason, coal ash will be used in the making of a road
pavement mixture [37]. The use of coal ash in the road pavement, especially in the
HRS-WC mixture, has been widely studied. It was conducted to determine the effect
of fly ash on the performance of the HRS-WC mixture. Samples have been made
using a variation of fly ash 0, 6, 10 and 12% of the total aggregate mixture, with a
total sample of 60 samples. It can be concluded that the higher the ash content of the
coal used results in a decrease in the value of stability, flow and MQ. However, the
stability, flow and MQ values obtained still meet the specifications. The use of fly ash
as a filler in the HRS-WC mixture can still be used. The use of coal ash as a filler in
the HRS-WC mixture has a positive effect on the stability and the residual strength
index of the mixture using conventional fillers, namely rock ash alone. Hence, coal
ash is suitable for use as an alternative filler material in the HRS-WC mixture. The
higher the ash content of the coal used in the mixture, the values of VIM and VMA
tend to increase, while the values of Stability, Flow, and MQ decrease [3].
5 Conclusions
Based on the analysis and discussion that has been described. In this literature study,
conclusions can be drawn:
1. Several alternative materials were obtained that could be used to increase the
HRS-WC Mixture in terms of the Marshall test results. Among them are used
tire powder, concrete waste, Tras, bagasse ash, rice husk ash, palm oil shells,
natural rubber (latex) and coal ash.
2. From several alternatives obtained, the most recommended alternative for use
in the HRS-WC mixture is bagasse ash. This is based on the results of research
that shows all Marshall parameters meet specifications, and almost all Marshall
parameters have increased from the standard HRS-WC mixture. Sugarcane ash
has a very high silt content, so it is suitable for pavement mixtures. Besides
that, the material is also easy to obtain. If it is used, it will reduce the impact of
environmental pollution because there is no large-scale utilization of bagasse
ash waste.
3. In terms of cost, making mixtures using alternatives is more economical because
it utilizes waste, thereby reducing/replacing new aggregates (fresh aggregate).
Acknowledgements This research was supported by Engineering Faculty of University of Andalas

grant no. T/14/UN.16.17/PT.01.03/IS-RD/2021.
References
1. Darunifah N (2007) Pengaruh Bahan Tambahan Karet Padat Terhadap Karakteristik Campuran
hot rolled sheet wearing course (HRS-WC) (the effect of solid rubber additives on the
characteristics of the hot rolled sheet wearing course). Pascasarjana Universitas Diponegoro
2. Amiruddin AA, Sasmita SA, Ali N, Renta I (2012) Kajian Eksperimental Campuran HRS-WC
Dengan Aspal Minyak Dan Penambahan Aditif Lateks Sebagai Bahan Pengikat (experimental
study of HRS-WC mixture with oil asphalt and addition of latex additive as binder) proceeding
Konferensi Nasional Teknik Sipil, 6
3. Ambarwati L, Arifin MZ, Bawono H (2012) Pengaruh Kadar Abu Batubara Sebagai Filler
Terhadap Karakteristik dan Indeks Kekuatan Sisa (IKS) pada Campuran Hot Rolled Sheet
(HRS) (Effect of Coal Ash Content as a Filler on Words and Residual Strength Index in Hot
Rolled Sheet Mixtures). Rekayasa Sipil 3(2):131–140
4. Roberts FL, Kandhal PS, Brown ER, Lee D-Y, Kennedy TW (1991) Hot mix asphalt materials,
mixture design and construction
5. Sukirman S (1999) Perkerasan Lentur Jalan Raya (Highway Flexible Pavement) Nova,
Bandung, 2
6. Azizah N, Rahardjo B (2017) Kinerja campuran hot rolled sheet-wearing course (HRS-WC)
dengan filler abu ampas tebu (performance of a hot rolled sheet-wearing courseix with bagasse
ash filler). Bangunan: Teori, Praktek, Penelitian, dan Pengajaran Teknik Bangunan 22(2)
7. Sulaksono S (2001) Rekayasa Jalan (highway engineering), Departemen Teknik Sipil, Penerbit
ITB, Bandung
8. Departemen Pekerjaan Umum.: Spesifikasi Umum Binamarga 2010 Revisi 3 (General Spec-
ifications Binamarga 2010 Revision 3). Dinas Pekerjaan Umum Bidang Binamarga, Jakarta
(2010)
9. Suryaman F (2009) Pengaruh Penggunaan Limbah Botol Plastik Sebagai Bahan Tambah (Addi-
tive) Terhadap Karakteristik Beton Aspal (the effect of use of plastic bottle waste as an addi-
tive on the characteristics of asphalt concrete) In: Skripsi teknik sipil Universitas Atma Jaya.
Yogyakarta
10. Dena T (2013) The literature review: a few tips on conducting it. Health Sciences Writing
Centre, University of Toronto
11. Septiawan TD (2018) Pengaruh Penggunaan Bahan Tambah Serbuk Karet Ban Pada Campuran
Lapis Aspal Beton (the influence of using additional material for tire rubber powder on concrete
asphalt mixture), Jurnal Rekayasa Sipil, Univ Islam Malang
12. Indriyati EW, Susanto HA (2015) Kajian Sifat-Sifat Reologi Aspal Dengan Penambahan
Limbah Ban Bekas (study of the rheological properties of asphalt with the addition of waste
tires)
13. Oktaviastuti B, Wijaya HS, Indrawan P (2018) Pengaruh penambahan serbuk ban bekas untuk
bahan tambah campuran ATB (asphalt treated base) (the effect of adding used tire powder to
add material to the ATB mixture) Reka Buana. Jurnal Ilmiah Teknik Sipil dan Teknik Kimia
3(1):16–29
14. Hardwiyono S (2012) Pengaruh Penambahan Parutan Karet Ban Gradasi Tipe 2 Terhadap
Parameter Marshall Pada Campuran hot rolled sheet wearing course (the effect of addition of
type 2 grated rubber tire to marshall parameters in hot rolled sheet wearing course mixture)
Jurnal Ilmiah Semesta Teknika
15. Kiswara CP, Saiful AA (2014) Studi Perilaku Campuran Lataston (HRS-WC) Dengan “Limbah
Karet Ban Luar” Sebagai Bahan Pengganti Sebagian Aspal (behavior study of mixed lataston
(HRS-WC) with “waste rubber tire” as a substitute for partial asphalt), Media Teknik Sipil
16. Gosali G, Jaya H, Wulandari PS, Padmadjaja H (2016) Pengaruh Penambahan Serbuk Ban
Karetmesh# 80 Pada Campuran Laston Untuk Perkerasan Jalan Raya (the effect of addition
of Rubbermesh # 80 tire powder to Laston mixture for highway pavement). Jurnal Dimensi
Pratama Teknik Sipil 5(1)
17. Faisal S (2014) Karakteristik Marshall Campuran Aspal Beton AC-BC Menggunakan Material
Agregat Basalt Dengan Aspal Pen. 60/70 Dan Tambahan Parutan Ban Dalam Kendaraan Roda
4 (characteristics of marshall asphalt concrete mixture AC-BC using basalt aggregate material
with asphalt pen. 60/70 and additional grater tires in 4-wheeled vehicles) Jurnal Teknik Sipil
3(3):38–48
18. Jatmiko M (2019) Pengaruh Penambahan Limbah Serbuk Ban Luar Sebagai Bahan Pengganti
Sebagian Agregat Halus Pada Campuran HRS-WC (hot rolled sheet-wearing course) Ditinjau
Dari Karakteristik Marshall (the effect of addition of outer tire powder waste as a substitute
for partial fine aggregate in the HRS-WC (hot rolled sheet-wearing course) mixture in terms
of marshall characteristics), Skripsi Tesis, ITN Malang
19. Suwastika PMW, Wedyantadji B, Erfan M (2019) Pemanfaatan Limbah Beton Sebagai Peng-
ganti Agregat Dalam Campuran Lapis Tipis Aspal Beton (hot rolled sheet-wearing course,
HRS-WC) (utilization of concrete waste as a substitute for aggregates in an HRS-WC mixture),
Skripsi Tesis, ITN Malang
20. Rahman ZA, Harnaeni SR (2016) Pemanfaatan Limbah Beton Pada Campuran hot rolled sheet
base Ditinjau Dari Aspek Propertis Marshall (utilization of concrete waste in hot rolled sheet
base mixtures in terms of marshall properties), Universitas Muhammadiyah Surakarta
21. Utomo B, Harnaeni SR (2017) Analisis Marshall properties asphalt concrete Dan hot rolled
sheet Menggunakan Limbah Beton Sebagai Pengganti Agregat Kasar (analysis of marshall
properties of asphalt concrete and hot rolled sheet using concrete waste as a substitute for
coarse aggregate). Universitas Muhammadiyah Surakarta
22. Manoppo MRE (2011) Pemanfaatan Tras Sebagai Filler Dalam Campuran Aspal Panas HRS-
WC (utilization of tras as filler in hot asphalt mixture HRS-WC). Jurnal Ilmiah Media Eng
1(2)
23. Achmad F, Maksud R (2014) Kajian Penggunaan Tras Lompotoo Sebagai Agregat Halus Pada
Lapis Pondasi Bawah Ditinjau Dari Spesifikasi Umum 2007 dan 2010 (study of the use of
Lompotoo tras as fine aggregate in the subbase layer)
24. Wijanarko FH, Mulia MB, Wicaksono Y, Purwanto D (2013) Pemanfaatan Limbah Ampas
Tebu Sebagai Pengganti Filler Untuk Campuran Aspal Beton Jenis HRS-WC (utilization of
sugarcane dregs waste as a filler substitute for HRS-WC for asphalt concrete mixtures) Jurnal
Karya Teknik Sipil 2(2):363–369
25. Fauziah M, Kushari B, Ranski F (2014) Pengaruh Abu Ampas Tebu Sebagai Filler Pengganti
Terhadap Karakteristik Marshall Campuran Superpave (effect of sugarcane bagasse ash as a
substitute filler on Marshall characteristics of superpave)
26. Hadiastari IG, Fauziah M (2018) Kinerja Campuran Split Mastic Asphalt (SMA) 0/11 Dengan
Filler Abu Sekam Padi Akibat Lama Rendaman Air Laut (performance of split mastic asphalt
(SMA) 0/11 with rice husk ash filler due to long seawater soaking)
27. Susanto HA, Indriyanti EW, Edison B (2014) Permeability Campuran Hot Rolled Sheet
Wearing Course (HRS-WC) Dengan Filler Abu Sekam Padi Untuk Jalan Perkotaan (perme-
ability mixture HRS-WC with rice husk ash filler for urban roads). Jurnal APTEK 6(1):17–32
28. Ismadarni R, Kasan M (2013) Karakteristik Beton Aspal Lapis Pengikat (AC-BC) Yang Meng-
gunakan Bahan Pengisi (Filler) Abu Sekam Padi (characteristics of AC-BC using rice husk ash
filler). MEKTEK 15(2)
29. Senolinggi RD (2018) Pengaruh Penambahan Serbuk Crumb Rubber Menggunakan Filler Abu
Sekam Padi pada AC-WC ditinjau dari Karakteristik Marshall (effect of crumb rubber powder
using rice husk ash filler on AC-WC in terms of Marshall characteristics). ITN Malang
30. Putri EE, Adrian A, Hariadi M (2014) Penggunaan Limbah Cangkang Kelapa Sawit (CKS)
sebagai agregat Kasar Pada Campuran Aspal HRS-WC (use of palm oil shell waste (CKS)
as coarse aggregate in HRS-WC asphalt mixture). In: Andalas Civil Engineering (ACE)
conference
31. Ali S (2009) Karakteristik Marshall Campuran Hot Rolled Sheet (HRS) Yang Mengandung
Cangkang Kelapa Sawit Sebagai Agregat Kasar (characteristics of Marshall hot rolled sheet
(HRS) mixtures containing palm oil shells as coarse aggregates). Rekayasa Sipil 5(1):9–19
32. Mukhlis L, Suardi E, Adibroto F (2018) Kinerja Marshall Immersion Pada Campuran asphalt
concrete wearing course (AC-WC) Dengan Penambahan Cangkang Sawit Sebagai Substitusi
Agregat Halus Marshall immersion performance in asphalt concrete wearing course (AC-WC)
mixture with the addition of palm shells as a substitute of fine aggregate. Jurnal Ilmiah Rekayasa
Sipil 15(2):99–105
33. Pataras M, Dewi R, Prasetya AD, Bazidno FD (2017) Pemanfaatan Karet Mentah Pada flexible
pavement Laston AC-WC dan Lataston HRS-WC (utilization of raw rubber in flexible pavement
Laston AC-WC and Lataston HRS-WC). Cantilever: Jurnal Penelitian dan Kajian Bidang
Teknik Sipil 6(1)
34. Agusomal R, Hasan MW, Taufik (2019) Pengaruh Penggunaan Getah Karet alam Terhadap
Karakteristik Campuran hot rolled sheet wearing course (HRS-WC) (the effect of the use of
natural rubber latex on the characteristics of the hot rolled sheet wearing course (HRS-WC)).
Abst Undergr Res Faculty Civ Plann Eng Bung Hatta Uni 1(1)
35. Susanto I (2019) Pemanfaatan Karet Alam Pada Campuran Beraspal AC-WC (utilization of
natural rubber in AC-WC asphalt mixture). Prosiding SNST Fakultas Teknik 1(1)
36. Syaiful, Mulyaman S (2013) Studi Penambahan Abu Batubara sebagai Filler Pada Campuran
Beraspal (study of addition of coal ash as filler to asphalt mixtures)
37. Shalahuddin M (2014) Kinerja Abu Batu Bara Bercampur Kayu Pada Stabilitas dan Rongga
Pori Aspal Hot Mix HRS-WC (the performance of coal ash mixed with wood on stability and
cavity of hot mix asphalt pores HRS-WC). Jurnal APTEK 2(1):96–106
Marshall Characteristics of Asphalt
Mixture with Water Hyacinth Ash
as Filler
Dony Rohmad Dony, Bagus Hario Setiadji , and Bambang Riyanto
Abstract Road pavement is a quite important to support the mobilization of goods

or services. Asphalt pavement is the main type of road pavement constructed in
Indonesia. This kind of pavement consists of several materials, such as coarse aggre-
gate, fine aggregate, asphalt as binder, and filler. Cement as filler is generally used
in Indonesia, however, because of high cost, ones are more preferable to seek the
substitute of cement. One type of material that could be used as cement substitute
is water hyacinth, due to chemical properties it possesses. Water hyacinth contains
chemical cellulose from organic polymers, and in the form of ash, it has a potential
to increase the quality of asphalt. The purpose of this study was to evaluate the possi-
bility of the use of water hyacinth ash (WHA) as a part of material of asphalt mixture.
The method used was an empirical design method by conducting laboratory works
to obtain primary data. The coarse and fine aggregates used came from Gringsing
District, Batang Regency, Central Java Province, while the asphalt was the one with
penetration grade 60/70. In this study, the content of WHA used as filler was 0, 25,
50, 75 and 100% of the total weight of filler. The results showed that the increasing
use of WHA will decrease the stability and durability. Therefore, it is recommended
to limit the use of WHA to a maximum of 25% as filler of the asphalt mixture, as a
part of the implementation of green technology in practice.
Keywords Water hyacinth ash · Asphalt mixture · Stability · Durability
1 Introduction
An asphalt mixture is generally designed to be composed of constituent materials

such as aggregates, asphalt, and fillers and made using standardized methods. The
asphalt mixtures also have to be tested to ensure that the properties could fulfill
the required specifications. It indicates that to obtain a road pavement structure that
meets minimum performance standards, the required specifications to the material
D. R. Dony · B. H. Setiadji (B) · B. Riyanto

Universitas Diponegoro, Jl. Prof. Soedarto, S.H., Semarang 50275, Indonesia
1110 D. R. Dony et al.
and supported by standard manufacturing procedures should be sufficient to pro-duce

an asphalt mixture that is recognized for its quality.
However, with the reason to reduce materials that are not renewable and not envi-
ronmentally friendly, there have been many studies conducted to add additives to
an asphalt mixture with various objectives, such as increasing the mixture’s stability
and increasing the durability of the mixture. Likewise, the materials used are gener-
ally processed materials, such as rice husk ash, crumb rubber, and fly ash [1–3]. The
performance of asphalt mixtures with used material which must be removed is gener-
ally compared to that of asphalt mixtures without additive material as a reference,
with the main hypothesis is that the additive can improve the performance of one or
more properties of the asphalt mixture.
In this study, water hyacinth ash, or called WHA in the rest of the paper, processed
from water hyacinth plants, i.e., a plant that is often considered a nuisance at a
pond or lake, is used as a partial replacement of cement fillers in asphalt mixtures.
Compared to the use of WHA in asphalt mixtures, research on the use of WHA
in concrete construction has been widely carried out [4–6]. The high potential for
the use of WHA was found in work by Widyaningsih and Sutanto [7], in which a
simulation was conducted to find the optimum asphalt content of the mixtures with
Portland cement, WHA, or a combination of both as filler. However, the research
did not provide detailed analysis results regarding the suitability of the WHA as a
filler material in asphalt mixtures. Therefore, this study aims to explore in-depth the
opportunities for using WHA as filler material in asphalt mixtures in terms of the
physical aspects expressed in Marshall characteristic.
2 Water Hyacinth Ash
Water hyacinth (Eichhornia crassipes) is a floating tropical plant that grows in still or
slow-moving fresh water bodies discovered by Carl Friedrich Philipp von Martius,
a German botanist, in 1824 while on an expedition on the Brazilian Amazon River
[8] Water hyacinth has a high growth speed, so that this plant is considered a weed
that can damage the aquatic environment, as shown in Fig. 1.
Water hyacinth plants have the benefit of absorbing nutrients, metals, and trace
organic substances from water [8], but due to uncontrolled growth of water, hyacinth
causing adverse impacts, such as: (i) increasing evapotranspiration due to their broad
leaves and rapid growth; (ii) decreasing the level of dissolved oxygen due to the
reduced amount of oxygen entering the waters; (iii) accelerating the silting process,
especially by dead water hyacinth plants. In order to overcome this negative impact,
some actions are taken such as eradication using herbicides, predatory fish or lifting
the water hyacinth and using it to support other useful things; one of them is by
processing water hyacinth into water hyacinth ash (WHA) for added materials in
construction in the field of Civil Engineering.
WHA was produced by collecting, washing, and drying water hyacinth in an oven
for a certain period to convert the organic matter into an inorganic substance. The
Marshall Characteristics of Asphalt Mixture … 1111
(a) (b)
Fig. 1 Water hyacinth plant: a at Rawa Pening lake, Indonesia [9]; and b in the form of ash (WHA)
[10]
dried water hyacinth then was ground and passed through a 150-micron sieve to
enable the material used as the replacement for cement [4] (see Fig. 1b).
Based on the results of research by Sukarni et al. [11], the top five chemical
proper-ties of water hyacinth are O (49.5 ± 6.71%), C (14.4 ± 6.81%), K (8.26
± 2.62), Si (5.33 ± 4.52) and Fe (4.71 ± 4.32). After drying, one of the major
constituents found in WHA is silicon dioxide or silica [10], which is one of the raw
materials for Portland Cement. It is a strong reason why WHA has the potential
to be investigated for its use as a partial replacement for Portland Cement in Civil
Engineering structures. Although WHA has some similar constituent materials like
cement, the specific gravity of WHA is much smaller than cement, which is less than
2.5, while the specific gravity of cement is equal to or more than 3.1 [5], and this
could affect the calculation of voids in the mixture.
The methodology of this research consists of several steps as follows:

a. Testing of material properties
In this research, an asphalt mixture called asphalt course wearing course (AC-
WC) was used. The mixture was composed of dense-graded aggregate and asphalt
Penetration 60/70. All materials have to be tested to ensure that their properties could
conform to the specification used [12].
b. Mixture preparation
All materials were used to prepare the mixture briquettes in which the optimum
asphalt content of the original asphalt mixture (i.e., with no WHA) was determined
as the references for the subsequent works. Similar to Widyaningsih and Sutanto [7],
four percentages of WHA (25, 50, 75, and 100 as ratio to ordinary cement) were
selected in this study.
iii. Testing of the mixtures
The mixture were tested to determine the Marshall characteristics (Marshall
stability, flow, VMA, VIM, and VFA) as indicators of the performance of the
mixture with the addition of WHA. The mixtures also underwent a durability test by
immersing the mixtures in the water at 60° for 1, 3, and 7 days, respectively.
iv. Analysis
The analyses conducted in this study were: (i) comparison of mechanistic and
volumetric characteristics of each mixture with WHA against those of the original
mixture, (ii) deep analysis of Marshall characteristics, especially volumetric ones, to
see a possibility to answer the phenomenon occurred, and (iii) recommendation of
the use of WHA in the actual project.
Three fractions of local aggregate from Gringsing District, Central Java Province,
Indonesia, i.e., ¾ and ½ in. maximum-size coarse aggregate fraction and fine aggre-
gate fraction, and one type of asphalt, i.e., Pertamina Pen 60/70 asphalt, were used in
this study. To ensure that aggregate and asphalt properties conform to the specifica-
tions, all materials were tested using ASTM/AASHTO or SNI (Indonesian national
standards). The results of the tests are shown in Tables 1 and 2.
Table 1 Results of aggregate tests

No. Properties Result Specification
Fine aggregate
1 Passing sieve no. 200 (%) 3.31 Max. 8
2 Water absorption (%) 2.48 Max. 3
3 Specific gravity 2.57 Min. 2.5
4 Sand equivalent value (%) 98.21 Min. 50
Coarse aggregate
1 Loss Angeles abrasion value (%) 12.52 Max. 40
2 Affinity for asphalt (%) 97.9 Min. 95
3 Flakiness and elongation indices (%) 8.12 Max.10
4 Passing sieve no. 200 (%) 0.39 Max. 1
5 Water absorption (%) 1.46 Max. 3
6 Specific gravity 2.63 Min. 2.5
7 Angularity with one/two fractured faces (%) 98.21/95.49 Min. 95/90
Table 2 Results of asphalt tests

No. Properties Result Specification
1 Penetration 60/70 Asphalt (0.1 mm) 68.33 60–70
2 Softening point (°C) 50.5 Min. 48
3 Ductility (cm) 127.5 Min. 100
4 Solubility of Trichloroethylene (%) 99.67 Min. 99
5 Specific gravity 1.015 Min. 1
6 Loss on heating (%) 0.4 Max. 0.8
7 Penetration after loss on heating (%) 77.52 Min.54
8 Ductility after loss on heating (cm) 91.5 Min. 50
Tables 1 and 2 indicate that the aggregate and asphalt properties can conform to
the specifications. An aggregate gradation has to be selected carefully based on the
aggregate gradation specification used; that is, aggregate gradation for the mixture of
Asphalt Concrete Wearing Course (AC-WC) [12]. The aggregate gradation selected
in this study is shown in Fig. 2.
Using the selected aggregate gradation, the mixtures can be designed using the
following variation of WHA—PC compositions (by weight): 0–100%, 25–75%, 50–
50%, 75–25%, and 100–25%, which is the similar composition to Widyaningsih and
Sutanto [7]. The obtained optimum asphalt content and the Mar-shall characteristics
of the mixture with 0–100% composition was used as the reference for the rest of
the mixtures.
Fig. 2 Aggregate gradation used in this study

The optimum asphalt content of the mixture with 0–100% of WHA—PC com-
position is 5.5%, and the Marshall characteristics of that mixture at optimum asphalt
content are as follows: Marshall stability 1157 kg, Flow 3.77 mm, VMA 17.49%,
VIM 5%, and VFA 71.48%. All Marshall properties of the mixture can fulfill the
Indonesian specification for the AC-WC mixture [12]. Next, using asphalt optimum
content 5.5%, four sets asphalt mixture with different percentages by weight of WHA
were prepared. Five Marshall properties (Marshall stability, flow, VMA, VIM, and
VFA) were measured and calculated. The comparison among the four-set asphalt
mixtures’ property values against the reference mixture (i.e., mixture with no WHA)
is shown in Fig. 3.
Figure 3 shows that only a mixture with WHA content equals 25% among mixtures
with four different WHA content, which can satisfy all the Marshall properties. VMA
is the only Marshall property that can be fulfilled by all mixtures containing WHA.
Surprisingly, this result was supported by the similar findings by Widyaningsih and
Sutanto [7], although the similarity of Marshall property trends between this study
and the research by Widyaningsih and Sutanto [1] is only on the stability property,
i.e., the highest the WHA content, the lowest the Marshall stability will be. This
fact is also supported by the decrease of mixture density with the increase of WHA
content.
From Fig. 3, it can be summarized that the mixture with a WHA content equal or
more than 50% is not recommended due to the Marshall stability of the mixture is
lower than the threshold specified by the specification. Marshall stability is one of
the important parameters for AC-WC mixture because the advantage of this mixture
type is the high Marshall stability value due to the ability to withstand high loads.
Besides, high WHA levels also cause high VMA, but this finding is unfavorable,
because the voids between aggregates are mostly filled with air voids. The high air
void is a consequence of the low VFA in the asphalt mixture with high WHA at the
same bitumen content, and this can cause the asphalt mixture to be less durable. The
low VFA is also an indicator that the aggregates are not properly bonded to produce
sufficient stiffness.
A durability test was also conducted on the mixtures with different WHA content.
In this test, the mixtures were immersed for three different immersion periods (1,
3, and 7 days), and at the end of the periods, the mixtures were tested to measure
the Marshall stability (see Fig. 4). Asphalt mixtures with a WHA content of 25%
surprisingly experienced a significant decrease in stability value initially; however,
the stability value did not change much after seven days of immersion and still met the
mini-mum requirement of 800 kg. Meanwhile, asphalt mixtures with a WHA content
of more than 50% experienced a slow decline in stability; however, due to the high
air voids, it contributed to the lack of stiffness, therefore, it is not recommended to
be used in practice.
As a closing remark, the use of WHA as a substitute for cement is a good option
because it will reduce the cement material, which is an unrenewable material and
a significant contributor to greenhouse gases. However, this study’s results indicate
that the use of WHA as a substitute for cement should be limited because excessive
Fig. 3 Comparison of Marshall properties for asphalt mixtures with various WHA as filler: a
Marshall stability; b Flow; c VMA; d VIM; e VFA; f density
use of WHA will reduce the stiffness of the mixture and absorb more asphalt causing
increased air voids so that it reduces the durability of the mixtures.
25% WHA
50% WHA
75% WHA
100% WHA
Fig. 4 Comparison of the results of durability test
5 Conclusions
This paper presented an evaluation of water Hyacinth Ash (WHA) used as a filler in
asphalt mixture. To find out the influence of WHA on Marshall mixture’s properties,
five combination WHA—PC, i.e., 0, 25, 50, 75, and 100% WHA, were used. The
results of this study indicated that the increasing use of WHA will decrease the
stability. The use of WHA also could make the mixtures less durable due to high
asphalt absorption and causing the increase of VIM. As a part of the implementation
of green technology in practice, this study recommended a limit use of WHA to a
maximum of 25% as filler of the asphalt mixture.
References
1. Hoe GB, Sunasee P, Hon CK, Gyoo KB, Ti KS (2014) Utilisation of rice husk ash in asphaltic
concrete pavement. Adv Mater Res 1030–1032:961–964
2. Mashaan NS, Ali AH, Karim MR, Abdelaziz M (2014) A review on using crumb rubber in
reinforcement of asphalt pavement. Sci World J
3. Mistry R, Roy TK (2016) Effect of using fly ash as alternative filler in hot mix asphalt. Perspect
Sci 8:307–309
4. Murugesh V, Balasundaram N (2019) Experimental behavior of water hyacinth ash as partial
replacement of cement on short column. Int J Recent Technol Eng 7(65S):648–651
5. Murugesh V, Balasundaram N (2017) Experimental investigation on water hyacinth ash as the
partial replacement of cement in concrete. Int J Civ Eng Technol 8(9):1013–1018
6. Neelu D, Shashikant S (2016) Evaluation of water hyacinth stem ash as Pozzolanic material
for use in blended cement. J Civ Eng Sci Technol 7:1–8
7. Widyaningsih N, Sutanto B (2018) Influence of hyacinth plant as filler on mixed AC-WC

(Asphalt concrete—wearing course) with Marshall Test. In: International conference on
design, engineering and computer sciences on IOP conference series: materials science and
engineering, vol 453. IOP Publishing, Jakarta
8. National Research Council (1981) Food, fuel and fertilizer from organic waste. National
Academy Press, Washington, D.C.
9. Media Indonesia Homepage. https://mediaindonesia.com/. Last accessed 29 Feb 2020
10. Larra-Serrano JS, Rutiaga-Quinones OM, Lopez-Miranda J, Fileto-Perez HA, Pedraza-Bucio
FE, Rico-Cerda JL, Rutiaga-Quinones JG (2016) Physicochemical characterization of water
hyacinth ((Eichhornia crassipes (Mart.) Solms). BioResources 11(3):7214–7223
11. Sukarni S, Zakaria Y, Sumarli S, Wulandari R, Permanasari AA, Suhermanto M (2019) Physical
and chemical properties of water Hyacinth (Eichhornia crassipes) as Sustainable Biofuel Feed-
stock. In: International conference on condensed matters and advanced materials (IC2MAM
2018) on IOP conference series: materials science and engineering, vol 515. IOP Publishing,
Malang
12. Directorate General of Highway (2020) 2018 General specification for road and bridge
construction (Revision 2). Ministry of Public Works and Public Housing, Jakarta
The Setting Time of Portland Composite
Cement Mixed with Calcium Stearate
A. Maryoto, P. Hardini, and R. Setijadi
Abstract This study aims to determine the effect of calcium stearate in setting time
of Portland Composite Cement setting time. Tests carried out include the consistency
of normal PCC, initial setting time, final setting time, mortar compressive strength
and phase hydration. The calcium stearate content is used as additive for PCC is 0,
1, 2 and 3% of the weight of cement. Used cement is PCC type without added fly
ash and with additional 10% fly ash by cement weight. Mortar compressive strength
specimens are cubes with size of 50 × 50 × 50 mm. Mortar compressive strength
test was performed at 3, 7, 14 and 28 days. The phase of hydration test object is a
powder which is drilled from a harden cement paste. This powder is tested for its
compound content using an X-Ray Diffraction (XRD) tool. The test results found
that the use of calcium stearate accelerated the initial and final setting time of PCC
around 78 min. Mortar compressive strength decreases when the use of calcium
stearate in the mortar mixture increases. This tendency is in line with the results of
the XRD test that the amount of Crystalline formed at 28 days becomes smaller when
the amount of calcium stearate added to PCC increased.
Keywords Calcium stearate · Fly ash · Final setting time · Initial setting time ·
Portland composite cement
1 Introduction
Concrete damage caused by corrosion attacks endangers the existence of reinforced

concrete structures and prestressed concrete [1]. Corrosion attacks can be caused
by corrosive ion infiltrations such as chloride ions, sulfates and carbonation. These
A. Maryoto (B) · P. Hardini

Department of Civil Engineering, Universitas Jenderal Soedirman, Purwokerto, Indonesia
P. Hardini
R. Setijadi
Department of Geology Engineering, Universitas Jenderal Soedirman, Purwokerto, Indonesia
1120 A. Maryoto et al.
corrosive ions enter the concrete through concrete capillaries which form during the
hydration process. There are two methods of reinforcement in concrete, i.e. treatment
of reinforcement and treatment of the concrete. Treatment on the reinforcement can
be done by coating on reinforcement or by cathodic protection [2]. While treatment
on concrete can be done by making the concrete more waterproof. The increase in
water resistance can be treated in several ways, namely by using the coating method
on the concrete surface, by installing a layer of membrane on the concrete surface
or by adding admixture or additive concrete.
Treatment for increasing permeability of concrete can be performed by adding
waste material such as fly ash and waster rubber [3] to the concrete matrix. Particle
size of fly ash is smoother when compared to particle size of cement. The use of
fly ash in concrete can increase concrete compressive strength [4, 5], reducing the
penetration of chloride ions contained in seawater [6, 7], decreasing the corrosion rate
[8], decreasing the amount of pore in the concrete [9] and reducing water absorption
[10]. Porosity in concrete can be reduced by reducing the amount of water used in
concrete. Unfortunately, using a small amount of water in concrete causes a decrease
in the workability of fresh concrete.
Meanwhile, poor workability of fresh concrete can be reduced by adding super-
plasticizers [11]. It can be produced using either lignosulfonate or naphtalene
sulfonate. The addition of lignosulfonate compounds can accelerate the setting time
of concrete [12]. Transportation and handling of the concrete, and also removal of
form work are very dependent on the setting time of concrete. Some concrete addi-
tives and chemical compounds have an influence on the concrete time setting. The
long duration of fresh concrete transportation requires added ingredients which can
slow down the initial concrete time setting. Triethanolalamine [13], Carbamide [14],
and styrene butadiene [15] have been studied to slow down the initial concrete setting
time.
The use of lime in mortar [16], calcium stearate and superhydrophobic nano-
silica [17] as additive in concrete can increase compressive strength and reduce water
absorption [18], improve water repellents of concrete [19], chloride ion infiltration
[20–22]. In this ragard, the accelerated corrosion attack on concrete reinforcement
by chloride ions in concrete with calcium stearate addition, calcium stearate added
to concrete is equal to 0, 1, 5 and 10 kg per 1 m3 of concrete. It concluded then
calcium stearate usage in concrete reduced significant corrosion rate ion attack. It
can be observed in the Fig. 1a, b concrete without and with calcium stearate. The
higher the calcium stearate, the higher the resistance of the concrete to the attack of
corrosion by chloride ions. Unfortunately, the effect of calcium stearate in the initial
and final setting time, phase of hydration and compressive strength in mortar have
not been well known.
Therefore, this study aims to determine the effect of calcium stearate utilized with
Portland Composite Cement (PCC) and fly ash in properties of mortar. The cement
paste properties tested included the initial and final settings time of cement paste. In
addition, the compressive strength of the mortar with variations in the addition of
calcium stearate was also tested at 3, 7, 14 and 28 days. The type of solid constituents
of PCC paste due to addition of fly ash and calcium stearate is investigated as well.
The Setting Time of Portland Composite … 1121
Fig. 1. Concrete without calcium stearate (a), concrete with calcium stearate (b)
Table 1 Chemical content of cement and fly ash

Chemical composition SiO2 Al2 O3 Fe2 O3 CaO MgO SO3 Na2 O K2 O
Cement (%) 18.5 4.9 2.9 59.2 1.2 2.0 0.2 0.3
Fly ash (%) 45.2 17.9 10.9 10.0 2.8 0.3 1.6 0.7
2 Method
2.1 Material and Equipment
Materials for research include PCC, fly ash, calcium stearate, sand and water. The
chemical content of PCC, fly ash and calcium stearate properties are shown in Table
1. The used calcium stearate contains 6.65 of calcium, 70.35 of carbon, 11.65 of
hydrogen, and 10.6% of oxygen.
The equipment utilized consists of cement paste/mortar mixer, beaker glass, vicat
apparatus, weighing balance 1000 gr with accuracy 1 gr, vicat mold, glass plate,
the plunger of 10 mm diameter, hand trowel, stop watch, cube mold with the size
50 mm, compactor with the dimension. 13 × 25 × 120 mm, flow table, compression
machine, drill and X-Ray Diffraction machine. The used Vicat apparatus is based on
SNI 20,149:2015 [23] requirements.
2.2 Specimens
The amount of test specimens are 3 pieces for each type of code and testing. The code
of specimens are C0F0, C1F0, C2F0, C3F0, C0F10, C1F10, C2F10, and C3F10. The
type of testing are normal consistency of cement, initial and final setting time, phase
of hydration and compressive strength. C represents calcium stearate, the number
after C is the percentage of calcium stearate by weight of cement, F is a symbol for
fly ash, and number after F is the content of fly ash by weight of cement. There are
5 types of testing namely normal consistency of cement, initial setting time, final
setting time, compressive strength of mortar, phase of hydration.
2.3 Testing Procedures
Normal Consistency of Cement

This test is conducted based on the Indonesia National Standard, SNI 2049:2015.
Initial and Final Setting Time of Cement
The initial setting time is the time needed by cement and water to react to form cement
paste and begin losing its plasticity. It can be measured by penetration of the needle.
The initial setting time for penetration is the 1 mm diameter reaches 25 mm depth.
Procedures of the those test is conducted and refer Indonesia National Standard, SNI
2049:2015.
Compressive Strength of Mortar
Specimen for compressive strength of mortar is cube with dimension 50 × 50 × 50
mm. The compressive strength test was carried out when the mortars were 3, 7, 14
and 28 days old. The proportion material for compressive strength of mortar is shown
in the Table 3. The procedurs of making and testing the specimens of compressive
strength of mortas is based on the Indonesia National Standard, SNI 2049:2015.
Table 2 shows the proportion of material for setting time test.
The compressive strength test was conducted at 3, 7, 14 and 28 days old. Compres-
sion test is performed using UTM. The amount of compressive strength of mortar
(MPa) is obtained by dividing the maximum load (N) and surface area of the specimen
(mm2 ). Figure 2 shows the process of compressive strength test.
Table 2 Proportion material for setting time test

Code Weight (gr)
Cement Water Calcium stearate Fly ash
C0F0 650.0 148.5 0 0
C1F0 643.5 148.5 6.5 0
C2F0 637.0 148.5 13.0 0
C3F0 630.5 148.5 19.5 0
C0F10 585.0 148.5 0 65
C1F10 578.5 148.5 6.5 65
C2F10 572.0 148.5 13.0 65
C3F10 565.5 148.5 19.5 65
Table 3 Proportion material for mortar

Code Weight (gr)
Cement Sand Water Calcium stearate Fly ash
C0F0 500 1375 235 0 0
C1F0 495 1375 235 5 0
C2F0 490 1375 235 10 0
C3F0 485 1375 235 15 0
C0F10 450 1375 235 0 50
C1F10 445 1375 235 5 50
C2F10 440 1375 235 10 50
C3F10 435 1375 235 15 50
Phase of Hydration
Cement paste is poured in the molding 50 × 50 × 50 cm. After the specimen was
48 h old, then demolding the cement paste the specimen. Next, specimens are cured
for 28 days. Drilling the hard specimen of cement paste (Fig. 3). The drilling results
are in the form of cement paste powder (Fig. 4). Perform testing using XRD (Fig. 5)
machine to obtain compound content and the type of solid in the cement paste powder.
Fig. 2 Compression test

Fig. 3 Drilling process
Fig. 4 Powder of cement

paste
Fig. 5 X-ray diffraction
3.1 Normal Consistency of Cement
Needle penetration depth on cement paste with 20, 25 and 30% water amount from
the cement weight shown in Fig. 6. By relating the required water and penetration
depth in Fig. 6, the water content for consistency of normal cement is obtained when
the penetration of needle reaches 10 mm by interpolating the data. The required water
in the normal consistency of cement is 22.9%. This water content of 22.9% is then
used to investigate the initial and final setting, phase of hydration as well.
3.2 Initial and Final Setting Time
Initial and final setting time cement without fly ash and with fly ash is shown in
Fig. 7. It is found that additional of calcium stearate in cement reduces initial setting
time of cement. The higher content of calcium stearate, the lower of initial setting
time. Additional of calcium stearate 3% by weight of cement reduce the initial setting
time from 129 to 111 min. This tendency is also occured in the final setting time
of cement paste with calcium stearate. The usage of calcium stearate in the cement
paste accelerate the final setting time. It can be observed in the Fig. 7 that the setting
time of cement paste with 3% of calcium stearate is faster around 35 min than the
final setting time of cement paste without calcium stearate.
The descent of setting time in the cement paste containing fly ash is obtained due
to the added of calcium stearate. The initial and final setting time of cement paste
with calcium stearate 3% is declined around 50 min and 78 min compared to the
setting time of cement and fly ash paste without calcium stearate. This trends may
be caused by cluster of stearate react with calcium hydroxide to form and mix with
calcium silicate hydrate that the properties is like a wax [20]. Those reaction require
some higher water and as the effect is the setting time is accelerated.
It is also can be confirmed from the Fig. 8 that the additional of fly ash in the cement
paste cause the longer setting time. It is commonly understood that the prolonged
setting time is affected by calcium rich gel. Appearance of calcium rich gel formation
decrease the silica content. As the effect is the setting time to be retarded [24].
Fig. 6 Normal consistency 20

of cement 18
Depth of penetration (mm)
16
14 y = 0.89x - 10.35
12
10
8
6
4
2
0
20 21 22 23 24 25 26 27 28 29 30
Water content (% of cement weight)
350
300
Setting Time (minutes)

250
200
150
100 Cement without Fly Ash (Initial)
Cement without Fly Ash (Final)
50 Cement with Fly Ash (Final)
Cement with Fly Ash (Initial)
0
0 1 2 3
Calcium Stearate Content (% by cement weight)
Fig. 7 Setting time of cement
3.3 Compressive Strength of Mortar and Phase of Hydration
Figure 8 shows the relationship between calcium stearate content as a horizontal axes
and compressive strength of mortar without fly ash as a vertical axes. It can be found
that the additional of calcium stearate in the mortar reduce the compressive strength
not only at 3 days old but also at t 7, 14 and 28 days old. The usage of calcium
stearate of 3% of cement weight decrease the compressive strength of mortar at
3 days old around 16.1% compare to mortar without calcium stearate. The reduction
of compressive strength of mortar is also occurred at 7, 14 and 28 days old of 19.1,
49.0 and 54.5% respectively.
Figure 9 presents the correlation between calcium stearate and compressive
strength of the mortar with fly ash 10%. Usage of calcium stearate in the mortar
with fly ash 10% decrease the compressive strength. The higher calcium stearate
content in the mortar increase a little of compressive strength of mortar. It is may be
caused by reaction between silicon content in fly ash with calcium in the calcium
stearate to form more calcium silicate hydrate (CSH) in the mortar.
Fig. 8 Compressive strength 350

Compcressive strength (kg/cm2)
of mortar without fly ash

300
250
200
150
100
3 days 7 days
50 14 days 28 days
0
0 1 2 3
Calcium stearate content (% of cement weight)
Fig. 9 Compressive strength 350
Compcressive strength (kg/cm2)

of mortar with fly ash 10%
300
250
200
150
100
3 days 7 days
50 14 days 28 days
0
0 1 2 3
Calcium stearate content (% of cement weight)
The result of X-ray diffraction of cement paste non fly ash with calcium stearate
0, 1, 2 and 3% at 28 days age as an example. The raw data in the Fig. 10 is then
analyzed using Xpowder software and the result is shown in the Table 4.
According to Table 4 on cement paste with and without fly ash shows that the
percentage of crystalline solid has an effect of compressive strength of mortar. Crys-
talline solid is a material in the solid form. Increasing of calcium stearate content tend
to decrease the content of crystalline solid. Moreover, when the tobermorite (calcium
silicate hydrate/C-S-H) in the cement paste increase, the compressive strength of
mortar containing calcium stearate is reduced. This may be caused by some alike
wax constituent united with C-S-H in the cement paste. The alike wax constituen
have a better effect in the concrete to avoid the corrosion attack by chloride ion, but
reduce the compressive strength of concrete.
On the other hand, specimens of cement paste containing fly ash with and without
calcium stearate show that the percentage of crystalline solid have a similar tendency
to the compressive strength. The crystalline solid is around 60%. It may be affected
by Portland Composite Cement (PCC). This cement have been already mixed with
fly ash during the production process. When PCC is then mixed again with fly ash,
the some silika in the fly ash can not react with calcium hydroxide. As the effect,
the compressive strength of mortar containing fly ash at this case little go down
or same. This is also influenced by that the percentage of amorphous. A higher
content of amorphous, the compressive strength of mortar tend to decrease, because
amorphous solid usually is in the form of gel. The usage of calcium stearate both in
concrete and mortar around 1% of cement weight have many advantages to reduce
water absorbsion, corrosion rate, infiltration of chloride ion in the concrete. Even
though the application calcium stearate in concrete with high dosage decrease the
compressive strength, the dosage of 1% of calcium stearate in the concrete is still
safe to be used.
Fig. 10 X-ray diffraction result of cement paste powder: without calcium stearate (a), with calcium
stearate 1, 2 and 3% (b, c, d)
Fig. 10 (continued)
Table 4 Mineral content of cement and fly ash

Compound Code of cement paste
C0F0 C1F0 C2F0 C3F0 C0F10 C1F10 C2F10 C3F10
Calcium sulphate hydrate (%) 17.4 17.4 24.9 22.8 19.8 30.2 26.0 23.2
Tobermorit (%) 3.6 4.6 7.1 4.0 4.1 5.8 4.8 3.7
Portlandite (%) 3.1 2.2 4.4 2.4 2.9 4.2 2.9 2.7
Ikaite (%) 75.8 75.8 63.5 70.8 73.2 59.8 66.3 70.4
Crystalline (%) 64.5 62.5 63.6 38.1 60.9 60.9 60.0 60.6
Amorphous (%) 35.5 37.5 36.4 41.9 39.1 39.1 39.1 39.4
4 Conclusions
According to the results and discussion above, the main conclusions are summarized
as follows.
a. The consistency of normal paste of portland composite cement (PCC) is reached
by adding of water around 22.9% by weight of PCC.
b. The addition of calcium stearate in cement paste accelerates the initial setting
time of PCC. By adding 3% of calcium stearate by weight of cement, the initial
setting time is faster around 19 min compared to non calcium stearate cement
paste. The usage of calcium stearate 3% can reduce final setting time of PCC
from 165 min to be 115 min.
c. In the cement paste of PCC with fly ash, addition of calcium stearate 3% decrease
the inital setting time from 260 min to be 225 min. And the final setting time is
reduced from 318 min to be 240 min.
d. In the mortar with PCC without fly ash, addition of calcium stearate 3% reduces
the compressive strength around 17%. It also happens in the mortar with PCC
and fly ash, addition of calcium stearate 3% in the mortar, the compressive
strength is decreased 8.9%.
e. Reduction of compressive strength due to addition of calcium stearate is
supported by forming of smaller crystalline solid in the cement paste. The higher
calcium stearate in the mortar, the more reduction of compressive strength of
mortar is produced.
Acknowledgements This study was conducted completely due the assistance of Jihan, Deni
and PT. Holcim Indonesia—Cilacap plant. Also, thank you to Universitas Jenderal Soedirman,
Purwokerto for funding this research through Hibah Riset Unggulan Terapan, 2018, Contract No.
3787/UN23.14/PN/2018.
References
1. Maryoto A, Shimomura T (2017) Effect of prestressed force and size of reinforcement on

corrosion crack width in concrete member. J Eng Sci Technol 12(10):2664–2676
2. Laurens S, Francois F (2017) Cathodic protection in reinforced concrete structures affected
by macrocell corrosion: a discussion about the significance of the protection criteria. RILEM
Tech Lett 2:27–32
3. Maryoto A, Nastain N, Supriyanto H (2019) The bond response of concrete brick with recycled
tire chip as partial replacement of aggregate applied in the non-structural masonry wall. Int
Rev Civ Eng 10(1):33–44
4. Ashadi HW, Aprilando BA, Astutiningsih S (2015) Effects of steel slag substitution in
geopolymer concrete on compressive strength and corrosion rate of steel reinforcement in
seawater and an acid rain environment. Int J Technol 2:227–235
5. Hefni Y, Zaher YA, Wahab MA (2018) Influence of activation of fly ash on the mechanical
properties of concrete. Constr Build Mater 172:728–734
6. Moffatt EG, Thomas MDA, Fahim A (2018) Performance of high-volume fly ash concrete in
marine environment. Cem Concr Res, Article in Press
7. Yu Y, Yu J, Ge Y (2016) Water and chloride permeability research on ordinary cement mortar
and concrete with compound admixture and fly ash. Constr Build Mater 127:556–564
8. Wang W, Lu C (2018) Time-varying law of rebar corrosion rate in fly ash concrete. J Hazard
Mater, Article in Press
9. Jiang P, Jiang L, Zha J, Song Z (2017) Influence of temperature history on chloride diffusion
in high volume fly ash concrete. Constr Build Mater 144:677–685
10. Nagrockiene D, Daugela A (2018) Investigation into the properties of concrete modified with
biomass combustion fly ash. Constr Build Mater 175:369–375
11. Khudair MH, Youbi MS, Elharfi A (2018) Data on effect of a reducer of water and retarder
of setting time admixtures of cement pastes and mortar in hardened stat. Accepted Article,
Elsevier
12. Colombo A, Geiker MR, Justnes H, Lauten RA, Weerdt KD (2017) On the effect of calcium
lignosulfonate on the rheology and setting time of cement paste. Cem Concr Res 100:435–444
13. Yaphary YL, Yu Z, Lam RHW, Lau D (2017) Effect of triethanolamine on cement hydration
toward initial setting time. Constr Build Mater 141:94–103
14. Gong C, Zhou X, Dai W, Liu Y, Lu L, Cheng X (2018) Effects of carbamide on fluidity
and setting time of sulphoaluminate cement and properties of planting concrete from
sulphoaluminate cement. Constr Build Mater 182:290–297
15. Li L, Wang R, Lu Q (2018) Influence of polymer latex on the setting time, mechanical properties
and durability of calcium sulfoaluminate cement mortar. Constr Build Mater 169:911–922
16. Mydin MAO (2017) Prelimenary studies on the development of lime-based mortar with added
egg white. Int J Technol 5:800–810
17. Chen H, Feng P, Du Y, Jiang J, Sun W (2018) The effect of superhydrophobic nano-silica
particles on the transport and mechanical properties of hardened cement pastes. Constr Build
Mater 182:620–628
18. Ma C, Chen B (2016) Properties of foamed concrete containing water repellents. Constr Build
Mater 123:106–114
19. Legazzo A, Vicini S, Cattaneo C, Botter R (2016) Effect of fatty acid soap on microstructure
of lime-cement mortar. Constr Build Mater 116:384–390
20. Maryoto A, Gan BS, Hermanto NIS, Setijadi R (2018) The compressive strength and resistivity
toward corrosion attacks by chloride ion of concrete containing type I cement and calcium
stearate. Int J Corros 2018:1–9
21. Maryoto A, Gan BS, Aylie H (2017) Reduction of chloride ion ingress into reinforced concrete
using a hydrophobic additive material. J Teknologi 79(2):65–72
22. Maryoto A, Gan BS, Hermanto NIS, Setijadi R (2017) Protection of corrosion attack in rein-
forced concrete due to chloride ion using calcium stearate. J Eng Appl Sci 12(special issue
6):7965–7970
23. Indonesia National Standard, SNI 2049:2015. Cement Portland, National Standardization
Bureau
24. Kaja AM, Lazaro A, Yu QL (2018) Effects of Portland cement on activation mechanism of
class F fly ash geopolymer cured under ambient conditions. Construction and Building Material
189:1113–1123
Evaluation of Silt Content in Natural
Sand Used as Building Materials:
A Statistical Analysis Approach
Jauhar Fajrin, Mulyadi, Hariyadi, and Agung Prabowo
Abstract The excessive amount of silt within natural sand may reduce the bonding
between sand and other constituent materials when used as building materials. The
maximum amount allowed varies from 4 to 10%, and the allowable amount according
to the Indonesian Standard (SNI S-04-1998-F) should be less than 5%. This study
aims to evaluate the amount of silt in natural sand from several sand quarries in
Lombok, Indonesia. The natural sand samples were collected from eight sand quarries
in Lombok and the silt content was determined based on weight measurement. The
study was designed to assess the silt content of each sand quarry against the SNI
and also compared the silt content among the quarries. Two appropriate statistical
analysis methods were applied; one-sample t-test and analysis of variance (ANOVA).
The main conclusion drawn was that most of the natural sand in Lombok contains
high silt content. The one-sample t-test results showed that the P-value of seven out of
eight quarries was greater than the significance level (α), except for the sample taken
from Sukadana. This value indicates that the silt content pretty close to the maximum
value allowed by the SNI, which is already critical to be used as a building material.
Meanwhile, the ANOVA result shows that the silt content from all sand quarries does
not have a significant difference, which is indicated by the P-value greater than 0.05.
Nevertheless, the result of a further analysis using Fisher’s test shows that the sand
collected from Sukadana quarry has fairly low silt content and differs significantly
from the others. It is strongly recommended to give a proper pre-treatment such
cleaning process before using sand in this area as a building material.
J. Fajrin (B) · Hariyadi · A. Prabowo

Department of Civil Engineering, Mataram University, Mataram City, Indonesia
Hariyadi
A. Prabowo
Mulyadi
Ministry of Public Works, Mataram, West Nusa Tenggara Province, Indonesia
1134 J. Fajrin et al.
Keywords Silt content · Inferential statistics · Material assessment
1 Introduction
One of the issues related to infrastructure development in Indonesia is the quality

of building materials. Although the regulations in the form of standards have been
issued, the implementation has not been well managed. Based on field observations,
a lot of building materials especially those originating from nature or traditionally
handmade, are directly delivered to users without sufficient quality checking process.
As a result, many infrastructures built up using those materials were damaged earlier
than their intended service life. Silt content is one of the important parameters to
define the quality of sand. The maximum amount allowed varies between 4 and 10%,
with the Indonesian Standard (SNI S-04-1998-F) stating that the maximum amount
authorized should be less than 5%. Meanwhile, the maximum amount allowed by the
British Standard (BS) is 4%, but the American Standard for Testing and Materials
(ASTM) allows up to 10% [1].
Research related to the assessment of sand quality has not been widely reported,
especially in Indonesia. Among a few reported studies, Qomaruddin et al. [2] who
conducted a comparative study of river sand characteristics in Jepara reported that
the silt content ranges from 10.55 to 14.37%, which means 2–3 times exceeds the
maximum standard required by SNI. In addition to silt content, the study also eval-
uated several other parameters such as fine aggregate gradation, organic material
content, density, and specific gravity. Endroyo [3] reported that the quality of sand
depends on the time and the location of the sand quarry. Further, Koraila and Joshi [4]
reported the results of their research on sand quality and its management in Nepal,
which indicated that the quality of sand is also greatly influenced by how it is being
managed. It was also observed in the study reported in Ref. [4] that the awareness of
using sand that meets the technical requirements is very low among local contractors.
Likewise, the attention of the local government to ensure the use of appropriate sand
is also very low.
Some previous research reports the contribution of sand when it has been used
as a certain building material, for example after being used as a mortar or concrete.
Triadi et al. [5] conducted a compressive strength analysis of concrete made of fine
aggregate from Rokan river sand, which showed that the compressive strength of
concrete produced ranged from 12 to 14 MPa. Meanwhile, Ginting [6] conducted a
comparative study of the utilization of Ciliwung, Cisadane, and Cikeruh river sand
as normal fine concrete aggregates. The results reported in reference [6] show that
concrete made with Cikeruh river sand has the highest compressive strength, which
is 18.74 MPa. Furthermore, Ngugi et al. [7] reported that the high content of silt
and organic material within sand reduces the bonding strength between concrete and
reinforcement. It was also found that 86.2% of the observed sand samples in the study
reported in Ref. [7] did not meet the requirements stated in the BS 882 standard.
Gashahun [8] studied the impact of sand quality on concrete strength and found
Evaluation of Silt Content in Natural Sand … 1135
that using sand with silt levels over the permitted limit affected concrete strength
significantly. In addition, Obafaye et al. [9] stated that the percentage of silt content
that is unfit for engineering usage, notably concrete, is around 7%. Meanwhile,
Kirthika et al. [10] proposed a fine aggregate content of less than 5% for alternative
fine aggregates to make concrete with satisfactory performance.
The application of statistical methods in research in the field of civil engineering
materials is still limited to descriptive statistics where the data are presented in such a
way that they are easily read and understood by readers. The use of statistical methods
for decision-making or inference is still limited. Fajrin et al. [11] reported the results
of their research related to the development of composite materials by applying
inferential statistical methods, namely the t-test method. The application of other
inferential statistical tools, which is the F-test method, in the development of civil
engineering materials was comprehensively reported in Refs. [12–14]. Moreover,
the analysis of variance method was also applied by Fajrin et al. [15] to study the
physical and mechanical properties of mortars with silica fume as additional material.
Furthermore, Parmar et al. [16] applied the independent sample t-test method in their
study of ready mix materials in the Gujarat region, India.
To further ensure the quality of the sand used as building material currently in
Lombok, it is important to conduct a comprehensive investigation and laboratory
testing. This article reports the results of an investigation into the silt content within
the sand used as a building material in Lombok as a place for case studies. Further-
more, the results of laboratory tests in this study were statistically analyzed using
Minitab software.
The initial stage of this research is determining the sampling locations that repre-
sent the area in Lombok. Lombok is divided into 5 regencies/cities. Based on the
mapping results it was decided that the sampling locations were in 4 districts namely:
Gebong and Keling (West Lombok), Srojan and Pondok Omak (Central Lombok),
Lenek Daya and Ijo Balit (East Lombok), and Luk and Sukadana (North Lombok).
This study was conducted in the form of field surveys and laboratory testing which
were further analyzed using inferential statistical methods. There are two research
questions in this study that also become the focus of research. First, how much is the
silt content at each sand quarry location, and how significant is the difference with
the maximum content required by SNI. Second, how significant the difference of silt
content among sand quarry locations. The measurement of silt content within the
sand can be done in two ways, which are based on volume or weight. In this study,
the second method was chosen which is based on weight measurement.
The suitable statistical method to elaborate the first research question is the one-
sample t-test method. This method is commonly used to compare the average of a
single variable or a set of data with a certain value [17]. In the case of this study, the
average silt content at each sand quarry location was compared to the maximum value
specified in SNI S-04-1998-F, which is 5%. There are two prerequisites for applying
this statistical test method [17]; the data should be quantitative and satisfied a normal
distribution assumption. The decision to accept or reject the null hypothesis (H0 ) is
based on the following rules; (1) The null hypothesis (H0 ) is accepted if the P-value
is greater than the confidence level (α). If the value is smaller, then the alternative
hypothesis (H1) should be accepted. (2) The null hypothesis (H0 ) is accepted if the
calculated t value (tscore ) is smaller than the t value obtained from the table of t
distribution (ttable ).
The second research question is related to several (more than two) sets or groups
of data, so the appropriate statistical method used is the analysis of variance (Anova).
This method is commonly used to test mean differences in more than two groups.
According to Reference [17], three important assumptions must be satisfied to apply
this method; (1) The groups must be independent, (2) There has an equal variance
among the groups (homogeneous data), and (3) Data are normally distributed. The
null hypothesis (H0 ) in Anova states that there is no difference between the means
of the n groups. In contrast, the alternative hypothesis (H1 ) states that there is a
difference between the means of the n groups. ANOVA results are expressed in the
form of F values, which are further called the F-test or F-score. An inference or
decision is drawing by comparing the F-score to the F value obtained from the F
distribution table (Ftable ). The conditions are; if the F-score is greater than F-table,
then H0 is rejected, and as a consequence, the alternative hypothesis (H1 ) is accepted.
Another way to draw inference is to observe the P-value; if the value is less than the
confidence level (α), then H0 is accepted.
The average percentage of silt content for each sand quarry location is shown in
Fig. 1. It is clearly shown that the percentage of silt content lies at around 5%, which
is the maximum limit allowed according to SNI S-04-1998-F. There are four sand
quarry locations with the percentage of the silt content exceeds this limit, while the
other three are pretty close to this number. There has only a single location that has
a fairly low level of silt content, which is safe to be used as a building material.
Overall, the average silt content from the eight sample locations in Lombok is
4.8%. The silt content at 4 locations has exceeded the maximum limit, with an average
percentage of 5.58%. The other 3 locations are pretty close to the maximum limit
determined by SNI, with an average value of 4.56%. Although the silt content at
3 sand quarry locations is currently still below 5%, the average value is very close
to the limit and if the samples are checked one by one, there are samples whose
values have exceeded 5%. To further confirm the results of the analysis, it is also
necessary to do a more in-depth analysis using inferential statistics. As stated in the
research methods section, the statistical method that can be used to assess whether
the average silt content at a sand quarry location is the same or not with the maximum
limit allowed in the SNI is the one-sample t-test method.
Fig. 1 The average percentage of silt content at each sand quarry location
Like many other statistical analysis methods, the first thing that needs to be done
as a condition for applying the one-sample t-test method is that the data must be
normally distributed. For this reason, a normality test had been performed using
Minitab software based on three models; namely Anderson–Darling (AD), Ryan-
Joiner (RJ), and Kolmogorov–Smirnov (KS). The results are presented in Table 1,
which clearly shows that the data are normally distributed that indicated by higher
P-values over the significance level (α).
Having been convinced that the data were normally distributed, a one-sample
t-test was performed for all sample locations and the summary of results is shown
in Table 2. The results of the one-sample t-test show that almost all locations have
Table 1 The summary of normality test results

Sample Anderson–Darling Ryan-Joiner Kolmogorov–Smirnov
location AD P-value RJ P-value KS P-value
Srojan 0.341 0.321 0.954 >0.1 0.239 >0.150
Pondok 0.256 0.539 0.968 >0.1 0.218 >0.150
Omak
Lenek Daya 0.432 0.170 0.931 >0.1 0.255 >0.150
Ijo Balit 0.600 0.053 0.892 0.075 0.349 0.046
Luk 0.365 0.272 0.936 >0.1 0.286 >0.150
Sukadana 0.316 0.378 0.957 >0.1 0.227 >0.150
Gebong 0.275 0.497 0.968 >0.1 0.212 >0.150
Keling 0.227 0.640 0.974 >0.1 0.186 >0.150
Table 2 The summary of one sample t-test results

Sample location Significance (α) 5% Significance (α) 1%
t-score P-value t-score P-value
Srojan 0.24 0.824 0.24 0.824
Pondok Omak 1.27 0.275 1.27 0.275
Lenek Daya −0.39 0.715 −0.76 0.488
Ijo Balit −0.39 0.716 −0.39 0.716
Luk −0.67 0.538 −0.67 0.538
Sukadana −4.59 0.010 −4.59 0.010
Gebong 0.27 0.800 0.27 0.800
Keling 0.47 0.662 0.47 0.662
a P-value greater than the significance level (α), except the sand quarry location
in Sukadana, North Lombok. It has a meaning that the H0 hypothesis which states
that the silt content is equal to 5% must be accepted or the alternative hypothesis
H1 is rejected. Another way to conclude is by looking at the value of the t-score
with the t-table. Based on the t distribution table, the value of the t-table is 2.31 and
3.75 for the value of 5 and 1%, respectively. It can be seen clearly in Table 2 that,
except for the Sukadana location, the t-score value for all other locations is smaller
than the value of the t-table, so that the H0 is accepted. As the hypothesis null (H0 )
is accepted, this means that the level of silt within the sand in Lombok is mostly
pretty close to the maximum value tolerated by SNI, hence the sand must be cleaned
up before be further used as a building material. The results of the homogeneity
test which is based on two models; Bonett’s test and Levene’s test, are displayed in
Table 3.
Ideally, a building material should have the same quality although it is provided or
obtained in different ways. Hence, it is necessary to further analyze the consistency
of the silt content observed at all sand quarries. A quick look at the data presented
in Fig. 1 indicates that the silt content differs from each other. But how significant
the differences have to be further analyzed using the appropriate method. The most
relevant statistical method to apply for this case is the analysis of variance method.
One of the prerequisites for data to be analyzed by the Anova method is that variants
between groups must be homogeneous. To find out whether the data meets these
requirements, it is necessary to do a data homogeneity test.
The homogeneity test result demonstrates that the variations (variances) of the silt
content data are equal which is indicated by a P-value greater than the significance
level (α = 0.05) for all locations. Another requirement for applying the ANOVA
method is that the sample must be independent of each other. This requirement
has been satisfied because the samples were obtained from different locations. The
third requirement is the data have a normal distribution and this has been confirmed
as shown in Table 1. The results of statistical analysis using Minitab software are
presented in Table 4.
Table 3 The summary of

Sample location Homogeneity test
homogeneity test results
P-value P-value
Bonett Levene
Srojan - P Omak 0.412 0.531
Srojan – L Daya 0.380 0.586
Srojan – I Balit 0.154 0.532
Srojan – Luk 0.795 0.636
Srojan – Sukadana 0.580 0.733
Srojan – Gebong 0.182 0.260
Srojan - Keling 0.307 0.447
P Omak - L Daya 0.898 0.980
P Omak – I Balit 0.550 0.786
P Omak - Luk 0.335 0.306
P Omak - Sukadana 0.246 0.318
P Omak - Gebong 0.645 0.607
P Omak - Keling 0,730 0.827
L Daya – I Balit 0.442 0.780
L Daya – Luk 0.300 0.362
L Daya – Sukadana 0.182 0.391
L Daya – Gebong 0.525 0.617
L Daya – Keling 0.636 0.820
I Balit - Luk 0.116 0.395
I Balit - Sukadana 0.077 0.429
I Balit - Gebong 0.876 0.967
I Balit - Keling 0.854 0.906
Luk – Sukadana 0.859 0.812
Luk – Gebong 0.139 0.141
Luk – Keling 0.250 0.273
Sukadana - Gebong 0.091 0.131
Sukadana - Keling 0.194 0.288
Gebong - Keling 0.962 0.809
Table 4 The summary of statistical analysis using Minitab software

One-Way ANOVA Tukey Test Fisher LSD Test
F-Value: 1.42 All sand quarries labeled as a single letter A 4 quarries labeled as A
P-Value: 0.23 3 quarries labeled as AB
1 quarry labeled as B
The most important information from the data shown in Table 4 is the P-value
and the F-score, which are 0.23 and 1.42, respectively. As the P-value is greater than
the significance level of 0.05, then the hypothesis H0 is accepted. The conclusion to
accept the null hypothesis is also confirmed by the value of the Fscore (1.42), which is
smaller than the value of the Ftable (2.26). This indicates that the average sand content
at all sand quarry locations is just basically the same. However, there is a little doubt
about the Anova results because they equate the highest average of mud content at
the Pondok Omak (6.29) with the lowest one (2.29) at Sukadana. For this reason, it
is essential to carry out further tests (simultaneous tests) which usually employing
Tukey, Fisher, and Dunnet tests. The simultaneous tests provide two information in
the form of numbers and grouping information. The result of means grouping based
on the Tukey and Fisher test is also shown in Table 4. The result of Tukey’s test laid
all the sand quarries within only one group labeled with the A letter, which means
that there has no distinct difference in the silt content for all sample locations. The
LSD Fisher test showed a more comprehensive result where the sand quarries were
categorized into 3 levels that were labeled with a different letter. Meanwhile, the
result of another simultaneous test using Dunnet’s test provides a similar result with
the Tukey test. It seems that the Fisher LSD method provides more accurate results.
4 Conclusions
The main conclusion drawn from this study is that most of the sand in Lombok has a
silt content of approximately 5%, which is a maximum limit allowed by the National
Indonesian Standard (SNI) S-04-1998-F. More specific findings are as follows:
(1) Most of the sand in Lombok has a fairly high silt content with an overall average
of 4.8%. The result of inferential statistical analysis using the one-sample t-test
method proves that the silt content in 7 out of 8 sand quarry samples already
within the maximum limit allowed by SNI S-04-1998-F.
(2) The result of ANOVA shows that the silt content in the sand collected from
8 sand quarry location do not have any significant difference. However, the
results of the simultaneous or pairwise test using the LSD Fisher method show
that the sand sample taken from the Sukadana region in North Lombok has
a fairly low silt content and differs significantly from the rest sand quarry
locations.
(3) As the sand in Lombok contains fairly high silt, in which some of which have
exceeded the maximum limit allowed by SNI, the sand must be cleaned up
before being used as a building material.
References
1. Demissie T (2019) Evaluation of the quality of river sand on properties of concrete in the
context of Addis Ababa, Ethiopia, A thesis of Master of Science in Addis Ababa Science and
Technology University
2. Qomaruddin M, Ariyanto A, Umam K, Saputro YA (2018) Studi komparasi karateristik pasir
sungai di Kabupaten Jepara. Jurnal Ilmiah Teknosains 4(1):6–10
3. Endroyo B (2015) Kualitas pasir Muntilan (Jawa Tengah) ditinjau dari tempat pengambilan
dan musim. Wahana Teknik Sipil 12(1):1–8
4. Koirala MP, Joshi BR (2017) Construction sand, quality and supply management in an
infrastructure project. Int J Adv Eng Sci Res 4(4):1–15
5. Triadi Y, Marliansyah J, Rani A (2017) Analisa kekuatan tekan beton dengan bahan agregat
pasir dan kerikil sungai Rokan kanan Kabupaten Rokan Hulu. Jurnal Teknik Sipil UPP 2(1):1–8
6. Ginting ER, Erizal (2014) Pemanfaatan pasir sungai Ciliwung, Cisadane dan Cikeruh yang
melalui wilayah Bogor sebagai agregat halus pembuatan beton normal. Skripsi, Departemen
Teknik Sipil dan Lingkungan, Institut Pertanian Bogor
7. Ngugi HN, Mutuku RN, Gary ZA (2014) Effects of sand quality on bond strength of concrete:
a case study in Nairobi and its environments Kenya. Int J Civ Struct Eng Res 2(1):119–129
8. Gashahun AD (2020) Investigating sand quality effect on concrete strength: a case of Debra
Markos and its vicinities. Int J Constr Manag. https://doi.org/10.1080/15623599.2020.1774838
9. Obafaye BJ, Olukotun A, Audu MT, Ndububa EE (2020) Determination of tolerance limit
of silt and clay impurities in the strength of concrete made with Abuja sand. Int J Eng Sci
Invention 9(3):10–15
10. Kirthika SK, Surya M, Singh SK (2019) Effect of clay in alternative fine aggregates on the
performance of concrete. Constr Build Mater 228
11. Fajrin J, Zhuge Y, Bullen F, Wang H (2011) The implementation of statistical inference to
study the bending strength of sustainable hybrid sandwich panel composite. Adv Mater Res
250–253:956–961
12. Fajrin J, Zhuge Y, Bullen F, Wang H (2011) Flexural strength of sandwich panel with ligno-
cellulosic composites intermediate layer-a statistic approach. Int J Prot Struct 2(4):453–464.
ISSN 2041-4196
13. Fajrin J, Zhuge Y, Bullen F, Wang H (2013) significance analysis of flexural behavior hybrid
sandwich panels. Open J Civ Eng 3:1–7
14. Fajrin J (2016) The application of statistical design of experiments to study the in-plane shear
behavior of hybrid composite sandwich panel. J Civ Eng Dimension 18(1):25–30
15. Fajrin J, Pathurahman P, Pratama LG (2016) Aplikasi metode analysis of variance (Anova)
untuk mengkaji pengaruh penambahan silica fume terhadap sifat fisik dan mekanik mortar.
Jurnal Rekayasa Sipil 12(1):11–23
16. Parmar CR, Makwana AH, Pitroda J, Bhavsar JJ (2013) A study on ready mixed concrete using
independent t-test through SPSS software in Charotar region of central Gujarat. J Int Acad Res
Multidisc 1(11):292–300
17. Irianto A (2012) Statistik; konsep dasar, aplikasi dan pengembangannya. In: Handbook of
statistics, Kencana Prenada Media Group, Indonesia
Compressive and Flexural Strength
Behavior of Banana Tree Fiber Hybrid
Concrete
Fadillawaty Saleh, Fanny Monika, Hakas Prayuda, Bella Lutfiani Al Zakina,

Martyana Dwi Cahyati, Adira Aldi, and Feri Adri Wibowo
Abstract The purpose of this study is to discuss the utilization of banana tree fiber
as additional material to improve hardened concrete properties. The existence of
banana tree waste from the agriculture business has not been optimally utilized.
Using banana tree waste as a constituent of concrete is expected to improve the
quality of the concrete as well as to reduce waste. The fibers used in this study are
placed in the tensile part of the specimen because concrete has low resistance to
tensile forces, so it is hoped that the fibers can increase the tensile capacity of the
concrete. The test consisted of checking the workability level of fresh concrete, while
the hardened properties test consisted of compressive and flexural strength. Banana
tree fibers are varied from 0 to 0.24% of the weight of cement. The results of the
fresh properties show that the workability of concrete is reduced when the amount of
fiber is increased. The hardened properties result show that the increased amount of
F. Saleh · F. Monika · H. Prayuda (B) · M. D. Cahyati · A. Aldi · F. A. Wibowo

Yogyakarta, Tamantirto, Kasihan, Bantul, Yogyakarta 55182, Indonesia
F. Saleh
F. Monika
M. D. Cahyati
A. Aldi
F. A. Wibowo
B. L. Al Zakina
Department of Civil Engineering, Faculty of Science and Engineering, Universitas Bojonegoro,
Glendeng, Kalirejo, Bojonegoro, East Java 62119, Indonesia
1144 F. Saleh et al.
fiber is not affect the compressive strength. However, the bending test results show
that the flexural strength is improved when the amount of fiber increased.
Keywords Fiber concrete · Banana tree · Hybrid concrete · Compressive

strength · Flexural strength
1 Introduction
Concrete is one of the most widely used construction materials in the world. Concrete
has advantages including high ability to withstand compressive forces, durable, and
is economical compared to other construction materials. However, one of the disad-
vantages of concrete is its low ability to withstand tensile forces. When the tensile
strain exceeds the tensile strain capacity, the concrete will begin to crack, causing
a decrease in the ability of the concrete to withstand loads, as such, it increases the
corrosion of reinforcement. One method that can be used to increase the ability of
concrete to withstand tensile forces is to add fiber, especially to areas of concrete
that resist tensile forces [1].
High-strength concrete is known to have high brittle properties making it
extremely dangerous if the load exceeds its capacity. Utilization of fiber in high-
strength concrete can increase the tensile strength thereby reducing the brittle prop-
erties which is one of the weaknesses of high-strength concrete [2]. Normal fiber
concrete can be defined as a composite material consisting of cement, water, coarse
aggregate, fine aggregate, and fiber in the form of small pieces [3]. The fiber used
in the manufacture of concrete comes from various resources, generally divided into
two parts, namely artificial fiber, and natural fiber [4]. Examination of concrete prop-
erties using natural fiber has been carried out by many previous researchers whose
study involved the use of bamboo fiber [5, 6], henequen fiber [7], Pineapple leaf fiber
[8, 9], Abaca fiber [10, 11], and several classifications of other fibers [12].
In this research, banana tree fiber is used as natural fiber in the process of making
concrete. Examination of the characteristics of fiber concrete using banana tree fiber
has been done by many previous researchers, this is because banana tree fiber is
easily available in tropical countries such as Indonesia, Malaysia, the Philippines,
and Thailand [13]. Other researchers explained that the use of banana fiber as a
reinforcement material significantly increased resistance to cracking and spalling
and increased flexural strength [14, 15]. The utilization of banana fiber as additional
materials of concrete can also reduce the use of cement in the manufacturing, thereby
reducing the release of carbon dioxide into the atmosphere [16, 17].
Excessive use of natural fiber in concrete can also cause several problems, espe-
cially reducing the durability of concrete. This causes the service life of the concrete
to be shorter. Thus, one possible method to increase the flexural and tensile capacity
of composite concrete using banana fibers is by implementing a hybrid system. The
hybrid system is achieved by placing fiber concrete only in the tensile part, while the
compressive part is filled with normal concrete with the same compressive strength
Compressive and Flexural Strength Behavior … 1145
planning. Additionally, it is believed that the fibers in the concrete do not signifi-
cantly affect the compressive capacity of the concrete but have a significant effect
in increasing its tensile strength. Although the properties of concrete using a hybrid
system have been discovered by several previous researchers [18–22], none had
utilized fibers from banana trees.
In this study, a fiber concrete properties test was carried out using a hybrid system.
Tests carried out on concrete with a compressive strength and flexural strength were
tested on concrete with ages 3, 7 and 28 days. The binders used in this study consisted
of cement and silica fume. All specimens were cured using water curing.
The concrete material in this study consisted of a binder in the form of cement
with the type of Portland pozzolan cement based on ASTM C595 [23] and Silica
fume from PT. Sika Indonesia with a classification from ASTM C1240 [24]. Coarse
aggregate, fine aggregate, water, Sikament-NN type superplasticizer aims to increase
workability in fresh concrete to facilitate the processing process [25], and banana
tree fiber. The fiber used in this study comes from a banana tree with a length of 5 cm
and a diameter of 1 mm as shown in Fig. 1. Before the fiber is used, the fiber must
be dried to remove the water contained in the fiber, then the surface of the fiber is
coated with resin so that the fiber becomes waterproof and durable. The composition
of the fiber used in this study varied from 0, 0.08, 0.16, and 0.24% out of the weight
of cement. The mix proportions used can be seen in Table 1. The water to binder
ratio used in this study was 0.65. The type of curing used is water curing carried out
1 day after the casting process. The flexural strength and compressive strength tests
were carried out on concrete of 3, 7, and 28 days.
Fig. 1 Banana tree fiber

Table 1 Mix proportions for 1 m3

Materials Fiber variations
S1 (0%) S2 (0,08%) S3 (0.16%) S4 (0.24%)
Sand (kg/m3 ) 836.29 836.29 836.29 836.29
Cement (kg/m3 ) 287.98 287.98 287.98 287.98
Coarse aggregate (kg/m3 ) 916.64 916.64 916.64 916.64
Silica fume (kg/m3 ) 44.02 44.02 44.02 44.02
Fiber (kg/m3 ) 0.00 0.265 0.531 0.797
Water (l/m3 ) 217.05 217.05 217.05 217.05
Superplasticizer (l/m3 ) 12.54 12.54 12.54 12.54
Table 2 Properties of fine

Properties Unit Fine aggregate Coarse aggregate
and coarse aggregate
Specific - 2.58 2.63
gravity
Water % 2.10 1.80
absorption
Mud content % 4.70 1.50
Water % 2.20 2.00
content
Mass density g/cm3 1.34 1.35
Roughness % - 20.90
content
Properties examination of fine aggregate and coarse aggregate is also carried

out to determine the quality of the material used. The results of the mechanical
properties of the aggregates can be seen in Table 2. The results show that both types
of aggregate can be used as a mixture of concrete. The compressive strength test
was carried out on cube-shaped specimens with specimen sizes of 100 × 100 ×
100 mm referring to the ASTM C39 standard [26] as shown in Fig. 2. Meanwhile the
flexural strength test was carried out referring to ASTM C239 [27] with specimen
sizes of 150 × 150 × 600 mm. The flexural strength testing used the 4-point method
without a reinforcement inside the concrete. Variation of 0% fiber was a test object
as a control to determine the effectiveness of adding fiber to concrete. Fiber placed
at the bottom of the specimen up to 50% of the total height of the specimen. The
compressive strength test was taken from the average results of 3 specimens, while
the flexural strength was taken from the average results of 2 specimens. Apart from
the compressive strength and flexural strength tests, slump testing was also carried
out to determine the workability level of fresh concrete.
Fig. 2 Setting up for

compressive strength test for
hybrid concrete (bottom:
fiber concrete; top: normal
concrete)
3.1 Slump Test
Slump testing is carried out on all variations of specimens to determine the level of
workability in concrete using fiber as additional materials. The results of the slump
test for each variation can be seen in Fig. 3a the test result data and Fig. 3b the process
of the slump inspection. The test results showed that normal concrete without fiber
obtained a slump value of 8.5 cm, while normal concrete used fiber 0.08%, 0.16%,
and 0.24% of 7.0 cm, 5.5 cm, and 4.5 cm, respectively. The more increasing amount
of fiber used, the less slump value will result, causing the level of workability in the
concrete to consequently be lower. The decrease in slump value in concrete with a
high amount of fiber is due to is due to the ongoing water absorption by the fiber
during the casting process, causing the resulting workability to be reduced.
10
8.5
8 7.0
Slump test (cm)
5.5
6
4.5
4
0
0 0.08 0.16 0.24
(a) Fiber Variations (%) (b)
Fig. 3 a Relationship between fiber variations and slump test value; b View of slump test process
3.2 Compressive Strength
One of the hardened properties tests of concrete is the compressive strength test.
Tests were carried out on test concrete aged 3, 7, and 28 days. Concrete testing was
carried out at the age of 3 and 7 days to determine the characteristics of the concrete
at the initial age. Such measure was done because the bond between the concrete
and the fibers at the early-age concrete is feared to affect the compressive strength of
the concrete. The type of curing used in this test is water curing during the waiting
period of the concrete test. The results of the concrete compressive strength test can
be seen in Fig. 4. The results show that the increasing age of the concrete results in
higher compressive strength, this is because the bond between the aggregate or the
fiber and the paste is getting stronger.
The test results at the initial age of normal concrete are 27.03 MPa, while in hybrid
concrete it is 26.16 MPa, 27.26 MPa, and 27.88 MPa correspond to 0.08%, 0.16%,
and 0.24%, respectively. The results of the early-age concrete testing done on day 3
show that the concrete has developed high resistance to withstand compressive forces
due to the addition of Silica Fume to the concrete which can increase the compressive
strength of the early-stage concrete. Silica fume particles which are much smaller
than cement be able to cover small pores in the Allowing the concrete to become
denser and hardens faster so that it can increase the compressive strength of the
concrete at the early-age stage. The function of the high compressive strength in the
initial age is to help speed up the work process in concrete construction, with high
initial compressive strength. It does not require a long waiting time for the concrete
construction process because the resulting compressive strength is considered to be
able to withstand loads during the construction process.
In normal concrete, the one without fiber, a test on concrete aged 28 days produces
a compressive strength of 38.40 MPa. Whilst when done with the concrete containing
50.00
Compressive Strength (MPa)
3 Days 7 Days 28 Days

40.00
30.00
20.00
10.00
0.00
0.00 0.08 0.16 0.24
Fiber Variations (%)
Fig. 4 Compressive strength of concrete with different amount of fiber

0.08, 0.16, and 0.24% fiber, it produces compressive strength of 38.11; 38.88; and
38.55 MPa. In the 28-day concrete test, it can be concluded that the hybrid system
in concrete does not affect the compressive strength as seen from the results of
the compressive test for each variation which does not have a striking difference.
However, it is necessary to pay attention to the duration during the casting process of
fiber concrete with normal concrete. It is feared that the use of too lengthy duration
between the two types of concrete will result in cold joints in hybrid concrete. Cold
joints in concrete can reduce the value of compressive strength and flexural strength
in concrete. The results of cold joint testing on concrete can be seen in the following
studies [28, 29].
20.00
Flexural Strength (MPa)

15.00
10.00
5.00
0.00
0.00 0.08 0.16 0.24
(a) Fiber Variations (%)
20.00
Displacement (mm)

15.00
10.00
5.00
0.00
0.00 0.08 0.16 0.24
(b) Fiber Variations (%)
Fig. 5 Bending test results with different amount of fiber a Flexural strength; b displacement at
the peak load
Fig. 6 Before and after bending test
3.3 Flexural Strength
The variation carried out in the flexural strength test is the same as the compressive
strength test. The results of the flexural strength test can be seen in Fig. 5a, while the
resulting maximum displacement can be seen in Fig. 5b. The results of the flexural
strength test show that the increased fiber content results in higher flexural strength
and the higher displacement value obtained. Figure 5a shows that the flexural strength
at the initial age (3 days) of normal concrete is 2.44 MPa, while the concrete with the
addition of fiber produces a successive flexural strength of 3.15 MPa (0.08% fiber),
4.15 MPa (0.16% fiber), and 4.81 MPa (0.24% fiber). However, the flexural strength
results of concrete at 28 days are 5.89 MPa (0% fiber), 7.28 MPa (0.08% fiber),
7.65 MPa (0.16% fiber), and 8.55 MPa (0.24% fiber). It can be concluded that the
hybrid system in flexural testing is very effective in increasing the flexural strength
of concrete. This is because the concrete is weak in resisting the tensile force if the
fibers are placed in the tensile area allowing the fibers to increase the capacity of the
concrete. Figure 6 shows the result of the test object before and after testing.
4 Conclusion
Based on the aforementioned result and discussion, conclusion can be presented as

follows:
a. The addition of fiber to concrete affects the workability of fresh concrete, a high
amount of fiber reduces the level of workability in fresh concrete.
b. A hybrid system in fiber concrete combined with normal concrete does not affect
the compressive strength value for the fiber is only placed in the bottom area of
the specimen.
c. The high amount of fiber in the hybrid system increases the flexural strength of
the concrete. This is because the fibers work effectively in resisting the tensile
forces that occur during the bending test.
References
1. Afriandini B, Monika F, Saleh F, Prayuda H, Cahyati MD, Djaha SIK (2020) Fresh and hardened
properties of self fiber compacting concrete (SFCC) incorporated with zeolite and nylon. IOP
Conf Ser Mater Sci Eng 771:1–9
2. Saad M, Agwa IS, Abdelsalam A, Amin M (2020) Improving the brittle behavior of high
strength concrete using banana and palm leaf sheath fibers. In: Mechanics of advanced materials
and structures, pp 1–10
3. Navilesh J, Rahul BK, Shankar BK, Patil S, Gutteder VA (2017) A study on hybrid fiber
reinforced concrete. Int Res J Eng Technol 4(06):1647–1656
4. Chandak A, Agrawal N, Thakur D, Titiksh A (2016) Analysis of self-compacting concrete
using hybrid fibers. Int J Trend Res Dev 3(2):641–645
5. Manoharan S, Kumar GS, Balasubramainan K, Dhanasakkaravarthi B (2016) Investigation on
properties of banana and bamboo fiber composites by experimentation. Middle East J Sci Res
24:16–18
6. Kavitha S, Kala TF (2016) Effectiveness of bamboo fiber as a strength enhancer in concrete.
Int J Earth Sci Eng 9(3):192–196
7. Castillo-Lara JF, Flores-Johnson EA, Valadez-Gonzalez A, Herrera-Franco PJ, Carrillo JG,
Gonzalez-Chi PI, Li QM (2020) Mechanical properties of natural fiber reinforce foamed
concrete. Materials 12(3060):1–18
8. Abirami R, Vijayan DS, John SJ, Albert A, Alex AK (2020) Experimental study on concrete
properties using pineapple leaf fiber. Int J Adv Res Eng Technol 11(6):913–920
9. Todkar SS, Patil SA (2019) Review on mechanical properties evaluation of Pineapple Leaf
Fiber (PALF) reinforce polymer composites. Compos B 174:1–16
10. Tampi R, Parung H, Dhamaluddin R, Amiruddin AA (2020) Reinforced concrete mixture using
abaca fiber. IOP Conf Ser Earth Environ Sci 419:1–7
11. Suhelmidawati E (2016) Tensile tests of abaca fiber as one of alternative materials for retrofitting
of Unreinforced Masonry (URM) Houses. Rekayasa Sipil 13(2):22–29
12. Geremew A, Winne PD, Demissie TA, Backer HD (2021) Treatment of natural fiber for
application in concrete pavement. Adv Civ Eng 2021:1–13
13. Shankar PS, Reddy KT, Sekhar VC (2013) Mechanical performance and analysis of banana
fiber reinforced epoxy composites. Int J Recent Trends Mech Eng 1(4):1–10
14. Elbehiry A, Elnawawy O, Kassem M, Zaher A, Uddin N, Mostafa M (2020) Performance of
concrete beams reinforced using banana fiber bars. Case Stud Constr Mater 13:1–13
15. Kumaat EJ, Mondoringin MRIAJ, Manalip H (2018) Basic behavior of natural banana stem
fiber reinforced concrete under uniaxial and biaxial tensile stress. Int J Geomate 14(44):166–175
16. Tirkey N, Ramesh GB (2018) Experimental study on the banana fiber reinforced concrete. Int
J Pure Appl Math 119(18):2053–2056
17. Musthafa S, Basha SM, Bhavani S, Anudeep P, Chakravarhy P, Vali SN (2019) Experimental
study on partial replacement of cement by banana leaves ash and glass fiber. Int J Res Advent
Technol (Special Issue):650–655
18. Libre NA, Shelarchi M, Mahoutian M, Soroushian P (2011) Mechanical properties of hybrid
fiber reinforced lightweight aggregate concrete made with natural pumice. Constr Build Mater
25:2458–2464
19. Dawood ET, Ramli M (2011) Contribution of hybrid fibers on the hybrid fibers on the properties
of high strength concrete having high workability. Procedia Eng 14:814–820
20. Oucief H, Habita MF, Redjel B (2006) Hybrid fiber reinforced self-compacting concrete:
hardened properties. Int J Civ Eng 4(2):77–85
21. Yuan T, Lee J, Min K, Yoon Y (2019) Experimental investigation on mechanical properties of
hybrid steel and polyethylene fiber reinforced no-slump high strength concrete. Int J Polym
Sci 2019:1–11
22. Fard SG (2020) Hybrid fiber reinforced concrete with pozzolanic materials. J Cement Based
Compos 1:16–24
23. ASTM International, ASTM C595/C595M-20 (2020) Standard specification for blended
hydraulic cements, American Standards Testing Materials, West Conshohocken, United States
24. ASTM International, ASTM C1240-15 (2013) Standard specification for silica fume used in
cementitious mixtures, American Standards Testing Materials, West Conshohocken, United
States
25. ASTM International, C494/C4949M-13 (2015) Standard specification for chemical admixtures
for concrete, American Standards Testing Materials, West Conshohocken, United States
26. ASTM International, C39/C39M-18 (2019) Standard test method for compressive strength of
cylindrical concrete specimens, American Standards Testing Materials, West Conshohocken,
United States
27. ASTM International, C293/C293M-16 (2016) Standard test method for flexural strength of
concrete (using simple beam with centre point loading), American Standards Testing Materials,
West Conshohocken, United States
28. Zega BC, Prayuda H, Monika F, Saleh F, Wibowo DE (2021) Effects of cold joint and its
direction on the compressive and flexural strength of concrete. Int J Geomate 20(82):86–92
29. Illangakoon GB, Asamoto A, Nanayakkara A, Trong LN (2019) Concrete cold joint formation
in hot weather conditions. Constr Build Mater 209:406–415
Investigation of Polypropylene Fiber
Reinforced Concrete After Elevated
Temperature Using Color Quantification
and Alkalinity Method
Ni Nyoman Kencanawati, Suryawan Murtiadi, and Zul Aida Nur
Abstract Concrete properties change when it is exposed to high temperature. Fire

can reduce the strength and the stiffness of the concrete. Another visible thing is the
change in the color of the concrete. Furthermore, a decrease in the alkaline degree
of concrete can also occur at high temperatures. This study combines quantifying
the value of the color change, quantifying the alkaline level, and evaluating the
residual strength of post-fire concrete. Two concrete grades were prepared: 25 and
45 MPa and exposed to a high temperature of 300 and 700 °C. The amount of 1.8, 2
and 2.5 kg/m3 of polypropylene fiber were added to the concrete. According to the
temperature rise analysis, the farther the position from the concrete surface causes
a decrease in temperature. The 2.5 kg/m3 fiber content shows more excellent post-
fire mechanical performance because it provides the pores needed by concrete when
water vapor pressure occurs due to high temperature; therefore, cracks in the cement
paste can be avoided. Visually, the color change between grade 25 and 45 MPa are the
same in each temperature; however, they have different coordinates as observed by
the color test. Furthermore, the addition of 2.5 kg/m3 fiber to the concrete contributes
a superior effect to prevent the decrease of the alkaline level during the fire.
Keywords Concrete color quantification · Concrete pH · Elevated temperature ·

Polypropylene fiber
1 Introduction
Concrete is often used as material construction worldwide because of many advan-

tages. Concrete is relatively easy in terms of producing and maintenance and is more
N. N. Kencanawati (B) · S. Murtiadi · Z. A. Nur

Postgraduate Study Program of Civil Engineering, Mataram University, Jl. Majapahit 62,
Mataram 83125, Indonesia
S. Murtiadi
1154 N. N. Kencanawati et al.
resistant to climate change. However, the quality of the ingredient materials deter-
mines the strength of concrete. Fiber is also often added to concrete mixtures to
improve its properties [1, 2]. Polypropylene fiber is one of the most widely used
fibers in concrete mixtures. These fibers are in the form of filaments that will spread
into the mixture when mixed in the concrete mix. When the temperature increases,
polypropylene fibers can increase permeability by forming pore cavities to release
concrete moisture. The existence of this water vapor release path can reduce spalling
in the concrete when a fire occurs. European standards require the addition of a
minimum of 2 kg/m3 polypropylene fibers to reduce spalling in fire-induced concrete
[3, 4].
After being exposed to high temperatures, the concrete will change its proper-
ties depending on the fire’s temperature. Fires can cause cracking or spalling of
the concrete, which can reduce its compressive strength. Another visible thing is
the change in the color of the concrete. In general, the concrete’s color becomes
slightly reddish when the temperature reaches 300 °C, turns greenish grey when the
temperature becomes 600 °C and even turns yellow when the temperature increases
to 1200 °C [5–8]. A decrease in the alkaline degree of concrete can also occur at
high temperatures. A decrease in the concrete’s pH is a sign that there is carbonation
of the concrete due to a fire [9].
Assessment of post-fire concrete condition, primarily to determine the residual
strength, is mostly done using non-destructive tools on the concrete. The hammer
test and ultrasonic pulse velocity methods are often used to estimate the residual
strength after a fire. The rebound number indicating the surface hardness of the
concrete is related to the compressive strength of the existing concrete based on
the conversion value or the tool’s calibration graph. Likewise, the velocity of the
ultrasonic wave propagation that is transmitted into the concrete body is collected by
the receiver sensor. It enables us to describe the quality of the material in the concrete
body, whether cracks have occurred or are still sound after being exposed to high
temperatures. Research on the application and analysis of non-destructive testing on
concrete at elevated temperatures has been well established [10–12].
During assessment after a fire, we also need to investigate the temperature that
has been reached by the concrete. Visual assessment by evaluating the color change
can lead to the demand. Various methods have been used to describe the color of
post-fire concrete. Spectrophotometer method, polarization microscopy, colorimetry,
and digital applied concepts have been studied. Samples are prepared either from a
concrete surface, heated concrete powder, or image capture directly on the concrete
surface. These methods are continually evolving to quantify post-fire concrete color
[13–16].
Examination of the alkaline content of the concrete also benefits in the assessment.
Concrete that has experienced a decrease in pH certainly indicates carbonation in
the concrete and requires proper follow-up repair. To date, checking the concrete’s
pH has been done by spraying phenolphthalein on the surface of the concrete. If the
concrete surface does not change color to magenta, carbonation has occurred in the
concrete [9].
Investigation of Polypropylene Fiber Reinforced Concrete … 1155
Table 1 Concrete mixture proportion

Grade (MPa) Weight (kg/m3 )
Water Cement Fine aggregate Coarse aggregate
25 205 410 694 1041
45 205 586 624 936
To the end, research has only emphasized each method on the behavior of post-
burn concrete mainly the investigation on mechanical properties. This study enriches
the investigation by applying the series methods of quantifying the value of the color
change, the method of quantifying the alkaline level in concrete, and evaluating the
residual strength value of post-fire concrete. Combining these methods is useful to
assess concrete after a fire and to produce more detailed conclusions.
2 Methodology
2.1 Material
Two concrete grades were prepared, which were 25 and 45 MPa. The water-cement
ratios were 0.46 and 0.35, respectively, for grade 25 and 45 MPa. Natural aggregate
was utilized as fine aggregate with a maximum size of 4.75 mm. Meanwhile, crushed
stone was used as coarse aggregate with a maximum size of 20 mm. The mixture
proportion of each concrete grade is shown in Table 1.
The mixture proportion shown in Table 1 is regarding as the control mixture
proportion without polypropylene fiber. Then the series of 1.8, 2 and 2.5 kg/m3 of
polypropylene fiber were added to each control mixture. The fiber length was 12 mm
and the diameter was 0.006 mm. Specimens were cylinder concretes with a size of
150 mm in diameter and 300 in height.
2.2 Temperature Growth
After curing time for 60 days, the concrete was heated to 300 and 700 °C. Concrete
subjected to elevated temperature can be carried out if the water content in the
concrete pores has decreased significantly. When there is still excess water in the
concrete pores, a water vapor pressure explosion will inevitably occur during exposed
high temperatures. This explosion can cause cracks in the cement paste due to the
high tensile stress. Therefore, exposure should be carried out at a concrete age of
more than 28 days.
After reaching the intended temperatures, they were kept until one hour. The
growth of temperature was according to ISO 834 standard furnace test [17] as shown
Fig. 1 Electric furnace
Fig. 2 Temperature growth

[17]
in Eq. 10. The furnace was an electric furnace, as shown in Fig. 1 The maximum
temperature capacity was 1200 °C. The relation between time and temperature from
the standard furnace test is given in Eq. 1 and Fig. 2.
θg = 20 + 345log(8t + 1) (1)
where t is time measured in minute.
2.3 Procedures of pH Level Measurement
The digital pH meter using the Ezdo brand was used to measure the alkaline content
of concrete after the fire. This tool can measure the pH of the material from a range
of 0–14. The pH testing process began with extracting concrete samples after the
compression test with a thickness of approximately 2 cm from the surface. A sample
of 30 g of concrete cement paste was crushed to a powder form and dissolved in
50 ml distilled water (Aquades).
Fig. 3 Temperature rise in concrete for 700 °C exposed to fire
The solution is mixed with a stirrer to ensure homogeneity during measuring pH.
At the same time, the pH measurement is carried out by inserting the sensor into the
solution and the pH value was read on the display.
3.1 Temperature Rise
The temperature rise on the surface and some concrete depth using Wickstrom’s
method for a maximum temperature of 700 °C is given in Fig. 3. The temperature of
the concrete surfaces approaches 657 °C; meanwhile, in the inner layer, the temper-
ature is 82 °C. This temperature rise occurs in concrete, which has been exposed
to the furnace test for one hour. The farther the position from the concrete surface
causes a decrease in temperature.
3.2 Concrete Strength at Elevated Temperature
At ambient temperatures (20 °C), it appears that concrete with the addition of
polypropylene fibers of 1.8 kg/m3 reaches the highest compressive strength among
other concrete with an additional proportion of other polypropylene fibers. This
occurs either in the grade of 25 or 45 MPa. The addition of 1.8 kg/m3 polypropylene
fiber has a compressive strength greater as much as 1.4 times compared to that of
control concrete without fiber, both on concrete grade 25 and 45 MPa.
However, when the temperature increased to both 300 and 700 °C, the concrete
containing 2 kg/m3 polypropylene fiber has the highest residual compressive strength
compared to concrete with other fiber content either in grade 25 or 45 MPa. For grade
25 MPa, concrete with a content of 2.0 kg/m3 only decreased by 30% from its initial
strength when being exposed to a fire of 700 °C. For concrete with grade 45 MPa,
it shows even better performance, which only decreases by 20%. Even at ambient
temperatures, concrete with a fiber content of 1.8 kg/m3 has the optimum strength,
yet concrete with a fiber content of 2 kg/m3 shows better performance when the
temperature increases. When the temperature was high, the polypropylene fibers
melted and provided more pores so that water subjected to vapor pressure due to
high temperatures can pass and does not push the surrounding cement paste. This
fiber content provides the pores needed by concrete when water vapour pressure
occurs due to high temperature; therefore, cracks in the cement paste can be avoided.
Thus, the concrete’s performance with a fiber content of 2 kg/m3 is better during the
fire. Figure 4 and Fig. 5 present the concrete compressive strength at ambient and
elevated temperature for concrete grade 25 MPa and 45 MPa respectively.
Fig. 4 Post-fire compressive strength of grade 25 MPa
Fig. 5 Post-fire compressive strength of grade 45 MPa

3.3 Color Identification
The change of concrete surface color after exposed to 300 and 700 °C is presented in
Fig. 6. The color test was carried out qualitatively and quantitatively. Qualitatively,
the examination was conducted by visual inspection. Meanwhile, the quantitative
test was performed using a spectrophotometer-based from the Color HunterLab tool.
Color systems are represented by numbers in x, y, and z coordinates. L is a black to
white color with a value range between black (0) and white (+100) and is located on
the z-axis. The a value lies between the colors magenta (+100) and green (−100)
and is located on the x-axis. The value of b is assigned between yellow (+100) and
blue (−100) and lies on the y-axis. The color test results are presented in Table 2 and
Table 3 for grade 25 MPa and 45 MPa respectively. Even though the color change
between Grade 25 and 45 MPa are the same in each temperature visually, however
they have different coordinates as observed by color test.
When the coordinates are plotted in a 3D curve (Figs. 7 and 8), the color pattern
can be more understood clearly. The ambient temperature shows the color which
lies mostly in the bottom, meanwhile due to exposed to the highest temperature,
the concrete color is positioned more at top of the curve. This pattern is valid for
Ambient temperature 300°C 700℃
Fig. 6 Visualization of pre and post-fire concrete
Table 2 Color test for grade 25 MPa

Polypropylene Visual examination Color quantification
(kg/m3 ) 20 300 °C 700 °C 20 °C 300 °C 700 °C
°C
L a b L a b L a b
0 Grey Brownish Greyish 34.83 0.17 3.98 62.71 0.70 5.22 68.50 0.07 3,29
grey white
1.8 Grey Brownish Greyish 42.78 0.65 5.17 65.85 0.25 5.66 63.26 0.16 3,51
grey white
2 Grey Brownish Greyish 43.45 0.28 6.41 64.33 0.17 3.24 66.61 0.24 3,43
grey white
2.5 Grey Brownish Greyish 47.63 0.65 5.83 65.75 0.19 3.51 65.38 0.02 5,08
grey white
Table 3 Color test for grade 45 MPa

Polypropylene Visual examination Color quantification
(kg/m3 ) 20 300 °C 700 °C 20 °C 300 °C 700 °C
°C
L a b L a b L a b
0 Grey Brownish Greyish 37.42 0.86 4.32 54.56 1.06 5.10 56.14 0.13 3.37
grey white
1.8 Grey Brownish Greyish 48.65 0.88 5.52 57.97 0.33 3.13 57.95 0.26 5.55
grey white
2 Grey Brownish Greyish 42.72 0.57 4.87 57.76 0.73 3.13 58.98 0.19 4.97
grey white
2.5 Grey Brownish Greyish 51.03 0.22 6.00 56.28 0.97 5.63 58.14 0.30 6.02
grey white
Fig. 7 Color visualization for grade 25 MPa
all concrete grades and each fiber contents. The color of red, green, and blue in
in the figures refers to the temperature of 20 °C, 300 °C and 700 °C respectively.
Meanwhile, the shape of square, diamond, circle, and triangle in the curves refers to
the content of fiber of 0 kg/m3 , 1.8 kg/m3 , 2 kg/m3 and 2.5 kg/m3 respectively.
Fig. 8 Color visualization for grade 45 MPa
3.4 Alkali Level Identification
The alkaline level of concrete can be indicated by the pH value. The pH value is
measured using a pH meter on concrete samples that have experienced high temper-
atures. The pH level of normal concrete was also measured as a control. The concrete
alkalinity before and after fire is shown in Fig. 9 for grade 25 MPa and Fig. 10 for
grade 45 MPa.
It appears that the properties of concrete are alkaline, with the top alkaline levels
ranging from a pH of 12.3. However, with the increase in temperature, the concrete’s
alkaline content decreases due to the chemical process. Considering the polypropy-
lene fiber content in concrete, the greatest the polypropylene content can be the
best in maintaining the pH value at high temperatures compared to other polypropy-
lene levels. It occurs either in grade 25 or 45 MPa. The addition of this fiber to the
Fig. 9 Alkaline level for

concrete grade 25 MPa
Fig. 10 Alkaline level for concrete grade 45 MPa
concrete is able to prevent the decrease of the alkaline level of the concrete because
the polypropylene fiber is made of plastic; therefore, it does not affect the alkalinity
level of the concrete during a fire. The alkalinity in concrete in concrete after a
fire is very important property to provide passive protection to steel reinforcement
continuously.
4 Conclusion
The 2 kg/m3 fiber content shows greater performance because it provides the pores
needed by concrete when water vapor pressure occurs due to high temperature;
therefore, cracks in the cement paste can be avoided. Visually, the color change
between grade 25 and 45 MPa are the same in each temperature; however, they
have different coordinates as observed by color test. The addition of this fiber to the
concrete is able to prevent the decrease of the alkaline level during the fire.
References
1. Tjokrodimuljo K (2007) Concrete technology (in Indonesian). Yogyakarta: Nafiri

2. Woodson RD (2009) Concrete structures: protection, repair and rehabilitation. Butterworth-
Heinemann
3. Purkiss JA (2007) Fire safety engineering design of structures. Elsevier, Second
4. Serrano R, Cobo A, Prieto MI, de las Nieves González M (2016) Analysis of fire resistance of
concrete with polypropylene or steel fibers. Constr Build Mater vol 122, pp 302–309. https://
doi.org/10.1016/j.conbuildmat.2016.06.055
5. Yüzer N, Aköz F, Öztürk LD (2004) Compressive strength–color change relation in mortars
at high temperature. Cem Concr Res 34(10):1803–1807. https://doi.org/10.1016/j.cemconres.
2004.01.015
6. Hager I (2014) Colour change in heated concrete. Fire Technol 50(4):945–958
7. Lee J, Choi K, Hong K (2009) Color and material property changes in concrete exposed to high
temperatures. J Asian Archit Build Eng 8(1):175–782. https://doi.org/10.3130/jaabe.8.175
8. Malik M, Bhattacharyya SK, Barai SV (2020) Thermal and mechanical properties of concrete
and its constituents at elevated temperatures: a review. Constr Build Mater 121398. https://doi.
org/10.1016/j.conbuildmat.2020.121398
9. Kencanawati N, Mahmud F, Merdana N, Ngudiyono N (2015) Alkaline onset detection of
concrete after fire as a preliminary estimation of steel reinforcement corrosion. Spektrum Sipil
2:1858–4896
10. Amini K, Jalalpour M, Delatte N (2016) Advancing concrete strength prediction using non-
destructive testing: development and verification of a generalizable model. Constr Build Mater
102:762–768. https://doi.org/10.1016/j.conbuildmat.2015.10.131
11. Uva G, Porco F, Fiore A, Mezzina M (2013) Proposal of a methodology for assessing the
reliability of in situ concrete tests and improving the estimate of the compressive strength.
12. Lin Y, Hsiao C, Yang H, Lin Y-F (2011) The effect of post-fire-curing on strength–velocity
relationship for nondestructive assessment of fire-damaged concrete strength. Fire Saf J
46(4):178–185. https://doi.org/10.1016/j.firesaf.2011.01.006
13. Annerel E, Taerwe L (2011) Methods to quantify the colour development of concrete exposed
to fire. Constr Build Mater 25(10):3989–3997. https://doi.org/10.1016/j.conbuildmat.2011.
04.033
14. Short NR, Purkiss JA, Guise SE (2001) Assessment of fire damaged concrete using colour
image analysis. Constr Build Mater 15(1):9–15. https://doi.org/10.1016/S0950-0618(00)000
65-9
15. Du S, Zhang Y, Sun Q, Gong W, Geng J, Zhang K (2018) Experimental study on color change
and compression strength of concrete tunnel lining in a fire. Tunn Undergr Sp Technol 71:106–
114. https://doi.org/10.1016/j.tust.2017.08.025
16. Annerel E, Taerwe L (2009) Revealing the temperature history in concrete after fire exposure
by microscopic analysis. Cem Concr Res 39(12):1239–1249. https://doi.org/10.1016/j.cemcon
res.2009.08.017
17. Lim L, Wade C (2016) Experimental fire tests of two-way concrete slabs, 2020. J 2(5):99–110
The Effect of 12.5% Metakaolin
and Variations of Silica Fume on Split
Tensile Strength and Modulus of Rupture
of High Strength Self-Compacting
Concrete (HSSCC)
Endah Safitri, Wibowo, Halwan Alfisa Saifullah, and Farhan Gilang Septian
Abstract High-strength self-compacting concrete (HSSCC) is significantly used

for the construction of roads and buildings due to its high compressive strength and
ability to compact without a vibrator. This research aims to determine the effect of
adding metakaolin and silica fume to increase HSSCC’s tensile strength and modulus
of rupture. The materials used were 12.5% metakaolin and 0, 9, 11, 13, and 15%
variations of silica fume by weight of cement. For the split tensile strength test, a
cylinder with a diameter of 15 cm and a height of 30 cm was used, while a beam
with dimensions of 10 × 10 × 40 cm was used to test the modulus of rupture. The
specimen was tested after 28 days. The result showed that the split tensile strength
and modulus of rupture of HSSCC concrete with 12.5% metakaolin and variation of
silica fume 0, 9, 11, 13, and 15% were 3.93, 4.76, 5.27, 4.94, 6.80, 7.66, 7.82, and
7.69 MPa.
Keywords HSSCC · Modulus of rupture · Split tensile strength
1 Introduction
High Strength Concrete is a concrete innovation with better workmanship and higher
requirements than the conventional which is impossible to achieve only with the use
of conventional materials and normal mixing [1]. Further research on admixtures and
additives for concrete mixtures, especially water reducers or plasticizers and super-
plasticizers, has produced high-strength concrete capable of enhancing all concrete
into a modern high-performance material [2]. Self-Compacting Concrete (SCC) is
an innovation with a high slump value and is capable of compacting itself without
using tools [3]. It has a high flowability value, able to flow under its weight, fill the
formwork, and reach its own highest density [4].
E. Safitri (B) · Wibowo · H. A. Saifullah · F. G. Septian

Department of Civil Engineering, Sebelas Maret University, Kota Surakarta, Indonesia
H. A. Saifullah
1166 E. Safitri et al.
Silica fume and metakaolin are pozzolanic admixtures often used in the manu-
facture of concrete to achieve certain characteristics. These materials are used as
additives or substitutes for cement due to their ability to increase the resistance and
strength of concrete compared to ordinary Portland cement [5]. Concrete added with
metakaolin increases its compressive strength, split tensile strength, flexural strength,
and modulus of elasticity [6].
From observation, the addition of metakaolin to the tensile strength of reinforced
concrete prevents cracking due to shrinkage and temperature changes. Meanwhile,
for unreinforced concrete, it is used in the planning and construction of roads, airports,
and prestressed concrete. A direct tensile strength test needs to be carried out to
determine its value on concrete. However, this test is difficult to conduct due to
frequent failures during testing, thereby making it impossible to obtain the tensile
strength value. Therefore, an indirect or split tensile strength commonly known as
the Brazilian Test or Splitting Test is used [7]. Another method used to determine
the value of the tensile strength of concrete is the Modulus of Rupture (MOR) test.
Based on ASTM C293/C293M, the modulus of rupture is used to determine the
tensile strength of concrete using the center point loading method due to its higher
value [8].
2 Experimental Investigation
The mix design of HSSCC concrete with 12.5% metakaolin and variations of Silica
Fume was made with 0.25 water-cement factors and 1.5% superplasticizer, as shown
in Table 1.
The split tensile strength test of concrete is based on Indonesian National Standard
(SNI) 03-2491-2002 [9]. The specimens were tested at the age of 28 days using the
Brazilian or Splitting Test method as shown in Fig. 1. Meanwhile, Fig. 2 shows that
the specimens used were concrete cylinders with a diameter of 15 cm and a height
of 30 cm. A compression testing machine was used to test the split tensile strength
of concrete.
Table 1 Recapitulation of mix design of high-strength self-compacting concrete (HSSCC)

Specimen code Fine Coarse Cement Metakaolin Silica Superplasticizer Water
aggregate aggregate (kg/m3 ) (kg/m3 ) fume (l/m3 ) (l/m3 )
(kg/m3 ) (kg/m3 ) (kg/m3 )
HSSCC-SF0% 915.30 777.90 525 75 0 9 150
HSSCC-SF9% 905.72 769.76 471 75 54 9 150
HSSCC-SF11% 903.59 767.95 459 75 66 9 150
HSSCC-SF13% 901.46 766.14 447 75 78 9 150
HSSCC-SF15% 899.33 764.33 435 75 90 9 150
The Effect of 12.5% Metakaolin and Variations of Silica Fume … 1167
Fig. 1 The Brazilian

splitting test
Fig. 2 Split tensile strength test setup
Fig. 3 Modulus of rupture test using Universal Testing Machine
The test object for the modulus of rupture was a concrete block with a size of
10 × 10 × 40 cm. The center point loading method was used to determine the
effective span length of the beam at 30 cm. The test was carried out at the age of
28 days using the Universal Testing Machine as shown in Fig. 3. The modulus of
rupture of concrete is determined by SNI 2847-2019 at 10–15% f’c [10].
3 Experimental Results and Discussion
Concrete is categorized as self-compacting concrete supposing it qualifies the

requirements by EFNARC 2005 such as filling and passing abilities, and segregation
resistance. These attributes are tested using the slump flow, L-Box, and V-Funnel
methods [4]. The results of the recapitulation of the self-compacting concrete test
are shown in Table 2.
Based on EFNARC 2005, the requirements for slump flow are T500 average
between 2 and 5 s and d average of 650–800 mm. For the L-Box H2/H1 required is
0.8–1, with a V-Funnel of 6–12 s [4]. The test results of the three methods showed
that the specimens with 12.5% metakaolin and variations of silica fume qualify the
requirements as self-compacting concrete.
The split tensile strength of concrete is calculated using Eqs. 1 and 2.
P πLD
f ct = where A = (1)
A 2
2P
f ct = (2)
πLD
where f ct is the split tensile strength (MPa), P is the maximum test load (N), A is
the cross-sectional area (mm2 ), L is the specimen length (mm) and D is the diameter
(mm).
The results of the concrete split tensile strength test are shown in Table 3 and
Fig. 4.
The addition of silica fume to HSSCC with 12.5% metakaolin is able to increase
the split tensile strength of the concrete as shown in Table 3 and Fig. 4. Metakaolin
and silica fume fill the voids between the aggregates, thereby making the concrete
denser and able to support larger loads. The results showed that the maximum and
optimum split tensile strengths of concrete were 5.78 MPa at 11% silica fume, and
6.02 MPa at 11.11% silica fume, respectively.
Table 2 Recapitulations of self-compacting concrete test result

Specimen code Slump flow L-box V-funnel
T500 average d average (mm) H2 /H1 Time (s)
(s)
HSSCC-SF 0% 4.7 690 0.81 12
HSSCC-SF 9% 4.6 710 0.9 10
HSSCC-SF 11% 4.1 725 0.91 9.5
HSSCC-SF 13% 3.9 730 1 7.5
HSSCC-SF 15% 3.5 740 1 6.5
Table 3 Split tensile strength test result

Specimen code f ct (MPa) Percentage increase in split tensile strength of concrete (%)
HSSCC-SF0% 3.93 –
HSSCC-SF9% 4.76 21.25
HSSCC-SF11% 5.78 47.31
HSSCC-SF13% 5.27 34.16
HSSCC-SF15% 4.94 25.74
8
Split Tensile Strength of
6
Concrete (MPa)
4
y = -110.74 2 + 24.61 + 3.89
2
R² = 0.73
0
0% 9.00% 11.00% 13.00% 15.00%
Silica Fume content (%)
Fig. 4 Relationship between split tensile strength and silica fume content
The modulus of rupture is the maximum flexural strength of the concrete before it
collapses due to loading. The tests showed the load limit that the concrete or structure
is capable of supporting before the collapse. Furthermore, the center point loading
method was used, where a 10 × 10 × 40 cm beam is loaded at one point in the middle
of the span, as shown in Fig. 5.
Modulus of rupture can be calculated using Eq. 3.
3P L e f
R= (3)
2B H 2
Fig. 5 Center point loading method

Table 4 Modulus of rupture and compressive strength test result

Specimen code R (MPa) Percentage increase in MOR (%) fc (MPa) R/fc × 100 (%)
HSSCC-SF0% 6.80 – 41.99 16.19
HSSCC-SF9% 7.66 12.64 55.39 13.83
HSSCC-SF11% 7.97 17.21 67.75 11.76
HSSCC-SF13% 7.82 15.08 61.80 12.65
HSSCC-SF15% 7.69 11.60 58.03 13.25
10
Modulus of Rupture (MPa)
8
6 y = -78.977x2 + 18.116x + 6.7898
4 R² = 0.9556
2
0
0% 9.00% 11.00% 13.00% 15.00%
Silica Fume content (%)
Fig. 6 Relationship between modulus of rupture and silica fume content
where R is the modulus of rupture (MPa), P is the maximum test load (N), L ef is the
effective length of test beam (mm), B is the width of test beam (mm) and H is the
height of test beam (mm).
The result showed that the addition of silica fume to HSSCC with 12.5%
metakaolin was able to increase the concretes modulus of rupture, as shown in
Table 4 and Fig. 6. The greater the silica fumes content, the higher the modulus
of rupture, which decreases in value after reaching a maximum point. Furthermore,
the maximum modulus of rupture of HSSCC with 12.5% metakaolin occurred at
11% silica fume content of 7.97 MPa. The results of the regression analysis showed
an optimum modulus of rupture of 7.83 MPa at 11.47% silica fume content. Table 4
shows the Modulus of Rupture value ranging from 11 to 16% fc , which is close to
the value of concrete that has been determined by SNI 2847-2019 of 10–15% of its
compressive strength.
A type of collapse considered valid in the modulus of rupture test is a crack in
the middle of the test object. It is found in both concrete with and without silica
fume, which means that all test objects can be considered valid. The collapse of the
modulus of the rupture test object is shown in Fig. 7.
Fig. 7 Concrete collapse

when testing the modulus of
rupture
4 Conclusion
In conclusion, the addition of 12.5% metakaolin and variations of silica fume up

to 15% content still maintains the properties of the concrete as SCC. The optimum
split tensile strength was achieved at 11.11% silica fume with a split tensile strength
value of 6.02 MPa. Meanwhile, the optimum Modulus of Rupture at a value of
7.83 MPa was achieved at 11.47% silica fume with a value of compressive strength
of 11–16%. The addition of metakaolin and silica fume increases the density of
the concrete through the ball bearing effect, therefore, it has good workability and
reduces the possibility of bleeding and segregation. This makes the concrete denser
and increases its capacity to support loads.
Acknowledgements The authors are grateful to Sebelas Maret University for supporting this
research through the Fundamental Research Scheme from the 2021 funds of non-APBN, under
the contract number: 260/UN27.22/HK.07.00/2021.
References
1. Nawy EG (1996) Prestessed concrete a fundamental approach. PrenticeHill, New York

2. Joseph A, Mathew LA, John R (2017) Performance of metakaolin on high strength self
compacting concrete. Int J Sci Technol Eng 3(12)
3. Wibowo, Mediyanto A, Dharmawan EA (2018) Study on effect of variations of metakaolin
addition on selfcompacting parameter of high strength concrete. Int J Integr Eng 10(2):93–97
4. European Federation of National Association Representing Concrete Specification and
Guidelines for Self-Compacting Concrete (EFNARC): The European Guidelines for Self-
Compacting Concrete Specification—Production and Use, Europe (2005)
5. Siddique R, Khan MI (2011) Supplementary cementing materials. Springer Berlin Heidelberg,
New York
6. Justice JM, Kennison LH, Mohr BJ, Beckwith SL, McCormick LE, Wiggins B, Zhang ZZ,
Kurtis KE (2014) Comparison of two metakaolins and a silica fume used as supplementary
cementitious materials. In: Proceedings seventh international symposium on utilization of high-
strength/high performance concrete
7. Indriyantho BR (2014) Nuroji: finite element modeling of concrete fracture in tension with the
Brazilian splitting test on the case of plane-stress and plane-strain. Procedia Eng 95(2014):252–
259
8. American Standard Testing and Material: ASTM C293/C293M—16 standard test method for
flexural strength of concrete (using simple beam with center-point loading), Amerika Serikat
(2016)
9. Badan Standardisasi Nasional: SNI 03-2491-2002 Metode Pengujian Kuat Tarik Belah Beton,
Jakarta, Indonesia (2002)
10. Badan Standardisasi Nasional: SNI 2847-2019 Persyaratan Beton Struktural untuk Bangunan
Gedung dan Penjelasan, Jakarta, Indonesia (2019)
The Strength and Modulus of Elasticity
of High Strength Self-compacting
Concrete (HSSCC) with 12.5%
Metakaolin and Variations of Silica Fume
Endah Safitri, Wibowo, Halwan Alfisa Saifullah,
and Fernanda Sarwatatwadhika Perdana
Abstract The strength and modulus of elasticity of concrete are essential in deter-
mining the quality of concrete. In general, concrete with high strength receives a
greater load that reduces the dimensions of the structural elements. The use of high-
strength self-compacting concrete (HSSCC) with 12.5% metakaolin and variations
of silica fume is perceived to be an innovation in construction. Metakaolin and
silica fume work as pozzolans with a smaller size than cement to fill voids between
aggregates and increase concrete density. This research used 12.5% metakaolin and
variations of silica fume 0, 9, 11, 13 and 15% of weight of cement as an addition
(admixture). A cylinder with a diameter of 15 cm and a height of 30 cm was used to
test the strength and modulus of elasticity of concrete after 28 days. The test results
showed that the concrete could be categorized as HSSCC using 12.5% metakaolin
and variations of silica fume 0, 9, 11, 13, and 15%. The strength and modulus of
elasticity of HSSCC concrete with 12.5% metakaolin and variation of silica fume
0%, 9%, 11%, 13%, 15% were 41.99 MPa; 55.39 MPa; 67.75 MPa; 61.80 MPa;
58.03 MPa and 25,495.7 MPa; 29,980.8 MPa; 34,952.5 MPa; 32,067.0 MPa and
31,147.5 MPa, respectively.
Keywords HSSCC · Modulus of elasticity · Strength of concrete
1 Introduction
The heavy loads supported by high-rise buildings require significant steel reinforce-
ment in their structural elements, including columns, beams, and foundations. Most
reinforcements lead to a small distance between reinforcement, meaning that fresh
concrete can hardly flow to flow to all parts of the structural elements. Since concrete
compaction is difficult, much air is trapped in the concrete, decreasing its strength
E. Safitri (B) · Wibowo · H. A. Saifullah · F. S. Perdana

Department of Civil Engineering, Sebelas Maret University, Kota Surakarta, Indonesia
H. A. Saifullah
to support the load. Therefore, high-strength self-compacting concrete (HSSCC) is

used because of good workability [1]. Furthermore, high-strength concrete reduced
the size of the structural element dimensions to lessen the structure’s total weight.
Self-compacting concrete can be compacted without tools [2] and depends on the
proportion of aggregate size and superplasticizer admixture to attain the required
viscosity. The use of superplasticizers allows water in the mixture to be reduced
while maintaining workability and flowability. Metakaolin is a waste product of
kaolin calcination at a temperature of 450–900 ◦ C with a smaller size than silica
fume with high SiO2 and Al2 O3 content [3]. Silica fume with 1/100 diameter of
cement fills the space between aggregate and cement paste. Both elements are the
main constituents of cement, hence metakaolin can be used as a substitute for cement
[4, 5]. Metakaolin as a cement substitute with a concentration of 12.5% increases
the compressive strength of the concrete [6].
The compressive strength and modulus of elasticity of the concrete are measured
to define the performance of concrete. The compressive strength of concrete is the
capability to receive compressive forces per unit area. The modulus of elasticity is the
ratio of the pressure to the change in shape per unit length. The greater the compres-
sive strength and modulus makes the performance of concrete better. The addition
of metakaolin and silica fume is expected to increase the compressive strength and
modulus of elasticity of HSSCC concrete.
2 Experimental Investigation
Mix design of HSSCC concrete with 12.5% metakaolin and variations of silica fume
was made based on EFNARC [1]. Although the water-cement factor of 0.25 was
used, the optimum superplasticizer content was 1.5% of the weight of the binder.
Table 1 shows the proportions of each material.
The compressive strength and modulus of elasticity of concrete were tested using
cylindrical specimens with a diameter of 15 cm and a height of 30 cm. Compression
Testing machines were used to obtain the maximum load of concrete, as shown in
Fig. 1.
Table 1 Recapitulation of mix design of high-strength self-compacting concrete (HSSCC)

Specimen code Fine Coarse Cement Meta-kaolin Silica Superplasticizer Water
aggregate aggregate (kg/m3 ) (kg/m3 ) fume (l/m3 ) (l/m3 )
(kg/m3 ) (kg/m3 ) (kg/m3 )
HSSCC-SF0% 915.30 777.90 525 75 0 9 150
HSSCC-SF9% 905.72 769.76 471 75 54 9 150
HSSCC-SF11% 903.59 767.95 459 75 66 9 150
HSSCC-SF13% 901.46 766.14 447 75 78 9 150
HSSCC-SF15% 899.33 764.33 435 75 90 9 150
The Strength and Modulus of Elasticity of High Strength … 1175
Fig. 1 The test specimen and compression testing machine
To test the modulus of elasticity of concrete, a dial gauge was installed on the test
object in addition to the Compression Testing Machine to obtain axial deformation
data (Fig. 2).
Fig. 2 Dial gauge

3 Experimental Results and Discussion
Fresh concrete testing based on EFNARC [1] includes Slump Flow Test, L-Box,
and V-funnel to define whether the test object qualifies the requirements as self-
compacting concrete (SCC). Tables 2, 3 and 4, shows the test results.
Tables 2, 3 and 4 show that adding 12.5% metakaolin and variations of silica fume
increased the workability. The test objects were still included in the self-compacting
concrete category.
Table 2 Slump flow test results

Specimen Slump flow test
code T500 average Requirements Test d average Requirements Test
(second) results (mm) results
HSSCC-SF 4.7 2–5 s Qualify 690 650–800 mm Qualify
0% EFNARC EFNARC
HSSCC-SF 4.6 2005 Qualify 710 2005 Qualify
9%
HSSCC-SF 4.1 Qualify 725 Qualify
11%
13%
15%
Table 3 L-box test results

Specimen code h2 (mm) h1 (mm) h2/h1 Requirements Test results
HSSCC-SF 0% 9 11 0.81 0.8–1 Qualify
HSSCC-SF 9% 9 10 0.9 EFNARC 2005 Qualify
HSSCC-SF 11% 10 11 0.91 Qualify
HSSCC-SF 13% 9 9 1 Qualify
HSSCC-SF 15% 9 9 1 Qualify
Table 4 V-funnel test results

Specimen code V-funnel (s) Requirements Test results
HSSCC-SF 0% 12 6–12 s Qualify
HSSCC-SF 9% 10 EFNARC 2005 Qualify
HSSCC-SF 11% 9.5 Qualify
The compressive strength test was carried out at 28 days of concrete based on SNI
1974-2011 [7]. Equation 1 shows the formula used in calculating the compressive
strength of concrete.
P
f c = (1)
A
where fc’ is compressive stress (MPa), P is compression force (N), and A is cross-
sectional area (mm2 ).
The results of the concrete compressive strength test can be seen in Table 5.
Table 5 and Fig. 3 show that the addition of silica fume increases the compressive
strength of high-strength self-compacting concrete with 12.5% metakaolin. Further-
more, the maximum compressive strength of 67.75 MPa occurred at 11% silica fume.
The analysis shows that the optimum content of silica fume was at 11.54%, with a
compressive strength of 61.62 MPa.
Load and axial deformation of the test object was obtained from CTM and dial
gauge. From these data modulus of elasticity of each specimen can be calculated
using Eq. 2 to Eq. 4.
Table 5 Compressive strength test result

Specimen code fc (MPa) Percentage increase in compressive strength of concrete (%)
HSSCC-SF0% 41.99 –
HSSCC-SF9% 55.39 31.91
HSSCC-SF11% 67.75 61.35
HSSCC-SF13% 61.80 47.19
HSSCC-SF15% 58.03 38.20
Fig. 3 Relationship between 80

Compressive Strength of Concrete (MPa)
compressive strength and

70
silica fume content
60
50
40
y = -1503.9x2 + 347.14x + 41.596
30 R² = 0.81
20
10
0
0% 9% 11% 13% 15%
Silica Fume Content
Table 6 The modulus of

Specimen code Modulus of elasticity
elasticity test results
Eurocode 2-1992 ASTM C469
HSSCC-SF0% 19,827.02 19,907.00
HSSCC-SF9% 24,430.58 25,443.33
HSSCC-SF11% 33,600.07 29,571.67
HSSCC-SF13% 28,242.92 26,616.00
HSSCC-SF15% 27,178.14 26,161.67
P
Str ess σ = (2)
A
L
Strain ε = (3)
L
σ
Modulus of Elasticity E = (4)
ε
where σ is stress (MPa), ε is strain (mm/mm), L is axial deformation (mm), L is

specimen length (mm), and E is modulus of elasticity of concrete (MPa).
The value of the modulus of elasticity can be calculated using the standard
Eurocode 2-1998 and ASTM C469 (Eqs. 5 and 6), as shown in Table 6 and Fig. 4
[8, 9].
0,4 f c
Eurocode 2-1992 E c = (5)
ε(0,4 f c )
35000
33000
31000
Modulus of Elasticity (MPa)
29000 Eurocode 2-1992

27000
25000
23000
21000 ASTM C469
19000
17000
15000
0% 9% 11% 13% 15%
Silica Fume Content
Fig. 4 Relationship between modulus of elasticity and silica fume content

0,4 f c − σ1
ASTM C469 E c = (6)
ε(0,4 f c ) − ε1
where fc’ is maximum compressive stress (MPa), ε(0,4 fc’) is strain at 0,4fc’ (mm/mm),
σ 1 is stress at specific strain (MPa), and ε1 is specific strain (mm/mm).
Table 6 and Fig. 4 show that the addition of silica fume increases the modulus
of elasticity of concrete. The maximum modulus of elasticity occurred at 11% silica
fume content. After the maximum replacement level, the value of modulus of elas-
ticity is reduced. However, it was greater than the modulus of elasticity of HSSCC
without the addition of Silica Fume.
4 Conclusions
High-strength concrete with 12.5% metakaolin and up to 15% silica fume is still
included in the self-compacting concrete (SCC) category with good workability.
Optimum compressive strength occurred with the addition of 12.5% metakaolin
and 11.54% silica fume with a compressive strength of 61.62 MPa. The optimum
modulus of elasticity occurred at 11.47% silica fume with a modulus of elasticity
of 32,354.51 MPa. The physical properties of silica fume with a cement diameter of
1/100 fill the space between the aggregate and cement paste. Therefore, silica fume
increases the density of the concrete for a greater load.
Acknowledgements The authors express gratitude to Sebelas Maret University for supporting this
research through the Fundamental Research Scheme from the 2021 funds of non-APBN, under the
contract number: 260/UN27.22/HK.07.00/2021.
References
1. European Federation of National Association Representing Concrete Specification and Guide-

lines for Self-Compacting Concrete (EFNARC): The European Guidelines for Self-Compacting
Concrete Specification—Production and Use, Europe (2005)
2. Wibowo Mediyanto A, Dharmawan EA (2018) Study on effect of variations of metakaolin
addition on self- compacting parameter of high strength concrete. Int J Integr Eng 10(2):93–97
3. Barbhuiya S, Chow P, Memon S (2015) Microstructure, hydration and nanomechanical
properties of concrete containing metakaolin. Constr Build Mater 95:696–702
4. Sambowo KA (2001) Engineering properties and durability performance of metakaolin and
metakaolin-PFA concrete.Thesis, Faculty of Engineering at University of Sheffield, Sheffield
5. Siddique R, Khan MI (2011) Supplementary cementing materials. Springer Berlin Heidelberg,
New York
6. Joseph A, Mathew LA, John R (2017) Performance of metakaolin on high strength self
compacting concrete. Int J Sci Technol Eng 3(12)
7. Badan Standardisasi Nasional: SNI 1974:2011 Cara Uji Kuat Tekan Beton dengan Benda Uji
Silinder, Jakarta, Indonesia (2011)
8. European Committee for Standardization (2004) Eurocode 2-design of concrete structures—part

1-1 general rules and rules for buildings. Europe, Brussels
9. ASTM International: ASTM C469/C469M-14e1, Standard test method for static modulus of
elasticity and Poisson’s ratio of concrete in compression, West Conshohocken, PA, www.ast
m.org (2014)

ICRMCE

Uploaded by

Copyright:

Available Formats

ICRMCE

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ICRMCE

Uploaded by

Copyright:

Available Formats

Lecture Notes in Civil Engineering

Stefanus Adi Kristiawan

More information about this series at https://link.springer.com/bookseries/15087

Buntara S. Gan Mohamed Shahin

Proceedings of the 5th

Mohamed Shahin Akanshu Sharma

ISSN 2366-2557 ISSN 2366-2565 (electronic)

The International Conference on Rehabilitation and Maintenance in Civil

In the process of organizing this conference, we received invaluable motivation,

Halwan Alﬁsa Saifullah

• Part VIII: Sustainability aspects in water resources management

Surakarta, Indonesia Stefanus Adi Kristiawan

Factors Affecting Performance of Buildings and Infrastructures

The Inﬂuence of the Number and Height Steps of Terraced Model on

Assessment, Protection, Maintenance, Repair, and Retroﬁtting

A Proposed Method of FRP Anchorage for FRP Conﬁned

Maintenance Management of Buildings and Infrastructures

Development of Preventive Maintenance Guidelines

The Stability of a Slope on Soft Soil Using the Hardening

Simulation of Debris Flow Using “SIMLAR” in the Watershed

Risk Reduction and Safety Management

Analysis of the Behaviour Model of Foreign Motorcyclists in Tourism

Sustainability Aspects in Transportation Engineering

Sustainability Aspects in Construction Projects

Development of Blockchain and Machine Learning System in the

Sustainability Aspects in Water Resources Management

Analysis of Leb Irrigation Patterns of Pipe System in Sorgum Plants

Construction Materials for Sustainable Infrastructures

Mechanical Properties of Fine-Grained Concrete Using Fine-Red

The Strength and Modulus of Elasticity of High Strength

R. W. Triadji, M. A. Berawi, and M. Sari

Abstract Designers usually use Building Performance Simulation (BPS) to support

Keywords Machine learning · Building performance · Energy performance ·

The importance of digital technology and automation systems in playing a role in

R. W. Triadji (B) · M. A. Berawi · M. Sari

Fig. 1 Conceptual diagram of machine learning and existing BPS

3 Result and Discussion

3.1 Energy Performance

3.2 Environmental Performance

Environmental performance is also one of the criteria that need to be considered in

Jonie Tanijaya and Robby S. Kwandou

Abstract The structural analysis programs have become increasingly sophisticated

Keywords P-Delta · P-Delta plus large displacements · Lateral displacement

Fig. 1 P-Delta on column

sufficiently large as to break from linear compatibility relationships, then Large-

The force–displacement relationship of bending element is given by the following

Fig. 2 Secondary moment

The inertia I in Eq. (5) for column element is expressed as:

where, Ec = concrete modulus (MPa), Ig = moment of inertia of gross concrete

Fig. 3 Analysis model

Table 1 Parameter model

Table 2 Buckling analysis result

Table 3 Force and stiffness applied in model

4 Result and Discussion

Nonlinear result (P-Delta or P-Delta plus large displacement) refers to as NL while

P-Delta vs Linear Analysis

NL-P1 NL-P2 NL-P3 NL-P4

Fig. 4 Lateral displacement for P-Delta model

P-Delta plus Large Displacement vs Linear Analysis

NL-P1 NL-P2 NL-P3 NL-P4