Advances in Structural Engineering: K. V. L. Subramaniam Mohd Ataullah Khan Editors

Lecture Notes in Civil Engineering
K. V. L. Subramaniam
Mohd Ataullah Khan Editors
Advances in
Structural
Engineering
Select Proceedings of FACE 2019
Volume 74
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:
– Mr. Pierpaolo Riva at [email protected] (Europe and Americas);
– Ms. Swati Meherishi at [email protected] (Asia - except China,
and Australia, New Zealand);
– Dr. Mengchu Huang at [email protected] (China).
Indexed by Scopus and Compendex
More information about this series at http://www.springer.com/series/15087

K. V. L. Subramaniam Mohd Ataullah Khan
•
Editors
Advances in Structural
Engineering
Select Proceedings of FACE 2019
123
Editors
K. V. L. Subramaniam Mohd Ataullah Khan
Indian Institute of Technology Hyderabad Mahindra Ecole Centrale
Sangareddy, India Hyderabad, India
ISSN 2366-2557 ISSN 2366-2565 (electronic)

ISBN 978-981-15-4078-3 ISBN 978-981-15-4079-0 (eBook)
https://doi.org/10.1007/978-981-15-4079-0
© Springer Nature Singapore Pte Ltd. 2020
This work is subject to copyright. All rights are reserved 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
Preface
The primary objective of this book is to bring excerpts from the structural engineering
community comprising researchers and practitioners, inculcating the state of the art
and state of the practice. This communication was established through the conference
Futuristic Approaches in Civil Engineering (FACE) 2019 held at Mahindra Ecole
Centrale, Hyderabad, during August 30–31, 2019. The papers submitted to the FACE
2019 are published in two volumes covering the subareas of structural, geotechnical
and environmental engineering. These papers were peer-reviewed and concurrently
independently reviewed by the editors for quality. Subsequently, Springer as a
publisher went through a general quality and plagiarism check for the accepted
papers. Finally, a total of twenty papers passed through the scrutiny of a thorough
peer-review process to become a part of this book. This book is one of the two
volumes from the papers accepted through a peer-review process, and it presents
high-quality research papers for researchers and practitioners in the area of structural
engineering.
It has been a delightful experience editing the vital contributions from FACE
2019 in the form of a book. The experimental evaluation has always been a key to
capturing the mechanics of existing and novel structures while the structural
modeling and simulation capabilities have been the catalyst in establishing their
response. The state of the practice hinges on the confidence of the practicing
engineers in the existing and developed techniques. In this book, research devel-
oped in the areas of construction materials, structural health monitoring, earthquake
engineering, fiber-reinforced polymer composites is covered.
In the area of construction materials, the book covers a range of new knowledge
on raw materials and manufacture of cement, mixing, rheology and hydration,
admixtures, structural characteristics and performance of cement-based materials,
characterization techniques and modeling, use of fiber in cement-based materials,
degradation and repair of cement-based materials, novel testing techniques such as
digital image correlation (DIC) and applications. In addition, performance assess-
ment and suitability of sustainable alternatives to the present building materials are
also covered. For example, a few chapters suggest geopolymer as a replacement for
cement and one chapter presents the usage of recycled aggregates for concrete
v
vi Preface
production. Techniques such as fuzzy logic and artificial neural networks to

understand the experiments and establish the response are also discussed.
In the area of earthquake engineering, papers related to the areas of seismic
response of load-bearing unreinforced masonry wall, reinforced concrete frames,
buildings with dampers are covered. The focus of these chapters is the numerical
modeling driven by the material response of such structural components under
earthquake stimuli. Additionally, there are chapters on structures subjected to
vehicular impact and fire. Fundamental contributions toward bond behavior
between steel and concrete at elevated temperatures and moment–curvature rela-
tionship are also included.
From the organizers and the editors of FACE 2019, we would like to express
heartfelt gratitude toward the authors, for contributing chapters. The reviewers
played a vital role in ensuring the quality of the papers by providing critical
reviews. We would also like to thank Mahindra Ecole Centrale (MEC) in providing
us the opportunity to organize FACE 2019. Lastly, we would like to thank our
publisher, Springer, for accepting our proposal to publish the proceedings from
FACE 2019 as a part of this book. The support from Dr. Akash Chakraborty
(Associate Editor, Applied Science and Engineering, Springer), Mr. Daniel Joseph.
G (Project Coordinator, Book Production, Springer) and Rekha Nimesh (Editorial
Assistant, Applied Science and Engineering, Springer) is gratefully acknowledged.
Sangareddy, India Dr. K. V. L. Subramaniam

Hyderabad, India Dr. Mohd Ataullah Khan
Contents
Seismic Fragility Assessment of Unreinforced Masonry

Shear Walls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Jayaprakash Vemuri and K. V. L. Subramaniam
Axial Behaviour of Corroded CFST Columns Wrapped with GFRP
Sheets—An Experimental Investigation . . . . . . . . . . . . . . . . . . . . . . . . . 15
S. Vijaya Bhaskar Reddy and S. Sivasankar
Tensile Membrane Structures: An Overview . . . . . . . . . . . . . . . . . . . . . 29
Allan L. Marbaniang, Subhrajit Dutta, and Siddhartha Ghosh
Investigation of Cold-Formed Steel Members Subjected to Extreme
Low Temperatures Relevant to the Arctic Environment . . . . . . . . . . . . 41
Gaurav S. Chobe and Mahendrakumar Madhavan
Experimental Investigation on Crack-Arresting Mechanism
of Steel Fibre-Reinforced Concrete Prism Specimens Using DIC
and AE Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
Chandrashekhar Lakavath, Aniket Bhosale, and S Suriya Prakash
Estimation of Fly Ash Reactivity and Dissolution Characteristics . . . . . 67
G. V. P. Bhagath Singh and K. V. L. Subramaniam
Numerical Modelling of Continuous Composite Beam
Under Fire Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Priya S. Natesh and Anil Agarwal
Effect of Heating Rate on Bond Behavior Between Steel and Concrete
at Elevated Temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
Ira Banoth and Anil Agarwal
Idealised Bilinear Moment-Curvature Curves of Reinforced Masonry
(RM) Walls . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Jacob Alex Kollerathu
vii
viii Contents
Performance of Self-flowing Concrete Incorporated with Alumina

Silicates Subjected to Elevated Temperature . . . . . . . . . . . . . . . . . . . . . 111
Manish S. Dharek, Prashant Sunagar, K. Harish,
K. S. Sreekeshava, S. U. Naveen, and Bhanutej
Response of RC Short Column Under Combined Fire
and Axial Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
Chinthapalli Hemanth Kumar, B. Subash, and Anil Agarwal
Numerical Analysis of Post-earthquake Fire Resistance
of Concrete-Filled Tubular Steel Columns . . . . . . . . . . . . . . . . . . . . . . . 133
Smita Singh and Anil Agarwal
Mathematical Model for Prediction of Compressive Strength
of Normal, Standard and High Strength SCC with RCA . . . . . . . . . . . . 145
Srinivas Vasam, K. Jaganadha Rao, and M. V. Seshagiri Rao
Seismic Response of UHPC Strengthened Reinforced Concrete
Frame Using Concrete Damaged Plasticity Model . . . . . . . . . . . . . . . . . 159
Nidhi Sannametla and Jyosyula Sri Kalyana Rama
Mix Design Methodology for Fly Ash and GGBS-Based
Geopolymer Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
G. Mallikarjuna Rao and M. Venu
Experimental and Analytical Behavior of Recycled Aggregate
Concrete Using ANN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
B. Suguna Rao, Vithal Joshi, and Srikanth M. Naik
Effect of Vehicle Impact on Reinforced Concrete Structures . . . . . . . . . 195
Mohammed Mujeeb, Venkata Dilip Kumar Pasupuleti,
and Archanaa Dongre
Seismic Behaviour of G+7 RC Open Ground Storey Buildings
with Fluid Viscous Dampers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
Shaik Shayza and Bodige Narender
Mathematical Model for the Compressive Strength and Elastic
Properties of Triple-Blended Steel Fiber Self-Compacting Concrete
Based on the Experimental Investigation . . . . . . . . . . . . . . . . . . . . . . . . 217
S. Vijaya Kumar, B. Dean Kumar, and B. L. P. Swami
Prediction of Concrete Compressive Strength Using Fuzzy Logic
and Model Tree . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Vardhan Nagarkar, Preeti Kulkarni, and Shreenivas Londhe
About the Editors
Prof. K. V. L. Subramaniam is a Professor in the Department of Civil

Engineering at Indian Institute of Technology Hyderabad (IITH). Prior to joining
IITH, he was Professor and Catell Fellow in Department of Civil Engineering at
the Grove School of Engineering, the City College of New York (CCNY).
Dr. Subramaniam obtained a B.Tech. in Civil Engineering from IIT Delhi and Ph.D.
in Structural Engineering and Materials from Northwestern University, Evanston.
After graduation, Dr. Subramaniam worked as Research Associate at the NSF
Center for Advanced Cement Based Materials. Dr. Subramaniam was awarded the
Early Career Award from the National Science Foundation of USA for investi-
gating the early-age changes in cementitious materials. He received the James
Instrument Award from the American Concrete Institute (ACI) in 1999 for his
research on nondestructive evaluation of concrete. He is the recipient of the out-
standing young researcher award at the Grove School of Engineering in 2006. He
was the Chairman, committee 215 on Fatigue of Concrete of the ACI. In 2009, he
was elected Fellow of the American Concrete Institute for notable contribution in
the area of fracture and fatigue of concrete. He has served on the Editorial Boards
of the Journal of Materials in Civil Engineering and the Journal of Bridge
Engineering of the American Society of Civil Engineers. He is a life member of the
Indian Concrete Institute. Prof. Subramaniam is currently the Director of the Center
of Excellence in Sustainable Urban Development of the Ministry of Human
Resource Development at IIT Hyderabad. He is the Program leader for the
Sustainable Development thrust in the Friendship program of the Japan
International Cooperation Agency. He was the recipient of excellence in teaching
award at IIT Hyderabad in 2011 and 2015.
Dr. Mohd Ataullah Khan is an Assistant Professor in the Department of Civil

Engineering, Mahindra Ecole Centrale (MEC), Hyderabad, India. He has a master’s
in Structural Engineering from the Indian Institute of Science (IISc), Bangalore,
India and a PhD in Interdisciplinary Engineering from the Khalifa University of
Science and Technology (KUST), Abu dhabi. His PhD was funded by the Masdar
institute (a part of KUST) – Massachusetts Institute of Technology (MI-MIT)
ix
x About the Editors
cooperative agreement. The doctoral work of Dr. Mohd was focused on theoretical
and computational modeling, 3D printing and experimental evaluation of material-
tailored multilayers. Dr. Mohd has published articles in the top quartile mechanics
journals along with peer-reviewed conferences. His research interest lies in the
areas of mechanics of adhesively bonded joints, micro-/nano- fibre-reinforced
composites, novel bio-inspired composites and concrete 3D printing. Dr. Mohd also
had experience working as structural engineer for the design of steel structures,
prior to joining PhD.
Seismic Fragility Assessment
of Unreinforced Masonry Shear Walls
Jayaprakash Vemuri and K. V. L. Subramaniam
1 Introduction
The Kutch region in the western part of India is seismically active and has been
struck by destructive earthquakes in the past. The Mw 7.6 earthquake which struck
Kutch on 26 January 2001 was the largest intraplate earthquake to strike in India [1].
Post-earthquake damage surveys reported intensity levels up to X on the Modified
Mercalli Intensity (MMI) scale [2]. The epicentre lay about 9 km southwest of the
village of Chobari in the Kutch district of Gujarat, India [3]. The earthquake caused
widespread damage in the province of Gujarat. Overall, over a million structures
were damaged, and the consequent economic loss was close to US$10 billion [4]. In
the Kutch region, the earthquake destroyed several towns causing major devastation
of low-rise unreinforced masonry (URM) structures [5]. Typically, URM structures
in the region consist of load-bearing masonry shear walls made of low-strength
bricks. The performance of shear walls greatly affects the vulnerability of masonry
buildings, and it is imperative to assess their seismic fragility.
In India, 69% of buildings are constructed using unreinforced brick masonry [6].
However, barring a few classical studies [7, 8] the in-plane behaviour of URM walls
has not received much attention [9, 10]. The research reported varies widely due
to inconsistency in test practices [11]. While researchers have conducted nonlinear
time-history analyses of URM structures, there is no available literature presenting
the correlation between ground motion parameters and URM structural response.
Also, the lack of recorded seismic data from major earthquakes of the Kutch region
presents a major challenge to the seismic assessment. Even the available information
J. Vemuri (B)
Mahindra Ecole Centrale, Hyderabad, India
e-mail: [email protected]
Indian Institute of Technology Hyderabad, Sangareddy, India
© Springer Nature Singapore Pte Ltd. 2020 1
K. V. L. Subramaniam and Mohd. A. Khan (eds.), Advances in Structural
Engineering, Lecture Notes in Civil Engineering 74,
https://doi.org/10.1007/978-981-15-4079-0_1
2 J. Vemuri and K. V. L. Subramaniam
following the major Kutch earthquake (M w 7.6 at 8:46 a.m. on 26 January 2001)
is limited to recorded peak ground accelerations (PGA) at selected towns. Since
the accuracy of seismic assessment depends on capturing the local soil effects, the
absence of seismic records from the Kutch region complicates our understanding of
the actual seismic hazard.
In this study, ground motions are generated at rock level by the modified stochastic
finite-fault method. The PGAs of these synthetic ground motions generated at rock
level are compared with bedrock-level PGAs, which are available at thirteen stations.
After validation, strong ground motions are generated at surface level by accounting
for varied local topography by using site amplification functions for various towns
available in the literature [12]. The ground motions are generated at bedrock level
and amplified to surface level using the soil amplification functions corresponding
to each particular site. The cyclic response of masonry, which is generated using an
analytical-based model for the shear response of masonry, is used to investigate the
vulnerability of the masonry to different parameters of the ground motions.
2 Material and Geometric Properties of URM Walls
Under lateral forces, a URM wall may fail due to sliding, flexure/rocking or shear.
Sliding occurs along bed and head joints under low values of normal stress. Masonry
can also crack under direct tension. Walls with high aspect ratios typically exhibit
flexural or rocking behaviour while those with moderate aspect ratios fail in shear.
In shear failures of URM walls, pre-cracking behaviour is characterised by mild
hysteresis. Flexural cracks originate first, and then diagonal cracks start close to the
ultimate shear strength. Diagonal cracking is visible when the peak lateral force is
achieved, followed by a degradation in strength. In the post-peak regime, strength
and stiffness drop rapidly, and high level of energy is dissipated. The wall collapses
by stepped cracking along mortar joints. Cyclic response of the wall is bounded
by the monotonic shear–deformation (V − ) envelope. To obtain the monotonic
envelope for URM walls, an incremental procedure available in the literature [13,
14] is used, where the lateral force is increased in increments, and at each step, the
shear and flexural deformations are combined to obtain the load–deformation curve
(V − curve). The flexural strength is computed [7] as:

p Dt p
Vr = 1− (1)
2αv k fu
where p = P/H t is the vertical compressive stress and P is the vertical compressive
force. H , D and t represent the height, length and thickness of the wall, f u is the
masonry strength in compression, k is a coefficient which accounts for vertical stress
distribution at the compressed toe, αv = M/(V D) is the shear ratio. The shear
strength can be computed [7] as:
Seismic Fragility Assessment of Unreinforced Masonry Shear Walls 3
Vd = Dtτu ; τu = minimum(τcs ; τws ; τb ) (2)
τcs is the shear stress corresponding to the cracked section and is given by:
1.5c + μp
τcs = (3)
1 + 3cα
p
v
τws is the shear stress of the complete section and is given by:
c + μp
τws = (4)
1 + αv
τb is the shear stress based on brick cracking and is given by:

f bt p
τb = 1+ (5)
2.3 f bt
where c is the joint cohesion, μ is the joint friction coefficient, f bt is the tensile
strength of brick. The expressions for joint cohesion and joint friction have been
defined [15] as c = kc and μ = kμ where k is:
1
k=
(6)
1 + μ2 xy
where x and y are the height and length of the brick unit, respectively. Due to
the absence of experimental data on shear walls from India, two shear walls [16,
17] were selected from literature, which are representative of the typical masonry
construction in Kutch [18]. Table 1 shows the properties of the two URM walls. The
monotonic backbone envelope is computed for both the walls.
The cyclic behaviour was modelled in SAP2000 [19] where the link element is
chosen to model the URM wall. The multilinear-plastic pivot model is used to define
the hysteresis loop. The four pivot parameters, α1 , α2 , β1 and β2 , govern the shape of
the cyclic loops and the degradation in strength and stiffness. The optimum values
for the four parameters were obtained using the least-squares method to minimise the
error concerning the values obtained from experiments. For the wall with lower axial
stress, α1 , β1 , α2 and β2 are 0.90. The symmetrical response and no pinching resulted
in the parameters to be equal in magnitude. The hysteresis behaviour obtained from
the experiment matches well with the loops obtained from the pivot model (Fig. 1a).
For the shear wall having higher axial stress, α1 , β1 , α2 and β2 were computed to
be 0.90. Figure 1b shows the match between the experimental hysteresis loops of
the wall and the loops obtained from the pivot model. Limit states corresponding to
cracking, ultimate strength and collapse are noted for both walls.
Table 1 Properties of the two URM shear walls considered in the present study
Property Wall-1 Wall-2
H (mm) 1350 3000
D (mm) 1000 1500
t (mm) 250 380
p (MPa) 0.6 1.24
f b (MPa) 16 19.7
f bt (MPa) 1.22 1.07
f m (MPa) 3.31 4.33
f u (MPa) 6.2 7.9
c (MPa) 0.23 0.21
c (MPa) 0.17 0.14
μ 0.58 0.81

μ 0.43 0.55
strength
250
strength
a 100 b
cracking
collapse
cracking
collapse
Force (kN)
Force (kN)
80 200
60 150
40 100
20 50
0 0
-8 -6 -4 -2 0 2 4 6 8 -20 -15 -10 -5 0 5 10 15 20
-20 Displacement (mm) -50 Displacement (mm)
-40 -100
-60 Experiment -150 Experiment

-80 Pivot Model Pivot Model
-200
-100 -250
Fig. 1 Comparison between experiment and pivot model. a URM wall with diagonal shear cracking
through mortar joints; b URM wall with shear–tensile cracking of bricks
3 Nonlinear Time-History Analyses Using Synthetic

Ground Motions
Seismic records for the 2001 Kutch earthquake are not available. In such regions
where a rich database of historical strong motions is unavailable, modified records
may be obtained either by generating artificial records or by altering natural records.
The modified stochastic finite-fault method [20] has been used by researchers [21]
to generate surface-level synthetic ground motions for the 2001 Kutch earthquake.
From these synthetic accelerograms, ground motion characteristics such as PGA,
PGV, predominant period, mean period, significant duration and response spectra
are computed. Figure 2 shows the ground motions generated at surface level. At
each site, ten synthetic ground motions were generated by conducting trials. The
modified stochastic finite-fault method is extended to derive synthetic ground motions
at surface level, for towns across various seismic zones. These synthetic ground
motions are used for nonlinear time-history analyses.
The earthquake hazard in the province of Gujarat ranges from low to high. Based
on the estimated hazard level, the seismic zoning map of India IS1893:2016 [22]
has assigned four seismic zones to Gujarat, ranging from Zone 2 (low) to Zone 5
(high). The Kutch region has the highest earthquake hazard and lies in Zone 5 (SA
at DBE level of 0.45 g), and intensity levels more than IX may be expected in this
region. Zone 4 (SA at MCE level of 0.30 g) comprises regions from Saurashtra and
Kutch, within a belt with a width of 60–70 km around Zone 5 and intensity VIII
may be expected in this zone. Mainland Gujarat is assigned Zone 3 (SA at DBE
level of 0.20 g) where intensity level of VII may be expected. Table 2 shows the
characteristics of the surface-level synthetic ground motions generated for various
towns in Gujarat. These towns are situated across various seismic zones as per the
seismic zonation map of India (IS1893:2016). The epicentral distance of the station,
the PGA, PGV, the significant duration and, the mean displacements for both walls
are tabulated. The MMI data as observed from the earthquake isoseismals [23] to
indicate the overall damage to structures in each town is also tabulated in Table 2.
Additionally, in this study, two scalar measures, the predominant period, Tp , and the
mean period, Tm , were used to assess frequency content in ground motion. Both scalar
frequency parameters exhibited a weak correlation with the structural displacements
of the two walls. Nevertheless, it is observed that ground motions at most sites
exhibited very low predominant periods. From Table 2, it is observed that the mean
periods of the ground motions are higher than predominant periods but higher than the
natural periods of the walls. Observation of experimental hysteresis loops for wall-1
shows that the stiffness decreased by nine times, causing the fundamental period
to increase three times. For wall-2, experimental hysteresis loops indicate that the
stiffness decreased by three times. The increased natural period of both walls is close
to the frequency content, indicated by Tm , in the ground motion. The low values of
both frequency parameters indicate that the seismic energy of the ground motions
generated due to the 2001 Kutch earthquake was concentrated in the low time period
range. This makes the masonry structures vulnerable to collapse once there is an
increase in time period due to damage in the walls.
From Table 2, it is observed that URM walls in Seismic Zone 5 witnessed high
displacements which caused collapse. Most ground motions from Zone 5 have PGV
values > 0.1 m/s and PGA values > 0.1 g and caused the walls to reach the collapse
displacement levels. In Figs. 3 and 4, the PGA and the PGV are correlated with
structural displacements of the two walls. The three limit states are marked for
reference. Ground motions in the epicentral region, with PGAs > 0.1 g, caused a
1.00
Acceleration (g) 0.50 Time (sec)

0.00
0 5 10 15 20 25 30 35 40
-0.50
Bhachau
-1.00
0.20
Acceleration (g)
0.10
Time (sec)
0.00
0 5 10 15 20 25 30 35 40
-0.10
Adesar
-0.20
1.00
Acceleration (g)
0.50
Time (sec)
0.00
0 2 4 6 8 10 12 14 16 18 20
-0.50
Dhori
-1.00
1.00
Acceleration (g)
0.50
Time (sec)
0.00
0 5 10 15 20 25 30
-0.50
Dudhai
-1.00
0.40
Acceleration (g)
0.20
Time (sec)
0.00
0 5 10 15 20 25
-0.20
Khavda
-0.40
0.40
Acceleration (g)
0.20
Time (sec)
0.00
0 2 4 6 8 10 12 14 16 18 20
-0.20
Mandvi
-0.40
0.30
Acceleration (g)
0.20
0.10 Time (sec)
0.00
-0.10 0 2 4 6 8 10 12 14 16 18 20
-0.20 Nakhatrana
-0.30
0.30
0.20
Acceleration (g)
0.10 Time (sec)

0.00
-0.10 0 5 10 15 20 25 30 35 40
-0.20 Morbi
-0.30
Fig. 2 Synthetic ground motions generated at surface-level using site amplification functions
Table 2 Results from nonlinear time-history analyses of URM walls using surface-level synthetic ground motions
S. No. Station Dist. (km) MMI PGA (g) PGV (m/s) T p (s) T m (s) Significant Disp. wall-1 Disp. wall-2
duration (s) (mm) (mm)
Sites in Zone 5 (IS1893:2016)—PSA(DBE) = 0.45 g
1 Dudhai 15 X 0.74 0.87 0.21 0.55 12.81 7.5 16
2 Bhachau 19 X 0.64 0.54 0.10 0.31 18.08 7.5 16
3 Suvai 34 IX 0.38 0.32 0.22 0.35 21.38 7.5 16
4 Lakadia 36 IX 0.42 0.53 0.18 0.50 21.73 7.5 16
5 Rapar 46 IX 0.33 0.18 0.11 0.19 22.80 7.5 16
6 Dhori 49 X 0.74 0.63 0.26 0.40 5.92 7.5 16
7 Khavda 73 IX 0.35 0.18 0.15 0.23 12.24 7.5 16
8 Bela 77 IX 0.12 0.12 0.19 0.41 21.39 7.5 16
9 Adesar 78 VIII 0.15 0.13 0.19 0.38 22.88 7.5 16
10 Mandvi 111 VIII 0.24 0.17 0.14 0.26 9.26 7.5 16
11 Dayapur 141 VIII 0.13 0.09 0.16 0.29 7.80 7.5 13.6
12 Nakhtarana 155 VIII 0.20 0.16 0.13 0.33 8.31 7.5 16
13 Naliya 192 VIII 0.14 0.11 0.13 0.39 8.07 7.5 16
Seismic Fragility Assessment of Unreinforced Masonry Shear Walls

14 Dwarka 183 VIII 0.03 0.03 0.25 0.46 9.95 3.4 3.99
15 Lalpur 122 VIII 0.09 0.05 0.11 0.18 13.67 6.11 6.19
16 Morbi 93 VIII 0.16 0.12 0.17 0.25 21.13 7.5 16
17 Radhanpur 148 VIII 0.06 0.07 0.30 0.49 24.33 7.5 10.81
18 Sipu 235 VIII 0.02 0.01 0.10 0.25 25.13 1.73 1.90
(continued)
7
8
Table 2 (continued)
S. No. Station Dist. (km) MMI PGA (g) PGV (m/s) T p (s) T m (s) Significant Disp. wall-1 Disp. wall-2
duration (s) (mm) (mm)
19 Amreli 256 VII 0.017 0.006 0.08 0.13 20.51 2.91 1.04
20 Gandhinagar 248 VI 0.012 0.010 0.15 0.32 25.44 1.91 1.38
21 Jhagadia 354 VI 0.002 0.001 0.24 0.36 25.68 0.17 0.18
22 Junagadh 235 VII 0.019 0.010 0.09 0.17 18.45 1.71 1.37
23 Kadana 369 V 0.004 0.002 0.11 0.15 27.53 0.18 0.20
24 Kevadia 396 V 0.002 0.001 0.13 0.19 26.96 0.08 0.08
25 Rajkot 129 VII 0.065 0.074 0.45 0.54 19.16 7.5 9.80
26 Surendranagar 158 VII 0.047 0.041 0.21 0.36 23.34 5.21 5.41
27 Ukai 422 V 0.002 0.001 0.07 0.15 26.14 0.10 0.10
28 Una 280 V 0.009 0.005 0.17 0.24 19.61 1.41 0.63
29 Vadodara 321 V 0.004 0.003 0.22 0.32 25.61 0.33 0.35
J. Vemuri and K. V. L. Subramaniam
8 8
a b
7 7
Displacement (mm)
Displacement (mm)
6 6
Collapse Collapse
5 5
Cracking Cracking
4 4
Ultimate Ultimate
3 3
2 2
1 1
0 0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 1
PGA (g) PGV (m/s)
Fig. 3 a Displacement versus PGA (wall-1), b displacement versus PGV (wall-1)
a 18 b 18
16 16
14 14
Displacement (mm)
Displacement (mm)
12 12
10 Collapse 10 Collapse
8 Cracking 8 Cracking
6 Ultimate 6
Ultimate
4 4
2 2
0 0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 1
PGA (g) PGV (m/s)
Fig. 4 a Displacement versus PGA (wall-2), b displacement versus PGV (wall-2)
collapse of the walls: this observation matches well with reconnaissance survey
which reports catastrophic damage to low-rise masonry structures in this region.
Figure 3 and Table 2 show that most towns in Seismic Zone 5 were assigned high
MMI levels (X, IX and VIII) due to extensive damage observed to masonry structures
[2, 5]. On the other extreme, ground motions from Zone 3 had very low PGA values
and did not cause the collapse of the two URM walls. Some ground motions caused
the walls to reach the cracking limit. This also corresponds well with data from
reconnaissance surveys [2, 21] which show little or no damage to low-rise masonry
construction in the 2001 Kutch earthquake. It is to be noted that the MMI levels for
some towns in Zone 3 are higher due to high damage levels of mid-rise concrete
buildings due to poor construction practices [2]. Ground motions in towns situated
in Zone 4 caused the walls to exhibit displacements not exceeding the limit state
of peak strength. Substantial damage to masonry structures was reported [5] in one
particular town, Morbi, which is situated in Zone 4. In the current study, it is observed
that surface-level ground motions in Morbi had high PGA values of 0.16 g, causing
both URM walls to collapse. The high PGAs observed in Morbi town are primarily
due to site amplification [21].
The method of maximum likelihood has been recommended by several researchers
[24–26] for fitting fragility curves for displacement data. The fitted fragility function
is a lognormal cumulative distribution function, with a median, θ , and dispersion, β.
Table 3 shows that for wall-1, the median values for cracking, ultimate strength and
collapse are 0.01 g, 0.02 g and 0.06 g, respectively. For wall-2, the median values for
cracking, ultimate strength and collapse are 0.01 g, 0.03 g and 0.12 g, respectively. In
IS1893:2016, the peak spectral acceleration for design basis earthquake for seismic
Zones 3, 4 and 5 are prescribed as 0.20 g, 0.30 g and 0.45 g, respectively. The median
values are lower than expected seismic intensity indicating the high vulnerability of
both walls.
Figure 5a, b show seismic demand fragility curves for both walls. The seismic
intensity is represented by PGA. Figure 6 shows the fragility curves for the two walls,
with PGV as the seismic intensity measure. The graphs provide a means to visually
assess the vulnerability in terms of probability of exceedance of various limit states
of the masonry shear walls, considering variability in demand. The damage state
displacement thresholds were identified earlier, in Fig. 2, from the cyclic hysteric
curves. The median values for achieving strength limit states for wall-1 and wall-2
Table 3 Fragility curves: median and dispersion

Damage state Wall-1 Wall-2
Median, θ Dispersion, β Median, θ Dispersion, β
Cracking 0.01 0.10 0.01 0.41
Strength 0.02 0.40 0.03 0.23
Collapse 0.06 0.52 0.12 0.26
1 b 1
a
0.8 0.8
0.6 0.6
Pf
Pf
Cracking Cracking
0.4 Strength 0.4 Strength
Collapse Collapse
0.2 0.2
0 0
0 0.05 0.1 0.15 0.2 0 0.1 0.2
PGA (g) PGA (g)
Fig. 5 a Seismic fragility curves (w.r.t. PGA) for wall-1, b seismic fragility curves (w.r.t. PGA) for
wall-2
a 1 b 1
0.8 0.8
0.6 0.6
Pf
Pf
Cracking Cracking
0.4 0.4
Strength Strength
Collapse Collapse
0.2 0.2
0 0
0 0.05 0.1 0.15 0 0.05 0.1 0.15 0.2
PGV (m/s) PGV (m/s)
Fig. 6 a Seismic fragility curves (w.r.t. PGV) for wall-1, b seismic fragility curves (w.r.t. PGV)
for wall-2
are 0.05 g and 0.1 g, respectively. The median values for achieving cracking limit
for wall-1 and wall-2 are 0.01 g in both walls. The seismic fragility of the wall with
higher axial load (wall-2) is only slightly lower than the wall with lower axial load
(wall-1).
4 Anomalous Displacements/Outliers
In general, the wall displacement is observed to increase linearly with PGA, till
collapse. However, it is observed that for two cases, Radhanpur (Zone 4) and Rajkot
(Zone 3), the wall with lower axial load, differs slightly from the overall trend as
the limit state of peak strength is reached at low PGA. These two outliers were
shown, marked in Fig. 3. Figures 7 and 8 show velocity pulses from Radhanpur and
Rajkot. Although the two records had low PGA, the occurrence of sharp velocity
pulses triggered the collapse of the URM. The displacement time history in both
cases shows that the velocity pulse imposed a high strength demand on the walls.
Subsequently, there is a degradation in strength and stiffness of walls, leading to high
displacements.
5 Summary and Conclusions
The 2001 Kutch earthquake was one of the most destructive intraplate earthquakes
ever recorded. It had a moment magnitude of 7.6 and caused heavy damage to
URM structures in the Kutch region. Unfortunately, no strong ground motions were
0.08 10
0.08 8
0.06 8
0.06 6
Displacement (mm)
6
Displacement (mm)
4 0.04
Velocity (m/s)
0.04 4
Velocity (m/s)
2 0.02
0.02 0 2
0
0 -2 0 10 20 30 40 0
0 20 40 -0.02 Time (sec) -2
-0.02 -4
Time (sec) -6 -0.04 Velocity -4
-0.04 Displacement -6
-8 -0.06
-0.06 Velocity -10 -0.08 -8
-0.08 Displacement -12
-0.1 -10
(a) (b)
Fig. 7 a Radhanpur velocity pulse and displacement (wall-1), b Rajkot velocity pulse and
displacement (wall-1)
0.08 8 0.08 10
0.06 6 0.06 8
Displacement (mm)
4 6
Displacement (mm)
0.04 0.04
Velocity (m/s)
2
Velocity (m/s)
4
0.02
0.02 0 2
0
0 -2 0 10 20 30 40 0
0 20 40 -0.02
-0.02 -4 Time (sec) -2
Time (sec) -0.04
-6 -4
-0.04 Velocity -0.06 Velocity
-8 -6
Displacement Displacement
-0.06 -10 -0.08 -8
-0.08 -12 -0.1 -10
(a) (b)
Fig. 8 a Radhanpur velocity pulse and displacement (wall-2), b Rajkot velocity pulse and
displacement (wall-2)
recorded in this earthquake. Only PGA values from structural response recorders at
thirteen sites are available. In this paper, nonlinear time-history analyses of URM
walls are performed using surface-level synthetic ground motions. Characteristics of
the ground motions such as PGA, PGV, frequency content and duration are tabulated.
It is observed that ground motions have low values of predominant period. Such high-
frequency ground motions could be particularly damaging to low-rise structures. A
mechanics-based technique is used to develop the monotonic load envelope for two
masonry shear walls. Using stress-based criteria, various displacement limit states
are identified. Results from the dynamic analyses performed on calibrated hysteretic
models were used to derive seismic fragility curves. Examination of the fragility
curves indicates that the collapse limit state is reached for most towns in Zone 5, the
highest seismic zone as per IS1893:2016. Most ground motions from Zone 5 have
PGV values > 0.1 m/s and PGA values > 0.1 g and cause the walls to reach the collapse
displacement levels. These results confirm satisfactorily with reconnaissance studies
which report catastrophic damage to masonry structures in Zone 5, particularly in the
epicentral region. Results from NTHA for towns situated in Seismic Zones 3 and 4
vary between no damage to collapse, which correlate well with post-earthquake field
surveys carried out in these two zones, where varying levels of damage intensities
were observed. The wall displacement is observed to vary linearly with PGA and
PGV, until the collapse of the wall. Pulse ground motions from two distant sites,
Radhanpur and Rajkot, had a destructive effect on the walls. It is concluded that
URM structures with shear walls are highly vulnerable for expected ground motions
in the Kutch region. Such URM structures existing in the Kutch region need to be
upgraded to resist the expected level of seismic forces.
References
1. Nagamani, D., & Mandal, P. (2017). Estimation of earthquake source parameters in the Kachchh
seismic zone, Gujarat, India, using three component S-wave spectra. Journal of Earth System
Science, 126(5), 1–14.
2. Sinvhal, A., Bose, P. R., Prakash, V., Bose, A., Saraf, A. K., & Sinvhal, H. (2003). Isoseismals
for the Kutch earthquake of 26th January 2001. Journal of Earth System Science, 112(3),
375–383.
3. Chopra, S., Kumar, D., & Rastogi, B. K. (2010). Estimation of strong ground motions for 2001
Bhuj (M w 7.6), India earthquake. Pure and Applied Geophysics, 167(11), 1317–1330.
4. Rastogi, B. K., Gupta, H. K., Mandal, P., Satyanarayana, H. V. S., Kousalya, M., Raghavan,
R., et al. (2001). The deadliest stable continental region earthquake occurred near Bhuj on 26
January 2001. Journal of Seismology, 5(4), 609–615.
5. Jagadish, K. (2003). Behaviour of masonry structures during the Bhuj earthquake of January
2001. Journal of Earth System Science, 112(3), 431–440.
6. WHE-PAGER Survey. (2007). Retrieved from http://www.world-housing.net/related-projects/
whe-pager-project/about-this-project. Accessed July 2015.
7. Magenes, G., & Calvi, G. M. (1997). In-plane seismic response of brick masonry walls.
Earthquake Engineering and Structural Dynamics, 26, 1091–1112.
8. Abrams, D. P. (1992). Strength and behavior of unreinforced masonry elements. In Proceedings
of the Tenth World Conference on Earthquake Engineering, Madrid, Spain, July 19–24, 1992.
9. Vemuri, J., Ehteshamuddin, S., & Kolluru, S. (2018). Numerical simulation of soft brick
unreinforced masonry walls subjected to lateral loads. Cogent Engineering, 5(1), 1551503.
10. Vemuri, J. P., & Kolluru, S. (2017). Seismic analysis of unreinforced masonry walls. IDRiM
Journal, 6(2), 102–115.
11. Federal Emergency Management Agency (FEMA). (1999). Evaluation of earthquake damaged
concrete and masonry wall buildings: Technical resources (FEMA P-307). Washington, DC.
12. Chopra, S., Kumar, D., Rastogi, B. K., Choudhury, P., & Yadav, R. B. S. (2012). Estimation of
site amplification functions in Gujarat region, India. Natural Hazards, 65(2), 1135–1155.
13. Penelis, G. R. G. (2006). An efficient approach for pushover analysis of unreinforced masonry
(URM) structures. Journal of Earthquake Engineering, 10(3), 359–379.
14. Vemuri, J., Ehteshamuddin, S., & Kolluru, S. V. L. (2018). Evaluation of seismic displacement
demand for unreinforced masonry shear walls. Cogent Engineering, 5(1), 1480189.
15. Mann, W., & Muller, H. (1982). Failure of shear stressed masonry. An enlarged theory, tests
and application to shear walls. Proceedings of the British Ceramic Society, 30, 223–235.
16. Anthoine, A., Magonette, G., & Magenes, G. (1995). Shear compression testing and analy-
sis of brick masonry walls. In Proceedings of the 10th European Conference on Earthquake
Engineering, August 28–September 02, 1994. Rotterdam: Balkema.
17. Magenes, G., & Calvi, G. M. (1992). Cyclic behaviour of brick masonry walls. In Proceedings
of the 10th World Conference on Earthquake Engineering, Madrid, Spain, July 19–24, 1992.
18. BIS (Bureau of Indian Standards). (1991). Handbook on masonry design and construction,
SP:20. New Delhi, India.
19. CSI (Computers and Structures Incorporated). (2012). SAP2000 version 15, software. Walnut
Creek, California, USA.
20. Motazedian, D., & Atkinson, G. M. (2005). Stochastic finite-fault modelling based on dynamic
corner frequency. Bulletin Seismological Society of America, 95, 995–1010.
21. Vemuri, J., Kolluru, S., & Chopra, S. (2018). Surface level synthetic ground motions for M7.6
2001 Gujarat earthquake. Geosciences, 8(12), 429.
22. Bureau of Indian Standards (BIS). (2016). Indian Standard criteria for earthquake resistant
design of structures: Part 1—General provisions and buildings, IS: 1893. New Delhi, India.
23. Narayan, J. P., & Sharma, M. L. (2004). Effect of local geology on damage severity during
Bhuj, India earthquake. In Proceedings of the Thirteenth World Conference on Earthquake
Engineering (Paper No. 2042), Vancouver, Canada, August 1–4.
24. Baker, J. W., & Cornell, C. A. (2005). A vector-valued ground motion intensity measure con-
sisting of spectral acceleration and epsilon. Earthquake Engineering and Structural Dynamics,
34(10), 1193–1217.
25. Cutfield, M. R., & Ma, Q. T. (2018). Solution strategies for three problems in empirical fragility
curve derivation using the maximum likelihood method. Journal of Earthquake Engineering,
22(3), 435–453.
26. Straub, D., & Der Kiureghian, A. (2008). Improved seismic fragility modeling from empirical
data. Structural Safety, 30(4), 320–336.
Axial Behaviour of Corroded CFST
Columns Wrapped with GFRP
Sheets—An Experimental Investigation
S. Vijaya Bhaskar Reddy and S. Sivasankar
1 Introduction
Corrosion causes severe deterioration both in reinforced concrete (RC) and steel
structures like buildings, bridges and marine structures. Once they got corrosion, their
structural performance and safety get reduced significantly. Recent days concrete-
filled steel tubular (CFST) members are widely served as structural elements (beams,
columns and slabs) both in onshore and offshore applications. Since, composite mem-
bers have concrete and steel, resulting in a member that has the beneficial qualities
of both materials [1]. Also, they have very good strength and ductility properties
compare to reinforced concrete structures. Offshore structures are more prone to
corrosion, since they are directly contacted with very severe environment like sea
water. In offshore CFST members, steel is directly exposed to sea water, and there-
fore, they get deteriorated by corrosion due to long-term chloride ingress. As the
amount of corrosion increases, CFST member starts loosing design strength, and
finally, the structure fails as a whole. In addition, maintenance losses such as cost
and time needed to make the replacements of corroded objects are incurred. The
ancient and traditional method of retrofitting is providing additional plates on cor-
rosion affected region by welding techniques. They increase the self-weight of the
section further affecting the design strength. Also, the welded plates get further cor-
rosion when it is exposed to the sea water. So, we need a permanent, easy to use
and cost-effective solution to overcome these issues. For the past few decades, a lot
of RC structural members were failed by corrosion and other severe environmental
deterioration, and they are successfully repaired by fibre-reinforced polymers (FRP)
S. V. B. Reddy · S. Sivasankar (B)

Department of Civil Engineering, CMR Technical Campus, Medchal, Hyderabad, Telangana
501401, India
S. V. B. Reddy
16 S. V. B. Reddy and S. Sivasankar
composites [2–4]. FRP composites are made up of high-tensile strength fibres embed-
ded in an epoxy matrix; they afford superior resistance to corrosion, low weight, high
strength-to-weight ratio, impact resistance, high mechanical strength, fast and eco-
nomical way of rehabilitation or repair of beams, columns and slabs [6]. Due to the
superior mechanical and physical properties, FRP composites are widely used for
external strengthening of RC columns [4–6], beams [2, 3, 8], slabs [7, 9], bridge
decks [8], impact [10] and fatigue strength [11]. After the successful use of FRP
composites in RC structures, researchers and engineers are now using FRP in steel
structures for strengthening and retrofitting applications. Al-Saidy et al. [12] per-
formed a study on corrosion damaged beam retrofitted with CFRP plates. Results
showed that the strength and stiffness of beam retrofitted by CFRP plates increased
significantly. Photiou et al. [13] studied the applications of high modulus CFRP
composites for beams artificially corroded. Four beams were wrapped, two utilizing
U-shaped prepreg units, whereas the other two beams used a flat plate prepreg. From
the results, it was found that the beam wrapped by high modulus CFRP composite
showed higher load-carrying capacity and ductility compared to unwrapped beam.
Sundarraja and Sivasankar [14] did an experimental study on square steel tubular
members strengthened by CFRP strips. It was found that the specimen wrapped by
CFRP strips exhibited higher strength and ductility compared to control specimen.
Teng and Hu [15] suggested the use of FRP sheets for circular steel specimens. Steel
members were wrapped with one, two and three layers of GFRP sheets around the
periphery throughout the height. Test results indicated that the specimen wrapped
by three GFRP layers showed higher axial load-carrying capacity and stiffness com-
pared to that of specimens wrapped with one and two layers of GFRP sheets. Also,
it was found that the GFRP sheets significantly reduce the local buckling. Seica and
Packer [16] performed study on retrofitting of steel members using CFRP composites
for underwater and in-air applications. Two specimens were wrapped in normal con-
dition and the remaining four were wrapped in underwater. Test results showed that
the CFRP sheets significantly increase the load and stiffness of both in-air and under-
water wrapped specimens. Also, they strongly recommended the CFRP sheets for
underwater applications. Keykha et al. [17] done both investigational and numerical
analyses on steel columns reinforced by CFRP sheets. Experimental and numerical
studies results are close to each other.
From the previously published journal results, it has been found that the FRP is
the most suitable composite materials for strengthening and retrofitting of RC and
steel structures. To date, only a limited number of studies have specifically reported
on corroded CFST members repaired with GFRP sheets. Nowadays, CFST members
were broadly used in the marine and other structural applications. The prime goal
is to experimentally explore the usefulness of GFRP (full wrapping) on corroded
CFST column members subjected to axial compression.
Axial Behaviour of Corroded CFST Columns Wrapped with GFRP … 17
2 Material Properties
2.1 Steel Tube
Circular hollow steel tube confirming to IS 1239-1983 having length and diameter
600 and 101.6 mm was used in this experiment. Thickness and height of the CCFST
members are 3.6 mm and 600 mm, respectively. The yield strength provided by the
manufacture is 250 N/mm2 .
2.2 Concrete Mix
Concrete strength of M25 was achieved by using commercially available materials

with mixing using vibrators and simple curing methods. The design proportion of
1: 1.62: 2.58 was used to attain a maximum strength of 25 N/mm2 with w/c ratio
as 0.45. The average compressive strength of 26.29 N/mm2 was found at the end of
28 days.
2.3 Glass Fibre-Reinforced Polymer (GFRP) Sheet
The unidirectional webbed glass fibre was used in the experimental investigation.
GFRP has Young’s modulus of 50 GPa and tensile strength can be up to 1600 N/mm2 .
The thickness and width of the glass fibre were 0.360 mm and 600 mm, respectively.
2.4 Adhesive
The most suitable adhesive material with GFRP fabrics with steel is epoxy resin.
The adhesive material consists of two compounds, namely resin and hardener. These
have a mixing ratio of 100: 10 (B: H) as specified by the manufacturer.
3 Experimental Investigation
A total of eight columns were used for this investigation. All the specimens are under-
gone corrosion by artificial corrosion process. Artificial corrosion was performed by
adding 70 g of sodium chloride (NaCl) and 100 ml of acidic acid with one litre of
portable water. Thereafter, all samples are immersed in the solution for a period of six
weeks. To accelerate the corrosion, samples were taken out and cured in the ambient
Fig. 1 Samples before

corrosion
temperature for alternate days. Corrosion rate was calculated by weight loss method
at an interval of seven days. Samples before corrosion and after corrosion processes
were presented in Figs. 1 and 2. Corrosion rate calculation at seven days interval is
presented in Table 1.
3.1 Test Specimen Details
GFRP fabrics were used as continuous sheets wrapping around the peripheral of
the circular concrete-filled columns. Among the eight CCFST columns, two were
reference and rest of the specimens were confined with GFRP continuous sheets. For
easy identification, specimens were named such as CCFST-C, CCFST-1L, CCFST-2L
and CCFST-3L. For specimen CCFST-1L, CCFST denoted the corroded concrete-
filled steel tubes and 1L denoted the specimen confined by single layer of GFRP
sheets.
Fig. 2 Samples after

corrosion
3.2 Application of GFRP Sheets on CCFST Columns
Prior bonding of CFRP strips surface preparation of specimen was done by sand-
blasting [8]. This could avoid the debonding of CFRP strips. Before bonding, surface
of the specimens was cleaned by acetone. This removes rust and impurities present
on the surface. Resin and hardener were mixed well as per the mix ratio given by
the manufacturer. Before wrapping the first layer, mixed resin is applied around the
specimen using brush. Immediately after application of resin, initial layer is wrapped
around the specimen and is presented in Fig. 3. Steel ribbed rollers were used for
removing excess air present between the layers which are presented in Fig. 4. Similar
procedure was repeated for second and third layer GFRP wrapping. Wrapped spec-
imens were dried at room temperature for 2–3 days. The crushing tests were carried
out in a 2000 kN universal testing machine shown in Fig. 5.
4 Experimental Results and Discussion
4.1 Failure of Specimens
Specimen with no wrapping (control specimen) failed by local buckling located at

the top which is shown in Fig. 5. GFRP fabrics do not alter the actual failure mode;
however, it is capable to postpone buckling and permit crushing strains reaching
upper values therefore resulting in greater load-carrying capacity. Buckling of all
20
Table 1 Corrosion rate analysis

Sl. No. Sample Initial weight Corrosion rate
details in kg Wt. loss after Wt. loss after Wt. loss after Wt. loss after Wt. loss after Wt. loss after Total weight
7 days 14 days 21 days 28 days 35 days 42 days loss in kg
1 CCFST-C-1 14.630 14.609 14.567 14.546 14.518 14.486 14.458 0.171
2 CCFST-C-2 14.680 14.649 14.616 14.571 14.532 14.482 14.447 0.233
3 CCFST-1L-1 14.500 14.479 14.428 14.336 14.310 14.279 14.234 0.266
4 CCFST-1L-2 14.530 14.522 14.508 14.477 14.449 14.427 14.408 0.122
5 CCFST-2L-1 14.570 14.554 14.527 14.481 14.457 14.431 14.403 0.167
6 CCFST-2L-2 14.600 14.581 14.562 14.529 14.492 14.461 14.428 0.172
7 CCFST-3L-1 14.470 14.455 14.431 14.399 14.356 14.332 14.310 0.160
8 CCFST-3L-2 14.500 14.481 14.454 14.422 14.394 14.367 14.332 0.168
S. V. B. Reddy and S. Sivasankar
Fig. 3 GFRP wrapping
Fig. 4 Air removal using

rollers
Fig. 5 Failure of control

column
confined columns is effectively resisted by GFRP; it is evident from the load incre-
ment associated with normal specimen is shown in Table 2. All confined columns,
rupture of GFRP and the peak load occur simultaneously. This ensures the effective-
ness and bonding of GFRP on corroded steel columns. Also it was observed from
failure pattern, all confined columns failed by rupture of GFRP alone and not by
fibre delamination. After the fibre rupture, there was a sudden drop in load occurs
in all confined columns. Photograph in Figs. 6, 7 and 8 shows the failure modes of
samples wrapped by unique, two and three GFRP layers.
Table 2 Ultimate compressive load and its increment comparison

Sl. No. Specimen identification Ultimate compressive % Development in
strength (kN) compressive strength
compared to normal
sample
1 CCFST-C-1 677.00 –
2 CCFST-C-2 671.00 –
3 CCFST-1L-1 702.85 3.82
4 CCFST-1L-2 713.00 5.32
5 CCFST-2L-1 728.85 7.66
6 CCFST-2L-2 734.00 8.42
7 CCFST-3L-1 746.00 10.19
8 CCFST-3L-2 741.00 9.45
Fig. 6 Failure of CCFST-1L

4.2 Ultimate Compressive Load
The gain in ultimate compressive load for all strengthened columns including control
columns is given in Fig. 9 and Table 2. The percentage improvement in compres-
sive load of GFRP confined columns compared with normal column was given in
Table 2. For specimen CCFST-1L (one layer), the ultimate compressive load was
713 kN and the percentage gain compared control column was 5.32%. The ultimate
compressive load observed for specimen CCFST-2L (two layers) was 734 kN which
is 8.42% higher when related to normal sample. Similarly, the compressive strength
for CCFST-3L (three layers) was 746 kN and the percentage gain was 10.19% com-
pared to control column. The ultimate compressive strength of all confined columns
increased with an increase in number of GFRP layers. But the increases are not twice
compared with the ratios increased by single layer, similarly for three layers also.
The reason was resin breakage takes place between the GFRP layers. Overall, there
was an increase in strength compared to control column. These results confirm the
effectiveness of GFRP sheets on corroded steel members.
4.3 Load–Deformation Relationship
Figure 10 presented the load–deformation curve of all tested specimens. It was

observed that all the curves are similar (linear) in the initial stages. But in the second
1000
800 734 746

713
677
Ultimate Load (kN)
600
400
200
0
CCFST-C-1 CCFST-IL-2 CCFST-2L-2 CCFST-3L-1
Circular Specimens
Fig. 9 Ultimate load comparison
80000
60000
Load (N)
40000 CCFST-C-1
CCFST-1L-2
CCFST-2L-2
20000
CCFST-3L-1
0
0 2 4 6 8 10 12
Deformation (mm)
Fig. 10 Load versus deformation
stage, curves are different (nonlinear) because of stiffness provided by GFRP sheets.
This phenomenon indicates that the load increment and reduction in deformation as
the number of layers increased until the GFRP rupture. Deformation values for all
tested specimens are presented in Fig. 11.
8
7
Deformation (mm)
6
5
4
3
2
1
0
CCFST-C-1 CCFST-IL-2 CCFST-2L-2 CCFST-3L-1
Specimen Details
Fig. 11 Deformation curve for all specimens
4.4 Stress–Strain Relationship
Figure 12 represented the stress–strain relations of control column and GFRP con-
fined columns. All confined columns exhibited same approach until reach as the
normal samples highest strain value afterwards a massive difference in strain. Strain
value for control column at peak is 0.01148, and for confined columns CCFST-1L-2,
CCFST-2L-2 and CCFST-3L-1, the strain value is 0.01102, 0.010733 and 0.010467,
respectively. The percentage of strain for confined columns (CCFST-1L-2, CCFST-
2L-2 and CCFST-3L-1) compared to control column was 4%, 7.14% and 8.88%,
respectively. From the above details, it can be seen that the percentage strain increase
was not linear with increasing number of GFRP layers. This was mainly because of
the resin breakage between CFRP layers.
Fig. 12 Stress–strain 140

relationship
120
100
Stress (N/mm2 )
80
CCFST-C-1
60 CCFST-IL-2
40 CCFST-2L-2
20 CCFST-3L-1
0
0 0.004 0.008 0.012 0.016 0.02
Strain
5 Conclusions
Six CCFST columns wrapped with different number of GFRP layers and two
unwrapped specimens were tested in UTM machine. From the investigation, the
following conclusions were summarized:
• Adding more number of GFRP sheets to steel samples significantly delayed the
buckling and in that way raises the compressive strength.
• Failure of columns wrapped by GFRP sheets is purely by rupture only and no sign
of delamination. This result concludes the superior attachment of GFRP sheets
and CCFST tubes.
• All GFRP wrapped columns shown higher ultimate compressive load than the
bare corroded steel column.
• Ultimate compressive load gain of columns wrapped with one, two and three
GFRP layers CCFST-1L-2, CCFST-2L-2 and CCFST-3L-1 compared to CC2
was 5.32%, 8.41% and 10.19%, respectively.
• Axial deformation control for GFRP wrapped columns CCFST-1L-2, CCFST-
2L-2 and CCFST-3L-1 was 4.06, 6.53 and 8.85% higher compared to CC2. The
corresponding deformation at peak load was 6.61 mm, 6.44 mm and 6.28 mm,
respectively.
• The strain values for columns CCFST-1L-2, CCFST-2L-2 and CCFST-3L-1 are
0.01102, 0.010733 and 0.010467, respectively at the peak load.
• It is recommended that the columns corroded by corrosion, wrapping with GFRP
sheets are considered to be an effective technique to improve the compressive
strength and stiffness.
Conflict of Interests The authors declare that there is no conflict of interests regarding the
publication of this paper.
References
1. Gupta, P. K., Sarda, S. M., & Kumar, M. S. (2007). Experimental and computational study
of concrete filled steel tubular columns under axial loads. Journal of Constructional Steel
Research, 63, 182–193.
2. Pham, H., & Al-Mahaidi, R. (2004). Experimental investigation into flexural retrofitting of
reinforced concrete bridge beams using FRP composites. Composite Structures, 66(1), 617–
625.
3. Soudki, K., El-Salakawy, E., & Craig, B. (2007). Behavior of CFRP strengthened reinforced
concrete beams in corrosive environment. Journal of Composites for Construction, 11(3),
291–298.
4. Saadatmanesh, I. I., & Ehsani, M. R. (1997). Repair of earthquake damaged reinforced concrete
columns with FRP wraps. ACI Structural Journal, 94(2), 206–215.
5. Demers, K. W. (1999). Neale confinement of reinforced concrete columns with fiber reinforced
composite sheets—An experimental study. Canadian Journal of Civil Engineering, 34, 226–
241.
6. Li, J., & Hadi, M. N. S. (2003). Behaviour of externally confined high-strength concrete columns
under eccentric loading. Composite Structures, 62, 145–153.
7. Mosallam, A. S., & Mosalam, K. M. (2003). Strengthening of two-way concrete slabs with
FRP composite laminates. Construction and Building Materials, 17, 43–54.
8. Ghosh, K. K., & Karbhari, V. M. (2007). Evaluation of strengthening through laboratory testing
of FRP rehabilitated bridge decks after in-service loading. Composite Structures, 77, 206–222.
9. Rochdi, E. H., Bigaud, D., Ferrier, E., & Hamelin, P. (2006). Ultimate behavior of CFRP
strengthened RC flat slabs under a centrally applied load. Composite Structures, 72, 69–78.
10. Erki, M. A., & Meier, U. (1999). Impact loading of concrete beams externally strengthened
with CFRP laminates. Journal of Composites for Construction, 3, 117–124.
11. Tavakkolizadeh, M., & Saadatmanesh, H. (2003). Fatigue strength of steel girders strengthened
with carbon fiber reinforced polymer patch. Journal of Structural Engineering, 129(2), 186–
196.
12. Al-Saidy, A. H., Klaiber, F. W., & Wipf, T. J. (2004). Repair of steel composites beams with
carbon fibre reinforced polymer plates. Journal of Composites for Construction, 8(2), 163–172.
13. Photiou, N. K., Hollaway, L. C., & Chryssanthopoulos, M. K. (2006). Strengthening of an
artificially degraded steel beam utilising a carbon/glass composite system. Construction and
Building Materials, 20, 11–21.
14. Sundarraja, M. C., & Sivasankar, S. (2012). Axial behaviour of HSS tubular sections strength-
ened by CFRP strips—An experimental investigation. Science and Engineering of Composite
Materials, 19, 159–168.
15. Teng, J. G., & Hu, Y. M. (2007). Behaviour of FRP-jacketed circular steel tubes and cylindrical
shells under axial compression. Construction and Building Materials, 21, 827–838.
16. Seica, M. V., & Packer, J. A. (2007). FRP materials for the rehabilitation of tubular steel
structures, for underwater applications. Composite Structures, 80, 440–450.
17. Keykha, A. H., Nekooei, M., & Rahgozar, R. (2016). Numerical and experimental investiga-
tion of hollow steel columns strengthened with carbon fiber reinforced polymer. Journal of
Structural and Construction Engineering, 3(1), 49–58.
Tensile Membrane Structures:
An Overview
Allan L. Marbaniang, Subhrajit Dutta, and Siddhartha Ghosh
1 Introduction and History
Tensile membrane structures (TMS) are excellent solutions for covering long spans
efficiently. These days, TMS are commonly adopted as roofing solutions for different
types of large open areas and are mostly seen in stadia, amphitheatres and airports
(Fig. 1). Membrane tents have also been used as makeshift shelters and roof shades
since ancient history, being used extensively for military purposes since the Roman
times till the present day. Modern tensile membrane structures, however, have a level
of sophistication in their design, construction technology and material properties that
separate them considerably from their historical counterparts.
Most of the tension structures developed during the late nineteenth century were in
bridges, and they were mostly buildings during the twentieth century [1]. The devel-
opment of such structures in the early twentieth century was attributed to circus tents
and the need for larger structures after the industrial age started [2]. The advantages
for lightweight construction are plentiful with an obvious decrease in dead load, low
costs and an efficient use of the material. Frei Otto was one of the pioneers in the field
of lightweight structures, with his exploration of different types and forms of tensile
structures [3]. One of the earliest notable tensile structures was his German Pavilion
at the Expo 1967 in Montreal. His research at the Institute for Lightweight Structures
in Germany created a firm base of experience, knowledge and direction, for future
A. L. Marbaniang (B) · S. Ghosh

Indian Institute of Technology Bombay, Mumbai 400076, India
S. Ghosh
S. Dutta
National Institute of Technology Silchar, Assam 788010, India

30 A. L. Marbaniang et al.
Fig. 1 Denver International Airport. Source Wikimedia commons
research and designs. Although physical models were popular those days, they are,
however, costly and inaccurate for the use in large structures. This led to a push for
computational methods for modelling tensile structures. During the construction of
the large cable net roof for the Munich Olympic Stadium, different computational
methods were used alongside physical models.
Digital designing has provided architects with powerful tools to come up with
different shapes and forms. The focus has now shifted to construction materials and
technologies that are more sustainable, cost-effective and which leave a lower carbon
footprint. Tensile membranes are a popular choice due to their flexibility in shape,
light material, free flowing and aesthetically pleasing nature. The increase in use
of tensile membranes was complemented with better materials, computational mod-
elling tools and construction techniques. The adoption of tensile membranes has,
however, not been as common as expected when compared to some other popular
materials like concrete or steel. Structural failures like the collapse of the fabric
roof over Montreal’s Olympic Stadium in 1999 [4], lack of technical know-how
and an inadequate availability of code provisions slow down the wide adoption of
such structures. Currently, there is a slight shift of architectural designs to flatter
membrane shapes like the Millennium Dome [5]. The attempt to remove the anti-
clastic shapes may lead to unforeseen circumstances [6] which further complicate
the design process. The few code provisions available and the limited number of
architects or structural engineers with the knowledge of implementing effective and
safe designs create a gap for fully utilizing such structures. The aim of this study is
to therefore present a review on the design of TMS. The design of such structures
is unconventional as the initial shape is not known beforehand. This central concept
Tensile Membrane Structures: An Overview 31
of TMS is presented, along with a review of past research on form finding, analysis
and patterning. The other challenges that need to be included during the design are
emphasized which could improve the design of TMS in the future.
2 Tensile Membrane Structures
Tensile membranes are lightweight, efficient membranes, possessing no bending

stiffness, and they resist any external loading only through in-plane stresses. The
ability of the full cross section to be stressed uniformly over the whole domain
results in the most efficient use of the material. The two most common fabrics used
are the PVC-coated polyester and the PTFE Teflon-coated glass cloth [7]. The fabric
material is generally formed by weaving fabrics with the weft yarns passed over and
under fairly straight warp yarns, causing the material to behave differently along the
yarn directions as seen in Fig. 2.
Cables behave similarly to tensile membranes and are many a times used along the
TMS boundary to apply prestress. The membrane can also be stretched to fit a fixed
boundary or a cable perimeter with a desirable prestress. This initial prestress used
to stretch the membrane should be high enough so that wrinkling or slack is avoided
without over-stressing or causing a tear in the material. The shape of the membrane
changes on the application of external loads. External loading can occur in the form
of wind or snow/rain loading, which may lead to higher stresses or ponding if not
properly designed. Unique shapes are formed under different loads and boundary
conditions. These shapes and corresponding prestress are not known beforehand but
need to be found [6]. This process is different from conventional structures, where the
initial shape is already known. The required shape is sometimes forcibly acquired
Fig. 2 Fibre material with

weft and warp yarns
Fig. 3 Anticlastic shape under prestress σ i providing stability and stiffness under loads in both
directions
by applying uneven initial prestresses, which may be undesirable. Therefore, an

inverse process of form finding is needed so that the initial equilibrium geometry
is found under a particular prestress and the given boundary conditions. The form
found shape should be anticlastic (i.e. principal curvatures are of opposite signs) [8],
which provides the necessary stability as seen in Fig. 3.
The final design has to further undergo a patterning stage so that different pieces
of membrane fabric can be seamed together. The patterned membrane needs to be
compensated so that it forms the right geometry on the application of prestress [9].
The form found shape should also be analysed to check whether it can withstand
environmental loads. These structures, however, possess highly nonlinear behaviours,
that further complicate the form finding, analysis and patterning stages. A typical
design of a tensile membrane structure from the initial form finding to the final
patterning phase can be seen in Fig. 4.
3 Form Finding
A majority of research done on TMS is in finding different ways of solving the ini-
tial equilibrium problem by using different techniques. Frei Otto was a pioneer who
obtained shapes for membranes and cable nets by physically using soap film mod-
els. The interest in moving toward computational methods started with the Munich
Olympic Stadium, when scaled physical models of the cable net roof could not pro-
vide a desirable level of accuracy [10]. The use of physical models, however, gives
a good intuition to the problem and should be used alongside numerical methods,
according to many practitioners [6].
3.1 Formulation
In the form-finding stage, the membrane continuum is discretized using finite ele-
ments. This discretization can be done using equivalent cable elements. This approach
Fig. 4 Design process of tensile membrane structures
has, however, shown to be inaccurate due to the inability to model shear [9, 11]. The
use of constant strain triangle elements (CST) is a popular choice to discretize the
membrane domain as shown in Fig. 5b.
The equilibrium at each node i is given as
T x
Ri = Pi + (1)
L
where at each node i, Ri is the residual due to the imbalance between external loads
Pi and the prestress T in the neighbouring elements and x/L is the direction cosine
of the prestress force from each neighbouring element. During form finding, the
residual is only due to the internal prestress without any external loading and material
dependency [11]. The basis of form finding is then to numerically find a geometry
in which the residual force at all the nodes is close to zero for a particular pretension
and boundary condition, as shown in Fig. 5a.
Fig. 5 a Equilibrium at node i without external loads, and b the final form found shape of a hypar
TMS
Minimal surfaces having constant stresses are, however, more efficient as all
sections in the membrane surface carry an equal amount of stress. This reduces the
chance of occurrence of wrinkles and local stress concentrations. Physical soap films
have been used to acquire minimal shapes over a certain boundary. The boundary
conditions can sometimes not allow for minimal surfaces to be formed, and so pseudo-
minimal surfaces are acquired. In that case, the prestress applied to acquire the shape
is not constant which is not desirable.
3.2 Computational Methods
The force density method was used to find the shape for the cable net roof for the
Munich Olympic Stadium [10, 12]. At the same time, Argyris et al. [13] used the nat-
ural form-finding method to also find the equilibrium shape of the same structure. The
force density method linearized the nonlinear equilibrium equations by assuming the
unknown parameters as force densities while the natural form-finding method used
the stiffness method. Apart from static approaches to form-finding, dynamic meth-
ods such as the dynamic relaxation method have also been used [14]. The method is
popular and robust where the system oscillates towards its equilibrium state under
damping. Lewis [15] compared the transient stiffness method with dynamic relax-
ation methods. The dynamic relaxation method was found to be more efficient with
the stability of iterations assured. Veenendaal and Block [11] performed an overall
comparison of different types of form-finding methods within the same formulation.
The different methods were categorized into stiffness matrix, geometric stiffness and
dynamic equilibrium methods. It was found that geometric methods fare better when
compared to dynamic and stiffness methods.
4 Analysis and Design
Analysis of tensile membrane structures under external loading is done by consider-

ing the material properties, external loading and finding the equilibrium shape. Wind
loading and snow loading are usually the dominant forces with self-weight being
largely negligible due to the lightweight material [4]. The form should also allow for
drainage to avoid ponding under different load conditions. Wind loading is updated
with the geometrically nonlinear deformation [14]. Barnes [14] used the dynamic
relaxation method to find the static response. The use of viscous damping is more
preferable with extreme environmental loads if it causes slackening which slows
the overall convergence [16]. Tabarrok and Qin [17] used large deformation finite
element analysis for the form finding and load analysis of structures made of cables,
frames and membrane elements. Bridgens and Birchall [5] showed the response of
different types (hypars, conic and saddles) of tensile membrane structures. Hypars
show a low sensitivity to shear stiffness especially when the patterning is such that
the yarns are in the direction of the opposite corners of the structure. Dutta et al.
[18, 19] showed that by optimizing the initial prestress in the membrane, one could
get shapes with better performance, particularly under uncertain wind loading. The
dynamic relaxation method was used [20] along with heuristic optimization methods
and metamodels to save computational time.
Since these structures show high geometric nonlinearity, a limit state approach
is not suitable for designing as the geometry of the structure is dependent on the
magnitude and the distribution of the loading [21]. This approach is followed in
codes like ASCE [22] where stress factors are used as permissible limits. Gosling
et al. [4] had kept the requirements of a design to fulfil the requirements of permissible
stress limits, deflection and avoidance of ponding and wrinkles.
5 Patterning
The form found membrane shapes are doubly curved and cannot be easily projected
on a two-dimensional plane. The cloth pieces used for the structure have to be seamed
together from sheets of fabric material such that the desired shape can be formed as
shown in Fig. 6. Some of the methods used to find the patterned shapes are physical,
geometrical or based on equilibrium shapes [23]. Patterning can be done by finding
the unstressed side lengths from the equilibrium shape along with cable slack lengths
which can be independently unstressed [2]. The cutting pattern can be found using
geodesic lines on the form found shape with compensation to allow for the actual
stretching when it is installed [9]. If non-minimal forms are used, varying levels of
compensation should be applied in different directions, which is rarely followed [21].
Fig. 6 Patterning along geodesic lines. Adapted from [24]
6 Challenges
6.1 Form Finding
Form finding is varied in its approaches and is dependent on the choice of different
discretizations and the solving method used. Minimal shapes may not form within
all boundary conditions. Bridgens and Birchall [5] showed that for conic structures,
the minimal surface may neck off with unfeasible boundary conditions. This leads to
pseudo-minimal surfaces with higher stresses at high curvature points, where the soap
film would fail. In a round-robin exercise, it was found that a high variation in stresses,
displacements and reactions was obtained from all the participants highlighting the
challenges faced with benchmarking different methods and software packages [25].
The different computational methods available also present their own challenges
that have to be considered such as computational time, stability and simplicity. Com-
putational time is an important aspect of any numerical scheme, which can be inten-
sive in some cases. The dynamic relaxation method improves with better descriptions
of the initial shape close to the formed shape [6]. This may not always be known
for surfaces with complex shapes and boundary conditions. Different parameters
are required to control stability and convergence of dynamic equilibrium methods
[11]. The stiffness methods unnecessarily require material properties which may be
excluded. The linear force density method, which is in itself elegant and simple, can
produce impractical or unwanted results, necessitating a nonlinear solution phase for
proper results.
6.2 Uncertainty
The woven fabric material is highly nonlinear with uncertain mechanical properties.
Assumptions in the material properties for a given fabric, that are usually adopted,
can be very approximate [5]. Tensile membrane structures are highly sensitive to
wind loading due to their lightweight nature. Wind loading in itself is stochastic in
nature, and the current practice of taking critical static wind loads is unrepresentative
of the actual behaviour. The nonlinear behaviour of the structure and wind loading
is coupled, and this requires detailed fluid–structure interaction studies [26].
6.3 Design
The design process is made more difficult due to an inadequate availability of code
provisions. While there is a standard guideline on TMS in the USA [22], there is no
code yet established in Europe, while other existing codes provide only broad design
principles [4]. The step to a Eurocode provision is, however, ongoing with the release
of the “Science and Policy Report” [27] which provides a base for the formulation
of a new standard in the near future [28]. India does not have any standard for TMS,
and there are only a few capable practitioners.
The current general design practice is to provide a single stress reduction factor
which tries to account for the uncertainty in material property, loading, durability
and construction errors, which is unreasonable. The consideration of displacement as
ultimate states within a reliability approach should be considered [4]. The reliability
against uncertain loading can be increased by optimizing the initial prestress resulting
in more stable structures [18]. Investigating the erection process is also important as
the whole system may not be stable until the full process is completed and stresses
during erection may also reach failure limits [9]. The prestress has to be maintained
throughout the life of the structure, which needs to be monitored for a long-term
creep behaviour [21].
6.4 Other Constraints
Membrane structures should satisfy various other constraints and requirements sur-
plus to the general structural requirements. The structure should be able to discharge
water (due to snow or rainwater) safely. One of the major advantages of tensile mem-
branes is their capability of transmitting light. The translucency of the fabric should
be an important factor while choosing the material for the structure. Uninsulated
membranes are sensitive to changes in radiation, and they quickly heat up or cool
down. This can make the structure behave as a hot or cold radiator over an area which
can be uncomfortable [7]. The fabric is thin and may not provide good acoustic insu-
lation. Fire safety is also an important aspect, and while most membrane fabrics are
at least fire-resistant, they should be tested [29]. Toxic fumes from PTFE fabrics at
high temperatures require proper ventilation and sprinkler systems [7].
7 Conclusion
Tensile membrane structures are efficient and lightweight and are gaining in popular-
ity for roofing large open areas. The adoption of TMS is, however, not straightforward
due to an unfamiliar design process and lack of codes and resources. In this paper, a
review of past research on the design of TMS is presented. Form finding is an integral
part of the design of TMS, with numerical methods used to solve the nonlinear initial
equilibrium problem. These methods are also used for analysis under external loads.
The fabric needs to be patterned properly in order to acquire the desired prestress
or shape upon construction. While computational methods are effective, most of the
software packages and methods used in the design process need to be benchmarked
and regularized. The computational cost associated with the existing numerical meth-
ods of form finding and load analysis is expected to drive future research towards more
efficient algorithms. The future scope of TMS can be improved through the develop-
ment of codes, resources and reliable software. Further research into the inclusion of
material uncertainty and nonlinearity, stochastic loading and fluid–structure interac-
tion need to be explored. The comparison of minimal shapes over non-minimal ones
can follow with a detail investigation between various shapes.
References
1. Bradshaw, R., Campbell, D., Gargari, M., Mirmiran, A., Tripeny, P. (2002). Special struc-
tures: Past, present, and future. Journal of Structural Engineering, 128. https://doi.org/10.
1061/(asce)0733-9445(2002)128:6(691).
2. Topping, B., Ivanyi, P. (2007). Computer aided design of cable membrane structures. Saxe-
Coburg Publications.
3. Drew, P. (1976). Frei Otto: Form and structure. London: Granada Publishing Limited.
4. Gosling, P., Bridgens, B., & Zhang, L. (2013). Adoption of a reliability approach for membrane
structure analysis. Structural Safety, 40, 39–50. https://doi.org/10.1016/j.strusafe.2012.09.002.
5. Bridgens, B., & Birchall, M. (2012). Form and function: The significance of material properties
in the design of tensile fabric structures. Engineering Structures, 44, 1–12. https://doi.org/10.
1016/j.engstruct.2012.05.044.
6. Lewis, W. (2003). Tension structures-form and behaviour. London: Thomas Telford Publishing.
7. Chilton, J. (2010). Tensile structures—Textiles for architecture and design. In: G. Pohl
(ed.), Textiles, polymers and composites for buildings, Woodhead publishing series in textiles
(pp. 229–257). Woodhead Publishing. https://doi.org/10.1533/9780845699994.2.229.
8. Meek, J. L., & Xia, X. (1999). Computer shape finding of form structures. International Journal
of Space Structures, 14(1), 35–55. https://doi.org/10.1260/0266351991494687.
9. Wakefield, D. S. (1999). Engineering analysis of tension structures: Theory and practice.

Engineering Structures, 21(8), 680–690. https://doi.org/10.1016/S0141-0296(98)00023-6.
10. Gruendig, L., Moncrieff, E., Singer, P., Strbel, D (2000). A history of the principal developments
and applications of the force density method in Germany 1970–1999. In 4th international
colloquium on computation of shell and spatial structures (IASS-IACM) (pp. 1–13). Chania-
Crete, Greece.
11. Veenendaal, D., & Block, P. (2012). An overview and comparison of structural form finding
methods for general networks. International Journal of Solids and Structures, 49(26), 3741–
3753. https://doi.org/10.1016/j.ijsolstr.2012.08.008.
12. Schek, H. J. (1974). The force density method for form finding and computation of general
networks. Computer Methods in Applied Mechanics and Engineering, 3, 115–134. https://doi.
org/10.1016/0045-7825(74)90045-0.
13. Argyris, J. H., Angelopoulos, T., & Bichat, B. (1974). A general method for the shape finding
of lightweight tension structures. Computer Methods in Applied Mechanics and Engineering,
3(1), 135–149. https://doi.org/10.1016/0045-7825(74)90046-2.
14. Barnes, M. R. (1999). Form finding and analysis of tension structures by dynamic relax-
ation. International Journal of Space Structures, 14(2), 89–104. https://doi.org/10.1260/
0266351991494722.
15. Lewis, W. J. (1989). The efficiency of numerical methods for the analysis of prestressed nets
and pin-jointed frame structures. Computers & Structures, 33(3), 791–800. https://doi.org/10.
1016/0045-7949(89)90254-X.
16. Barnes, M. R. (1988). Form-finding and analysis of prestressed nets and membranes. Computers
& Structures, 30, 685–695. https://doi.org/10.1016/0045-7949(88)90304-5.
17. Tabarrok, B., & Qin, Z. (1992). Nonlinear analysis of tension structures. Computers &
Structures, 45, 973–984. https://doi.org/10.1016/0045-7949(92)90056-6.
18. Dutta, S., Ghosh, S., Inamdar, M.M. (2017). Reliability-based design optimization of frame-
supported tensile membrane structures. ASCE-ASME Journal of Risk and Uncertainty in
Engineering Systems, Part A: Civil Engineering, 3(2). https://doi.org/10.1061/ajrua6.0000866.
19. Dutta, S., Ghosh, S., & Inamdar, M. M. (2018). Optimisation of tensile membrane struc-
tures under uncertain wind loads using PCE and kriging based metamodels. Structural
and Multidisciplinary Optimization, 57(3), 1149–1161. https://doi.org/10.1007/s00158-017-
1802-5.
20. Dutta, S., Ghosh, S. (2014). Numerical form-finding of a tensile membrane structure using
dynamic relaxation method. In Proceedings of the 5th International Congress on Computational
Mechanics and Simulation (ICCMS), Chennai, India.
21. Bridgens, B., Gosling, P., & Birchall, M. (2004). Tensile fabric structures: Concepts, practice
and developments. Structural Engineer, 82, 21–27.
22. ASCE: ASCE/SEI 55-10. (2010). Tensile membrane structures. American Society of Civil
Engineers.
23. Moncrieff, E., & Topping, B. (1990). Computer methods for the generation of membrane
cutting patterns. Computers & Structures, 37(4), 441–450. https://doi.org/10.1016/0045-
7949(90)90034-Y.
24. Veenendaal, D. (2017). Design and form finding of flexible formed concrete shell structures
(Ph.D. Thesis).
25. Gosling, P. D., Bridgens, B. N., Albrecht, A., Alpermann, H., Angeleri, A., Barnes, M.,
et al. (2013). Analysis and design of membrane structures: Results of a round robin exercise.
Engineering Structures, 48, 313–328. https://doi.org/10.1016/j.engstruct.2012.10.008.
26. Dutta, S., Ghosh, S. (2019). Analysis and design of tensile membrane structures: Challenges and
recommendations. Practice Periodical on Structural Design and Construction, 24(3). https://
doi.org/10.1061/(asce)sc.1943-5576.0000426.
27. Stranghner, N., Uhlemann, J., Bilginoglu, F., Bletzinger, K. U., Bgner-Balz, H., Corne, E.,
Gibson, N., Gosling, P., Houtman, R., Llorens, J., Malinowsky, M., Marion, J. M., Mollaert,
M., Nieger, M., Novati, G., Sahnoune, F., Siemens, P., Stimpfle, B., Tanev, V., Thomas, J. C.
(2016). Prospect for European guidance for the structural design of tensile membrane structures,
science and policy report (Sap-report). https://doi.org/10.1016/j.proeng.2016.08.027.
28. Stranghner, N., Uhlemann, J., & Mollaert, M. (2016). Background to the science and policy
report for tensile membrane structures. Procedia Engineering, 155, 256–264. https://doi.org/
10.1016/j.proeng.2016.08.027.
29. Huntington, C. G. (2013). Tensile fabric structures. American Society of Civil Engineers. https://
doi.org/10.1061/9780784412893.
Investigation of Cold-Formed Steel
Members Subjected to Extreme Low
Temperatures Relevant to the Arctic
Environment
Gaurav S. Chobe and Mahendrakumar Madhavan
1 Introduction
Metals tend to become hard and brittle under low-temperature condition. To meet
the increasing demand of oil and natural gas, the exploration activities taking place
in the Arctic region have expanded in the past decade due to the presence of 13% of
the world’s undiscovered oil and gas in the Arctic Circle [1]. However, the optimal
design and operation of Polar Offshore structures face the risk of brittle fracture of
the equipment due to extremely low temperatures (down to −50 °C).
Cold-formed steel (CFS) has an excellent combination of good strength-to-weight
ratio and high corrosion resistance [2], which makes it suitable for the rapid and
demountable construction of low-rise residential and industrial structures. Tempera-
ture induces a change in the mechanical properties of steel which affects the ultimate
load-carrying capacity of a steel structure [3].
Much of the research study has been conducted to understand the behavior of
the CFS at ambient and elevated temperature [4–8]. However, limited literature is
available on its performance at low temperatures relevant to the structural applications
[9–12]. According to Yu [3], when the temperature is lowered, the ultimate strength,
yield strength, and the elastic modulus generally increase, and the ductility decreases.
Hence, it is generally recommended to design CFS structures with great care.
In this study, results from a numerical investigation on the effect of temperature,
slenderness ratio, and cross-sectional geometry on the ultimate load-carrying capac-
ity of the unlipped CFS channel column are presented. A FE model was developed
G. S. Chobe (B) · M. Madhavan

Department of Civil Engineering, Indian Institute of Technology Hyderabad, Hyderabad, India
M. Madhavan

42 G. S. Chobe and M. Madhavan
and analyzed using ANSYS 15.0 software package. Specimen temperature was var-
ied from −60 to 20 °C at an interval of 20 °C to encompass the temperature variations
experienced in Polar Regions. The model was verified against the test results pro-
vided by Young and Yan [7]. FE results were found to be in good agreement with
the experimental data; hence, the model was used for the further parametric study to
analyze the behavior of different CFS channel columns at low temperatures. Finally,
the results obtained from the FE analysis were compared with the design strength
calculated using the Direct Strength Method (DSM) with modified material proper-
ties at low temperatures. Suitable graphs are drawn to demonstrate the effect of low
temperature on the load-carrying capacity of the column.
2 Background
Inadequate literature is available which aims to study the performance of steel struc-
tures at sub-zero temperatures relevant to structural applications. Polyzois et al. [9]
studied the compressive behavior of cold-formed angles in a temperature range of −
45 to 25 °C. They found that the capacity of angles at temperatures below −40 °C was
8% higher than the capacity at room temperature. In a study conducted by Abdel-
Rahim and Polyzois [10], it was found that yield strength, ultimate strength, and
elastic modulus of the CFS coupon increased by as much as 13.3%, 14%, and 24%,
respectively, at −50 °C compared to the room temperature. However, the maximum
percentage elongation decreased by as much as 35% at −50 °C. The research con-
ducted by Ritakallio and Björk [11] demonstrated that the Charpy-V impact tough-
ness decreased by approximately 50% at −50 °C compared to room temperature for
both corner portion and the flat portion of the hollow section EN10219.
3 Finite Element Modeling Details
3.1 Description of Material
The mechanical properties of CFS at low temperatures were obtained from the
research conducted by Abdel-Rahim and Polyzois [10]. Stress–strain curves were
generated for the material at temperatures ranging from −60 to 20 °C at an interval
of 20 °C. Mechanical properties like ultimate strength, yield strength, elastic modu-
lus, and maximum elongation were obtained for each temperature. Table 1 lists the
mechanical properties of CFS at different low temperatures.
Bilinear elastic–plastic stress–strain curve was assumed for modeling the material.
Investigation of Cold-Formed Steel Members Subjected … 43
Table 1 Mechanical properties of CFS at low temperatures

Temperature Yield strength Ultimate Elastic modulus Maximum strain
(°C) (MPa) strength (MPa) (GPa) (%)
−60 514.16 608.82 235.68 17.11
−40 501.59 592.59 229.78 18.66
−20 489.03 575.51 224.32 20.13
0 475.57 561.05 218.42 21.76
+20 463.89 543.97 213.39 23.16
3.2 Development of Finite Element Model
A three-dimensional nonlinear FE model was developed in ANSYS Mechanical

APDL 15.0 commercial software package to simulate the compressive loading con-
dition for the unlipped channel column. The SHELL181 element was used to model
the thin walls of the channel to incorporate out-of-plane deformations. To incorporate
the inelastic strains that will be developed in the CFS sections, engineering stress (σ )
and engineering strain (ε) were converted into corresponding true stress (σ T ) and
true strain (εT ) according to the following equations:
σT = σ (1 + ε) (1)
εT = ln(1 + ε) (2)
The amplitude for geometric imperfection such as camber and sweep was taken
as L/1000 for length L of the column. The condition of fixed ends was simulated by
restraining all six degrees of freedom at both ends except the translational degree of
freedom at one end. The residual stresses were not incorporated while modeling since
they are deemed negligible [13]. The effects of large deflections were accounted for
by using NLGEOM, ON command. A static nonlinear analysis was carried out to
obtain a solution with better accuracy. Size of mesh and distribution of finite elements
throughout the body were decided based upon a rigorous mesh sensitivity analysis.
The axial force was applied exactly on the center of gravity (CG) of the cross
section perpendicular to the plane of the cross section. The displacement control
method was used for loading the model. Multi-point constraint (MPC) algorithm
was used to assign the force at the CG points. Initially, CG point at each of the ends
was defined as a reference point. Then, the nodes at the column ends were coupled
to the respective reference points (Fig. 1).
By applying coupling constraints between the nodes at the column end and the
corresponding reference point, the translational and rotational degrees of freedom of
the column end were coupled to the corresponding reference point. For the fixed-end
condition, displacement and rotation constraints (U x = U y = 0, θ x = θ y = θ z =
0) were applied at both ends, and additionally, the displacement constraint (U z =
0) was applied at reference point Y. For each load step, the normal penalty stiffness
Fig. 1 Couplings, constraints, and mesh size in the FE model
factor was redefined to obtain more accurate results. The analysis was an iterative
process which was terminated when the load value decreased for the increasing value
of deflection.
3.3 Verification of Finite Element Model
The accuracy of the FE analysis was evaluated by comparing numerically derived

load-carrying capacity and the mode of failure from FE analysis with the experimental
test results. The developed FE model is validated with the results obtained from seven
tests carried out by Young and Yan [7]. The tests were carried out at room temperature,
and the validation is also performed at the same room temperature. Figure 2 shows
the comparison of the buckled shape observed in the experimental test setup and FE
analysis.
Table 2 shows the comparison of the values of the ultimate load-carrying capacity
for the experimental test and FE analysis (FEA). The difference between the ulti-
mate load-carrying capacity is found to be within acceptable limits with mean and
coefficient of variation (CoV) of Pexp /PFEA as 0.96 and 0.042, respectively. Hence,
the model was used for further parametric analysis.
Difference in the failure modes observed in experiments and in simulations is
attributed to the facts that the residual stresses are neglected while modeling the
column, the amplitude of the geometric imperfections was assumed to be equal to
Fig. 2 Comparison of failure mode observed in experiment [7] and FE model
Table 2 Comparison of experimental and FEA results

Specimen Experimental FEA Comparison
Pexp Failure PFEA Failure Pexp /PFEA
P48F0300 66.0 L+F 69.3 L 0.95
P48F1000 62.7 L+F 61.3 L+F 1.02
P48F1500 55.5 L+F 59.6 L+F 0.93
P48F1850 47.2 L+F 49.0 L+F 0.96
P48F2100 43.6 L+F 47.5 L+F 0.92
P48F2500 38.5 L+F 41.4 L + F+FT 0.93
P48F3000 37.4 L + F+FT 36.9 L + F+FT 1.01
Mean 0.96
CoV 0.042
L/ 1000 for each length of beam, and the load was applied exactly at the CG point,
which may not happen in practical situations.
3.4 Parametric Study
To understand the effect of different parameters on the load-carrying capacity of

CFS channel column, the following parameters were considered for the numerical
analysis:
1. The temperature of the specimen
2. Slenderness ratio of the specimen
3. The thickness of the CFS channel
4. Flange width.
The maximum and minimum temperatures of the specimen were considered to
be −60 °C and 20 °C, respectively. Slenderness ratio of the specimen was varied
by changing the length of the specimen from 250 to 3250 mm at an interval of
500 mm. The shortest length of the column was in accordance with the Structural
Steel Research Council guidelines [14] for stub columns, whereas the longest length
complies with the AS/NZS 4600 specifications [15] for slender columns. Flange
width was varied as 30 and 50 mm. The thicknesses of 1 and 2 mm were considered.
Larger thicknesses were not considered since plates thicker than 2.5 mm would
produce cracks upon a 90° bend [16]. Ideal fixed support was considered as the end
condition for the column.
3.5 Labeling of a Specimen
In the format for labeling, P represents pinned end. W and L denote the width of
the flange and the unbraced length of the column, respectively. Figure 3 explains the
labeling in general format.
Fig. 3 Format for labeling the specimen

4 Design Rules and Comparison of Design Strengths
4.1 Direct Strength Method
The Direct Strength Method (DSM) as given in the supplement to the North Amer-
ican Specification [17] is used to determine the column strengths by substituting
modified material properties at low temperature. The DSM developed by Schafer
and Peköz [18] and Schafer [19] is used to predict the strength of CFS columns at
low temperatures in this study.
The column design guidelines of DSM consider the local, distortional, and over-
all buckling. The unlipped channel sections do not fail in distortional buckling
mode since they lack edge-stiffened flanges. Since the present study consists of only
unlipped channel sections, distortional buckling is not considered while calculating
the strength. The rules governing the design of columns are given in Eqs. (3)–(8).
These equations take into account the effect of temperature while calculating the
strength of the column.

PDSM = min Pne,T , Pnl,T (3)
⎧
⎨ 0.658λ2c,T Py,T for λc,T ≤ 1.5
Pne,T = (4)
⎩ 0.877
Py,T for λc,T > 1.5
λ 2
c,T
⎧
⎨ Pne,T
for λl,T ≤ 0.776
Pnl,T = 0.4 0.4 (5)
Pcrl,T Pcrl,T
⎩ 1 − 0.15 Pne,T Pne,T
Pne,T for λl,T > 0.776
where Py,T = A g f y,T (6)

λc,T = Py,T /Pcre,T (7)

λl,T = Pne,T /Pcrl,T (8)
In the above equations, Ag represents the gross cross-sectional area of the channel,
and f y,T represents the yield strength of steel at temperature T. The nominal axial
strength (PDSM ) is calculated for two cases. Pne,T represents nominal axial strength
for flexural, torsional, and flexural–torsional buckling mode of failure at low tem-
peratures. Pnl,T represents nominal axial strength for local buckling as well as the
interaction of local and global buckling at low temperatures. The critical elastic local
buckling loads (Pcrl,T ) were determined in accordance with Appendix 2 of North
American Specifications for CFS Structures. The nominal axial strength (PDSM ) is
taken as the minimum of Pne,T and Pnl,T . The unfactored design column strengths
(PDSM ) obtained are compared with the FE results in Table 3 for the series P-1-W30
Table 3 Comparison of
Specimen DSM FEA Comparison
DSM and FEA results
PDSM PFEA PDSM /PFEA
P-1-W30-L0250 30,640 29,924 1.02
P-1-W30-L0750 25,506 25,400 1.00
P-1-W30-L1250 20,051 18,753 1.07
P-1-W30-L1750 17,783 16,777 1.06
P-1-W30-L2250 13,154 12,435 1.06
P-1-W30-L2750 9357 8828 1.06
P-1-W30-L3250 6621 6399 1.03
Mean 1.07
COV 0.02
at −20 °C. From the table, it can be noted that, in general, DSM conservatively pre-
dicts the strength of CFS columns at low temperatures with a mean and coefficient
of variation (CoV) of PDSM /PFEA as 1.07 and 0.02, respectively.
4.2 Results and Conclusion
Overestimated load capacities obtained from FE analysis may be attributed to the

neglected residual stresses while modeling the column. Linear regression analysis
is performed as shown in Fig. 4 by comparing the ultimate load-carrying capacity
Fig. 4 Comparison between PFEA (T ) and PFEA (20 °C) for fix-ended columns
at normal temperature (20 °C) and different low temperatures. The results indicate
that on an average, the ultimate capacity of channel columns was approximately
1.9%, 4.4%, 6.8%, and 8.8% higher at temperatures 0 °C, −20 °C, −40 °C, and
−60 °C, respectively, than at room temperature of 20 °C. Higher increment in the
load-carrying capacity was observed at lower temperatures due to the improvisation
of material properties at lower temperatures.
Stub columns were observed to fail in local buckling while slender columns
were observed to fail primarily in flexural buckling. Intermediate columns failed
in combined local and global buckling mode of failure.
5 Future Work
So far, the work is limited to the behavior of columns at low temperatures. To fully
understand the performance, future study will be carried out on beams, frames, beam–
columns, and connections to capture the modification in the load-carrying capacity
and the failure modes. Design guidelines will be suggested for the design of such
members at low temperatures.
References
1. Gautier, D. L., Bird, K. J., Charpentier, R. R., et al. (2009). Assessment of undiscovered oil
and gas in the Arctic. Science, 324(5931), 1175–1179.
2. AISI (American Iron and Steel Institute). (2007). Galvanized steel framing for residential
buildings. (Research Report RP06-1). Washington, DC: AISI.
3. Yu, W. W. (1982). Cold-formed steel design. N.Y.: Wiley Inc.
4. Gunalan, S., Kolarkar, P., & Mahendran, M. (2013). Experimental study of load bearing cold-
formed steel wall systems under fire conditions. Thin-Walled Structures, 65, 72–92.
5. Laím, L., Rodrigues, J. P. C., & de Silva, L. S. (2014). Experimental analysis on cold-formed
steel beams subjected to fire. Thin-Walled Structures, 74, 104–117.
6. Kankanamge, N. D., & Mahendran, M. (2011). Mechanical properties of cold-formed steels at
elevated temperatures. Thin-Walled Structures, 49, 26–44.
7. Young, B., & Yan, J. (2002). Finite element analysis and design of fixed-ended plain channel
columns. Finite Elements in Analysis and Design, 38(April), 549–566.
8. Huang, Z. F., & Tan, K. H. (2007). Structural response of restrained steel columns at elevated
temperatures. Part 2: FE simulation with focus on experimental secondary effects. Engineering
Structures, 29(September), 2036–2047.
9. Polyzois, D., Charnvarnichborikarn, P., Rizkalla, S. H., & Wong, C. K. (1990). Effect of temper-
ature and galvanization on the compressive strength of cold-formed angles. Canadian Journal
of Civil Engineering, 17, 440–451.
10. Abdel-Rahim, A. B., & Polyzois, D. (1996). Effect of temperature and galvanization on cold-
formed steel. Journal of Materials in Civil Engineering, 8(3), 114–122.
11. Ritakallio, P., & Björk, T. (2014). Low-temperature ductility and structural behavior of cold-
formed hollow section structures—Progress during the past two decades. Steel Construction,
7, 107–115.
12. Yan, J. B., Liew, J. R., Zhang, M. H., & Wang, J. Y. (2014). Mechanical properties of nor-
mal strength mild steel and high strength steel S690 in low temperature relevant to Arctic
environment. Materials and Design, 61(September), 150–159.
13. Young, B., & Rasmussen, K. J. R. (1998). Test of fixed-ended plain channel columns. Journal
of Structural Engineering, 124(2), 131–139.
14. Galambos, T. V. (Ed.). (1988). Guide to stability design criteria for metal structures (4th ed.,
pp. 708–710). New York: Wiley Inc.
15. Standards Australia. (2005). Cold-formed steel structures. AS/NZS 4600. Sydney, Australia:
Standards Australia.
16. Hancock, G., Murray, T., & Ellifritt, D. (2001). Cold-Formed steel structures to the AISI
specification. Boca Raton: CRC Press.
17. Supplement to the North American specification for design of cold-formed steel structural
members. Washington (DC): American Iron and Steel Institute (2004).
18. Schafer, B. W., Peköz, T. (1998). Direct strength prediction of cold-formed steel members
using numerical elastic buckling solutions. In Proceedings of the 14th International Specialty
Conference on Cold-Formed Steel Structures. (pp 69–76) University of Missouri-Rolla.
19. Schafer, B. W. (2002). Local, distortional, and Euler buckling of thin-walled columns. Journal
of Structural Engineering, 128(3), 289–299.
Experimental Investigation
on Crack-Arresting Mechanism of Steel
Fibre-Reinforced Concrete Prism
Specimens Using DIC and AE Techniques
Chandrashekhar Lakavath, Aniket Bhosale, and S Suriya Prakash
1 Introduction
Steel fibres play an important role in improving the ductility of the concrete
matrix which results in changing brittle failure mode to ductile failure mode. Many
researchers [1–8] have observed that the steel fibre-reinforced concrete (SFRC)
beams showcase increase in fracture energy with increase in fibre dosage. In addition
to this, it is important to understand the crack-arresting mechanism in SFRC beams
at different stages. Hence, to understand this behaviour, both digital image correla-
tion (DIC) technique and acoustic emission (AE) technique are used. Surface strain
quantities and crack propagation depths at different CMOD points were investigated
using DIC. On the other hand, using the AE technique, internal crack formation
mechanisms, such as concrete matrix rupture and fibre pull-out, were studied. The
mechanism of each fracture point was classified based on AE parameters.
DIC is a non-contact full-field displacement measurement technique, which can
be used to monitor the local surface displacement of the test specimen. In this tech-
nique, three basic stages need to be followed which are (i) pre-processing in which
the specimen preparation and capturing of images were performed, (ii) correlation of
images, which includes the analysis of all captured images by VIC-2D™ package and
(iii) post-processing stage includes the extraction of strains, displacement and other
C. Lakavath · A. Bhosale (B) · S. S. Prakash

Civil Engineering Department, Indian Institute of Technology Hyderabad (IITH), Hyderabad,
Telangana 502285, India
C. Lakavath
S. S. Prakash

52 C. Lakavath et al.
Fig. 1 3D damage source location in beams: a working principle of three sensors for identification
of damage location; b tension crack identification; c shear crack identification
required data from analysed images. Acoustic emission (AE) is a widely used tech-
nique to monitor structural health performance. It works based on the phenomenon of
detection of elastic waves in solids that occur when a material undergoes irreversible
changes in its structure. In this experimental investigation, three wideband sensors of
range 100–1000 kHz and three narrowband sensors of range 50–300 kHz were used
for monitoring the damage in notched specimens. The 3D location of the fracture
point is obtained based on the time difference in signals and respective velocities
captured by sensors. In the AE system, an event generated once the same hit was
captured by at least three sensors and it is identified as the source location.
Based on previous studies carried by Ohno [9, 10], the cracks were classified using
elastic wave parameters such as duration, frequency, amplitude and rise time. The
tensile fracture mode is having high frequency, low duration elastic waves and high
amplitude. Whereas, the shear failure mode can be identified by its low frequency,
high duration elastic waves and low amplitude as detailed in Fig. 1b, c.
Earlier studies [3, 6] have shown consistent results in locating the fracture pro-
cess zone reinforced concrete prism specimens and quantifying AE energies. In
this experimental study, a 3D source location was identified for the various damage
mechanisms in the fracture process zone at different damage stages.
2 Research Significance
As a secondary reinforcement in the concrete matrix, the discrete hooked-end steel

fibres were added. Due to which the variation in fracture energy, strain redistribution
and crack propagation variations were investigated.
1. Variation in fracture energy with the addition of different volume fraction (V f )
of hooked-end steel fibres.
2. Role of steel fibres in strain redistribution and crack propagation across the depth
of notched beams using DIC technique.
3. Identification of three-dimensional fracture location using AE sensors, which
helps in indicating fracture points in FPZ.
Experimental Investigation on Crack-Arresting Mechanism … 53
3 Experimental Programme
The steel fibres of volume fraction (V f ) (0.0, 0.5, 0.75 and 1.0%) are the variables
and other parameters such as concrete strength, dimensions of specimen, initial crack
length and loading conditions were kept constant. The mix design details used in this
experimental test programme is shown in Table 1. The hooked-end steel fibres of
aspect ratio 50 were used in this test programme and other properties of the steel
fibres are enlisted in Table 2.
3.1 Raw Materials
Concrete The concrete of compressive strength 60 MPa was prepared as per

IS10262-2009 [11] used in this experimental programme. The details of concrete mix
proportions of 60 MPa grade concrete is presented in Table 1. In the same con-
crete matrix, steel fibres with different volume fractions were added for steel fiber
reinforced concrete mix.
Steel fibres To improve the post-peak response in plain concrete, hooked-end steel
fibres were added to the concrete matrix. Based on the previous research [12–14],
steel fibre was selected. The steel fibres of modulus of elasticity of 200 GPa with a
diameter of 0.6 mm and length of 30 mm were used in Table 2.
Table 1 Mix design proportions of ingredients (kg/m3 )

Ingredient Cement Fine aggregate Coarse aggregate Water Steel fibres (0, 0.5,
0.75 and 1.0%)
Proportion 320 668 1062 144 0, 39.25, 58.875,
78.5
Table 2 Properties of fibres

Specifications Properties Hooked-end fibre
Fibre type Hooked-end steel
Tensile strength 1000
(MPa)
Modulus of 200
elasticity (GPa)
Specific gravity 7.85
Length (mm) 30
Diameter (mm) 0.6
Aspect ratio 50
Density (kg/m3 ) 7850
Fig. 2 Specimen preparation sequence: a mixing using drum-type mixture, b cubes after 24 h. of
concreting, c prism specimens, d testing cube using compression testing machine
3.2 Casting and Testing Procedure
All raw materials were mixed by using drum-type mixture and steel fibres were spread
by hand in the mixing process. The sequence of specimen preparation is depicted
in Fig. 1. The concrete beams of size 150 mm × 150 mm × 500 mm were cast for
each dosage in the same batch. Similarly, the concrete cubes were cast to examine
the 28-day compressive strength of concrete.
The prism specimens were prepared as per EN14651-2005 [15] with a notch of
25 mm depth and 5 mm width at the middle of specimen perpendicular to the face
of casting direction [16, 17]. The specimens were tested with a rate of 0.05 mm/min
opening of CMOD by using a servo-controlled flexure testing machine as shown in
Fig. 2. To evaluate the response of prism specimens, one LVDT at mid-span, CMOD
gauge and CTOD gauge at notch location was used.
4 Experimental Results and Discussions
4.1 Compressive Strength of Concrete
In each mix, three numbers of cubes of size 150 mm × 150 mm × 150 mm were tested
at a rate of 4120 N/s using compression testing machine as per IS456-2000 [18]. From
test results, it is observed that there is not much improvement in compressive strength,
the improvement in compressive strength is from 61.63 to 65.36 MPa (Table 3).
Table 3 Concrete cubes of

Steel fibre dosage Compressive Standard deviation
compressive strength (MPa)
(V f ) strength (MPa) (MPa)
0.00% (Control) 61.63 1.97
0.50% 65.66 0.40
0.75% 63.09 1.83
1.00% 65.36 0.47
4.2 Residual Flexural Strength by RILEM Code

and Fracture Energy by JCI Code
The test configuration of a Mode-I fracture test is shown in Fig. 3a. The response of
load-CMOD curves represents that as increase in fibre volume fraction the ductility
of fibres was increased (Fig. 3b). From this load-CMOD curves of Mode-I fracture
test, the residual flexural tensile strengths (f R,1 , f R,2 , f R,3 and f R,4 ) are calculated
based on RILEM [19] recommendations. The residual flexural strength calculated
by Eq. (1) at 0.5 mm, 1.5 mm, 2.5 mm and 3.5 mm CMODs, respectively, to evaluate
the fracture response through serviceability limit to the ultimate limit.
3FR,i L
f R,i = (MPa) (1)
2b h 2sp
As per Japanese concrete institute standards [16, 17], the fracture energy (GF ) of
control beams (SF00) and steel fibre-reinforced concrete (SFRC) beams was calcu-
lated using Eqs. (2), (3) to evaluate and compare the effect of fibre addition on the
fracture behaviour.
20
18
16
14
Load (kN)
12
10
8
SF00 (Vf=0.0%)
6
SF50 (Vf=0.5%)
4 SF75 (Vf=0.75%)
2 SF100 (Vf=1.0%)
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
CMOD (mm)
(a) (b)
1. Loading Jig, 2. The notched specimen, 3. AE sensors, 4. Preamplifier, 5. DIC camera
Fig. 3 Fracture test of SFRC beams: a test set-up for Mode-I test, b load-CMOD response. 1
Loading Jig, 2 the notched specimen, 3 AE sensors, 4 preamplifier, 5 DIC camera
Table 4 Residual flexural SFRC prisms specimens

Beam ID Residual flexural tensile strength (MPa) Fracture Increment
energy in fracture
(Gf ) energy
(N/mm)
f R1 f R2 f R3 f R4 Gf (× Gf =
(CMOD = (CMOD = (CMOD = (CMOD = 10−3 ) Gf /Gfc
0.5 mm) 1.5 mm) 2.5 mm) 3.5 mm)
SF00 0.69 0.00 0.00 0.00 156.60 1.00
SF50 2.89 2.74 2.67 2.35 1656.8 10.58
SF75 3.36 3.57 3.34 2.97 1983.7 12.66
SF100 4.27 4.58 4.35 3.99 2697.9 17.23
0.75W0 + W1
GF = (2)
Alig

S
W1 = 0.75 m 1 + 2m 2 g.CMODc (3)
L
Fracture test results reveal that the fracture energy of steel fibre-reinforced con-
crete is improved to 10.6, 11.78 and 17.2 times in relation to control beams (SF00) for
SF50, SF75 and SF100 beams (Table 4) [20]. The improvement in fracture energy is
due to the steel fibres bridging action across the cracked plane. It has been observed
at all the stages that the residual flexural strengths increased significantly as the V f
was increased. This improvement in residual strengths was more significant just after
cracking resulting in improved stiffness of the beams.
From the load-CMOD response of SFRC notched beams, the improvement in
post-cracking load distribution was increased. This is due to the steel fibres bridging
action against the crack opening. The same mechanism was confirmed by observing
the surface of tested specimens. It was observed that the few fibres along the depth
of the beams were pulled out with some of the fibres hook ends were flattened;
this confirms the perfect bond between the concrete and fibre, which leads to stress
transformation. Due to this, the improvement in ductility is observed in steel fibre-
reinforced concrete beams. As the volume fraction increases, the number of fibres
along the depth of the beam was increased. Hence, the improvement in ductility is
more in case of a higher dosage of fibres.
4.3 Monitoring of Crack-Arresting Mechanism by Using DIC

Technique
In the pre-possessing stage, all the test prism specimens are painted using acrylic-
based white paint for having good correlation over the speckled area. The equal
intensity of light on the test zone was ensured throughout the test to retain uniform
illumination of the surface. Prior to the test starts, a calibration image (with scale)
and a reference image (un-deformed shape) were captured. Digital images were
captured at equal intervals throughout the test to capture deformed shapes. In post-
processing, DIC works by pattern matching principle where the speckle patterns of
the un-deformed and deformed images were correlated to obtain the corresponding
displacement and strains. The subset size directly determines the area of the subset
from the reference image to track in the target image. During analysis, the subset
size of 35 × 35 pixels was used for correlation [3, 21–23].
To understand the crack-arresting mechanism of notched SFRC specimens, sev-
eral CMOD points were considered throughout the test such as 0.5, 1.5, 2.5 and
3.5 mm. The surface strain redistribution across the crack was observed in SF75 and
SF100 as shown in Fig. 4. As a result of more number of fibres bridging the crack,
the strain localisation zone at 3.5 mm CMOD, which was observed at the tip of the
crack in SF00 and SF50, was found to have distributed over the depth of beam in
SF75 and SF100. This strain redistribution leads to absorb more energy and results
0.5 1 0.5 1.5 2.5 3.5

20 20
SF00 (Vf=0.0%) SF50 (Vf=0.5%)
18 18 0.5 1.5 2.5 3.5
16 16
0.5 1.0
14 14
Load (kN)
Load (kN)
12 12
10 10
8 8
6 6
4 4
2 2
0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
CMOD (mm) CMOD (mm)
(a) (b)
0.5 1.5 2.5 3.5 0.5 1.5 2.5 3.5

20 20
SF75 (Vf=0.75%) SF100 (Vf=1.0%)
18 18
16 16
14 14
Load (kN)
Load (kN)
12 12
10 10 0.5 1.5 2.5 3.5
8 0.5 1.5 2.5 3.5 8
6 6
4 4
2 2
0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
CMOD (mm) CMOD (mm)
(c) (d)
Fig. 4 Load versus CMOD of notched beams: a SF00, b SF50, c SF75, d SF100 specimen
Table 5 Crack length of steel fibre-reinforced notched beams at different CMOD points
Beam ID/CMOD Crack length (mm)
0.2 0.5 1.5 2.5 3.5
SF00 56.9 87.7 96.85 (CMOD = 1 mm) – –
SF50 48.6 81.0 102.04 102.04 102.04
SF75 31.9 80.0 99.70 103.20 108
SF100 10.56 64.28 108.44 112.24 115.14
in efficient use of the cross section. The crack propagation path was also observed to
be meandering as the fibre dosage was increased. It may be because of more fibres
affecting the weak link path as shown in Fig. 4. Due to which the crack length at
respective CMOD values were varied and same is presented in Table 5.
4.4 Monitoring of Crack Propagation Mechanism by Using

Acoustic Emission Technique
A total of six acoustic emission sensors were attached on the specimen at the same
coordinates of the beam as shown in Fig. 5. The sensors configuration was fixed to
improve the accuracy of crack detection. Also, the proposed sensor configuration
is required if the AE results are to be subjected to Simplified Green’s functions for
Moment Tensor Analysis (SiGMA).
Based on the distance between the sensors, the loss in the amplitude and scattering
in capturing of signals was recorded. After conducting trial tests, the amplitude
threshold was set to 40 dB to ignore the effect of environmental/electronic noise. In
AE system, the preamplifier amplitude gain was set as 40 dB. Before the actual test
starts, each AE sensor was calibrated by standard pencil lead test to ensure sensors
Fig. 5 Notched beam with AE sensors location: a position of AE sensors, b preamplifier, c AE

sensor
attachment to the specimen surface. In this study, AE data acquisition (DAQ) system
with eight channels was used for recording the signals at rate of 5 MHz.
As expected in fracture test, most of the fracture events captured by AE sensors
were in tension mode. Since the test set-up (three-point bending) configuration causes
the matrix to undergo Mode-I fracture. In case of control beams (SF00) at CMOD
of 0.5 mm, the events were effectively distributed up to 100 mm height from the
bottom-most fibre of the beam as shown in Fig. 6. The number of tension events
was comparatively much higher than the number of shear events. The distribution of
tension and shear events up to CMOD of 1 mm, as presented in Fig. 6b, represents
the fracture process zone length.
In the case of 0.5% volume fraction (SF50) beams at CMOD of 0.5 mm, the AE
activities clearly show the fracture process zone and nature of crack propagation.
Inner progressive damage was fairly captured by AE sensors and surface crack prop-
agation as well as strain redistribution was monitored by DIC. Figure 7a–d represents
the detailed 3D locations of all the tension and shear events. As CMOD value was
increased, the shift in FPZ was clearly showcased by the AE activities. The crack
length is varying from 81 to 102 mm for CMOD of 0.5–3.5 mm.
Similarly, in the case of SF75 and SF100 specimens, the number of events
increased with increase in respective CMOD values (Fig. 8 and Fig. 9). And, it was
noticed that with an increase in fibre dosage, number of shear events also increased
significantly. This may be due to the pull-out of steel fibres during the crack propa-
gation. To understand the clear crack propagation mechanism in FPZ of SFRC spec-
imens, further study is needed and a number of investigation points in load-CMOD
curve have to be considered.
The width of FPZ is almost the same irrespective of fibre volume fraction; the
90% of events lie in the range of 160 mm along the length of the specimen as shown
in Fig. 6 and Fig. 7. On contrary, the range of FPZ length along the crack propagation
direction varies as fibre dosage was increased [3]. Since AE events depend on the
number of fibres as well as the type of fibres added to the concrete matrix. To
understand the effect of fibre volume fraction and types of fibres in FPZ, further
study needs to be carried by using DIC and AE techniques.
5 Conclusions
To understand the influence of hooked-end steel fibres on crack propagation and

fracture energy of the concrete matrix, notched beams and cube specimens were
tested. The variable in this study is fibre volume fraction such as 0.0, 0.5, 0.75 and
1.0%. All experimental test specimens were monitored using acoustic emission (AE)
and digital image correlation (DIC) techniques. From this test series, the following
conclusions were observed,
1. With the addition of hooked-end steel fibres from 0.0 to 1.0% of volume fraction,
there is not much improvement in compressive strength.
60
Strain in Strain in Strain contour

SF00 Beam AE events location AE events location
X-direction X-direction in X-direction
CMOD of
CMOD of 1.0mm
0.5mm
(a) (b) (c) (d) (e)
Fig. 6 Acoustic emission hits location and strain in longitudinal direction of SF00 beam: a events up to CMOD of 0.5 mm, b strain at 0.5 mm of CMOD,
c events up to CMOD of 1.0 mm, d strain at 1.0 mm of CMOD, e strain in x-direction
C. Lakavath et al.
Strain in X- Strain in X- Strain contour
direction direction in X-direction
CMOD of CMOD of
0.5 mm 1.5 mm
(a) (b) (c) (d)
CMOD of CMOD of
2.5 mm 3.5 mm
(i)
Experimental Investigation on Crack-Arresting Mechanism …
(e) (f) (g) (h)
c events up to CMOD of 1.5 mm, d strain at 1.5 mm of CMOD, e events up to CMOD of 0.5 mm, f strain at 0.5 mm of CMOD, g events up to CMOD of
0.5 mm, h strain at 0.5 mm of CMOD, i strain in x-direction
61
62
Strain in X- Strain in X- Strain contour

direction direction in X-direction
CMOD of
0.5 mm CMOD of
1.5 mm
(a) (b) (c) (d)
CMOD of CMOD of
2.5 mm 3.5 mm
(i)
(e) (f) (g) (h)
C. Lakavath et al.
Strain in Strain in Strain contour
X-direction X-direction in X-direction
CMOD of
0.5 mm CMOD of
1.5 mm
(a) (b) (c) (d)
CMOD of CMOD of
2.5 mm 3.5 mm
(i)
Experimental Investigation on Crack-Arresting Mechanism …
(e) (f) (g) (h)
63
2. The three-point bending test reveals that the fracture energy is changing from
156.6 × 10−3 (N/mm) to 2697.9 × 10−3 (N/mm), i.e. energy absorption amount
was increased to 17 times with the addition of fibres from 0.0 to 1.0% of volume
fraction.
3. DIC investigation illustrates that the surface strain redistribution increased in
the fracture process zone due to fibre bridging action, resulting in reduced crack
length as well as crack width with an increase in fibre volume fraction.
4. The number of AE events increased with an increase in fibre dosage from 0.0 to
1.0%, which indicates a rise in the fibre pull-out activities in the fracture process
zone.
Acknowledgements This research work is carried out as part of the project funded through Utchat-
tar Avishkar Yojana (UAY) Scheme of Government of India. The authors thankfully acknowledge
the support.
References
1. Banthia, N., & Gupta, R. (2004). Hybrid fibre reinforced concrete (HyFRC): Fibre synergy in
high strength matrices. Materials and Structures, 37(10), 707–716.
2. Majdzadeh, F., Soleimani, S. M., & Banthia, N. (2006). Shear strength of reinforced concrete
beams with a fibre concrete matrix. Canadian Journal of Civil Engineering, 33(6), 726–734.
3. Bhosale, A., Rasheed, M. A., Prakash, S. S., & Raju, G. (2019). A study on the efficiency of
steel vs. synthetic vs. hybrid fibres on fracture behavior of concrete in flexure using acoustic
emission. Construction and Building Materials, 199, 256–268.
4. Das, S., Aguayo, M., Sant, G., Mobasher, B., & Neithalath, N. (2015). Fracture process zone
and tensile behavior of blended binders containing limestone powder. Cement and Concrete
Research, 73, 51–62.
5. Das, S., Aguayo, M., Dey, V., Kachala, R., Mobasher, B., Sant, G., et al. (2014). The frac-
ture response of blended formulations containing limestone powder: Evaluations using two-
parameter fracture model and digital image correlation. Cement & Concrete Composites, 53,
316–326.
6. Rasheed, M. A., Prakash, S. S., Raju, G., & Kawasaki, Y. (2018). Fracture studies on synthetic
fibre reinforced cellular concrete using acoustic emission technique. Construction and Building
Materials, 169, 100–112.
7. Lakavath, C., Pidapa, V., & Joshi, S. S. (2018). Shear behavior of steel fibre reinforced pre-
cast prestressed concrete beams. In International Conference on Composite Materials and
Structures.
8. Lakavath, C., Allam, R., & Kondraivendhan, B. (2019). Experimental and numerical studies
on the behaviour of broad-gauge railway sleepers in static bending condition. In Sustainable
construction and building materials (pp. 781–792). Singapore: Springer.
9. Ohno, K., & Ohtsu, M. (2010). Crack classification in concrete based on acoustic emission.
Construction and Building Materials, 24(12), 2339–2346.
10. Ohno, K., Shimozono, S., Sawada, Y., & Ohtsu, M. (2008). Mechanisms of diagonal-shear
failure in reinforced concrete beams analyzed by AE-SiGMA. Journal of Solid Mechanics and
Materials Engineering, 2(4), 462–472.
11. Indian standard concrete mix proportioning - Guidelines (First revision) IS 10262:2009, Bureau
of India Standard, New Delhi, India.
12. Joo Kim, D., Naaman, A. E., & El-Tawil, S. (2008). Comparative flexural behavior of four
fibre reinforced cementitious composites. Cement & Concrete Composites, 30(10), 917–928.
13. Lee, S. C., Oh, J. H., & Cho, J. Y. (2015). Compressive behavior of fibre-reinforced concrete
with end-hooked steel fibres. Materials, 8(4), 1442–1458.
14. Hsu, L. S., & Hsu, C. T. (1994). Stress-strain behavior of steel-fibre high-strength concrete
under compression. Structural Journal, 91(4), 448–457.
15. En, B. (2007). 14651, “Test method for metallic fibre concrete—Measuring the flexural tensile
strength (Limit of proportionality (LOP), residual). London, UK: British Standards Institution.
16. JCI-S-002-2003. Method of test for load-displacement curve of fibre reinforced concrete by
use of notched beam. 2003; (L):3–8.
17. JCI-S-001-2003. Method of test for fracture energy of concrete by use of notched beam. 2003;
2(2):1–14.
18. BIS: 456-2000 R 2005. Plain and reinforced concrete. Bur Indian Stand Dehli. 2005.
19. Vandewalle, L., Nemegeer, D., Balazs, L., Barr, B., Barros, J., Bartos, P., & Falkner, H. (2003).
RILEM TC 162-TDF: Test and design methods for steel fibre reinforced concrete’-sigma-
epsilon-design method-final recommendation. Materials and Structures, 36(262), 560–567.
20. Lakavath, C., Joshi, S. S., & Prakash, S. S. (2019). Investigation of the effect of steel fibres
on the shear crack-opening and crack-slip behavior of prestressed concrete beams using digi-
tal image correlation. Engineering Structures, 193, 28–42. https://doi.org/10.1016/j.engstruct.
2019.05.030.
21. Pan, B., Xie, H., Guo, Z., & Hua, T. (2007). Full-field strain measurement using a
two-dimensional Savitzky-Golay digital differentiator in digital image correlation. Optical
Engineering, 46(3), 033601.
22. Rajan, V. P., Rossol, M. N., & Zok, F. W. (2012). Optimization of digital image correlation
for high-resolution strain mapping of ceramic composites. Experimental Mechanics, 52(9),
1407–1421.
23. Srikar, G., Anand, G., & Prakash, S. S. (2016). A study on residual compression behavior
of structural fibre reinforced concrete exposed to moderate temperature using digital image
correlation. International Journal of Concrete Structures and Materials, 10(1), 75–85.
Estimation of Fly Ash Reactivity
and Dissolution Characteristics
G. V. P. Bhagath Singh and K. V. L. Subramaniam
1 Introduction
Fly ash is an industrial waste material produced by thermal power plants. In develop-
ing countries like India, there is an increasing gap between production and utilization
of fly ash. Currently, 184 million tones of fly ash are available, but only 55% is being
used for various applications [1]. Fly ash is rich in an aluminio-silicate source, and it
is often used in the production of geopolymers and in binary fly ash cement system
[2–5]. The alumina and silica in fly ash are present in both amorphous and crystalline
phase [6]. Crystalline phases are indicative of the non-reactive portion, and it typ-
ically consists of quartz and mullite in predominant quantities [6]. The amorphous
phase is representative of the glassy phase. The reactivity of fly ash depends on the
glassy phase and release of reactive species into the solution.
Previous research has been shown; calcium and silicon were the first elements to
dissolve followed by aluminium in a low alkaline environment [7]. The dissolution of
aluminium and silicon was increased in the solution with an alkaline concentration
and higher curing temperatures [8]. A study by Li et al. [9] reported that higher
the molarity level, curing time and temperatures favour the dissolution of silicon
and aluminium from the glassy phase. The lower reactive glassy content fly ashes
are typically found in India, and large quantities of silica and alumina are present
in crystalline portion [6]. Successful development of fly ash-based geopolymers
requires an understanding of the dissolution characteristics of the reactive species
from the glassy phase.
G. V. P. Bhagath Singh (B)

Laboratory of Construction Materials, IMX, EPFL, 1015 Lausanne, Switzerland
Department of Civil Engineering, Indian Institute of Technology
Hyderabad (IITH), Hyderabad, Telangana 502285, India
68 G. V. P. Bhagath Singh and K. V. L. Subramaniam
The current paper explores the estimation of reactivity and dissolution of sili-
con, and aluminium elements from low glassy content fly ash at various alkaline
environments and at different curing temperatures are evaluated. The link between
the dissolution of low reactive glass and releasing of elements into the solution is
established.
2 Materials and Methods
Fly ashes used in this study are directly collected from power stations without any
prior treatment and labelled as B and D. The chemical composition of fly ashes is
listed in Table 1. Both the ashes are classified as siliceous fly ash as per BIS 3812
[10] and as per ASTM C 618 [11]. Due to the very low calcium content, further fly
ashes are classified as low calcium siliceous fly ashes.
2.1 Dissolution Experiments
The dissolution mechanism of fly ash was analysed in two different alkaline envi-
ronments and at two different curing temperatures. Sodium hydroxide (NaOH) was
used as an alkaline activator. 1M and 6M NaOH alkaline solution and 25 and 60 °C
temperatures were used. The experimental work is as follows: 1 gm of ash was added
to 100 ml alkali solution in a 150 ml polypropylene container. Dissolution measure-
ments were carried out at 3, 24 and 48 h. At a particular time, the solution was filtered
through a Whatman filter paper. The solution was used for analyses purpose. In all
the samples, Si and Al elements were evaluated, as these are the major elements for
geopolymerization process as well as for pozzolanic reaction. The glassy content
present in fly ashes was determined using two different techniques such as 1% HF
acid dissolution [12] and Rietveld-based XRD method [13]. The dissolved elements
Table 1 Chemical
Composition Fly ash B Fly ash D
composition of fly ashes
(wt%) SiO2 55.42 55.57
Al2 O3 27.10 29.01
Fe2 O3 7.30 5.47
CaO 2.46 1.65
K2 O 2.55 2.12
TiO2 1.91 2.28
MgO 1.20 0.63
SO3 0.56 0.66
P2 O5 0.54 0.78
Estimation of Fly Ash Reactivity and Dissolution Characteristics 69
present in the solution were determined using inductively coupled plasma optical
emission spectrometry (ICP-OES).
D2 Phaser (Bruker) benchtop XRD was used to measure the intensity pattern
of the samples. The readings were taken from 10° and 70° at a step size of 0.02°
and 0.6 step per second, resulting in a total measurement time of 30 min per scan.
Corundum was used as an external standard to determine various phases. The total
glassy content present in the fly ash is determined by subtracting all the crystalline
phases from unity [13]. The reactive Si and Al elemental contents present in the
glassy phase were determined by subtracting the crystalline phases from the total Si
and Al [9, 12, 13].
3 Results and Discussion
3.1 Identification of Glass Type
The XRD pattern of fly ash D is shown in Fig. 1. A wide hump is identified between
2θ angles equal to 15° and 30° along with various crystalline phases. The hump
is the representation of the glassy phase present in fly ash. The crystalline phases
primarily consist of quartz and mullite along with other minor phases. The XRD
pattern of synthesized low and high calcium aluminosilicate glass is shown in Fig. 1b.
By comparing the XRD pattern of fly ash amorphous portion with the synthesized
glasses, the glass present in the fly ash is low calcium aluminosilicate glass, and it is
centred at a lower 2θ angle approximately 22°.
Fig. 1 XRD pattern of a fly ash D and b synthesized glasses

Fig. 2 Comparison of a glassy content and b reactive Si and Al contents using both the
methodologies
3.2 Comparison of Glassy Content and Reactive Species

Using Different Methodologies
As shown in Fig. 2, both the methodologies are shown similar in values with a little
difference. The glassy content obtained from the Rietveld refinement and elemental
contents obtained from XRF-XRD methods are shown as lower values compared
to 1% HF treatment method. The difference in the values is due to some of the
crystalline phases sensitive to acid treatment, and few particles may be left on top of
the filter paper.
3.3 Influence of Molarity and Temperature on Dissolution

of Glassy Portion
The change in the XRD pattern of glassy portion in 6M alkaline solution at both the
curing temperatures of fly ash B is shown in Fig. 3. It can be seen that the alkaline
concentration clearly influences the dissolution process, and it enhanced with the
curing temperature. During the dissolution process, crystalline phases remain inert
in the system, and glass is completely depleted by 48 h at higher curing temperature.
3.4 Influence of Alkaline Concentration on Dissolution

of Elements
The changes in the Si and Al elemental concentration at different molarities cured at

25 °C in both the fly ashes are shown in Fig. 4. Elemental concentration in the solution
is increasing with time and molarities. At a given time, the elemental concentration
in the solution is high at higher molarity. There is an increase in the elemental
concentration from 1M to 6M. This is indicative of higher molarity that favours a
more dissolution of the glassy phase. By comparing Fig. 2b, still Si and Al elements
need to be dissolved, and it means that the dissolution process is still progressing in
the system.
Estimation of Fly Ash Reactivity and Dissolution Characteristics 71
Fig. 3 XRD patterns of dissolved glassy content in 6M alkaline solution at 48 h
Fig. 4 Dissolved Al and Si elements with time at different molarities. a fly ash B and b fly ash D
Fig. 5 Dissolved Al and Si elements at two different curing temperatures with time. a fly ash B,
b fly ash D
3.5 Influence of Temperature on Dissolution of Elements
The changes in the elemental proportions in the 6M solutions at 25 and 60 °C for fly
ashes B and D are shown in Fig. 5. As per the results, temperature clearly enhances
the elemental quantities in the solution at any particular time. Higher Si and Al
elements were dissolved at 60 °C compared to 25 °C within 3 h. Samples cured at
60 °C, significant elemental quantities are observed at 48 h. By comparing Fig. 2b,
Si and Al elements are completely dissolved at 60 °C by 48 h.
4 Conclusions
The dissolution behaviour of low glassy content siliceous fly ash system is evaluated.
Dissolution process depends on the glassy content present in the fly ash. The exper-
imental work shows that the dissolution of the glassy phase leads to releasing the
reactive species into the solution, primarily Si and Al without affecting the crystalline
phases. Molarity level and temperature enhance the rate of the dissolution process.
References
1. Central Electricity Authority. (2015). Annual report on fly-ash utilization. India.

2. Provis, L., & Van Deventer, J. S. J. (2009). Geopolymers: Structure, processing, prosperities
and industrial applications (41p). Cambridge, UK: Woodhead Publishing Limited.
3. Davidovits, J. (1999). Geopolymers: Inorganic polymeric new materials. Journal of Thermal
Analysis, 37(8), 1633–1656.
4. Palomo, A., Grutzeck, M., & Blanco, M. (1999). Alkali-activated fly ashes: A cement for the
future. Cement and Concrete Research, 29(8), 1323–1329.
5. Bhagath Singh, G. V. P., & Subramaniam, K. V. L. (2014). Activation of siliceous fly ash at
very high levels of cement replacement. Indian Concrete Journal, 88, 23–32.
6. Bhagath Singh, G. V. P., & Subramaniam, K. V. L. (2018). Characterization of Indian fly ashes
using different experimental techniques. Indian Concrete Journal, 92, 10–23.
7. Lee, W. K. W., & van Deventer, J. S. J. (2002). Structural reorganization of class F fly ash in
alkaline silicate solutions. Colloids Surfaces A, 211(1), 49–66.
8. Mikuni, A., Komatsu, R., & Ikeda, K. (2007). Dissolution properties of some fly ash fillers
applying to geopolymeric materials in alkali solution. Journal of Material Science, 42, 2953–
2957.
9. Li, C., Li, Y., Sun, H., & Li, L. (2011). The composition of fly ash glass phase and its dissolution
properties applying to geopolymeric materials. Journal of American Ceramic Society, 94(6),
1773–1778.
10. BIS (Bureau of Indian Standards). (2003). Pulverized fuel ash-specification. IS 3812-part 1.
New Delhi: Manak Bhavan.
11. ASTM C 618 Standard. (2015). Specification for fly ash and raw or calcined natural Pozzolan
for use as a mineral admixture in Portland cement concrete.
12. Fernandez-Jimenez, A., De la Torre, A. G., Palomo, A., Lopez-Olmo, G., Alonso, M. M., &
Aranda, M. A. G. (2006). Quantitative determination of phases in the alkali activation of fly
ash. Part I. Potential Ash Reactivity. Fuel, 85(5), 625–634.
13. Bhagath Singh, G. V. P., & Subramaniam, K. V. L. (2016). Quantitative XRD analysis of binary
blends of siliceous fly ash and hydrated cement. Journal of Material in Civil Engineering
(ASCE), 28(8), 1–7 (04016042).
Numerical Modelling of Continuous
Composite Beam Under Fire Loading
Priya S. Natesh and Anil Agarwal
1 Introduction
Composite beams consist of a steel beam and concrete slab connected through shear
studs. The composite beam optimises the material by making use of concrete slab in
the compression zone and structural steel beam in the tension zone of the member.
A concrete slab can be of flat shape or trapezoidal metal deck in nature. Composite
floor systems are gaining popularity in India as the need for rapid construction is
increasing. Currently, there is a significant need to understand its behaviour, espe-
cially under various critical loading conditions such as earthquake, fire, and impact
loading. When a composite beam subjected to fire loading, it leads to very complex
behaviour. Usually, during a fire, the members will be subjected to service level grav-
ity load along with the thermal loading associated with fire. The members are said
to be healthy if they can resist the structural gravity load and thermal load without
any overall collapse. Experimental investigation of structural members subjected to
fire loading is very expensive and challenging. Experiments can be accompanied
by numerical models to develop a deeper insight into the behaviour of structural
members.
Wainman and Kirby [1] studied the functioning of the simply supported composite
beam of effective span 4.58 m under fire loading. The specimens consisted of 254 ×
146 × 43 UB steel beams that were connected to flat-reinforced concrete slabs using
headed shear studs. The width and depth of the slab were 0.642 m and 0.130 m,
respectively. 8 mm diameter reinforcement with 200 mm spacing in longitudinal
P. S. Natesh (B)
Faculty of Science and Technology, ICFAI Foundation for Higher Education (Declared as
Deemed to be University u/s 3 of the UGC Act 1956), Hyderabad 501203, India
A. Agarwal
Indian Institute of Technology Hyderabad (IITH), Hyderabad, Telangana 502285, India
74 P. S. Natesh and A. Agarwal
direction and 100 mm spacing in transverse direction is used in the concrete slab.
Once the plastic moment capacity was reached at elevated temperature, the beam
failed in flexure.
Zhao and Kruppa [2] performed an experimental investigation on 4.9 m long
composite beam under fire loading. They investigated the effect of varying the con-
crete slab width and influence of protection for steel beam. IPE300 steel beams were
connected to flat concrete slabs using 19 mm diameter headed shear studs. The slab
widths considered were 1.20 m, 1.60 m, 0.80 m, and 1.20 m, respectively. Crushing
of concrete leads to the failure of protected composite beams and plasticisation of
the steel beam leads to the failure of unprotected beams.
Anil et al. [3–6] studied in detail the effect of various parameters including shear
tab connection, in the stability of 10-storey office building subjected to a corner
compartment fire. Wellman et al. [7] conducted an experimental investigation on
the behaviour of lightweight composite floor assemblies subjected to fire loading.
Selden et al. [8] investigated fire behaviour of composite beams and associated sim-
ple shear beam to column connections. A series of partial composite beams of span
3.65 m were tested. The authors observed that if the beams are overloaded, concrete
compression failure is a potential failure mode for composite beams with flat slabs
exposed to medium temperature levels. Prediction of behaviour of composite beams
during the cooling phase makes it challenging, as the fracture of shear tab connection
observed in the cooling period of loading. Fischer et al. [9–11] studied the behaviour
of different types of the beam to column connections such as shear tab connection,
single angle connection, and double angle connection under fire loading. Composite
beams of span 3.81 m with thick lightweight concrete slab was considered for anal-
ysis. The authors observed that during the cooling phase of loading, in single angle
and double angle connections, prying action develops which permits the contraction
of steel beams, thereby avoiding the brittle failure.
Choe et al. [12] performed an experimental investigation on steel-concrete com-
posite beam with double angle connections. The authors considered 12.3 m long-span
beams (Test 2 and Test 3) subjected to combined structural and thermal loading. The
effect of slab continuity was one of the parameters under consideration. Authors
observed that local buckling of steel beam near the connection leads to an increase
in tensile force demand in continuity bars of the concrete slab.
This paper presents 3D finite element-based modelling approach of composite
beams. The model is validated against the fire tests on composite beams available
in the literature, and the parametric study is conducted under various fire scenarios,
continuity conditions, and amount of reinforcement provided.
2 Modelling Approach
Commercially available finite element-based computer programme ABAQUS [13] is

used for simulating the behaviour of composite beams in fire conditions. The problem
Numerical Modelling of Continuous Composite Beam … 75
is simplified into a sequentially coupled thermo-mechanical problem. The analysis

was conducted in two parts as follows
1. Nonlinear heat transfer analysis performed independently to obtain nodal
temperature histories.
2. Dynamic explicit stress analysis by incorporating nodal temperature histories
from heat transfer analysis.
2.1 Heat Transfer Modelling
Since the thermal and structural models need to be compatible, the same mesh size
is used for both models. The steel beam, concrete slab, insulation materials, rebars,
shear studs, profiled steel deck, and connections were modelled using 8-node linear
heat transfer elements (DC3D8), and a 3D heat transfer analysis is performed on the
entire system. 30 mm mesh size was adopted after performing convergence analysis.
The propagation of the heat along the member depends on thermal properties of
various materials used such as specific heat, conductivity, and thermal expansion,
which are the function of temperature and fire scenario. In all models, the fire sce-
nario is adopted from literature. For heat transfer analysis, a user-defined subroutine
developed by Cedeno et al. [14] used for finding the heat flux on various surfaces of
the model. This subroutine integrates the conduction-based heat transfer within the
solid elements with convection and radiation heat transfer on the surface.
2.2 Stress Analysis Modelling
3D structural modelling developed and 8-node linear brick elements with reduced
integration and hourglass control (C3D8R) were used for the modelling the assembly.
The coupled analysis conducted in two steps in ABAQUS. In the first step, the model
loaded with gravity load, and in the second step, thermal loading is assigned to it. Time
versus temperature response of each node obtained from the heat transfer analysis
is incorporated in the second step as the thermal load. The predefined field option
under loading module in ABAQUS was used for this purpose. Dynamic explicit
analysis procedure chosen for second steps as the analysis is of time-dependent
process and its ability to effectively handle severely nonlinear behaviour. Contact
between surfaces such as steel-concrete, and connection-beam was modelled by
using penalty contact with the coefficient of friction as 0.6 [15]. Shear stud—steel
beam connection was developed by defining tie between the bottom of the stud and
steel beam and connection between concrete and shear stud is developed by defining
embedded constraint. Reinforcement bars and welded wire mesh were embedded in
concrete. Welded connections of the angle with column were represented by giving
Fig. 1 Isometric view of the model (Choe_3)
displacement boundary conditions. Isometric view of the modelling the literature

Choe et al. [12] with shear studs and wire mesh highlighted is shown in Fig. 1.
2.3 Material Properties
As the beam subjected to combined structural and thermal loading, temperature-

dependent mechanical and thermal properties are adopted from Eurocode [16, 17].
Eurocode sketches various mechanical and thermal property models for structural
steel, normal weight concrete, and lightweight concrete. For thermal property assign-
ments, such as conductivity, specific heat, and thermal expansion the relationships
defined by Eurocode were taken directly.
Steel For defining strength, stress-strain relationship and elasticity from Eurocode
have been adopted. For both steel and concrete models, temperature-dependent
strength reduction factors from the code taken care of reduction in strength at elevated
temperature. From the published experimental works, the yield strength of steel taken
as the inputs for Eurocode distinct relationships for mechanical properties. Ductile-
damage model under plastic property, with damage strain 0.2 has been used for
simulating damage in steel. Figures 2, 3, 4, and 5 show the typical Eurocode stress-
strain curve for steel with yield strength 350 MPa, thermal conductivity, thermal
expansion, and specific heat of steel as a function of temperature.
Concrete Figures 6, 7, 8, and 9 show the typical Eurocode stress-strain curve for
concrete with characteristic compressive strength of 28 MPa, thermal conductivity,
600
20
500
100
400 200
Stress, MPa
300
300
400
200 500
600
100
700
0 800
0.000 0.050 0.100 0.150 0.200
Strain
Fig. 2 Typical Eurocode stress-strain curve for steel with yield strength 350 MPa as a function of
temperature
Thermal conductivity, W/mK
60
50
40
30
20
10
0
0 500 1000 1500
Temperature, 0C
Fig. 3 Thermal conductivity of steel as a function of temperature
0.02
0.018
Thermal Expansion
0.016
0.014
0.012
0.01
0.008
0.006
0.004
0.002
0
0 200 400 600 800 1000 1200 1400
Temperature (0C)
Fig. 4 Thermal expansion of steel as a function of temperature

1200
Specific Heat, (J/kgK)

1000
800
600
400
200
0
0 500 1000 1500
Temperature
Fig. 5 Specific heat of steel as a function of temperature
30
20
300
25
400
20 500
Stress, MPa
600
15
700
800
10
900
0
0.000 0.005 0.010 0.015 0.020 0.025 0.030
Strain
Fig. 6 Typical Eurocode stress-strain curve for concrete with characteristic compressive strength
28 MPa
thermal expansion, and specific heat as a function of temperature. Drucker–Prager

plasticity model was used for simulating damage in concrete. Characteristic com-
pressive strength of concrete for various models was taken from experimental data
provided in the literature. Eurocode suggests upper and lower limits for thermal con-
ductivity of concrete, as shown in Fig. 3b. For modelling purpose, the lower limits
were adopted as it was showing better agreement with experimental results. Specific
heat of concrete with 3% moisture content adopted for modelling purpose as shown
in Fig. 3c.
1.6
Thermal conductivity, W/mK

1.4
1.2
0.8
0.6
0.4
0.2
0
0 200 400 600 800 1000 1200
Temperature, 0C
Fig. 7 Thermal conductivity of concrete as a function of temperature
2500
2000
Specific heat, J/kgK
1500
1000
500
0
0 200 400 600 800 1000 1200 1400
Temperature, 0C
Fig. 8 Specific heat of steel as a function of temperature
3 Benchmarking of the Analysis Procedure
Both heat transfer and structural are benchmarked with experimental data available
in the literature. Tests conducted by Wainman and Kirby [1], Zhao and Kruppa [2],
and Choe [12] simulated the model parameters are calibrated accordingly.
0.000025
0.00002
Thermal Expansion 0.000015
0.00001
0.000005
0
0 500 1000
Temperature (0C)
Fig. 9 Thermal expansion of concrete as a function of temperature
3.1 Wainman and Kirby [1]
Test 15 from Wainman and Kirby [1] has been taken and modelled in ABAQUS. Test
15 is a composite beam with a bare steel beam and flat concrete slab. Figure 10 shows
the isometric view of model in ABAQUS. Furnace temperature reported by Wainman
Fig. 10 Isometric view of the model (Wainman Test Specimen 15)

Fig. 11 Comparison of steel 900

bottom flange temperature
(Wainman Test 15) 800
700
Temperature, 0C
600
500
Wainman and
400
Kirbby
300 Predicted
200
100
0
0 10 20 30 40 50
Time, minutes
Fig. 12 Comparison of 250

midspan deflection
Displacement, mm
(Wainman Test 15) 200
150
100 Wainman Test-15

PredicƟon
50
0
0 10 20 30 40 50
Time, minute
and Kirby [1] has been incorporated in the user subroutine to conduct the heat transfer
analysis. The results from the heat transfer analysis are used as the thermal input for
the structural analysis. Figure 11 shows the comparison of bottom flange temperature
of the steel beam. In structural analysis, the mechanical loading applied at four points
gradually and followed by the thermal load. The midspan displacement obtained from
this procedure shown in Fig. 12 indicates good agreement with experimental results.
3.2 Zhao and Kruppa [2]
Test 1 from Zhao and Kruppa [2] is also simulated to validate the model. Test 1
was a composite beam with a protected steel beam and flat slab and is subjected
to gravity load followed by fire. IPE 300 steel beam with 25 mm thick mineral
wool used as fire protection. ISO fire curve was used as furnace temperature; the
same is incorporated in the heat transfer analysis. Figure 13 shows the bottom flange
Fig. 13 Comparison of steel 800

bottom flange temperature
700
(Zhao Test 1)
600
Temperature, 0C
500
400
300
Zhao(Test 1) Predicted
200
100
0
0 20 40 60 80 100 120
Time, minutes
Fig. 14 Comparison of 250

midspan deflection (Zhao
Test 1)
200
Displacement, mm
150
100
Zha (Test 1)
50
Predicted
0
0 20 40 60 80
Time, minutes
temperature distribution of the model compared with experimental data. A total load
of 102.5 kN is applied as gravity load in the first step of structural analysis followed by
thermal analysis. In thermal loading, nodal thermal histories of heat transfer analysis
are used as input in the predefined field. Figure 14 shows the midspan deflection of
the model, which shows a very good correlation with experimental values and Fig. 15
shows a comparison of the deflected shape of the experimental model and numerical
model.
3.3 Choe et al. [12]
Test 3 from Choe et al. [12], Ramesh et al. [18] modelled. It is a 12.8 m long-span
composite beam. W18x35 steel beam with 16 mm thick Southwest Type 5MD, a
cementitious gypsum-based fire protection and 1.83 m wide profiled lightweight
concrete slab with metal deck modelled. Four 12.7 mm diameter reinforced bars of
yield strength 415 MPa used in hogging moment region to control crack width. The
Fig. 15 Comparison of deflected shape of specimen in a experiment and b simulation
reinforcements were extended beyond the slab and tied to the support, to give con-
tinuity to the slab. Figure 16a, b shows the isometric view and cross-sectional view
of the model. Furnace temperature distribution from the literature has been used in
the subroutine to get nodal temperature history. Figure 17a–c shows the comparison
of temperature distribution of steel beam and concrete slab of the model with exper-
imental results. A total load of 106 kN at six equally distributed points have been
applied gradually in the first step and step two, the thermal load applied. Around
Fig. 16 a Isometric view of model, b cross-sectional view of the model (Choe Test 3)
900
900
800 Choe-Test 3
800
700 Prediction
Temperature, 0C
700
Temperature ,0C
600
600
500 Choe_Test-3
500
400 model
400
300 Predicted
300
200
200
100
100
0
0 50 100 150 200 250 300 0
0 100 200 300
Time, minutes Time, minutes
180
160
140
Temperature ,0C
120
100
80 Choe-Test-3
60
40 Prediction
20
0
0 100 200 300
Time, minutes
Fig. 17 a Comparison of bottom flange temperature distribution, b bottom concrete slab

temperature distribution, and c top concrete slab temperature distribution (Choe Test 3)
30 min after commencing the experiment, local buckling of steel beam near to the
connection leads to an increase in tensile force demand in continuity bars of the
concrete slab. Figure 18 shows the comparison of the buckled shape of experimental
test specimen and prediction. Similar behaviour has been observed in the analytical
model also. The rebar was modelled using 3D brick element DC3D8, and slab con-
tinuity in the experiment was governed by anchoring the rebar to a hollow section
Fig. 18 Comparison of buckled shape of experimental test specimen and prediction (Choe Test 3)
800
700
Displacement, mm
600
500
400
Experiment
300
Prediction
200
100
0
0 50 100 150
Time, minutes
50
45
50 40
45
Rebar force, kN
35 Experiment
40
Rebar force, kN
Expperiment- 30 - rebar at
35
rebar at west East
30 25
25 Prediction 20 Prediction
20 15
15
10
10
5 5
0 0
0 50 100 150 0 50 100 150
Time, Minutes Time, Minutes
Fig. 19 Comparison of a midspan deflection versus time, b axial force versus time for rebar at
west, and c rebar at east of experimental test specimen and prediction (Choe Test 3)
connected to supporting column. In the numerical model developed, continuity was

modelled using connectors in interaction module of ABAQUS. Figure 19a–c shows
the comparison of the prediction with the experimental results of bottom midspan
deflection versus time and rebar force versus time. Gravity loading removed when the
midspan deflection exceeds an allowable limit of 620 mm (L/20) at around 65 min.
4 Parametric Study
A parametric study has been conducted to understand the influence of continuity

rebars in the global behaviour of the composite beams in fire conditions. The design
principles of the model have been discussed in Ramesh et al. [18]. In the model,
the rebar and welded wire mesh were designed primarily for crack control. For the
present study, the number of bars provided and their diameters are the parameters
considered. The number of bars, diameter of bars, and spacings considered in this
parametric study were within the limit of codal provisions. The numerical models
listed in Table 1 have been developed in ABAQUS to study the same. CB-1 has
the same property as Choe-3 experimental model. The failure point is taken from
the literature as when the midspan deflection reaches 620 mm (L/20). The midspan
Table 1 Model details used for parametric study

Model Diameter Number Time at Midspan Failure Maximum
notation of rebar of rebars local deflection at time deflection
(mm) provided buckling corresponding (min) under given
of steel time (mm) fire scenario
beam web (mm)
starts
(min)
CB-1 12.7 4 24 150 65 620
CB-2 8 4 24 148 56.3 743
CB-3 16 4 24 152 – 516
CB-4 12.7 2 24 154 57 715
deflection for all these models is shown in Fig. 12. Comparing CB-2 and CB-1, when
the diameter of the bar is reduced from 12.7 to 8 mm, the failure occurs at 56 min
(corresponding to the deflection of L/20), which is 9 min before the beams with
rebar of diameter 12.7 mm. The maximum deflection observed is 743 mm. As the
diameter of rebar is increased from 12.7 to 16 mm in CB-3, the maximum midspan
deflection was 516 mm, which is significantly lower than the limiting value. Whereas
in CB-4 with two number of 12.7 mm diameter rebars, the failure point occurs at
around 57 min. And the maximum deflection observed under the given fire scenario
is 713 mm.
Figure 20 shows the midspan deflection versus time for various models under
consideration. In all four models, at around 24 min after starting the experiment, at
30 °C, when the vertical midspan deflection reaches around 150 mm, local buckling
occurs in the web of steel beam near support and follows by a sudden increment
in the axial tensile force in rebar have been observed. From Table 1, it is clear that
the number of bars and its diameter affects the failure time and maximum deflection
under the given fire scenario as the stiffness of rebars changes with diameter. The
axial tensile force developed in west side and east side rebar located at 335 mm from
Fig. 20 Midspan deflection 800
700
Displacement, mm
600
500
CB-1
400
CB-2
300 CB-3
200 CB-4
100
0
0 20 40 60 80 100 120 140
Time, minutes
80 90
70 80
70
60
Rebar force, kN
CB-1 60 CB-1
Rebar force, kN
50
50 CB-2
40 CB-2
40
CB-3 CB-3
30
30
20 CB-4 CB-4
20
10 10
0 0
-10 0 50 100 150 0 50 100 150
Time, Minutes Time, Minutes
Fig. 21 Axial forces in continuity bars a west side, b east side
the central line of beam is shown in Figs. 21a and 15b. The presence of reinforcement
bars does not have much influence in the beam until initial buckling happens, but it
has a significant effect in maximum deflection value and thereby in failure time.
5 Conclusions
• Sequentially coupled two-step thermal-structural analysis procedure was devel-

oped using FEM tool ABAQUS for predicting the behaviour of steel-concrete
composite beam under combined gravity and fire loading.
• The numerical procedure is validated with various experimental data available in
the literature.
• A parametric study was conducted to understand the influence of continuity rebars
in the behaviour of long-span continuous composite beams under fire.
• It was observed that until the initial local buckling of steel beam web, the rebar
does not have a significant role in the overall fire behaviour of the beam. But
it has a significant effect on the final maximum deflection value. By providing
more quantity of rebar, we can increase the failure time of composite bars. Further
study has to be performed to understand the detailed behaviour of the beam by
increasing the variants.
References
1. Wainman, D. E., & Kirby, B. R. (1987). Compendium of UK standard fire test data, unprotected
structural steel—1. Reference Number RS/RSC/S10328/1/98/B, British Steel Corporation
(now Corus). Rotherham: Swinden Laboratories.
2. Zhao, B., & Kruppa, J. (1997). Fire resistance of composite slabs with profiled steel sheet and
of composite steel concrete beams, Part 2: Composite beams. ECSC—agreement No. 7210 SA
509, CTICM, France.
3. Agarwal, A., Selden, K. L., & Varma, A. H. (2014). Stability behaviour of steel building
structures in fire conditions: Role of composite floor system with shear-tab connections. Journal
of Structural Fire Engineering, 5(2), 77–96. https://doi.org/10.1260/2040-2317.5.2.77.
4. Agarwal, A., & Varma, A. H. (2014). Fire induced progressive collapse of steel building
structures: The role of interior gravity columns. Engineering Structures, 58, 129–140. https://
doi.org/10.1016/j.engstruct.2013.09.020.
5. Agarwal, A. (2011). Stability of steel building structures under fire loading (Ph.D. dissertation).
School of Civil Engineering, Purdue University.
6. Agarwal, A., & Varma, A. H. (2011). Design of steel columns at elevated temperatures due to
fire: Effects of rotational restraints. Engineering Journal, AISC.
7. Wellman, E. I., Varma, A. H., Fike, R., & Kodur, V. (2011). Experimental evaluation of thin
composite floor assemblies under fire loading. Journal of Structural Engineering, 137(9),
1002–1016.
8. Selden, K. L., Fischer, E. C., & Varma, A. H. (2015). Experimental investigation of composite
beams with shear connections subjected to fire loading. Journal of Structural Engineering,
142(2), 04015118. https://doi.org/10.1061/(ASCE)ST.1943-541X.0001381.
9. Fischer, E. C., Selden, K. L., & Varma, A. H. (2017). Experimental evaluation of the fire
performance of simple connections. Journal of Structural Engineering, 143(2), 04016181.
https://doi.org/10.1061/(ASCE)ST.1943-541X.0001664.
10. Fischer, E. C., & Varma, A. H. (2016) Fire resilience of composite beams with simple connec-
tions: Parametric studies and design. Journal of Construction Steel Research, 119–135. doi:
10.1016/j.jcsr.2016.08.004.
11. Fischer, E. C., & Varma, A. H. (2015). Fire behaviour of composite beams with simple con-
nections: Benchmarking of numerical models. Journal of Constructional Steel Research, 111,
112–125. https://doi.org/10.1016/j.jcsr.2015.03.013.
12. Choe, L., Ramesh, S., Seif, M., Hoehler, M., Grosshandler, W., Gross, J., & Bundy, M. (2018).
Fire performance of long-span composite beams with gravity connections. In Proceedings of
the 10th International Conference on Structures in Fire FireSERT, Ulster University, Belfast,
UK, June 6–8, 2018.
13. ABAQUS. (2012). ABAQUS/Standard Version 6.12 User’s Manuals, ABAQUS, Providence,
RI.
14. Cedeno, G., Varma, A. H., & Agarwal, A. (2009). Behaviour of floor systems under realistic
fire loading. In Proceedings on CD-ROM of the ASCE Structures Congress (pp. 1–10). Reston,
VA: ASCE. https://doi.org/10.1061/41031(341)224.
15. Gorst, N. J. S., & Williamson, S. J. (2003). Friction in temporary works. HSE BOOKS. ISBN
0 7176 2613 X http://www.hse.gov.uk/research/rrpdf/rr071.pdf.
16. European Committee for Standardization (CEN). (2005). Eurocode 4: Design of composite steel
and concrete structures, Part 1-2: General rules—Structural fire design. Brussels: European
Committee for Standardization.
17. European Committee for Standardization (CEN). (2002). Eurocode 1: Actions on structures,
Part 1.2: General actions—Actions on structures exposed to fire. European Committee for
Standardization.
18. Ramesh, S., Choe, L., Hoehler, M., Grosshandler, W., & Gross, J. (2018). Design and construc-
tion of long-span composite beam specimens for large structural-fire tests. In Proceedings of
Structures Congress 2018, Fort Worth, Texas, April 19–21, 2018.
Effect of Heating Rate on Bond Behavior
Between Steel and Concrete at Elevated
Temperatures
Ira Banoth and Anil Agarwal
1 Introduction
Bond behavior between steel and concrete at elevated temperatures is a very complex
phenomenon. Bond is the transfer of force between steel and concrete at the interface.
Bond strength is the main influencing parameter for the response of the structure in
terms of loss of structural integrity. The influencing parameters for bond strength are
concrete strength, concrete cover, aggregate type, rebar diameter, rebar size, surface
of rebar, and heating rate.
In the past years, number of studies have been conducted for the bond strength
between steel and concrete at ambient and after elevated temperature by varying dif-
ferent parameters (different concrete strengths, different types of concrete, different
cooling regimes, different concrete cover, the different diameter of rebar). But the
studies for the behavior of the bond during elevated temperatures are very less.
A. D. Edwards et al. studied the local bond stress-slip behavior under cyclic
loading. In this study, the varying parameters were concrete cover (25 and 35 mm)
and different load amplitudes (2, 3.9 and 5.6 MPa). The conclusions drawn from this
study for the bond behavior between 16 mm diameter hot-rolled rebar and concrete of
strength 43 MPa were (a) The destruction of bond strength was more with increasing
load amplitude in the first cycle compared to the prospering cycles, and (b) The
reduction of bond strength is more significant when the stress levels are well below
the peak stress [1].
Morley and Royles studied the bond behavior between steel and concrete at ele-
vated temperatures under different conditions: (a) Under constant stress and (b) Under
I. Banoth (B) · A. Agarwal

Department of Civil Engineering, Indian Institute of Technology Hyderabad (IITH), Hyderabad,
Telangana 502285, India
A. Agarwal
90 I. Banoth and A. Agarwal
no stress, and he concluded that the bond strength was more in under constant stress
condition than the unstressed condition [2].
Several studies have been carried out for the bond behavior between the steel and
fiber-reinforced concrete after elevated temperatures. Fibers are used to reduce the
crack width and spalling effect. Hadad et al. used different types of fibers hooked end
steel, a mixture of hooked end steel and basalt coated steel, a mixture of hooked end
steel and polypropylene and basalt coated steel. The inclusion of fibers prevents the
spalling under the temperature of 600 °C and increased the ductility of bond behavior
between the steel and concrete [3]. Eva Lubloy and Viktor Hlavicka have used hooked
end steel fibers and polypropylene fibers with different types of aggregates (quartz
gravel and expanded clay). The bond strength of concrete with quartz gravel and
steel fibers gives more bond strength than followed by the sequence of quartz gravel,
quartz gravel, and polypropylene fibers, expanded clay 2 and expanded clay 1. The
reduction of bond strength was more significant in the range of temperature 400–
500 °C for all the concrete, and this is due to the decomposition of Portlandite at the
temperature of 450 °C [4].
Ferhat and Rustum gul have considered the parameters were concrete strength
(20 and 35 MPa), embedment length (60, 100, and 120 mm), and cooling regimes
(water and air cooling conditions). The conclusions drawn from this study were
(a) The effect of cooling condition was not having a more significant effect for
the embedment length of rebar is 60 mm and, (b) For 100 and 120 mm embedment
lengths, water-cooled specimens were having more bond strength loss than air-cooled
specimens, and this was more significant for the concrete of strength 20 MPa. The
bond strength was increased with increasing concrete strength and embedment length
[5].
Lee et al. studied the bond behavior between steel and concrete after the elevated
temperature by considering different parameters like the heating rate, cooling con-
ditions, and different types of rebar (epoxy coated and uncoated). Bond strength
decreased with increasing of heating rate for both the cases (epoxy coated and
uncoated rebar). The effect of the cooling condition has a less significant effect
on the uncoated rebar [6].
The studies have been conducted for the bond behavior between steel and concrete
subjected to elevated temperatures are very less. Some researchers have been studied;
Schneider et al. studied the bond behavior of different types of rebars (plain (fresh
as rolled, heavily rusted), cold deformed steel, and pre-stressing steel). They have
concluded that the deformed steel was having more bond strength than the plain bar.
The bond behavior of cold deformed steel and pre-stressing steel was almost similar
because of both the bars having deformations [7]. Morley and Royles have studied the
bond behavior between steel and concrete at elevated temperatures under different
(stressed and unstressed) conditions, and the conclusions drawn were bond strength
was more in stressed conditions than the bond strength in unstressed condition [2].
In the previous studies, the bond behavior between steel and concrete during
elevated temperature is very limited. The effect of heating rate has a very significant
effect on the bond behavior, but the consideration of this parameter is very less
in literature and focuses only on elevated temperatures. So, in this present study,
Effect of Heating Rate on Bond Behavior Between Steel … 91
Table 1 Concrete mix design details (kg/m3 )

Cement Fine aggregate Coarse w/c ratio Super Compressive
aggregate plasticizer stress (MPa)
369.7 828 1173 0.42 4.43 25
considering the heating rate is the main parameter, and how it affects the bond
behavior at elevated temperatures.
There are no specific test methods for finding the bond behavior between steel
and concrete. But according to prEN 10080, there are two types of test methods
available; (a) Pullout test and (b) Beam test [8]. In this present study, pullout tests
were conducted for studying the bond behavior between steel and concrete at room
temperature and elevated temperatures.
2 Experimental Program
2.1 Material Properties
Concrete The compressive strength of concrete was 25 MPa calculated by the aver-
age of compressive strength of three cylindrical specimens of size (150 × 300 mm)
tested in compression. Ordinary Portland cement of grade 53, fine aggregate (river
sand), coarse aggregate (gravel of max size 20 mm), and admixture superplasticizers
were used. The concrete mix was designed according to the code IS 10262:2009
[9]. Mix design details are given in Table 1. Concrete was mixed in a drum mixer
of capacity 0.3 m3 . After mixing, concrete was casted in steel molds and left for
24 hours and then demolded and placed in water tank for curing. After 15 days of
curing, specimens were allowed to dry for 15 days.
Steel Fe500 grade deformed steel of 12 mm and 20 mm diameter rebars were used.
2.2 Heating Procedure
The specimens were heated in an electrical furnace assembled by four heating panels
of dimensions 350 mm wide and 480 mm long. The capacity of the furnace is 1100 °C.
The specimens were heated up to the desired temperatures at the concrete–steel
interface by following a heating rate of 2 °C/min and ISO 834 [10] standard fire
curve. The temperature versus time plots for two different heating rates are shown
in Fig. 1. Temperature versus time for different elevated temperatures by following
different heating rates is presented in Fig. 2.
Fig. 1 Temperature versus time
Fig. 2 Temperature versus time at different elevated temperatures
2.3 Pullout Testing
Pullout tests were carried out for finding the bond behavior between steel and concrete
at ambient and elevated temperatures. Tested specimens were the cubical size of 200
× 200 × 200 mm with a central pullout rebar. The middle portion (4 d mm length)
of the rebar was bonded, and the remaining length top and bottom portion of the
rebar embedded in the concrete was un-bonded by PVC pipes (shown in Fig. 3a).
Nine specimens of 12 mm diameter and twelve specimens of 20 mm diameter were
tested. Three specimens of 12 mm diameter at ambient temperature and remaining
Fig. 3 Details of the test setup: a Geometric details of pullout specimen. b Test setup for ambient
temperature and c Test setup for elevated temperatures
were tested at elevated temperatures. Four specimens of 20 mm diameter at ambient

temperature and remaining at elevated temperatures were tested.
The load was applied after reaching a desired temperature at the steel–concrete
interface by Fatigue Testing Machine of 500 kN capacity. LVDTs were placed on
the concrete surface resting against the actuator as shown in Fig. 3b. Test setup at
ambient and elevated temperatures are shown in Fig. 3b, c.
3 Test Results and Discussion
Figure 4 depicts the bond versus slip diagram for 12 mm diameter rebar at different
elevated temperatures by different heating rates. Three tests were conducted at ambi-
ent temperature for checking the consistency. All the three specimens were failed by
bond. The failure pattern is shown in Fig. 5a. The bond strength at ambient temper-
ature was 20.4 MPa (average of 3 specimens) and at elevated temperatures of 200,
300, 400, and 500 °C were 8.1, 6.3, 3.1, and 3.9 MPa, respectively, by following
the heating rate of 2 °C/min presented in Table 1. At all the temperature, the spec-
imens were failed by bond. At 200 °C temperature, no cracks and at 300 °C small
cracks appeared on the surface of the concrete. At 400 and 500 °C of temperature,
more number of cracks appeared on the concrete surface. Failure patterns at different
elevated temperature are shown in Fig. 5b.
The bond strength of rebar of 12 mm diameter at 100 and 200 °C was 8.7 and
2.1 MPa, respectively, by following heating rate according to ISO 834 standard fire
curve presented in Table 1. At both the temperatures, specimens were failed by bond.
After bond failure, there was a splitting failure formed in the form of longitudinal
cracks along the length of the rebar. The failure patterns are shown in Fig. 5c. The
bond strength decreased with increasing of the temperature by following the heating
Fig. 4 Bond stress versus slip of 12 mm rebar at different elevated temperatures
(a) (b) (c)
Fig. 5 Failure patterns of 12 mm rebars: a At ambient temperature. b Heating rate of 2 °C/min and
c Heating rate of ISO—834
rate of 2 °C/min and ISO standard fire curve, but in the case of ISO heating rate, the
bond strength decreased very rapidly. The results are presented in the Table 2.
Specimen at ambient temperature: DX-AY; here X is diameter of rebar, Y is
number of specimen.
Specimen at elevated temperatures DX-TY-HRZ; X is diameter of rebar, Y is
desired temperature at the steel–concrete interface, and Z is the heating rate in °C/min.
The bond stress versus slip diagram of 20 mm diameter rebar at different elevated
temperatures by different heating rates is presented in Fig. 6. Four specimens were
tested at ambient temperature, and all the specimens were failed by bond, and strength
was 13.6 (average of 4 specimens). The failure patterns are shown in Fig. 7a. The
bond strengths of specimen at elevated temperatures of 200, 300, 400, 500, and
600 °C were 9.8, 6.1, 3.5, 3.2, and 2.1 MPa, respectively, by following heating rate
of 2 °C/min. All the specimens were failed in bond followed by splitting cracks
formed in the longitudinal direction of the rebar, and failure patterns are shown in
Fig. 7b. With increasing temperature, bond strength was decreased, and a number of
cracks were increased on the concrete surface. The specimens heated by following
ISO curve also failed in bond followed by cracks formed in the longitudinal direction
Table 2 Test results of 12 mm rebar

Specimen Bond strength Slip at maximum RBS (% of Type of failure
(MPa) load (mm) 20 °C)
D12-A1 20.3 2.89 – Bond
D12-A2 17.9 2.99 – Bond
D12-A3 23.1 3.67 – Bond
D12-A 20.4 – 100 –
D12-T200-HR2 8.1 3.11 39.6 Bond
D12-T300-HR2 6.3 5.73 30.8 Bond
D12-T400-HR2 3.1 4.65 15.2 Bond
D12-T500-HR2 3.9 2.22 19.1 Bond
D12-T100-HRISO 4.8 4.47 23.5 Bond
D12-T200-HRISO 1.2 4.20 5.9 Bond
Fig. 6 Bond stress versus slip of 20 mm rebars at different elevated temperatures
(a) (b) (c)
Fig. 7 Failure patterns of 20 mm rebars: a At ambient temperature. b Heating rate of 2 °C/min and
c Heating rate of ISO—834
Table 3 Test results of 20 mm rebar

Specimen Bond strength Slip at maximum RBS (% of Type of failure
(MPa) load (mm) 20 °C)
D20-A1 11.1 4.2 – Bond
D20-A2 12.4 4.2 – Bond
D20-A3 15.8 4.5 – Bond
D20-A4 15.2 3.8 – Bond
D20-A 13.6 – 100 –
D20-T200-HR2 9.8 11.5 59.4 Bond
D20-T300-HR2 6.1 4.0 44.8 Bond
D20-T400-HR2 3.5 5.3 25.7 Bond
D20-T500-HR2 3.2 11.2 23.5 Bond
D20-T600-HR2 2.1 2.4 15.4 Bond
D20-T100-HRISO 1.9 9.5 13.9 Bond
D20-T200-HRISO 1.3 3.5 9.5 Bond
D20-T300-HRISO 0.7 2.0 5.0 Bond
of the rebar. The bond strength at different elevated temperatures of 100, 200, and
300 °C was 1.9, 1.3, and 0.7 MPa, respectively as shown in Table 3. With increasing
temperature, bond strength decreased, crack width increased, and number of cracks
appeared on the concrete surface were more.
Specimen at ambient temperature: DX-AY; here X is diameter of rebar, Y is
number of specimen.
Specimen at elevated temperatures DX-TY-HRZ; X is diameter of rebar, Y is
desired temperature at the steel–concrete interface, and Z is the heating rate in °C/min.
3.1 Residual Bond Strength
Percentage of residual bond strength of rebar of different diameters with different

heating rates is shown in Fig. 8. The residual bond strength of 12 mm rebar at different
elevated temperatures of 200, 300, 400, and 500 °C by following heating rate of
2 °C/min was 39.6, 30.8, 15.2, and 19.1%, respectively. By following the heating
rate of ISO curve, the bond strength at elevated temperatures of 100 and 200 °C was
23.5, and 5.9%, respectively. Residual bond strength of 20 mm diameter rebar when
heated at 2 °C/min was 59.4, 44.8, 25.7, and 23.5% at the interface temperatures of
200, 300, 400, 500, and 600 °C, respectively. However, when the specimens were
heated using the ISO standard fire curve, the measured residual bond strengths were
13.9, 9.5, and 5%, at the interface temperatures of 100, 200, and 300 °C, respectively.
The loss in bond strength was more by ISO fire curve. Hence, the heating rate has a
significant effect on the reduction of bond strength.
Fig. 8 Percentage residual bond strength of 12 and 20 mm rebar at different elevated temperatures
by following different heating rates
4 Conclusions
• Pullout tests were carried out for studying the bond behavior between concrete
and steel of different diameter of 12 mm and 20 mm at ambient and elevated
temperatures by following heating rates of 2 °C/min and ISO-834 standard fire
curve.
• Rate of heating has a very significant effect on the bond strength retention factors.
Faster rate of heating leads to a more rapid decline in the bond strength.
• Larger diameter bars underwent a more rapid bond strength decrease with
increasing temperature.
• The residual bond strength of 20 mm rebar was more than the residual bond
strength of 12 mm rebar by following both the heating rates. The reduction of
bond strength was very significant by following heating rate of ISO 834. So,
bond strength decreased with faster heating rate. Hence, the heating rate has a
significant effect on the reduction of bond strength.
References
1. Edwards, A. D., & Yannopoulos, P. J. (1978). Local bond-stress–slip relationships under

repeated loading. Journal Magazine Concrete Research, 30(103), 62–72. https://doi.org/10.
1680/macr.1978.30.103.62.
2. Royles, R., & Morley, P. D. (2009). Further responses of the bond in reinforced concrete to
high temperatures. Journal Magazine Concrete Research, 35(124), 157–163. https://doi.org/
10.1680/macr.1983.35.124.157.
3. Haddad, R. H., Al-Saleh, R. J., & Al-Akhras, N. M. (2008). Effect of elevated temperature
on bond between steel reinforcement and fiber reinforced concrete. Journal Fire Safety, 43(5),
334–343. https://doi.org/10.1016/j.firesaf.2007.11.002.
4. Lubloy, É., & Hlavička, V. (2017). Bond after fire. Journal of Construction and Building
Materials, 132, 210–218. doi:10.1016/j.conbuildmat.2016.11.131.
5. Bingöl, A. F., & Gül, R. (2009). Residual bond strength between steel bars and concrete after
elevated temperatures. Journal Fire Safety, 44(6), 854–859. https://doi.org/10.1016/j.firesaf.
2009.04.001.
6. Lee, J., Sheesley, E., JingY, Xi Y., & Willam, K. (2018). The effect of heating and cooling
on the bond strength between concrete and steel reinforcement bars with and without epoxy
coating. Journal Construction Building Materials, 177, 230–236. https://doi.org/10.1016/j.
conbuildmat.2018.05.128.
7. Diederichs, U., & Schneider, U. (2009). Bond strength at high temperatures. Journal Magazine
Concrete Research, 33(115), 75–84. https://doi.org/10.1680/macr.1981.33.115.75.
8. ENV 10080:1995, Steel for the reinforcement of concrete. Weldable reinforcing steel—Gen-
eral.
9. IS: 10262-2009, Concrete mix proportioning—Guidelines.
10. ISO: 834-11:2014, Fire resistance tests—Elements of building construction—Part 11: Specific
requirements for assessment of fire protection to structural steel elements.
Idealised Bilinear Moment-Curvature
Curves of Reinforced Masonry (RM)
Walls
Jacob Alex Kollerathu
1 Introduction
Most houses in rural India are masonry houses built with either burnt clay brick or
natural stone masonry. Technically, they are called Unreinforced Masonry (URM)
Houses; as they have masonry walls with no steel reinforcement embedded in them
to improve their behaviour during earthquakes. During an earthquake, unreinforced
masonry (URM) walls are pushed sideways, along their length (in-plane) and thick-
ness (out-of-plane) directions [8]. When shaken along their length, they develop
diagonal cracks along their length and/or separate at wall junctions. When walls col-
lapse, they bring down the roof along with them. This is the main reason for large
loss of lives during earthquakes that have occurred in different regions of the coun-
try. The poor performance of URM structures even under low to moderate seismicity
(Sikkim 2011 and Nepal 2015) has seen its use being banned in a few countries (New
Zealand) through techno-legal regulations.
The other variant of masonry structures involves the introduction of both vertical
and horizontal steel. The vertical steel is used to increase the flexural capacity of the
wall, while the introduction of horizontal steel increases the shear capacity of the
wall [1]. Nevertheless, Indian codes [2] have not formulated any design or detailing
code for reinforced masonry possibly as the quality of bricks in our country, on an
average, is not suitable for reinforced masonry. Bricks should be of high strength,
dense and need to have low rate of moisture [3]. Reinforced masonry (RM) struc-
tural walls form the main lateral load resisting system in low- to mid-rise buildings,
located in moderate/high seismic regions. Compared to unreinforced masonry walls,
RM walls have higher flexural and shear strengths and are expected to efficiently
resist earthquake shaking through inelastic actions. Proper design and detailing of
J. A. Kollerathu (B)
Department of Civil Engineering, Faculty of Engineering, Christ (Deemed to Be University),
Bengaluru, India
100 J. A. Kollerathu
Fig. 1 Detailing of longitudinal steel in RM walls
longitudinal and transverse reinforcements helps in achieving the required strength

and ductility. Two common ways of detailing longitudinal reinforcement in RC walls
are:
(a) Uniform distribution of reinforcement along the length of the wall, and (b)
reinforcement concentrated at the two ends of the wall (see Fig. 1a, b).
Although, in-plane flexural strength of walls with the two distributions is gener-
ally not too different for the same amount of total reinforcement, the former often
results in enhanced shear strength and improved shear behaviour (displacement and
ductility) as it resists the propagation of diagonal shear cracks in the wall [1]. The
flexural strength and curvature ductility of RM walls sections can be determined from
their axial load–moment interaction curves and nonlinear moment-curvature (M-)
curves. The M- response curve must represent the effective (cracked) flexural rigid-
ity, flexural strength and curvature ductility of the RM section. This paper proposes
methods to arrive at idealised multi-linear M- curves of RM wall sections with both
uniformly distributed reinforcement along the length of the wall and reinforcement
concentrated at the two ends of the wall, at different levels of axial stress.
2 Estimation of Axial Load, Flexural Capacity

and Curvature
The axial load and the flexural strength of RM walls can be estimated from basic
principles of mechanics considering equilibrium of forces, compatibility of strains
and constitutive relations of materials, given in Eqs. (1)–(3):
Σ( f sci − f msci )Asci + f m,avg t xu − Σ f sti Asti = P (1)
εm1 εst4
Compatibility conditions = (2)
xu L − xu
The design flexural compressive σ -ε curve of masonry is parabolic up to an

ultimate strain of 0.003 with maximum compressive stress f m . Also, strain limits εu
of 0.003 used are from experimental observations [4]. The constitutive relation of
masonry is defined by its design stress–strain curve given by:
Idealised Bilinear Moment-Curvature Curves … 101

ε ε 2
f m = 0.85 f m 2 − (3)
εu εu
where Asti is the area of reinforcement bars in ith layer under tension,
Asci is the area of reinforcement bars in ith layer under compression,
L is the distance of the extreme layer of reinforcement from the extreme layer in
compression, f m,avg is the average compressive stress in masonry,
f sti is the stress in ith layer of reinforcement bars under tension,
f sci is the stress in ith layer of reinforcement bars under compression (both
estimated from stress–strain characteristics of reinforcement bar),
f msci is the stress in masonry at the level of ith layer of reinforcement bars under
compression,
x u is the depth of neutral axis,
f m is the compressive strength of masonry,
εu is the compressive strain in masonry corresponding to f m ,
and ε is the ultimate strain in masonry at highly compressed edge at peak stress
(Fig. 2).
The limiting strain corresponding to yielding of reinforcement bars is given by:

0.87 f y
ε y = 0.002 + (4)
Es
where
Es : Youngs’ modulus of reinforcing steel (200 GPa)
Fig. 2 Typical stress and strain distributions across rectangular RM walls. a Entire section under
compression and b once tensile cracking commences
Once the stresses in the reinforcing steel and masonry are obtained, the axial load
is calculated as in Eq. (1). The flexural strength of the RM wall is calculated by con-
sidering the moments of the tensile and compressive forces along the centroidal axis.
Figure 4a shows a typical axial load–bending moment (P–M) interaction (normalised
to their respective capacities) of RM walls. The failure of a RM wall is defined by
values of the following limiting states of strains in masonry and the reinforcing steel.
• Tensile cracking of masonry
• Yielding of extreme of reinforcement on tension side
• Crushing of masonry
Based on the above-mentioned limit states, a P–M interaction envelope of RM
sections has three distinct regions (see Figs. 3 and 4);
• Balance point: It is defined as the point on the P–M interaction at which the strains
in concrete and steel reach corresponding limiting strains in crushing and tension
simultaneously. The axial load at this point is denoted as balanced axial load Pbal .
• Compression failure region (above balanced point): Region on the P–M interac-
tion where the failure is characterised by the strains in the extreme compressive
fibre reaching the crushing strain of masonry (εu ). The axial load in the region is
greater than Pbal .
• Tensile failure region (below balance point): It is the region on the P–M interaction
where the failure is due to the tensile strain in the extreme layer of reinforcement
reaching its yield. The axial load in this region is lesser than the balanced axial
load Pbal .
In the following paper, P–M interaction and nonlinear M- curves are developed
for two reinforced masonry walls (wall A with uniformly distributed steel and wall
B with reinforced concentrated at the two ends of the walls). The curvature ductility
of the walls has been examined at various levels of axial stress.
1
P>Pbal
0.8 P<Pbal
Normalised axial load
0.6
0.4
0.2
0
0 0.2 0.4 0.6 0.8 1
Normalised flexural strength
Fig. 3 Typical axial load–bending moment interaction of RM walls

t A B C
ε m1= ε u ε m1= ε u ε m1< ε u
εcs1 εcs1 εcs1
xu
xu εcs2 xu εcs2 εcs2
L φ
εts3 εts3 εts3
εts4 < εy εts4 = εy εts4 = εy
Net
tension
Fig. 4 Strain variations across RM sections under axial loads. a Above balance point, b at balance
point, c below axial load
3 Axial Load–Flexural Capacity (P–M) Interaction

and Moment-Curvature (M-)
The P–M interaction and M- curves of two reinforced masonry walls with rein-
forcing steel detailing as given in Fig. 1 is provided in these sections. The curves are
determined from simple principle of mechanics considering equilibrium of forces,
compatibility of strains and constitutive relations of materials. As per SP-7, the min-
imum reinforcement in each orthogonal should be at least 0.07% of the gross area of
the wall. In this paper, the P–M interaction and M- curves of a wall with dimensions
of 3270 * 3000 * 200 mm (l * h * t) are examined. The reinforcing steel is detailed
in two ways:
(a) Uniform distribution of reinforcement along the length of the wall (wall A), and
(b) Reinforcement concentrated at the two ends of the wall (wall B).
The P–M interaction of walls A and B is shown in Fig. 5. From Fig. 5 and Table 1,
one observes that the detailing of the reinforcement does not have an effect on the
axial load capacity of the wall. However, there is a 7.5% increase in the flexural
strength of the wall with reinforcement concentrated at the corner. The M- curve
of walls A and B subject to different levels of axial stresses is also investigated. In
this regard, M- curves of the walls at an axial load of 0.21 Pu (P = Pbal ), 0.4 Pu
(P > Pbal ) and 0.1 Pu (P < Pbal ) are studied and reported in Fig. 6 and Table 2. The
idealised bilinear curve is obtained by equating the energy dissipation capacity of
Fig. 5 Axial load–bending moment interaction diagram for the wall A and wall B
Table 1 Axial load and bending moment capacity of walls A and B

Wall # Axial load capacity, Pu Bending moment Balanced axial load, Pbal
(kN) capacity, M u (kN m) (kN)
Wall A 5593 1717 1174
Wall B 5589 1845 1173
Φy Φu
Fig. 6 Moment-curvature relationship of wall A
Table 2 Curvature ductility and effective flexural rigidity of walls A and B

Axial load Curvature Curvature Effective Effective
demand ductility, μ ductility, μ flexural rigidity, flexural rigidity,
(wall A) (wall B) EIeff (wall A) EIeff (wall B)
0.4 Pu 1.83 1.69 0.71 0.75
0.21 Pu 4.57 4.38 0.86 0.88
0.1 Pu 12.3 12.84 0.92 0.91
Φy Φu
Fig. 7 Moment-curvature relationship of wall B
the nonlinear and idealised bilinear M- curve. So the area below the nonlinear and
idealised bilinear M- curve must be the same. The moment capacity in the bilinear
M- curve does not imply the design value. An estimate of curvature ductility (μϕ )
and effective flexural rigidity (defined as the initial slope of the bilinear curve) of
walls A and B are studied and reported in Table 2. The flexural rigidity is expressed
as a function of the un-cracked flexural rigidity. The afore-mentioned parameters are
estimated at axial load demands of 0.4 Pu , 0.21 Pu and 0.1 Pu . Results from Fig. 5,
6 and Table 2 indicate that
• Curvature ductility of walls A and B subject to high axial loads (>Pbal ) is low since
their failure is characterised by the compressive failure (crushing) of masonry
rather than yielding of steel.
• A drop in the effective flexural rigidities is also observed at higher axial loads
since the failure of walls at high axial is characterised by crushing of masonry
rather than yielding of steel.
• Results also indicate that at medium to low axial loads, there is negligible effect
of the reinforcement detailing on either the curvature ductility or the effective
flexural rigidity (Fig. 7).
4 Idealised Moment-Curvature Curves
Analytical research by Sunitha et al. [5–7]) developed simple formulae based on

limiting states of strain to obtain idealised bilinear M- curve of RC sections. In
this section, simple formulae are proposed to determine idealised multi-linear M-
curves. The points on the M- curve are developed using point corresponding to
limit states of strain in masonry and steel (as described in Sect. 2). Hence, simple
formulae are proposed using the above limit states to develop idealised M- curves
of RM sections.
4.1 RM Walls with Axial Load, P < Pbal
The origin (0, 0) of the idealised bilinear M- curve of RM section in the presence
of axial compressive load corresponds to the case of uniform strain variation across
depth εm of the section, at that level of axial load. Point 1 on the M- curve away
from the origin is identified as the strain at which masonry cracks. Since masonry is
a material with negligible tensile strength, a small increment has to be provided at
the extreme compressed fibre to cause a strain gradient along the length of the wall.
In this study, 0.00005 is considered as the strain increment that gives a reasonably
accurate estimate of point 1 at all levels of axial loads. Strain at the opposite edge
is then estimated using principle of mechanics, discussed earlier. As described in
Fig. 8, point 2 corresponds to yielding of extreme layer of reinforcing steel in tension,
obtained only for axial loads P ≤ Pbal . Point 3 corresponds to compression failure
limit state, while point 4 represents yield point of the M- curve. The coordinates
of point 4 are determined as per Eqs. (5A)–(5C):

M 3 φ2 − M 2 φ3
φ4 = φ1 (5A)
φ1 (M3 − M2 ) − M1 (Φ3 − φ2 )
φ4
M4 = M1 (5B)
φ1
(Φ ,M ) 3
2 2 2 ε m = εu
4 (Φ4,M4) εm < εu
(Φ3,M3)
εsti> εy
εsti= εy
εm < εu
(Φ1,M1)
1
εt= εcr
Fig. 8 Proposed multi-linear moment-curvature relationship of walls with low axial load
Table 3 Comparison of curvature ductility and effective flexural rigidity

Wall A Wall B
Parameter Approach 1 Approach 2 Approach 1 Approach 2
Curvature ductility (μ ) 12.47 12.31 12.84 12.84
Effective flexural rigidity 0.921 0.921 0.911 0.911
Fig. 9 Comparison of M- curves (P < Pbal )
The curvature ductility is obtained as:

ϕ3
μϕ = (5C)
ϕ4
It is observed that the flexural strength of the RM wall obtained from the actual
nonlinear M- curve compares well with the flexural strength from Eqs. (5A) and
(5B). The curvature ductility and effective flexural rigidities of walls A and B are
compared with the respective values obtained in Sect. 3. Henceforth, the M- curve
obtained using the proposed method will be called Approach 1 and the M- curve
obtained using equivalence of area will be called Approach 2. Idealised M- curve
obtained from Approach 1 and Approach 2 of wall A and wall B is compared with
their respective nonlinear responses and is reported in Table 3 and Fig. 9. From the
results obtained, it is observed that both curvature ductility and effective flexural
rigidity obtained from Approach 1 compare very well with Approach 2.
4.2 RM Walls with Axial Load, P = Pbal
It is observed that in RM walls subject to an axial load that is equal to Pbal all the
limit states in masonry and reinforcing steel are mobilised and results presented in
Fig. 10 and Table 4 indicate that the equations prescribed in (5A) and (5B) hold good
even for RM with an axial load P = Pbal .
Fig. 10 Comparison of M- curves (P = Pbal )

Wall A Wall B
4.3 RM Walls with Axial Load, P > Pbal
In RM walls subject to very high axial loads, all the limit states may not be mobilised
(especially yielding of steel). So only three points are required to develop the idealised
M- curve that corresponds to limit states in masonry in cracking (point 1) and
crushing (point 3). The coordinates of the yield point are obtained from the following
formula:
M3
Φ2 = Φ1 (6A)
M1
M2 = M3 (6B)
And curvature ductility is

ϕ3
μϕ = (6C)
ϕ2
If the limit state due to cracking of masonry is not mobilised (observed at axial
loads close to and greater than 0.7 Pu ), then point 1 is identified at a small strain
increment εm at the highly compressed edge. The authors would like to point out
that Eqs. (6A) and (6B) are identical to proposed equations in Sunitha et al. [6,
7]. The idealised bilinear curve obtained from Eqs. (6A) and (6B) is compared to
the nonlinear response of wall (see Fig. 11). One observes that the proposed model
gives a good estimate of the initial elastic behaviour and the flexural strength of the
wall. The curvature ductility and flexural rigidity of wall A and wall B obtained from
Eqs. (6A)–(6C) are compared with values obtained in Sect. 3, and results are reported
in Fig. 12 and Table 5. From results obtained, it is observed that both the flexural
rigidities and curvature ductility obtained from the proposed equation compare well
with existing methods.
(Φ2,M2) 3
2 ε m = εu
(Φ3,M3)
1
εm < εu
(Φ1,M1)
εsti< εy
εt= εcr
Fig. 11 Proposed multi-linear moment-curvature relationship of walls with high axial load
Fig. 12 Comparison of M- curves (P > Pbal )


Wall A Wall B
5 Conclusions
The significant conclusions of this study are enumerated below:

• Simple hand-calculation-based multi-linear idealisation of actual nonlinear M-
curves proposed in this paper significantly reduces the computational time in
determining the initial flexural rigidity, moment capacity and curvature ductility
of RM sections subject to different levels of axial load without compromising on
its accuracy. The equations proposed in this paper serve as a powerful yet simple
to use tool to determine the multi-linear moment-curvature relation of reinforced
masonry walls.
• Results from this paper indicate that the method of detailing of flexural rein-
forcement in a masonry wall affects neither the curvature ductility nor the initial
flexural rigidity.
References
1. Gandhi, R., & Menon, A. (2014). Seismic performance of bed-joint reinforced solid brick
masonry walls. In Proceedings of the 9th International Masonry Conference, Guimaraes,
Portugal.
2. Indian Standard 1905 (1987), “Code of practice for structural use of unreinforced masonry”,
Bureau of Indian Standards, New Delhi.
3. Indian Standards-SP 7. (2016). National building code, Part 6: structural design—Section 4:
Masonry. Bureau of Indian Standards.
4. Kaushik, H. B., Rai, D. C., Jain, S. K. (2007). Uniaxial compressive stress–strain model for clay
brick masonry. Current Science, 92(4), 497–501.
5. Sunitha, P., Goswami, R., Murthy, C. V. R. (2015). Idealized bilinear moment-curvature curves
of slender rectangular RC wall sections. Journal of Seismology and Earthquake Engineering,
17(3).
6. Sunitha, P., Goswami, R., Murthy, C. V. R. (2016). Seismic behaviour of RC moment resisting
frame buildings designed and detailed as per first revision of IS 13920 draft provisions. The
Indian Concrete Journal, 90.
7. Sunitha, P., Goswami, R., Murthy, C. V. R. (2016). Idealized moment-curvature curves of RC
section for the pushover analysis of R.C. frame elements. The Indian Concrete Journal, 90.
8. Tomaževič, M. (1999). Earthquake-resistant design of masonry structures (pp. 159–173).
Imperial College Press.
Performance of Self-flowing Concrete
Incorporated with Alumina Silicates
Subjected to Elevated Temperature
Manish S. Dharek, Prashant Sunagar, K. Harish, K. S. Sreekeshava,

S. U. Naveen, and Bhanutej
1 Introduction
Concrete is one of the major building materials. With the accumulative need for
housing both in rural and urban areas, the demand for building resources has been
endlessly growing. The construction materials are produced by using natural wealth
thus affects the surroundings by exhausting natural properties, using energy, and
discharging toxins to the land and water. So in order to overcome from these problems,
the alternative building materials for conventional concrete are used incorporating
mineral admixtures, alumina silicates, etc.
Davidotis [1] in the year 1978 came up with a construction material as an green
alternative for cement which was innovative and ecofriendly material which is some-
times referred as “geo-polymer.” This was made by blending primarily silicon and
aluminum materials such as fly ash, silica fume, rice husk ash. This process was car-
ried under high alkaline conditions on Si-Al minerals, yielding polymeric Si-O-Al-O
bonds in amorphous form.
The fire-resisting performance of concrete when compared to other structural
materials such as wood and steel is by far superior. The main reason for the concrete
to be fire resistant is due to non-combustibility and its capacity to withstand higher
M. S. Dharek (B) · S. U. Naveen · Bhanutej

Department of Civil Engineering, BMS Institute of Technology and Management, Bengaluru,
India
P. Sunagar
Department of Civil Engineering, Ramaiah Institute of Technology, Bengaluru, India
K. Harish
Arvind Infrastructures Ltd., Bengaluru, India
K. S. Sreekeshava
Department of Civil Engineering, Jyothy Institute of Technology, Bengaluru, India

112 M. S. Dharek et al.
temperature gradients. It has been observed that the properties of concrete do not get
affected when the temperature is less than 100 °C. However, when the temperature
gradually goes on increasing beyond 100 °C, slight changes tend to occur inside the
concrete. When the temperature exceeds more than 300 °C, contraction of cement
paste and reduction in its compressive strength are observed. When the temperature
exceeds 500 °C, the bond between the cement paste and the aggregate tends to get
weaker in addition to rapid contraction of cement paste due to rapid loss of water
from the concrete resulting in expansion of concrete and strength degradation
Abdullah [2] studied the impact of molarities of NaOH solutions on the compres-
sive strength of fly ash-based geo-polymer paste. The samples which were calcined at
high temperature showed improved mechanical properties, low porosity, higher den-
sity, low water absorption when compared to samples that were calcined at low tem-
peratures. Dharek et al. [3] studied the impact of varying molarity of NaOH solution
on the workability and mechanical properties of self-flowing concrete incorporat-
ing alumina silicates. It was concluded that chemical admixtures such as high range
water reducers and viscosity-modifying admixtures played a vital role in improving
workability and the compressive strength also improved with increase in molarity
of NaOH solution. Mathew et al. [4] reported increase in compressive strength of
geo-polymer concrete samples which were oven cured at a temperature of 100 °C
and the compressive strength of the samples reduced when the temperature levels
were increased beyond 100 °C. The total aggregate content played a vital role in
increase of tensile strength of geo-polymer concrete. Ushaa et al. [5] studied the
impact of various mineral admixtures by varying the molarity of NaOH solution in
geo-polymer concrete. The GGBS-based geo-polymer mixes showed better worka-
bility properties when compared to mixes made with other mineral admixtures. The
mixes containing silica fume showed superior mechanical properties when compared
to mixes containing other mineral admixtures.
2 Materials and Methods
The program consisted of casting, curing and testing concrete cubes for compressive
strength subjected to elevated temperature. (200, 400, 600 °C) by varying the molarity
(6, 8, 10 M) maintained for a period of 60 min and air-cooled. The materials used
for investigation included GGBS, fly ash, alkaline liquids, manufactured sand. The
trial mix design for four proportions (fly ash: GGBS)—0:100, 10:90, 20:80, 30:70—
was carried out. The mix design was carried out as per the method proposed by
Rangan and Lloyd [6]. Since there was workability-related issues with 20:80 (fly
ash: GGBS) proportion, to arrive at the optimum dosage of superplasticizer, Marsh
cone test was carried out for the 30:70 proportion of fly ash to GGBS ratio. The
details of the same are shown in Fig. 1a, b. The superplasticizer dosage of 0.4% was
chosen based on trials from Marsh cone test. Initially, trials were conducted for 30:70
proportion with superplasticizer and later to obtain a mixing trials were conducted
using Viscosity modifying agents. The mix proportion (30:70) with admixture dosage
Performance of Self-flowing Concrete Incorporated … 113
Fig. 1 a Marsh cone test. b Variation of SP dosage (percentage) versus time (seconds)
of superplasticizer weight of cementitious material (0.4%) and viscosity-modifying

agent (0.1%) was chosen as it satisfied the workability criteria required for self-
flowing concrete. The mix proportion details are shown in Table 1. The materials used
in the present study were acquired locally from in and around Bengaluru except coarse
aggregate which was procured from Bijapur, India, and the engineering properties
were studied as per pertinent Indian Standards. The details of the same are given
below
Table 1 Details of the mix proportion adopted

Sl. no. Materials Quantity
1 Coarse aggregate—10 mm downsize 840/m3
2 M-Sand 840/m3
3 Na2 SiO3 160/m3
4 NaOH 64 g/m3
5 GGBS 348/m3
6 Fly ash 149/m3
7 Water 125 kg/m3 (25% of cementitious
materials)
8 Superplasticizer—PCE based 0.4% of cementitious material (GGBS +
fly ash)
9 Viscosity-modifying agent 0.1% of cementitious material (GGBS +
fly ash)
2.1 Fly Ash
Fly ash was obtained from RTPC, Raichur, for the present study. The fly ash obtained
belonged to low calcium F-grade conforming to IS 3812-1981 requirements [7].
2.2 GGBS
This was acquired from JSW Steel Plant Bellary, India, conforming to IS 16714-2018
requirements [8].
The color was found to be whitish.
2.3 Fine Aggregate
M-Sand was used as a fine aggregate obtained from an M-Sand plant in Chikka-
ballapur, India. The M-Sand was found to be clean and dust-free conforming to IS
383-2016 [9] requirements. The specific gravity and fineness modulus were found
to be 2.52 and 3.15, respectively. The M-Sand conformed to Zone II.
2.4 Coarse Aggregate
Coarse aggregate used for the present study was crushed basalt stone obtained from
Bijapur, India. The coarse aggregate was conforming to IS 383-2016 [9] require-
ments. The specific gravity and fineness modulus were found to be 2.83 and 7.4,
respectively.
2.5 Alkaline Liquids
Alkaline liquids used in the present study comprised of combination of sodium

hydroxide (NaOH) and sodium silicates (Na2 SiO3 ). The NaOH solution was prepared
for different molarities of 6, 8, 10 M. The solution was prepared by dissolving the
pellets in calculated amount of water. The sodium silicate solution was obtained
from a local authorized manufacturer in Bengaluru. Sodium silicate was available in
semisolid form.
2.6 Water
Water for the present study used was potable type conforming to requirements
specified in IS-456-2000 [13].
2.7 Chemical Admixtures
In the present study, since high workability was the requirement, poly carboxylic
ether based new age super plasticizer (PCE) was used. The dosage of the same was
arrived by conducting trials using Marsh cone test as shown in Fig. 1b.
Since self-compaction was the requirement as part of the study, in addition to
superplasticizer, viscosity-modifying agents were also used to modify the viscosity of
concrete incorporating alumina silicates and to reduce the segregation. The chemical
admixtures conformed to IS 9103-1999 [10].
3 Details of Casting and Experimentation
Sodium hydroxide solution was prepared 24 h before casting. The fine aggregates,
coarse aggregates, fly ash and GGBS were weighed as per the proportions obtained
through the mix proportion as per Rangan and Lloyd [6] given in Table 1. The
workability tests were conducted as shown in Table 2. The cube specimens were
kept in hot air oven at 60 °C for duration of 24 h. The specimens were taken out of
the oven after 24 h and demoulded. The specimen was then kept for air curing for
28 days.
Table 2 Results of workability tests for varying molarity of self-compacted geo-polymer concrete
Sl. no. Molarity (M) Slump flow (mm) V-funnel test (s) L-box test
1 6 600 10.31 0.81
2 8 640 9.85 0.83
3 10 650 9.92 0.82
Range as per EFNARC >600 mm 8–12 s >0.8
specifications [12]
Fig. 2 Details of slump flow, V-funnel and L-box test
3.1 Workability Tests
Tests such as slump flow, L-box, V-funnel test [12] were conducted for the samples
by varying the molarities of NaOH, viz. 6, 8, 10 M with 30:70 proportion The details
of the same are in Table 2 and Fig. 2.
3.2 Self-flowing Concrete Incorporating Alumina Silicates

Subjected to Elevated Temperature
The cubes were air-cured for duration of 28 days under ambient temperature. Later,
the cubes tested in a heavy-duty oven as shown in Fig. 3 for different temperature
ranges (200, 400, 600 °C) for a duration of 1 h after the temperature levels of 200, 400,
600 °C as retention period were obtained. After the retention period, the specimens
were removed and kept for air cooling to be brought to ambient temperature before
the samples were kept inside the heavy-duty oven as shown in Fig. 3, and the mass of
the cubes was noted down. After bringing the cubes to ambient temperature, again the
Fig. 3 Heavy-duty ovens used for subjecting the cubes to elevated temperature
mass of the cubes was noted down to get the percentage loss in mass after subjecting
the self-compacted geo-polymer concrete cubes to elevated temperature.
The percentage weight loss and compressive strength of cubes oven cured for 24 h
and subsequently air-cured subjected to elevated temperature ranges for a retention
period of 1 h after the temperature levels of 200, 400, 600 °C is shown in Tables 3,
4 and 5, respectively, for variation of molarity (6, 8, 10 M). Figure 4 shows the cube
samples after subjected to elevated temperature, respectively.
The percentage mass loss increases with increase in temperature which can be
clearly seen in Fig. 5. This value is higher for samples subjected to temperature of
400, 600 °C when compared to samples subjected to temperature of 200 °C. An
Table 3 Mass loss and compressive strength of self-flowing concrete incorporating alumina
silicates subjected to 200 °C temperature
Sl. no. Molarity (M) Specimen no. 1 2 3 Average
1 6 Mass loss (kg) 0.03 0.04 0.03 0.03
Compressive strength 50.0 39.6 49.6 46.4
(MPa)
2 8 Mass loss (kg) 0.02 0.03 0.03 0.04
(MPa)
3 10 Mass loss (kg) 0.02 0.03 0.02 0.03
(MPa)
Table 4 Mass loss and compressive strength of concrete incorporating alumina silicates subjected
to 400 °C temperature self-flowing
1 6 Mass loss (kg) 0.18 0.21 0.15 0.18
(MPa)
2 8 Mass loss (kg) 0.24 0.21 0.21 0.22
(MPa)
3 10 Mass loss (kg) 0.21 0.21 0.23 0.21
(MPa)
Table 5 Mass loss and compressive strength of self-flowing concrete incorporating alumina
silicates subjected to 600 °C
1 6 Mass Loss (kg) 0.22 0.38 0.39 0.33
Compressive Strength 62.0 58.5 57.6 59.3
(MPa)
2 8 Mass Loss (kg) 0.24 0.25 0.27 0.25
(MPa)
3 10 Mass Loss (kg) 0.23 0.27 0.29 0.26
(MPa)
Fig. 4 Concrete incorporated with alumina silicates samples subjected to elevated temperature
increase in average compressive strength can also be observed with the increase in
molarity as shown in Fig. 6.
5 Conclusions
The following conclusions are drawn based on the experimental investigations on

self-flowing concrete incorporated with alumina silicates.
• The percentage mass loss in samples increased with increase in temperature. The
reason attributed to this can be due to evaporations of calcium alumina silicate gel
and sodium aluminosilicate hydrate as noticed by Rao et al. [11]. This is shown
in Fig. 5. The percentage mass loss is more in the case of samples with 10 M.
Fig. 5 Variation of percentage mass loss of specimens subjected to elevated temperature
Fig. 6 Variation of compressive strength of specimens subjected to elevated temperature
• The increase in concentration of NaOH leads to an increase in compressive

strength of samples subjected to higher temperature indicating the importance
of molarity.
References
1. Davidovits, J. (2013). Geopolymer cement a review. In Geopolymer science and technics.

Saint-Quentin, France: Technical Paper #21, Geopolymer Institute Library.
2. Abdullah, L., Al Bakri, A. M. M., Kamarudin, H., Ruzaidi, C. M., Yahya, Z., Salleh, M. A. A.,
et al. (2014). Compressive strength and morphology of fly ash based geopolymer as artificial
aggregate with different curing temperature. Key Engineering Materials, 594–595, 151–155.
3. Dharek, M. S., Sunagar, P., Bhanu Tej, K. V., & Naveen, S. U. (2019). Fresh and hardened
properties of self-consolidating concrete incorporating alumina silicates. In B. Das & Neithalath
N (Eds.), Sustainable construction and building materials. Lecture notes in civil engineering
(Vol 25). Singapore: Springer.
4. Mathew, B. J., Sudhakar, M., & Natarajan C. (2013). Strength, economic and sustainabil-
ity characteristics of Coal ash-GGBS based geopolymer concrete. International Journal of
Computational Engineering Research, 3(1).
5. Ushaa, T. G., Anuradha, R., & Venkatasubramani, G. S. (2015). Performance of self-compacting
geopolymer concrete containing different mineral admixtures. Indian Journal of Engineering
& Materials Sciences, 22, 473–481.
6. Rangan, B. V., & Lloyd, N. A. (2010). Geopolymer concrete with fly ash. In Second
International Conference on Sustainable Construction Materials and Technologies, Italy.
7. IS 3812-1981. Specification for fly ash for use as pozzolana and admixture. New Delhi: Bureau
of Indian Standards.
8. IS 16714-2018. GGBS for use in cement, mortar and concrete-specifications. New Delhi:
Bureau of Indian Standards.
9. IS 383-2016. Coarse and fine aggregate for concrete—code of practice. New Delhi: Bureau of
Indian Standards.
10. IS 9103-1999. Concrete admixtures-specifications. New Delhi: Bureau of Indian Standards.
11. Mallikarjuna Rao, G., Gunneswara Rao, T. D., Siva Nagi Reddy, M., & Rama Seshu, D. (2019).
A study on the strength and performance of geopolymer concrete subjected to elevated tem-
peratures. In A. Rao, & K. Ramanjaneyulu (Eds.), Recent advances in structural engineering,
Volume 1. Lecture notes in civil engineering (Vol. 11). Singapore: Springer.
12. EFNARC. (2005). The European guidelines for self compacting concrete, specification,
production and use.
13. IS 456-2000. Plain and reinforced concrete—Code of practice. New Delhi: Bureau of Indian
Standards.
Response of RC Short Column Under
Combined Fire and Axial Loading
Chinthapalli Hemanth Kumar, B. Subash, and Anil Agarwal
1 Introduction
Columns are the main structural members that transfer loads of superstructure to
the foundation. RC structures are one of the majorly adopted construction practices.
Numerous characteristics of an RC column affect the overall strength of the members:
volume of longitudinal and confining reinforcement, length, load eccentricity, cross-
section area, end connections, type of concrete, and strength of the steel.
Mander et al. [1] studied the behavior of RC columns of different cross-section like
rectangular, square, and circular. For all section types, the most important parameter
influencing the stress–strain curve of confined concrete was the volume of confining
reinforcement. Confining reinforcement for square or rectangular columns may be
in the form of rectangular hoops or cross ties, spirals for circular columns. The
volumetric ratio of confining reinforcement increases the strength and decreases
(flattened) the slope of the falling branch. The circular, spiral-confined columns
performed better than rectangular or square columns. This results in both the strength
and the ultimate compression strain enhancement for a given volumetric confinement
ratio.
Cusson and Paultre [2] studied the behavior of high-strength-concrete (HSC)
columns confined by rectangular ties under axial loading. The lateral reinforcement
ratio increases from 2.0 to 2.8% showing 10% increase in the strength. Larger con-
fining pressure due to smaller tie spacing results in better confinement efficiency on
the concrete core. An increase in the strength of the confining reinforcement results
C. Hemanth Kumar (B) · A. Agarwal

B. Subash
Department of Civil Engineering, SASTRA Deemed to be University, Thanjavur, India

122 C. Hemanth Kumar et al.
increase in strength and toughness only for well-confined specimens with a large vol-
ume of confining reinforcement ratios. Some interactions between other variables
were noticed. Therefore, when evaluating the influence of an individual variable on
confined concrete behavior, this should be viewed as only a partial effect. Jun-Ichi
et al. [3] also observed that the columns with a high amount of lateral ties perform
better than the columns with a small number of lateral ties at ambient temperature.
Different parameters like column length, cross-section, longitudinal reinforcement
that influence the fire performance of RC Columns are studied by several researchers
[4–11].
The behavior of RC columns under fire conditions are different from ambient con-
ditions which result in thermal stresses and spalling of the concrete. Few researchers
such as Kodur et al. [10, 12] made comparitive study of the fire resistance perfor-
mance of high-strength-concrete (HSC) columns with that of normal-strength con-
crete (NSC) and fiber-reinforced HSC columns. The observations are different type
of fibers prevent the cracking implies an increase in the tensile strength and spalling
of the concrete, results in fire endurance of HSC concrete. The columns with stirrups
bent at 135° bend at ends performed better than the columns with stirrups at 90° bend
at ends. The confining pressure from stirrups does play a role in fire endurance of the
RC columns. The current design specifications such as Eurocode-2 and IS:1642 pro-
vide design guidelines for fire safety of concrete structures and strength calculation
methods based on the amount of longitudinal reinforcement, temperature contour of
the section, the percentage of design load acting on the column, and clear cover.
The main focus is to study the confining reinforcement spacing on the fire behavior
of the RC column capacity by keeping all other parameters same, i.e., using the same
concrete mix, volume of longitudinal reinforcement, shear reinforcement, and the
same moisture content in all the specimens. To achieve this, the specimens with
the same volume of lateral reinforcement with different spacing (by changing the
diameter of the stirrup) are heated for the required duration and tested for mechanical
loads to quantify the load carrying capacity of the RC columns.
1.1 Experimental program
1.1.1 Test Setup and Instrumentation
The concrete column tests were conducted at Indian Institute of Technology, Hyder-
abad, for exploitation; Fig. 1a shows the experimental test setup and loading mech-
anism and Fig. 1b shows the thermocouple locations along the cross-section at mid
height of the column. This heating test setup was designed to fix in the desired
configuration as a rectangular, square or heating wall in loading frames available at
Indian Institute of Technology, Hyderabad. The radiant heating panels can heat the
assembled chamber up to 1000 °C. The test set up consists of four radiant heating
panels; every heater is with 305 mm × 450 mm heating area with a PID controller.
A data acquisition system with two modules having 16-thermocouple channels for
Response of RC Short Column Under Combined Fire and Axial Loading 123
Fig. 1 a Test setup with loading frame b thermocouple locations in the specimen at mid height
each module is used to acquire temperature data from the specimen and chamber
cavity at desired intervals. The sample rate at which each module can acquire data
is 75 S/s at high speed and 1 S/s at high resolution.
A set of four heaters are arranged as a rectangular tube of internal size approxi-
mately 420 mm × 420 mm, around the column. These heaters are placed such that
all the four sides of the column are parallel and at equal distance from the face of
each radiant heater, i.e., the internal face of the rectangular tube chamber. To prevent
the electrocute and damage of the heater surface due to direct contact of the spalled
concrete portions, a clear space of 90 mm was ensured on each face of the specimen
and furnace internal heating surface. It should be noted that the full length of the
column was not subjected to heating, the specimen ends were not heated because
of their proximity to the hydraulic system and column supports on rigid platens. To
prevent the escape of the heat from the rectangular tube, ceramic glass wool fiber
blanket of density 128 kg/m3 is filled at the ends of furnace and opening which can
resist temperature up to 1260 °C. It was observed that the maximum temperatures
did not exceed 50 °C at the column ends during the experiment. Four linear variable
differential transformers (LVDTs) are used to measure the axial deformations at the
end of the column to trace the deformation at the center of the column. The test results
shown in this paper include axial loads, concrete and reinforcing steel temperatures
in the column and the furnace chamber temperatures for each specimen.
Table 1 Specimen heating duration and failure load

S. No. ID of the Spacing of Dia. of Heating f c (MPa) Failure
specimen stirrups stirrups duration loads (in
(mm) (mm) min (Hrs) kN)
1 S75H2 75 6 120(2) 30 830
2 S130H2 130 8 120(2) 30 908
3 S200H2 200 10 120(2) 30 800
4 S75H3 75 6 180(3) 30 724
5 S130H3 130 8 180(3) 30 729
6 S200H3 200 10 180(3) 30 664
1.1.2 Testing Procedure
The column specimens were tested as follows: (1) The column was centered in the
loading frame to prevent the load eccentricity and heaters are assembled around the
column with clear distance of minimum 90 mm from the face of the column as dis-
cussed in test setup and instrumentation section. The specimen is heated for specified
duration mentioned in Table 1 to achieve desired ISO fire curve in the furnace cham-
ber. For the safety of the instrument and its limitations, after the one-hour duration,
the average temperature in the furnace chamber was maintained constant at 950 °C
till the desired time duration. During the heating process, the columns are allowed
to expand freely, i.e., no axial restraints. (2) Axial load was increased monotonically
up to the failure of the column. The columns are deemed to fail when the load falls
by 20 percent of the peak load.
1.1.3 Test Matrix
Test matrix of the column tests, including the details of the lateral reinforcement bar
diameter, duration of heating are presented in Table 1. The test matrix includes six
reinforced concrete columns out of which all have the same amount of lateral and
longitudinal reinforcements.
The dimensional details of the column are 860 mm long, 230 mm × 230 mm2
cross-section and the column slenderness ratio (L/D) was approximately equal
to 3.73. All the specimens are reinforced with the same volume of longitudinal
reinforcement four numbers-16 diameter bars, i.e., 804.2 mm2 of grade Fe500.
The concrete mix proportion is designed as per IS 10262:2009 to achieve cylin-
drical strength of 30 MPa. The average (f c) strength is 31.76 N/mm2 . The concrete
mix proportion is presented in Table 2.
Table 2 Mix proportions

Grade of concrete W/C ratio Cement Sand Coarse aggregate
M30 0.4 1 1.92 3.6
1200
S200H2
1000
800
Temperature(
600
400
200
0
0 30 60 90 120
Time(min)
T-1 T-2 T-3 T-4 T-5 T-6 T-7 T-8 ISO
Fig. 2 Chamber time–temperature curves of cavity
1.1.4 Time–Temperature Responses in Furnace Chamber
Figure 2 shows the average temperature–time (T − t) responses measured in the

cavity. Minimum eight thermocouples are placed inside the cavity to measure the
average temperature in the furnace as per the requirements mentioned in ASCE-
E119 and IS: 3809-2000 for column tests apart from the PID-thermocouple and
limit thermocouple to control the furnace temperature. The location of the furnace
thermocouples is shown in Fig. 4.
The temperature is measured during the experiment from the eight chambers
thermocouple; we have shown in Fig. 2 for one of the specimens (S200H2). The
average furnace temperature from all the thermocouples for six specimens is shown
in Fig. 3 (Fig. 4).
1.1.5 Temperature–Time Responses of Specimens
The temperature–time (T − t) response measured from the specimens in reinforce-

ment and core of the concrete. The location of the thermocouples in the cross-section
for all the specimens is same, which is presented in Fig. 1b. The temperature–time
curves in the specimens S200H2 and S200H3 are presented in Fig. 5. It was observed
that the temperatures in all the section at a given location is almost same for all
the samples for a given furnace temperature–time curve, the slight variation of the
temperature profiles are due to the small difference in the average furnace cavity
temperature–time curves which were presented in Fig. 3.
1200
1000
800
Temperature(
S200H3
600 S130H3
S75H3
400 S200H2
S130H2
200 S75H2
ISO
0
0 30 60 90 120 150 180 210
Time(min)
Fig. 3 Average cavity temperatures in the furnace
(a) Plan (b) Elevation

Fig. 4 Schematic of the furnace showing the location of the chamber thermocouples
800 S200H2 1000 S200H3

1
1
800 2
Temperature(˚C)
Temperature(˚C)
600 2
3
3 600
4 4
400
5 400 5
6 6
200
7 200 7
8 8
0 0
0 30 60 90 120 150 0 30 60 90 120 150 180
Time(min) Time(min)
Fig. 5 Temperature–time curve of specimens

2 Test Results
In the present study, the volumetric ratio of transverse reinforcement is kept the
same in all the test specimens irrespective of the diameters of stirrups used (6, 8
and 10 mm). It was observed that the columns with 8 mm dia. stirrups at 130 mm
spacing having a high load carrying capacity 907 kN for 120 min and 728 kN for
180 min duration of heating under ISO fire curve, respectively. Whereas the column
with 6 mm stirrups at 75 mm spacing having a maximum load bearing capacity of
829 kN and 724 kN for 120 min and 180 min duration of heating under ISO fire
curve, respectively. Similarly, 10 mm stirrups at 200 mm spacing having maximum
load bearing capacity of 800 kN and 664 kN for 120 min and 180 in duration of
heating under ISO fire curve. The load–deformation curve is presented in Fig. 6; we
can clearly observe that stiffness and strength of specimens with 8 mm stirrups are
more than specimens with 6 and 10 mm stirrups.
From the failure loads, it is observed that 8 mm stirrup gave higher load capacity
at elevated temperature, which creates a mismatch with the ambient temperature
behavior. It was observed from the graph that the stiffness of the columns with 8 mm
dia. stirrup columns gave higher performance when compared with 6 and 10 mm
stirrup columns. The ductility of the columns with 6 mm dia. stirrups are executing
more than the other two columns.
As per the literature available on the RC columns at ambient temperatures, the
load bearing capacity of the column increases with decrease in spacing between the
stirrups, this may be due to increase in spacing between the stirrups, the effective con-
fining zone decrease, which cannot prevent the lateral bulging of the column results
in failure. Whereas in case of columns with closer spacing of stirrups, the effective
confined area by the lateral reinforcement is effected by the closer stirrups and results
in constant load bearing capacity after certain closer spacing. The hypothesis that
Load DeformaƟon Curves for 180 min Load DeformaƟon Curves for 120 min
HeaƟng HeaƟng
1000 1000
800 800
Load (kN)
Load (kN)
600 600
400 400
S75H3 S75H2
200 200 S130H2
S130H3
S200H3 S200H2
0 0
0 2 4 6 8 10 12 0 2 4 6 8 10 12
DeformaƟon (mm) DeformaƟon (mm)
a)180 min heating b)120 min heating
Fig. 6 Load–deflection response; a 180 min b 120 min duration of heating

the authors made here to explain the decrease in load carrying capacity with lesser
spacing between the stirrups is: due to overlapping of the confining zone or region by
a stirrup the effective confining pressure on concrete between the stirrups is limited.
Which may results in constant load bearing capacity, after certain decrease of spac-
ing between the stirrups. But when the confining reinforcement strength decreases at
high temperature due to material property degradation, the effective confining pres-
sure at higher temperatures on the same zone increases results in higher load carrying
capacity, which was observed in case of columns heated for 180 min duration under
ISO fire curve.
In Fig. 7, the peak loads are compared between different spacing of reinforcement
at 120 min and 180 min duration of heating. In the case of 180 min duration of heating,
the peak loads for specimens S75H3 and S130H3 are almost same but whereas for
specimen S200H3, the spacing of shear reinforcement affected the load carrying
capacity of the columns with increase in spacing of stirrups the loading capacity
decreases. This is true in case of S130H2 and S200H2, the increase in spacing of
stirrups from 130 to 200 mm spacing with same volumetric ratio.
This hypothesis made by the authors is also validated with the numerical analysis
developed by the authors to simulate the RC columns under fire conditions [13, 14].
To get understanding and insight into this problem, authors are conducting few more
experiments with wide range of parameters. The hidden parameter, the effective
cover, may play a role in better performance of the column in case of larger diameter
stirrups as the clear cover kept same to all the stirrups. This will be addressed by
studying a detailed parametric study using numerical model developed by the authors
[13, 14].
1000
800
600
Load (kN)
400
200
0
S75H3 S130H3 S200H3 S75H2 S130H2 S200H2
Fig. 7 Peak load comparison of specimens with different volume of stirrups

3 Conclusions
The photographs of damaged specimens after the test were presented in Fig. 8; it
was observed that the columns failed in pure compression at the center of the heating
portion. The columns with 8 mm diameter stirrup at 130 mm spacing performed
better than the columns with 6 mm and 10 mm stirrups at 75 and 200 mm spacing,
respectively.
The behavior observed in the tests leads to the following conclusions:
1. The moisture in the concrete evaporates at an initial 60 min duration of heating.
During the initial 60 min, the test zone reaches about 130 °C. The change in
concrete temperatures is very small in between 90 and 130 °C. In the outer
periphery of the column, the moisture effect on temperatures is not high due to
the escape of hot vapor from the periphery of the column.
2. All the columns failed in a compression mode at the mid height of the heating
zone.
3. The columns with 8 mm diameter stirrups at 130 mm spacing gave approximately
10% higher load capacity than 10 mm dia. stirrups at 200 mm spacing for 120 min
duration under ISO fire curve.
4. The columns with 6 mm and 8 mm dia. stirrups at 75 mm and 130 mm spacing,
respectively, carried 8% higher load compared to columns with 10 mm dia.
stirrups at 200 mm spacing for 180 min duration under ISO fire curve.
A more detailed experimental program (12 more columns) is planned by the
authors with a wide range of variables to get insight into the effect of spacing on the
fire behavior keeping the total volume of confining reinforcement constant.
S75H2 S130H2 S200H2
S75H3 S130H3 S200H3
Fig. 8 Photographs of the specimen after test for 120 and 180 min of heating
References
1. Mander, J. B., Priestley, M. J. N., & Park, R. (1988). Observed stress-strain behavior of confined
concrete. Journal of Structural Engineering, 114(8), 1827–1849.
2. Cusson, D., & Paultre, P. (1994). High strength concrete column confined by Rectangular ties.
Journal of Structural Engineering, 120(3), 783–804.
3. Hoshikuma, J., Kazuhiko, K., Kazuhiko, N., & Taylor, A. W.(1996) A model of confinement
effect on stress-strain relation of reinforced concrete columns for seismic design.
4. Kodur, V., Cheng, F.-P., Wang, T.-C., & Sultan, M. (2003). Effect of strength and fiber reinforce-
ment on fire resistance of high-strength concrete columns. Journal of Structural Engineering,
129(2), 253–259.
5. Wu, H., & Lie, T. T. (1992). Fire resistance of reinforced concrete columns: experimental
studies. NRC Publications Record, No. 632.
6. Lie, T., & Woollerton, J. (1988). Fire resistance of reinforced concrete columns: Test
results, national research council of Canada (Internal Report 569). Institute for Research
in Construction.
7. Lie, T. T. (1989). Fire resistance of reinforced concrete columns: A parametric study. Journal
of Fire Protection Engineering, 1(4), 121–129.
8. Dotreppe, J.-C., Franssen, J.-M., Bruls, A., Baus, R., Vandevelde, P., Minne, R., et al. (1997).
Experimental research on the determination of the main parameters affecting the behaviour of
reinforced concrete columns under fire conditions. Magazine of Concrete Research, 49(179),
117–127.
9. Kodur, V., & Phan, L. (2007). Critical factors governing the fire performance of high strength
456 concrete systems. Fire Safety Journal, 42(6), 482–488.
10. Kodur, V., Cheng, F., Wang, T., Latour, J., & Leroux, P. (2001). Fire resistance of high
performance concrete columns. NRC Publications Record (I R C - I R - 8 3 4).
11. Kodur, V., McGrath, R., Leroux, P., & Latour, J. (2005). Experimental studies for evaluating
the fire endurance of high-strength concrete columns.
12. Kodur, V., & Mcgrath, R. (2003). Fire endurance of high strength concrete columns. Fire
Technology, 39(1), 73–87.
13. Chinthapalli, H. K., & Agarwal, A. (2018). Role of confinement on fire resistance capacity
of RC columns. In 3rd R. N. Raikar Memorial International Conference and Gettu-Kodur
International Symposium on Advances in Science and Technology of Concrete, Mumbai.
14. Chinthapalli, H. K., & Agarwal, A. (2019). Effect of confining reinforcement on fire behavior
of reinforced concrete columns–An experimental and numerical study. Journal of Structural
Engineering, ASCE.
Numerical Analysis of Post-earthquake
Fire Resistance of Concrete-Filled
Tubular Steel Columns
Smita Singh and Anil Agarwal
1 Introduction
Earthquakes are one of the most devastating forms of natural hazard, which can have
a severely debilitating effect on structures. But if an earthquake event is followed by
fire, it may cause even more damage than the earthquake. In the past several cases
of fire following the earthquake, a large amount of damage has been documented.
In the early twentieth century, devastating fire following the San Francisco, Califor-
nia earthquake in 1906 and Tokyo earthquake in 1923 caused significant damage to
both the cities [1]. Although large-scale post-earthquake fires have not been reported
recently, there are numerous incidents of isolated post-earthquake fires causing sig-
nificant damages [2]. Hundreds of fires were reported after the 1994 Northridge, 1995
Kobe and 1989 Loma Prieta earthquakes. Recently, Davidson [3] reported several
conflagrations after Japan earthquake of 2011.
After a major earthquake, the ability of the structure to resist fire drastically
reduces. Also, the risk of fire after an earthquake increases significantly. Although
the fire following an earthquake has a devastating effect, the subject has not received
enough attention. There is a need to investigate the resistance capacity of structures
under the combined effect of earthquake and fire. Until now, several researchers have
studied the behaviour of CFST columns at fire and earthquake separately, but none
have studied the behaviour of CFST column under the effect of both incidents.
Several researchers have worked on the seismic behaviour of CFST columns and
indicated the desirable seismic performance of CFSTs. In the interest of brevity, only
one example is being presented here. Han and Yang [4] experimentally investigated
S. Singh (B) · A. Agarwal

A. Agarwal

134 S. Singh and A. Agarwal
the seismic behaviour of eight circular CFST columns. Specimens were tested under
constant axial load and cyclically increasing flexural loading. They also performed
the nonlinear full range analysis of these composite columns for simulating the
test results and did the parametric study. They analysed the influence of several
parameters on the behaviour of the moment versus curvature response and the lateral
load versus lateral displacement relationships for the composite columns.
A number of other studies have exhibited the superior fire resistance properties of
CFSTs in comparison with steel members [5, 6]. Hong and Varma [5] analytically
studied the behaviour of loaded CFT columns under standard fire. The sequentially
coupled thermal stress analysis was used to predict the behaviour of the CFT columns
subjected to standard fire loading. They conducted a detailed sensitivity analysis to
determine relevant input parameters for the analytical models. Consequently, they
used Poh 2001 σ–ε–T model for modelling steel tube and Lie and Irwin 1995 σ–ε–T
model for the concrete infill [7]. They used linear thermal expansion models for
steel and concrete materials and recommended full composite action with no local
buckling to model the CFT column specimens.
Very few researchers [8–11] have studied the post-earthquake fire behaviour of
different structural systems. Yassin et al. [8] conducted an analytical study on the per-
formance of moment-resisting steel frames in the post-earthquake fire. They found
that the performance of the structure highly depends upon the residual lateral defor-
mations present in structure caused by the earthquake. Fire induces asymmetrical
deformations due to uneven heating. This amplifies the previously present resid-
ual deformations in the structure and significantly reduces the fire resistance of the
structure, while the frames with no residue displacement from a prior seismic loading
exhibit better fire resistance properties.
Imani et al. [9, 10] demonstrated through both experimental and numerical studies
effect of post-earthquake fire resistance of concrete-filled double-skin tube (CFDST)
columns. After an initial simulation of lateral loading, they observed that concrete
damage plasticity (CDP) model for concrete failed to capture the pinching phenom-
ena of the hysteresis behaviour. Considering the shortcoming of the CDP model,
they modified the model by inserting the horizontal discrete crack [12] at the loca-
tion of maximum tensile stress. The numerical predictions of failure time-matched
preferably with the experimental results.
Talebi et al. [11] developed and validated a nonlinear three-dimensional finite
element model for investigating the behaviour of CFST columns subjected to post-
earthquake fire. They reported that CFT columns use benefits of the composite action
between the steel tube and in-filled concrete and could behave in a ductile manner.
Concrete-filled steel tube (CFST) column has shown good resistance under cyclic
as well as fire loading conditions. This study expands on these previous studies by
investigating the behaviour of CFSTs numerically when exposed to fire after an
earthquake. In this paper, a three-dimensional finite element model was developed
in ABAQUS [13]. The various key variable also included in modelling like mate-
rial degradation properties at elevated temperature, contact between steel tube and
concrete core, initial imperfection. The steps involved in numerical modelling were
cyclic, heat transfer and structural analysis. First effect of earthquake loading was
Numerical Analysis of Post-earthquake Fire Resistance … 135
simulated by cyclic loading of the specimen, and then, the outcome of the cyclic
loading in terms of residual deformation was applied as an initial condition to the
sequentially coupled thermal stress analysis. Results of the finite element model were
compared with the experimental results of cyclic loading test [4] and fire loading test
[14] separately to check the accuracy of the numerical model. Finally, validated
numerical model was used to study the behaviour of CFST column under cyclic and
fire loading as multi-hazard event.
2 Numerical Model Under Cyclic Loading
Effect of earthquake loading was studied by subjecting the CFST column under cyclic
loading. A three-dimensional finite element model was modelled in ABAQUS. The
validity of the numerical model was verified by the results of the experiment on
CFST columns conducted by Han and Yang [4]. They have experimentally studied
the behaviour of eight CFST column under cyclic loading, and columns labelled
as SC2-3 and SC2-4 modelled in this study for verification purpose. Details of the
specimens are enlisted in Table 1 where N o is the axial load applied on the composite
specimen, f c is the compressive strength of concrete; f y is the yield strength of steel,
E s is the modulus of elasticity of steel.
2.1 Loading, Boundary Condition, Contact at the Steel

Tube–Concrete Core Interface and Initial Imperfection
Two rigid steel plates of 16 mm thick were modelled at the columns endings to apply
the boundary condition and axial loading. Pinned-pinned boundary conditions were
modelled at both the endings, and one end was free to move longitudinally for appli-
cation of axial load. To apply the cyclic loading, a rigid steel stub of 150 mm length
was modelled at the centre of the CFST column to exactly replicate the experimental
setup. Loading was applied stepwise. First, the axial load was applied and kept con-
stant throughout the analysis, and then, cyclic loading is applied according to ATC-24
[15] loading protocol to simulate the seismic behaviour. Cycles were imposed in lev-
els of y,1.5 y,2 y,3 y,5 y,7 y, 8 y, where y was the yield displacement
of the column and y were used as obtained in the test [13]. Specimens are generally
not straight due to the manufacturing process and consist of some initial geometric
Table 1 Detail of the CFST columns for cyclic analysis

Column L (mm) D × t (mm) f c (MPa) f y (MPa) E s (GPa) N o (KN)
SC2-3 1500 114 × 3 38.9 308 200 286
SC2-4 1500 114 × 3 38.9 308 200 429
imperfection. So, to account for initial geometric imperfection, eigenvalue buckling

analysis was conducted in ABAQUS [13]. Then, first mode shape was multiplied by
an amplification factor of L/1000 was used as an initial imperfection. L is the length
of the column. Boundary conditions were chosen such that they could replicate the
experimental condition. To model, the interface between the steel tube and concrete
core normal behaviour and tangential behaviour were used. A “hard contact” defi-
nition was used to simulate the normal behaviour. Coulomb friction model with a
constant coefficient value of 0.2 used to simulate the tangential behaviour.
2.2 Material Model
For modelling the steel material properties, a bilinear material model consisting of
a linear elastic phase and post-yield linear hardening phase was used. The elastic
response of steel was modelled using elastic modulus as 200 GPa and Poisson’s ratio
as 0.3. Plastic behaviour of the structural steel was modelled by kinematic hardening
model. The behaviour of concrete in tension and compression was done according
to EN 1992 1-1. Initial elastic modulus of concrete was calculated according to EN
1992 1-1, and Poisson’s ratio was taken as 0.2. To model, the plastic behaviour of
concrete damage plasticity model [11, 13, 16] implemented in ABAQUS was used.
2.3 Element Model
The 3D solid element is used to model the concrete core, steel tube and steel plate
because it can capture large deformation and plastic strains. Because of the high
computational time of 20 node elements during analysis, a three-dimensional eight-
node solid elements with reduced integration are used to model in the present study
to decrease the computational time during analysis. Figure 1 shows the whole finite
element model of column.
Fig. 1 Whole FE model

2.4 Validation of Numerical Model Under Cyclic Loading
The cyclic performance of the columns is represented in terms of lateral load versus
mid-span deflection. Figures 2 and 3 show a comparison between the results of the
finite element model and experiment. Comparison of the hysteresis curve from the
numerical model and experiments shows the comparable results. Results also show
that by increasing axial load, lateral ductility and bearing also decrease.
3 Numerical Model Under Fire Loading
A three-step sequentially coupled thermal stress analysis approach was used for
modelling the fire resistance of the CFST column. The validity of the numerical model
was verified by the results of the experiment of CFST columns conducted by Lie
and Chabot [14]. They have experimentally investigated the behaviour of 44 CFST
columns, and among them, the columns labelled as C-26 and C-23 were selected
Fig. 2 Lateral deflection versus lateral load for column SC2-3 in a numerical model b experiments
Fig. 3 Lateral deflection versus lateral load for column SC2-4 in a numerical model b experiments
Table 2 Detail of the CFST columns for fire analysis

C-26 3810 323.9 × 6.35 24.3 350 200 1050
C-23 3810 273.1 × 12.7 27.4 350 200 525
to validate the numerical model. Length of the column was 3810 mm including
the endplate, but only middle 3048 mm length of column was exposed to fire. The
thickness of the endplates for columns C-26 and c-23 was 38 mm and 25 mm,
respectively. Details of the selected CFST column are described in Table 2.
3.1 Heat Transfer Analysis
A transient heat transfer analysis was conducted on the selected CFST columns
to simulate the effects of the standard ASTM E119 1-1 Fire curve [17]. First, the
recorded time–temperature history obtained from heat transfer analysis is applied at
the outer surface of the steel tube by setting the boundary condition. Two heat transfer
mechanisms conduction and radiation were defined for steel tube and concrete core
interface. In the heat transfer analysis, the model was meshed using three-dimensional
eight-node solid elements (DC3D8) for the steel tube and the in-filled concrete. The
results of heat transfer analysis consist of the nodal time–temperature history, which
was stored as an output data.
3.2 Stress/Deformation Analysis
In the stress analysis, three-dimensional eight-node solid elements with reduced inte-
gration (C3D8R) were used for meshing both the steel tube and the concrete core.
Two steel end plates were designed at the columns ending to apply the boundary con-
dition and axial loading. Loading was applied sequentially. First, axial loading was
applied at the top plate and kept constant throughout the analysis. In the second step,
fire loading was applied in terms of the nodal time–temperature curve a predefined
field, which was obtained from the heat transfer analysis.
3.3 Material Properties at Elevated Temperature
The thermal properties of concrete at an elevated temperature were taken as per EN

1992-1-2 [18]. For structural steel, thermal properties recommended by EN 1993-1-2
20 600
50.00
100 500 20
100
Stress (MPa)
40.00 200
Stress (MPa)
400 200
30.00 300
400 300 300
400
20.00 500 200 500
600
10.00 100 600
700
700
0.00 0
0.000 0.010 0.020 0.030 0.00 0.01 0.02 0.03 0.04 0.05
Strain Strain
(a) (b)
Fig. 4 Stress–strain curve at an elevated temperature a concrete b steel
[19] were adopted. The value of thermal expansion coefficient and thermal conduc-
tivity of steel and concrete were adopted from EN 1994 1-2 [20]. A classic metal plas-
ticity material model implemented in ABAQUS was used for modelling the plastic
behaviour of steel which follows von Mises yield and associated plastic flow. Elastic
behaviour of concrete was modelled by modulus of elasticity and Poisson’s ratio (μ
= 0.3). The nonlinear response of the concrete was modelled using Drucker–Prager
yield criteria available in ABAQUS. Figure 4 shows the stress–strain relationship for
concrete and steel at elevated temperatures.
3.4 Initial Geometric Imperfection of the Column
Specimens are generally not straight due to the manufacturing process and consist of
some initial geometric imperfection. So, to account for initial geometric imperfection,
eigenvalue buckling analysis was conducted in ABAQUS. Then, first mode shape was
multiplied by an amplification factor of L/1000 was used as an initial imperfection.
3.5 Loading, Boundary Condition and Contact at the Steel

Tube–Concrete Core Interface
Figure 5 shows the standard ASTM E119 [17] fire curve for thermal loading as the
same was used in the experiments. To exactly replicate the experimental condition, a
fixed–fixed boundary condition was used for both the selected CFST columns. The
modelling of steel tube and concrete core interface in the thermal stress analysis is
similar to that of cyclic loading case in Sect. 2.3. Friction coefficient value of the
Coulomb friction model is independent of thermal loading.
Fig. 5 Standard ASTM E119 fire curve
3.6 Validation of Numerical Model Under Fire Loading
The numerical model showed good agreement with the tests, as can be seen in Fig. 6.
Maximum fire resistance time and maximum axial displacement for the simulated
column C-23 were 163 min and 36 mm, respectively. The experimental graph shows
the comparable values as 143 min and 36 mm. Column C-23 experienced the maxi-
mum fire resistance time and maximum axial displacement as 102 min and 15.23 mm,
respectively. The corresponding experimental values were 93 min and 15.7 mm.
4 Numerical Model Under Post-earthquake Fire Scenario
The numerical models, thus calibrated, will be used to study the behaviour of the
CFST column subjected to post-earthquake fire loading. Detail of the CFST column
is presented in Table 3. The column selected for post-earthquake fire case is stocky.
For predicting the behaviour of the column under post-earthquake fire case, two
Fig. 6 Comparison of the axial displacement versus time curve for a C-23 b C-26
Table 3 Detail of the CFST column for post-earthquake fire analysis

C-26 3810 323.9 × 6.35 24.3 350 200 1050
analysis steps are required, namely cyclic analysis and sequentially coupled thermal
stress analysis. Boundary condition and loading were applied through the end plates.
4.1 Stepwise Analysis Procedure
To simulate the behaviour of the CFST column under post-earthquake fire scenario,
the analysis was performed stepwise as follows. First, a cyclic analysis was conducted
to simulate the seismic behaviour of the column. Analysis procedure and modelling
parameters are the same as mentioned in Sect. 2. Two damage levels in terms of drift
ratio were considered for this study. After the cyclic analysis, residual deformations
generated were used as an input for the thermal stress analysis. In the cyclic analysis,
seismic damage level was chosen such that there will not be any fracture on the steel
tube. In the next step, sequentially coupled thermal stress analysis was conducted
to simulate the fire loading. Thermal stress analysis process follows the techniques
defined in Sect. 3.
4.2 Result and Discussion
From Fig. 7, it can be seen that as the drift ratio increases even by small quantity, the
fire resistance time for column decreases. This can be attributed due to the residual
displacement of the column after cyclic loading and also because of the material
capacity being reached due to the axial loading. The column is also stocky in nature.
Fire resistance time after fire loading for the case of (a) without drift is 45 min, (b)
0.5% drift is 33 min and (c) 2% drift is 24 min.
5 Conclusions
Further numerical and experimental investigations on slender and stocky CFST

columns are required to understand the influence of the fire loading after an earth-
quake. Numerical study on one column cannot give a clear picture of the post-
earthquake fire behaviour. Thermal and general structures material properties taken
from the Eurocode were accurate enough to model the behaviour of the CFST
columns under cyclic loading and post-earthquake fire loading.
Fig. 7 Axial displacement vs time for column C-26 at different damage levels
References
1. Scawthorn, C., Eidinger, J. M., & Schiff, A. J. (2005). Fire following earthquake. American
Society of Civil Engineers.
2. Scawthorn, C. R. (2008). Fire following earthquake—The shake out scenario report. (SPA
Technical Report}. Berkeley, CA: SPA Risk, LLC.
3. Davidson, R. (2012). Post-earthquake fires in the March 2011 Japan earthquake and tsunami. In
Proceedings of 2012 Japan and New Zealand RAPID and Research Needs Workshop, Arlington,
VA: NSF, February 9–10.
4. Han, L. H., & Yang, Y. F. (2005). Cyclic performance of concrete-filled steel CHS columns
under flexural loading. Journal of Constructional Steel Research, 61(4), 423–452. https://doi.
org/10.1016/j.jcsr.2004.10.004.
5. Hong, S., & Varma, A. H. (2009). Analytical modelling of the standard fire behaviour of loaded
CFT columns. Journal of Constructional Steel Research, 65(1), 54–69. https://doi.org/10.1016/
j.jcsr.2008.04.008.
6. Kodur, V. K. R. (1998). Performance of high strength concrete-filled steel columns exposed to
fire. Canadian Journal of Civil Engineering, 25(6), 975–981.
7. Poh, K. W. (2001). Stress-strain–temperature relationships for structural steel. Journal
of Materials in Civil Engineering, 13(5), 371–379. https://doi.org/10.1061/(ASCE)0899-
1561(2001)13:5(371).
8. Yassin, H., Iqbal, F., Bagchi, A., & Kodur V. K. R. (2008). Assessment of post-earthquake
fire performance of steel-frame buildings. In The 14th World Conference on Earthquake
Engineering, 2008 (pp. 12–17). Beijing, China.
9. Imani, R., Mosqueda, G., & Bruneau, M. (2014). Experimental study on post-earthquake
fire resistance of ductile concrete-filled double-skin tube columns. Journal of Structural
Engineering, 141(8), 04015055. https://doi.org/10.1061/(ASCE)ST.1943-541X.0001168.
10. Imani, R., Mosqueda, G., & Bruneau, M. (2015). Finite element simulation of concrete-filled
double-skin tube columns subjected to post earthquake fires. Journal of Structural Engineering,
141(12), 04015055. https://doi.org/10.1061/(ASCE)ST.1943-541X.0001301.
11. Talebi, E., Korzen, M., & Hothan, S. (2018). The performance of concrete-filled steel tube
columns under post-earthquake fires. Journal of Constructional Steel Research, 150(1), 115–
128. https://doi.org/10.1016/j.jcsr.2018.07.013.
12. Chen, W. F. (2007). Plasticity in reinforced concrete (Vol. 474), J. Ross Publishing: Plantation,
FL.
13. ABAQUS Standard User’s Manual. (2013). The Abaqus Software is a product of Dassault
Systèmes Simulia Corp., Providence, RI, USA Dassault Systèmes, Version 6.14.1, USA.
14. Lie, T. T., & Chabot, M. (1992). Experimental studies on the fire resistance of hollow steel
columns filled with plain concrete (Internal Report No. 611). Ottawa (Canada): Institute for
Research in Construction, National Research Council of Canada NRCC.
15. ATC-24. (1992). Guidelines for cyclic seismic testing of components of steel structures.
Redwood City (CA): Applied Technology Council.
16. Lakavath, C., Allam, R., & Kondraivendhan, B. (2019). Experimental and numerical studies
on the behaviour of broad-gauge railway sleepers in static bending condition. In Sustainable
construction and building materials (pp. 781–792). Singapore: Springer. https://doi.org/10.
1007/978-981-13-3317-0_70.
17. ASTM. (1990). Standard ASTM E119–88: Standard Methods of Fire Test of Building
Construction and Materials, Philadelphia (PA), American Society for Testing and Materials.
18. EC2. (2004). EN 1992-1-1, Eurocode 2: Design of concrete structures—Part 1–1: General
rules and rules for buildings, European Committee for Standardization, Brussels (Belgium).
19. EC3. (2005). EN 1993-1-2, Eurocode 3: Design of steel structures—Part 1–2: General rules—
Structural fire design, European Committee for Standardization, Brussels (Belgium).
20. EC4. (2005). EN 1994-1-2, Eurocode 4: Design of composite steel and concrete structures—
Part 1–2: General rules-Structural fire design, European Committee for Standardization,
Brussels (Belgium).
Mathematical Model for Prediction
of Compressive Strength of Normal,
Standard and High Strength SCC
with RCA
Srinivas Vasam, K. Jaganadha Rao, and M. V. Seshagiri Rao
1 Introduction
Concrete is the most widely used man-made material and its consumption is now next
to water. Concrete has become popular because of its mouldability into any complex
shape, abundant availability of its ingredients, relative economy and high compres-
sive strength. Concrete is no mere a mixture of cement, coarse aggregates, fine
aggregates and water. The development of new admixtures has completely changed
the definition of concrete. Though there is a considerable research in the last few
decades on concrete-making materials and technologies, concrete has certain draw-
backs like the presence of voids due to improper mixing/compaction which affects
strength and durability. Self-compacting concrete (SCC) is one of the solutions to
minimize such voids in concrete. Self-compacting concrete finds several applica-
tions in the construction field due to its multiple advantages like ease of placement,
reduced noise pollution, less concreting time, etc. A lot of research is being carried
out in the last few decades on SCC with the use of different combinations of mineral
and chemical admixtures.
The focus on recycled aggregate concrete is gaining importance to address the
problems of depletion of natural resources and disposal of construction and demo-
lition waste to some extent. The use of recycled concrete aggregate (RCA) in SCC
S. Vasam (B)
Department of Civil Engineering, Siddhartha Institute of Technology & Sciences, Korremula,
Hyderabad 501301, India
K. Jaganadha Rao
Department of Civil Engineering, Chaitanya Bharathi Institute of Technology, Gandipet,
M. V. Seshagiri Rao
Department of Civil Engineering, Jawaharlal Nehru Technological University,
146 S. Vasam et al.
Fig. 1 Flow chart showing conversion of recycled concrete aggregate to a useful product
combines the advantages of both SCC and RCA. In general, the design of concrete
mix is a trial and error procedure and more so when new materials like RCA are
used. In the current paper, an empirical model is developed to predict the compres-
sive strength and other mechanical properties of SCC with RCA by carrying out
the multiple regression analysis on the test data obtained by carrying out extensive
laboratory investigations. The variables considered in the model are the grade of con-
crete and percentage of RCA. The empirical equation developed is found to predict
the compressive strength of SCC with RCA to an accuracy of about 90%. Thus, a
lot of effort, time and material can be saved by using the developed strength pre-
diction equation due to the reduced trial mixes. The following sections present the
methodology followed, results obtained, discussions and conclusions
1.1 C&D Waste Minimization
Minimization of C&D waste should commence from day one of the project. This
includes better planning, use of precast construction, reuse and recycling of C&D
waste materials on site with minimal import and export (Fig. 1).
1.2 Objective
The objective of the present work is to develop mathematical model to predict

the compressive strength of recycled aggregate self-compacting concrete (RASCC),
enabling the minimization of laboratory trial mixes for finalization of mix proportion.
1.3 Scope of the Work
The scope of the work is limited to the use of data pertaining to compressive strength
values with variables of study as grade of concrete (M20 to M70—normal to high
grades), percentage of RCA replacement ratio (0–100%). The data consist of 28 days
compressive strengths of 6 Grade of concrete with 5 Percentage replacement of RCA
as given in Table 6. NCSS and MATLAB software are used for modelling purpose.
Mathematical Model for Prediction of Compressive Strength … 147
The self-compatibility of the mixes is ensured by conducting various tests specified

by EFFNARC [1], viz., T50 Slump flow, L-Box, U-Box and V-funnel tests. Su et al.
[2] conducted research on the mix design for Self Compacting Concrete and arrived at
a simple mix design procedure. As this method is simple and already established, the
same method is adopted for SCC with recycled aggregate and suitable modification
factors are suggested. The model is expected to predict the compressive strength of
RAC. The study also aims at validating the model and estimating the percentage
error between theoretical (predicted) results and experimental results.
2 Experimental Programme
2.1 Materials
2.1.1 Cement
Ordinary Portland cement of Grade 53 confirming to IS: 12269 [3] with specific
gravity 3.15 was used in this investigation. The chemical composition of cement is
given in Table 1a.
Table 1 a Chemical
Chemical property Limits (as per IS) Results
composition of cement (as per
manufacturers’ test report). b (a)
Chemical composition of fly Lime saturation factor (%) 0.66–1.02 (max.) 0.82
ash
Alumina iron ratio Min. 0.665% 1.2%
Insoluble residue Max. 2% 0.95%
Magnesia (%) Max. 6 2.4
Sulphuric anhydride 2.5–35% 1.1%
Loss on ignition Max. 5% 2.2%
Chemical property Result (% mass)
(b)
Loss on ignition 0.43
Alumina (as Al2 O3 ) 16.31
Silica (as Sio2 ) 60.82
Iron (as Fe2 O3 ) 17.17
Calcium (as CaO) 4.64
Magnesium (MgO) Not found
Sodium(as Na2 O) 0.34
Potassium (as K2 O) 0.08
148 S. Vasam et al.
Table 2 Properties of sand, natural coarse aggregate and recycled concrete aggregate
Property Sand NCA RCA
Specific gravity 2.59 2.81 2.35
Total water absorption (%) 1.0 0.3 2.40
Moisture content (%) 0.15 0.8 0.45
Bulk density (Loose) (kg/m3 ) 1567 1380 1355
Bulk density (Compacted) (kg/m3 ) 1713 1530 1590
Fineness modulus 2.39 (Zone III) 6.36 6.35
Elongation index (%) – 7.10 15.5
Flakiness index (%) – 6.15 5.8
2.1.2 Mineral Admixtures
Fly ash conforming to IS 3812 (Part-1): 2003 [4] was used. The chemical composition
of fly ash is given in Table 1b.
2.1.3 Fine and Coarse Aggregates
The physical properties of sand, NCA and RCA used in the present experimental
investigations are tabulated in Table 2. The maximum size of coarse aggregate was
12.5 mm.
2.1.4 Chemical Admixtures
Chemical Admixtures were used in this investigation. Super plasticizer (polycar-

boxylate ether based) with specific gravity 1.01 and pH: 8 and viscosity modifying
admixture (Glenium B233 stream 2) with specific gravity 1.1 and pH: 6 were used
in this work.
2.1.5 Water
Locally available potable water was used for mixing and curing (Figs. 2, 3, 4 and 5;
Tables 3, 4 and 5).
Fig. 2 Sustainable concrete

approach
Fig. 3 Demolished concrete

material
3 Development of Model
Multi-linear regressions determine the relationship between two or more independent

variables and a dependent variable by fitting a linear equation to the observed data.
Every value of the independent variable is associated with a value of the dependent
variable as represented in the equation as per the general form of the multiple linear
regressions as below:
y = a0 + a1 x 1 + a2 x 2 + a3 x 3 + a4 x 4
. . . (multi linear regression quation)
Here, ‘y’ is dependent variable; X 1 , X 2 , X 3 and X 4 are independent variables repre-

senting various parameters. We measure a quantity y for several values of parameters
X 1 , X 2 , X 3 and X 4.
150 S. Vasam et al.
Fig. 4 LAB tests
Fig. 5 Concrete testing
The compressive strength of concrete after 28 days of curing depends upon the
grade of concrete. It also varies with varying the percentage of recycling aggregate
mixed in the concrete
Considering the grade of concrete and percentage of recycling aggregates as inde-
pendent variables and compressive strength of concrete after 28 days of curing as
dependant variable, the following formula was arrived at:
f cke = 17.04 + 0.85 × g c − 0.05 × pra

Table 3 Mix proportions and quantities of different grades of SCC
Mix prop. of concrete Cement Fly ash CA—12.5 mm (NA) CA—12.5 mm (RA) River sand Admixture (SP) VMA Water W/P
grade
NASCC-M20-0% 1.000 0.647 2.179 0.000 2.317 0.020 0.002 0.644 0.39
NASCC-M30-0% 1.000 0.394 1.812 0.000 1.927 0.017 0.002 0.521 0.37
NASCC-M40-0% 1.000 0.322 1.655 0.000 1.760 0.016 0.002 0.458 0.35
NASCC-M50-0% 1.000 0.226 1.508 0.000 1.603 0.015 0.001 0.407 0.33
NASCC-M60-0% 1.000 0.105 1.294 0.000 1.376 0.013 0.001 0.328 0.30
Mathematical Model for Prediction of Compressive Strength …
NASCC-M70-0% 1.000 0.057 1.039 0.000 1.105 0.013 0.001 0.261 0.25
151
152 S. Vasam et al.
Table 4 Acceptance criteria for self-compacting concrete

S. No. Testing method Typical Property Unit Range of
range of values
values as per obtained as
EFNARC per
experiment
Min. Max. Min. Max.
1 Slump flow by 650 800 Filling ability mm 658 800
abrams cone
2 T50 cm slump 2 5 Filling ability s 2 6.5
flow
4 V-funnel 6 12 Filling ability s 6.5 9.5
5 V-funnel at T5 0 +3 Segregation s 9.5 12.5
min resistance
6 L-box 0.8 1.0 Passing ability (h2/h1) 0.82 0.98
7 U-box 0 30 Passing ability (h2 − h1) mm 20 28
Table 5 Input–output data

Input/Output data Range
for the mathematical model to
predict the compressive Grade of concrete (gc —Input M20 to M70 (with an
strength of SCC data) increment of 10)
Percentage of replacement (0–100%) (with an increment
ratio (Pra —Input data) of 25%)
Compressive strength, MPa 29.483–76.625 Mpa
(f cke —Output data)
where ‘ f cke ’ is the estimated compressive strength of SCC after 28 days of curing,
in N/mm2
gc —The grade of concrete in N/mm2 (it is found to be suitable for ordinary,
medium and high strength SCC).
pra —The percentage of recycling aggregate mixed in the concrete (percentage of
recycled aggregate content varying from 0 to 100% in the increment of 25%).
Table 6 shows the compressive strength of concrete in N/mm2 after 28 days
of curing, computed as per the authors’ formula and the actual value obtained by
experimental investigation.
Independent Regression Standard error Lower 95% Upper 95%

variable coefficient b(i) Sb(i) conf. limit of conf. limit of
β(i) β(i)
Regression coefficients confidence intervals
Intercept 17.03507 1.214267 14.5436 19.52654
(continued)
(continued)
Independent Regression Standard error Lower 95% Upper 95%
variable coefficient b(i) Sb(i) conf. limit of conf. limit of
β(i) β(i)
C2 0.85128 0.02254838 0.8050146 0.8975455
C3 −0.04577333 0.0108919 −0.06812167 −0.023425
Figures 6, 7 and 8 give the comparison between the actual value obtained by
experimental investigation and the predicted values from the above equation for
normal, standard and high strength RASCC and NASCC after 28 days in N/mm2
with different percentages of replacement of NA by RCA.
4 Results and Discussions
The fresh state properties of concrete are shown in Table 4. The values of different
workability tests are within the acceptable limits as given by EFNARC. For normal
and medium strength self-compacting concrete, i.e. M20 to M50, it was observed that
50% replacement of NA with recycled aggregate gives the design strength at 28 days.
For high strength self-compacting concrete, i.e. M60 to M70, 25% replacement is
observed as optimum replacement value.
There is not much change in the compressive strengths of SCC with and without
RCA for M20 and M30 grade at three days and seven days for all percentages of
RCA replacements. However, at the ages of 28 days and beyond, loss of strength of
about 14% is observed for RCA 75 and 100%. For design mix M20 to M50 grade
concrete, all the replacements with RCA gained early strength and required strength
of 72% at 14 days age. The regression equation was developed for prediction of
compressive strength
Compressive strength ( f cke ) = 17.04 + 0.85 × gc − 0.05 × pra
The R2 value obtained in the regression analysis was 0.98, indicating a good
correlation between the experimental and theoretical compressive strengths. The
regression equation can predict the compressive strength values with a maximum
error of 10.9% for different mixes.
5 Conclusions
Based on the results and discussions presented above, the following conclusions are
drawn:
154 S. Vasam et al.
Table 6 Multiple linear regression equation values for the compressive strength of concrete for
normal, medium and high strength RASCC and NASCC after 28 days in N/mm2 with different
percentage of replacement of NA by RCA
S. No. Grade of RCA Compressive strength (Mpa) Percentage
concrete (M) replacement Experimental Theoretical error with
ratio model respect to
predicted theoretical
value value
1 20 0 34.25 34.061 −1
2 20 25 33.85 32.916 −3
3 20 50 32.93 31.772 −4
4 20 75 32.15 30.628 −5
5 20 100 31.45 29.483 −7
6 30 0 41.36 42.573 +3
7 30 25 39.77 41.429 +4
8 30 50 37.80 40.285 +6
9 30 75 35.30 39.140 +10
10 30 100 35.03 37.996 +8
11 40 0 53.60 51.086 −5
12 40 25 51.18 49.942 −2
13 40 50 50.30 48.798 −3
14 40 75 48.85 47.653 −3
15 40 100 48.55 46.509 −4
16 50 0 58.45 59.599 +2
17 50 25 56.50 58.455 +3
18 50 50 55.25 57.310 +4
19 50 75 54.10 56.166 +4
20 50 100 52.85 55.022 +4
21 60 0 70.25 68.112 −3
22 60 25 69.85 66.968 −4
23 60 50 68.10 65.823 −3
24 60 75 67.35 64.679 −4
25 60 100 66.70 63.535 −5
26 70 0 75.45 76.625 +2
27 70 25 73.55 75.480 +3
28 70 50 72.85 74.336 +2
29 70 75 72.15 73.192 +1
30 70 100 71.85 72.047 0
+ Indicate higher theoretical (predicted) values and
− Indicate lower theoretical (predicted) values
Fig. 6 Experimental and predicted value of compressive strength of normal grade (M20 and M30)
SCC with different RCA content
• Utilization of recycled concrete aggregate as natural aggregate replacement avoids

the environmental and ecological damages caused by quarrying and exploitation
of raw materials like coarse aggregate for making concrete. The substitution of
RCA will help to conserve the valuable natural resources.
• The developed mathematical model (by using multiple regression analysis) was
found to predict compressive strength of NASCC and RASCC in terms of grade
of concrete and percentage of recycled aggregate.
• A very good correlation is obtained between the experimental and theoretical
(predicted) compressive strength values with correlation coefficient (R2 ) value of
0.98.
156 S. Vasam et al.
Fig. 7 Experimental and predicted value of compressive strength of medium grade (M40 and M50)
Fig. 8 Experimental and predicted value of compressive strength of higher grade (M60 and M70)
References
1. Specifications and guidelines for self compacting concrete. 99 West Street, Farnham, UK:
EFNARC, Association House, February 2002.
2. Su, N., & Hsu, K. C. (2001). A simple mix design method for self-compacting concrete. Cement
and Concrete Research, 1799–1807.
3. Indian Standard Code IS: 12269, Specifications for 53 grade ordinary Portland cement.
4. Indian Standard Code 3812-2003 (Part-1), Indian standard pulverized fuel ash specification
Part-1; for use as pozzolana in cement, cement mortar and concrete.
5. Brouwers, H. J. H., & Radix, H. J. (2005). Self-compacting concrete: Theoretical and
experimental study. Cement and Concrete Research, 35, 2116–2136.
6. Jayaram M. A., et al. (2010). A fuzzy-neuro approach for design of light weight concrete mixes.
Journal of Computing in Civil Engineering.
7. Nashat, M., et al. (2012). Comparative study on fuzzy ınference systems for prediction of
concrete compressive strength. International Journal of the Physical Sciences.
8. Stremberk, P., et al. (2010). Fuzzy logic model of fibre concrete with effect of age. In 10th
International Conference on Modern Building Materials, Structures and Techniques, Luthiana.
9. Rasheeduzzafar, A. K., et al. (1984). Recycled concrete—A source of new concrete. Cement,
concrete and aggregates (Vol. 6, No.1, pp. 17–27). ASTM.
10. Ray, S. P., et al. (1991). Recycled aggregate concrete. Journal of Structural Engineering, 18(2),
67–75.
11. Vasam, S., Jagannadha Rao K, & Vasu K. (2012). Study on self compacting concrete (SCC) and
recycled coarse aggregates (RCA) in the present construction ındustry. In National Conference
on Advances in Civil Engineering (ACE-2012), May 2012.
12. Vasam, S., & Jagannadha Rao, K. (2013). Use of SCC and RCA for sustainable construction—
An overview. IJRET, 02(11), 629–633.eISSN 2319-1163, ISSN 2321-7308.
Seismic Response of UHPC Strengthened
Reinforced Concrete Frame Using
Concrete Damaged Plasticity Model
Nidhi Sannametla and Jyosyula Sri Kalyana Rama
1 Introduction
Numerical analysis is the study of algorithms which use numerical approximation

instead of symbolic manipulation to solve engineering problems. Finite element
analysis (FEA) is a numerical method typically used in structural engineering to solve
problems whose analytical solutions require solving for boundary values of partial
differential equations. It yields approximations for values of unknown components
at discrete points over the domain by dividing a large problem into small, simple
elements. Concrete is a highly heterogeneous, quasi-brittle material that shows
nonlinear mechanical behavior (post elastic limits). The major modes of failure of
concrete are tension cracking and compression crushing. Tension and low/unconfined
compression cause failure processes that result in irreversible deformations and
degradation of the material stiffness called strain softening. However, a high confined
compression is described by a ductile hardening response.
N. Sannametla · J. Sri Kalyana Rama (B)

Birla Institute of Technology and Science, Pilani, Hyderabad, India
N. Sannametla
J. Sri Kalyana Rama
Vignana Bharathi Institute of Technology, Hyderabad, India

160 N. Sannametla and J. Sri Kalyana Rama
1.1 Concrete Damaged Plasticity Model (CDPM)
The typical behavior of concrete can be determined using the concrete damage
plasticity (CDP) model which is available in ABAQUS, a FEA software. CDP is a
constitutive continuum-based model that combines plasticity and damage mechanics.
It is a modified form of the Drucker–Prager strength hypothesis in which the failure
of a material is determined by the non-dilatational strain energy, and the failure
surface is assumed to be of conical shape with circular cross section (not completely
consistent with the real behavior of concrete). In CDP, various parameters (scalar
damage variables) are used to describe the behavior of concrete in terms of yield
surface and flow.
1. Dilation/dilatancy angle (β or ψ): 36°–40°
2. Flow potential eccentricity (): Ratio of tensile to compressive strength. (~0.1)
3. fb0 /fc0 : Ratio of biaxial to uniaxial compressive yield stress (~1.16)
4. K c : Ratio of 2nd stress-invariant on tensile to compressive meridian (~2/3).
5. Viscosity parameter (υ or μ): 0.
The above five parameters along with stress–strain behavior of concrete in tension
and compression are the input parameters for CDPM in ABAQUS. Besides this,
compression damage (d c ) and tension damage (d t ) input are also to be defined (yield
stress vs. inelastic and cracking strain). Figures 1 and 2 depict the constitutive stress–
strain relationships of concrete used to describe the numerical structural response.
Fig. 1 Response of concrete to uniaxial compression loading

Seismic Response of UHPC Strengthened Reinforced … 161
Fig. 2 Response of concrete to uniaxial tension loading. Dassault Systèmes Simulia Corp.
ABAQUS/CAE Documentation. Version 6.8–3. Providence, RI, USA (www.simulia.com); 2008
This paper aims to explore the dynamic response of a CDP model of an RC frame
with and without slab subjected to real-time earthquake load through numerical FEA
using ABAQUS 6.14-5.
1.2 Literature Review
One of the most essential facets of analyzing failure of concrete structures is the
modeling of crack commencement and proliferation caused by continuous creation
and connection of several micro-cracks [1]. This may cause stiffness degradation in
case of dynamic loading which is hard to represent using classical plasticity. Hence,
CDP is adopted instead, which is a fracture energy-based model, predominantly
developed through constitutive modeling based on internal variable formulation of
plasticity theory for nonlinear analysis of concrete [2].
In order to develop the model further, identification of parameters of the concrete
damage plasticity (CDP) model was done [3]. Additionally, an equation for damage
parameters to capture damage behavior of concrete was suggested [4].
CDPM was used to assess the damage of various structures. Application and
calibration of the CDPM parameters for assessing the damage in an RC frame were
done [5]. Simulation and analysis of the responses of five concrete slabs, using CDP
model with material parameter calibration, through ABAQUS were done [6].
CDPM was also used to study the effects of structural response under various
loading conditions. The effect of quasi-static cyclic lateral load on the behavior
and failure mode of a conventional RC beam–column joint was investigated [7].
Numerous approaches to elucidate the performance of these masonry infill walls
subjected to dynamic loads have been presented [8]. Mechanical properties of
concrete depend significantly on the loading rate (dynamic, i.e., time-varying load). A
rate-dependent continuum model (based on visco-elastic constitutive relationships)
is presented for concrete modeling when subjected to dynamic load. Stefan effect
(tensile strength increasing with loading rate) and additional rate effect due to micro-
inertia are explained. The capabilities of the damage plasticity model in visualizing
localization, energy dissipation and hysteresis are described by Pedersen et al. [9].
CDPM was used to study the seismic response of various structures such as dams
and bridges. The seismic response of a bridge in Poland subjected to an earthquake
of moderate intensity was considered while using the CDP model to characterize the
inelastic behavior of concrete [10]. A 3-D analysis based on a continuous approach
was used to study the dynamic response of a suspended bridge which was subjected to
earthquake and moving vehicular loads both separately and simultaneously [11]. A
novel CDP model for the analysis of earthquake load on concrete dams through
the response of Koyna dam during 1967 earthquake was presented [12]. The same
was also investigated through a 2-D analysis [13]. A plastic damage model was
used to investigate the damage response of concrete gravity dams using various
damping mechanisms [14]. The damage of a concrete gravity dam subjected to
hydrostatic, gravity and earthquake loading was studied [15]. A micro-plane model
to describe the dynamic behavior of a concrete gravity dam was formulated and
implemented [16]. Push-over analysis was used to study the vulnerability of RC
structures to earthquake loading owing to irregularities in plan and elevation [17].
The behavior of beam–column joint retrofitted with FRP wrapping under seismic
loading was investigated [18]. The seismic response of RC columns retrofitted
with concrete jacketing, steel jacketing, CFRP wrapping, etc., was studied [19].
A novel methodology was proposed to evaluate the fracture properties of concrete
using concrete damage plasticity model in tandem with stress-strain model. Results
indicated that the proposed model is in good agreement with experimental values of
size independent fracture energy [20].
In this paper, novel CDP data is acquired from stress–strain curves, which are
derived from on-field destructive tests. These are then used as material properties for
concrete.
2 Methodology
The following steps were followed while modeling an RC frame subjected to real-
time earthquake loading in ABAQUS software. Dimensions (in mm) adopted are as
Fig. 3 Reinforcement in
concrete
follows: Beam of 4000 × 450 × 450, Column of 3500 × 450 × 450 and Slab of
4000 × 4000 × 450. Reinforcement is provided as shown in Fig. 3.
UHPC casing of 20 mm thickness was provided throughout the length of the
column. Ultra-high performance concrete (UHPC) is a cementitious material with
a minimum compressive strength of 120 MPa to over an excess of 200 MPa.
It is formulated by a combination of Portland cement, supplementary materials,
limestone/quartz powder, reactive powders, high range water reducers/plasticizers,
fine sand and water. Fine materials in the matrix lead to a dense and smooth surface.
In combination with metallic fibers, it can provide a tensile strength of 48 MPa,
hence sometimes eliminating the need for reinforcement in its applications, thus
simplifying construction. M40 concrete of density 25 kN/m3 , with Young’s modulus
of 31,623 N/mm2 and Poisson’s ratio of 0.2, varying Ψ ’s of 13, 25, 35, 40, of 0.1,
fb0 /fc0 of 1.16, K c of 0.667 and μ of 0, was adopted. These specific dilation angles
were chosen based on the following rationale. These four dilation angle values were
found to be critical values or points of trend change in previously conducted studies
[18]. Fe 415 steel of density 78.5 kN/m3 with Young’s modulus of 200,000 N/mm2
and Poisson’s ratio of 0.3 was adopted. All the boundary conditions and constraints
(detailed later) are set in this step and remain constant throughout the analysis, i.e.,
they are extended to further steps. The linear perturbation step is carried out to obtain
the natural frequencies of the first six modes of vibration of the structure. Then, the
dynamic implicit step is used for analysis of real-time earthquake data (Elcentro
earthquake, May 18, 1940) for a period of 31.18 s is specified in a tabular form;
the data gives acceleration of the ground. Reinforcement is embedded into concrete;
rebars are tied to each other and so are any connecting faces of members. The column
bases are fixed, except Z-translation, to permit definition of earthquake load.
In order to obtain a relatively medium-sized mesh, a 250 cm seed interval was
adopted for the entire model and all the individual elements. However, it was observed
that mesh sizing did not have a significant effect on the analysis, when comparing
100, 250 and 500 cm seed intervals. The slab is designated as an independent part
Fig. 4 Mesh in space frame
Table 1 Number of nodes and element details

Type/nodes Shape Order Without slab With slab
C3D8R hexahedral linear 480 880
T3D2 line linear 8064 9280
(in assembly), and the rest of the instances are designated as dependent instances
(to be meshed part by part). Figure 4 depicts the mesh adopted in the FEA software.
Table 1 gives element details.
2.1 Element for Concrete Modeling
The concrete is modeled with the brick elements for uniform stress distribution. Cube
Three Dimensional eight node Reduced integration (C3D8R) elements are utilized
as shown in Fig. 5. C3D8R uses linear interpolation in all directions, and hence, it is
called linear element of first order. They only have three degrees of freedom and are
Fig. 5 Concrete element

(C3D8R)
Fig. 6 Truss element

(T3D2)
stress-displacement elements. The convergence rate of quadrilaterals is better than

that of triangles and tetrahedrons; however, sensitivity of displacement calculations
to mesh orientation in regular meshes is higher.
2.2 Element for Rebar Modeling
Reinforcement or steel rebars are modeled as three-dimensional truss elements

(T3D2) as shown in Fig. 6. It is a two-node straight truss element that uses linear
interpolation for position displacements, with stress being a constant throughout. The
cross-sectional area of the truss element is detailed as part of the section definition.
When the displacements are large, the updated cross-sectional area is calculated
by assuming that the material is incompressible, regardless of the actual material
definition.
The concrete-rebar interface was modeled through ABAQUS by using the
embedded slave–master setting.
3 Results and Observations
3.1 With UHPC Improvement
The values of displacement and reactions observed were at the points where these
quantities attained their maximum values. These points are close to the base, for
columns, and close to the mid-span, for beams.
From Fig. 7, it is observed that there is a substantial increase in load-carrying
capacity with the increase in dilation angle. The increase from 25 to 35° is more
predominant. From Fig. 8, it is observed that there is a marginal decrease in the
displacement with the increase in the dilation angle.
From Fig. 7, it is observed that load-carrying capacity of space frame with slab
is slightly higher than space frame without slab. From Fig. 8, it is observed that
displacement in space frame with slab is slightly lower than space frame without
slab. From Table 2, the natural frequencies of space frame with and without slab for
the first five modes are in between the range of 0.08–0.32 Hz. The natural frequencies
of space frame with slab are almost comparable to those of space frame without slab.
Fig. 7 Reaction in kN 900

versus dilation angle in
degrees 800 With slab
Without Slab
Reaction in kN
700
600
500
400
300
0 10 20 30 40 50
Dilation angle in degrees
Fig. 8 Displacement in mm 15.4

versus dilation angle in
15.2
Displacement in mm
degrees
15
14.8
14.6
14.4 With slab
14.2 Without Slab
14
0 10 20 30 40 50
Table 2 Natural frequencies

Mode Frequency (Hz, cycles/time)
of space frame with slab
With slab Without slab
1 0.1 0.084
2 0.107 0.136
3 0.126 0.195
4 0.245 0.213
5 0.259 0.224
6 0.2633 0.314
From Fig. 9, correlating observed failure to modes of vibration, it is seen that the first
mode contributes the most to the response of the space frame justifying the failure
of column first. The damage then propagates to beam, as is visible in the second.
The later modes correspond to twisting of members. From Fig. 10: For space frame
with slab, it is observed that the first cracking load of concrete is increasing with
decreasing dilation angle, due to higher ductility. For space frame without slab, it is
Fig. 9 First six modes of vibration of space frame
Fig. 10 Concrete first crack 550 Without Slab

load in kN versus dilation
Concrete first crack load in kN
angle in degrees 500

With Slab
450
400
350
300
250
200
0 10 20 30 40 50
observed that the first cracking load of concrete has negligible change with dilation
angle. From Figs. 11 and 12, for space frame with and without slab, it is observed that
the first yielding loads of column and beam reinforcement increase with increasing
dilation angle, due to higher stiffness. For space frame with slab, it is observed that
the cracking of concrete is delayed at lower dilation angles. A delayed cracking
signifies that the concrete was able to sustain more deformation. This is a result of
decreased stiffness and improved ductility at lower dilation angles. From Table 3, it is
observed that, with UHPC improvement, the first cracking/yielding loads of concrete,
650
Column Rebar Yielding Lod

Without Slab
600
With Slab
550
in kN 500
450
400
0 10 20 30 40 50
Fig. 11 Column rebar yielding load in kN versus dilation angle in degrees
900
Beam Rebar Yielding Load in
Without Slab
800 With Slab
700
kN
600
500
400
0 10 20 30 40 50
Fig. 12 Beam reinforcement yielding load in kN versus dilation angle
Table 3 First cracking/yielding loads in kN and displacement in mm with and without UHPC for
space frame with slab at dilation angle 40
Case Concrete crack Colum rebar Beam rebar Displacement
load yielding load yielding load
No UHPC, 373.3 592.3 816.8 14.09
dilation angle 40
With UHPC, 558.7 887.9 927.6 14.23
dilation angle 40
column reinforcement and beam reinforcement are substantially increased, with a

negligible increase in displacement. From Figs. 13 and 14, it is observed that the
extent of damage is greatly reduced by UHPC casing. In general, the space frame
with slab has higher first cracking loads and thus a higher resistance to failure than
space frame without slab. In general, it is seen that the reaction developed at higher
Fig. 13 Damage in tension

without UHPC
Fig. 14 Damage in tension

with UHPC
dilation angles is higher in comparison with that of lower dilation angles, with the
exception of lateral reaction in X-direction, which reaches a peak at dilation angle
25° and then decreases, the displacement variation displaying no apparent pattern.
4 Conclusions
The natural frequencies of modes of vibration are similar in both space frames
with and without slab. The modes successfully describe the failure behavior of the
structure. Dynamic (seismic) analysis with parametric variation of dilation angle
was successfully performed on space frames with and without slab. Higher dilation
angles are superior to lower dilation angles, due to higher load-carrying capacity as
well as lower displacements attributed to higher stiffness. Space frame with slab is
superior to space frame without slab, due to higher load-carrying capacity as well as
lower displacements attributed to higher stiffness. At lower dilation angles, due to
higher ductility, resistance of concrete increases, and failure of concrete is delayed.
This advantage is only seen in space frame with slab. At higher dilation angles, due
to higher stiffness, the resistance to failure of reinforcement increases. This is the
case, in both space frames with and without slab. Earthquake being a displacement
loading, it is desirable to have deformability over strength, and sufficient warning
before failure is an important factor. The space frame without slab deforms more and
at an earlier time step without sufficient warning. Since the space frame with slab
(concrete and reinforcement) has a higher resistance to failure and delayed failure
than space frame without slab, providing ample warning, it is suggested to adopt
space frame with slab. UHPC column casing improvement greatly increases the load-
carrying capacity of the structure and reduces extent of damage, as computed using
the CDPM. As per the above conclusions, it is suggested to adopt space frame with
slab of concrete with higher dilation angles due to provision of ample warning, with
UHPC improvement due to reduction in extent of damage, applicable for construction
of structures resistant to seismic loads, or retrofitting of existing structures near
seismically active zones.
Acknowledgements The authors would like to acknowledge the Computer Center at BITS Pilani—
Hyderabad Campus. Second author would like to acknowledge the computational laboratory
sponsored by DST-FIST at Vignana Bharathi Institute of Technology.
References
1. Lee, J., & Fenves, G. (1998). A plastic damage model for cyclic loading of concrete structures.
Engineering Mechanics, 124(8), 892–900.
2. Lubliner, J., Onate, E., Oller, S., & Oliver, J. (1989). A constitutive model for cracking of
concrete based on the incremental theory of plasticity. Engineering Computations, 5(4), 309–
319.
3. Jankowiak, T., & Łodygowski, T. (2005). Identification of parameters of concrete damage
plasticity constitutive model. Foundations of civil and environmental engineering, 6(1), 53–69.
4. Sümer, Y., & Aktaş, M. (2015). Defining parameters for concrete damage plasticity model.
Challenge, 1(3), 149–155.
5. Michal, S., & Winnicki, A. (2015). Numerical simulations of corners in RC frames using
Strut-and-Tie method and CDP model. In XIII International Conference on Computational
Plasticity: Fundamentals and Applications (pp. 608–619). Barcelona: CIMNE ed.
6. Genikomsou, S., & Polak, M. (2015). Finite element analysis of punching shear of concrete
slabs using damaged plasticity model in ABAQUS. Engineering Structures, 98, 38–48.
7. Fadwa, I., Ali, T. A., Nazih, E., & Sara, M. (2014). Reinforced concrete wide and conventional
beam–column connections subjected to lateral load. Engineering Structures, 76, 34–48.
8. Ali Shah, S. A., Shahzads, K., Afzal, S., Shabab, M. E., & Khattak, N. (2013). Behaviour of
infill walls under lateral loads in reinforced concrete frames—A review. International Journal
of Advanced Structures and Geotechnical Engineering, 2(3), 106–109.
9. Pedersen, R., Simone, A., & Sluys, L. (2008). An analysis of dynamic fracture in concrete with
a continuum visco-elastic visco-plastic damage model. Engineering Fracture Mechanics, 75,
3782–3805.
10. Dulinska, J. M., & Szczerba, R. (2013). Simulation of dynamic behaviour of rc bridge with steel-
laminated elastomeric bearings under high-energy mining tremors. Key Engineering Materials,
531–532, 662–667.
11. Konstantakopoulos, T. G., Raftoyiannis, I. G., & Michaltsos, G. T. (2012). Suspended bridges
subjected to earthquake and moving loads. Engineering Structures, 45, 223–237.
12. Lee, J., & Fenves, G. (1998). A plastic-damage concrete model for earthquake analysis of
dams. Earthquake Engineering and Structural Dynamics, 27, 937–956.
13. Calayir, Y., & Karaton, M. (2005). Seismic fracture analysis of concrete gravity dams including
dam–reservoir interaction. Computers & Structures, 83(19–20), 1595–1606.
14. Omidi, O., & Vi, Lofti. (2012). Dynamic analysis of the Koyna Dam using three-dimensional
plastic-damage modelling. Dam Engineering, 22(3), 197.
15. Zappitelli, M. P., Villa, E. I., Fernández-Sáez, J., & Rocco, C. (2014). Cracking development
prediction in concrete. Mecánica Computacional, XXXIII, 23–26.
16. Indriyantho, B. R., Zreid, I., & Kaliske, M. (2017). Modeling of a concrete dam under
earthquake loading by a nonlocal microplane approach. Procedia Engineering, 171, 1010–
1018.
17. Ravikumar, C. M., KS, B. N., Sujith, B. V., & Reddy, V. (2012). Effect of irregular configurations
on seismic vulnerability of RC buildings. Architecture Research, 2(3), 20–26.
18. Ashish, U. K., & Harshalata, R. (2014). Investigation on behaviour of reinforced concrete
beam column joints retrofitted with FRP wrapping. International Journal of Civil Engineering
Research, 2278–3652.
19. Truong, G. T., Kim, J. C., & Choi, K. K. (2017). Seismic performance of reinforced concrete
columns retrofitted by various methods. Engineering Structures, 134, 217–235.
20. Kalyana Rama, J. S., Chauhan, D. R., Sivakumar, M. V., Vasan, A., & Murthy, A. R. (2017).
Fracture properties of concrete using damaged plasticity model-A parametric study. Structural
Engineering and Mechanics, 64(1), 59–69.
Mix Design Methodology for Fly Ash
and GGBS-Based Geopolymer Concrete
G. Mallikarjuna Rao and M. Venu
1 Introduction
With the continuous increase in human population, the housing sector and construc-
tion industry have gained boom to meet the current demand. The increased usage and
demand for cement production day by day lead to huge depletion of natural resources
which ultimately results in environmental issues. To overcome these problems, there
is a need to focus on new emerging binding materials such as industrial by-products,
which will reduce the environmental pollution. Emerging technology aids for the
utilization of industrial by-products into useful materials such as fly ash, GGBS,
metakolin and rice husk ash. The development of alkali-activated binders with supe-
rior engineering properties as well as longer durability has emerged as an alternative
to ordinary Portland cement (OPC). Geopolymer (inorganic polymer concrete) is an
emerging class of cementitious material and could be the next-generation concrete
for civil infrastructure applications. This innovative technology provides a new plat-
form for the sustainable growth of our urban society in the coming decades and helps
in building durable structures. These materials can be used as a replacement to the
binder in concrete as a major construction material.
Davidovits, 1970, reported the use of waste materials like fly ash and GGBS as
high alkaline solution activators to form geopolymer. The constituents of the com-
monly used alkaline solution are sodium hydroxide (NaOH) and sodium silicate
(Na2 SiO3 ). The alkaline solution helps to bind the loose aggregates in mixture to
form geopolymer concrete (GPC) which has high strength, durability and low creep
[1]. The alkaline solution activates silica and alumina to form aluminosilicate hydrate
in fly ash and forms calcium silicate hydrate (C–S–H) by reacting with calcium in
GGBS. The curing conditions, especially temperature, play a significant impact on
the polymerization process [2]. Researchers have [3] concluded that a combination
G. Mallikarjuna Rao (B) · M. Venu

Department of Civil Engineering, Vardhaman College of Engineering, Hyderabad, India
174 G. Mallikarjuna Rao and M. Venu
of sodium hydroxide and sodium silicate solutions can be a good application for
activators. High concentration of sodium hydroxide solution and curing temperature
enable the concrete compressive strength to be higher. Various authors have studied
the importance of molar ratio of Na2 SiO3 /NaOH and suggested that to achieve max-
imum compressive strength the binder content to be constant. The reported constant
binder content value is 2.5 [4]. Rangan [5] proposed a mix design methodology for
geopolymer concrete with fly ash and [6] suggested modifications in Indian Stan-
dard code for suitability of GPC. Black [7] carried out research on fly ash-based
geopolymer concrete by considering different mix proportions and developed a mix
design process by varying the water to geopolymer solids ratio with two different
molarities of NaOH, i.e. 8M and 12M. The research concluded that the flash set was
a significant problem for GPC mixes.
In the present study, a mix design procedure was developed for both normal-
and standard-grade geopolymer concrete by varying the ingredients. The ingredients
varied for all the mixes are quantity of fly ash, quantity of water, grading of fine
aggregate, fine aggregate-to-total aggregate ratio by maintaining sodium silicate-to-
sodium hydroxide ratio as 1 and sodium hydroxide molarity. Based on the past work
done on GPC, the present research work is planned by considering the parameters,
viz. type of binder, binder content, alkaline/binder ratio. The ratio of sodium silicate
to sodium hydroxide is kept at 2.5 with molarity of NaOH as 8M. This investigation
aims to study the influence of the above-considered parameters on strength and
durability of geopolymer concrete.
2 Research Significance
From the literature survey, to the best of authors’ knowledge, there is no proper mix
design for fly ash and GGBS-based geopolymer concrete till date, and it shows that
the study needs to focus on mix design aspects for geopolymer concrete. This will
help the structural engineers in design aspects to implement the same for the field
application. The present investigation aims a proper mix design methodology for fly
ash and GGBS-based GPC.
3 Materials
3.1 Fly Ash and GGBS
Materials used in this research are GGBS obtained from Andhra Cements, Vishakha-
patnam, India, and fly ash from Ramagundam Thermal Power Plant, India, with a
specific gravity of 2.90 and 2.17, respectively. The chemical compositions of GGBS
and fly ash were presented in Table 1.
Mix Design Methodology for Fly Ash and GGBS … 175
Table 1 Chemical
Chemical composition Fly ash GGBS
composition of both fly ash
and GGBS (% by mass) SiO2 60.10 34.02
Al2 O3 26.50 20.00
Fe2 O3 4.20 0.81
SO3 0.33 0.93
CaO 4.00 32.65
MgO 1.21 7.86
Na2 O 0.20 NIL
LOI 0.85 NIL
3.1.1 SEM and XRD for Fly Ash and GGBS
Scanning electron microscopy (SEM) is performed to study the microstructure of fly

ash and GGBS. The scanned images of fly ash and GGBS are shown in Figs. 1 and
2. Scanning electron microscopy image gives an approximate idea about the shape,
angularity, size and surface texture of fly ash and GGBS. The GGBS particles appear
Fig. 1 Scanning electron microscope of fly ash and EDXA of fly ash
Fig. 2 Scanning electron microscope of GGBS and EDXA of GGBS

Fig. 3 XRD analysis for fly ash and GGBS
to have straight, flaky-elongated shape with sharp-edged angularity, rough surface

texture and large variation in size (1–10 µm). The fly ash particles appear to have
spherical shape with smooth surface in a broad range of sizes (with the lower limit
being approximately 1/10 the maximum fly ash particle size, i.e. 1–10 µm in size).
The energy-dispersive X-ray spectroscopy of fly ash and GGBS is also shown in
Figs. 1 and 2. From the energy-dispersive X-ray spectroscopy, it is seen that GGBS
is enriched with more silica percent than other elements. Fly ash particles contain
more percent of silica and alumina (Fig. 3).
3.2 Alkaline Liquid
The alkaline solution is the combination of sodium hydroxide and sodium silicate
solutions. Sodium hydroxide is used in the present study because it is less expensive
than potassium hydroxide and widely available. Sodium hydroxide of 98% purity
available in pellets form is used in the investigation. These sodium hydroxide pel-
lets were dissolved in potable water and prepared the solution of required molarity.
Sodium hydroxide solution of required molarity and sodium silicate in liquid form
and the chemical composition of the sodium silicate is Na2 O—7.5–8.5%, SiO2 —
26.5% and remaining is H2O content are mixed and stored at room temperature for
24 h before its use (Fig. 4).
3.3 Fine Aggregate
A locally available river sand is used as a fine aggregate, and it conforms to Zone-2
according to IS: 383 [8]. The specific gravity and bulk density are 2.62 and 2.59,
respectively.
Fig. 4 Sodium hydroxide and sodium silicate solution
3.4 Coarse Aggregate
A locally available crushed rock maximum size of 20 mm is used as a coarse

aggregate, and it is according to IS: 383 [8].
3.5 Water
Potable water was used for entire research work to prepare NaOH solution.
3.6 Preparation of Concrete
The fly ash and GGBS are mixed well until once it reaches homogeneity, and there-
after, fine and coarse aggregate is added, and these are allowed to mix for 2 min in
Hobart electric mixer. For instance, the prepared alkaline activator is added to the
prepared mix which is a combination of fly ash, GGBS, fine aggregate, and it is
mixed well for another 3 or 4 min in Hobart electric mixer. The prepared fresh mixes
are cohesive and resistant to segregation. Afterward, GPC is poured into the molds
of size 100 mm × 100 mm × 100 mm by three layers consecutively with compaction
and vibration to avoid air bubbles in concrete. After 24 h of casting, the specimens
were demolded, and these were cured under ambient temperature until the age of
testing.
3.7 Experimental Program
An experimental program consisting of five mixes, in each mix fly ash—GGBS

proportion varied between 75FA:25GGBS, 50FA:50GGBS, 25FA:75GGBS and
0FA:100 GGBS for the binder. For the corresponding mixes, alkaline liquid-to-binder
ratios were selected as 0.45, 0.5, 0.55 and 0.6, respectively. Geopolymer concrete
density was noted, and it varied from 2200 to 2400 kg/m3 . Once the density of con-
crete is known, then the total amount of binder and alkaline solution can be calculated
easily. The ratio of sodium silicate to sodium hydroxide solution is kept constant as
2.5 for the total experimental program.
The density of the fly ash and GGBS based geopolymer concrete specimens are
shown in Fig. 5. With increase in density, there is an increase in the alkaline/binder
ratio, and the maximum density was observed for the mix of 0% FA+100% GGBS.
Maximum density was observed at alkaline/binder ratio 0.6 for all the combinations.
Geopolymer concrete is a new concrete, and the density of the geopolymer was quite
similar to that of the conventional concrete. The density of the geopolymer concrete
depends on the binder content, alkaline/binder ratio and method of compaction. The
density range observed for geopolymer concrete was 2300–2500 kg/m3 .
The UPV test is a measure to presence of voids and the consistency of concrete.
The 7 days UPV for different combinations of the fly ash and GGBS mix was found to
be 3.26–4.5 km/s for outdoor-cured specimens for different combinations, whereas
Density (300 kg/m 3 Binder content)

2550
75F+25G 50F+50G 75G+25F 100G+0F
2500
2450
Denity (Kg/m3)
2400
2350
2300
2250
2200
0.45 0.5 0.55 0.6
Alkaline/Binder ratio
Fig. 5 Density for fly ash and GGBS-based geopolymer concrete (300 binder content)
for higher GGBS content mixes, there is a little rise in the values compared to that
of lower GGBS content. Whitehurst classified fly ash-based geopolymer concrete
as excellent, good, doubtful, poor and very poor for UPV values of 4.5 km/s and
above, 3.5–4.5, 3.0–3.5, 2.0–3.0 km/s and below 2.0 km/s, respectively. Generally,
high-pulse velocity reading in concrete is indicative of concrete of good quality. The
presence of voids has been recognized to have an influence on the UPV transmis-
sion. The measured 28-day UPV values for all geopolymer concrete specimens are
presented in Fig. 6. In geopolymer concrete, for the attainment of strength, polymer-
ization process plays an important role. The quality of the concrete mainly depends
on the internal microstructure formed.
Figure 7 shows the compressive strength of different alkaline to binder ratios with
binder content 300 kg/m3 . In the present study, effect of fly ash and GGBS concrete
under outdoor curing was studied. The compressive strength of GPC cured under
outdoor curing (0% FA+100% GGBS) combination shows the higher strength than
the mixes prepared with the other combinations of GGBS. In order to eliminate the
oven curing, the specimens were casted with the fly ash and GGBS combination.
For many practical applications, it is very important that the concrete should be
cast and cured under normal temperatures. The maximum compressive strength of
45 MPa is attained at 100% GGBS even under outdoor curing. This concludes that
combination of fly ash and GGBS can produce satisfactory results even with 8M
NaOH. Hence, the obtained results proved that oven curing can be eliminated for
producing GPC, where there is difficulty in providing oven curing. At 7 days age
of curing, 100% GGBS achieved the maximum compressive strength. In the case
of 75% FA-25%, GGBS proportion indicates the lesser compressive strength. The
6
300 kg/m3 (Binder Content)
75F+25G
UPV
3
50F+50G
75G+25F
2 100G+0F
0
0.45 0.5 0.55 0.6
Alkaline/Binder
Fig. 6 UPV for fly ash and GGBS-based geopolymer concrete (300 binder content)
COMPRESSIVE STRENGTH (300)

45
COMPRESSIVE STRENGTH (MPA)
40
35
30
25 100G+0F
20 75G+25F
15 50G+50F
75F+25G
10
5
0
0. 45 0. 5 0. 55 0. 6
ALKALINE/BINDER
Fig. 7 Compressive strength for fly ash and GGBS-based geopolymer concrete (300 binder content)
increase in compressive strength is mainly due to high amount of calcium present

in GGBS. The improvement of strength of fly ash and GGBS-based GPC is due
to increased calcium content present in GGBS which significantly accelerates the
polymerization process to improve strength.
5 Conclusions
• The percentage increase of GGBS in binder content shows the increase in

compressive strength of geopolymer mortar.
• The combination of fly ash and GGBS is a possible solution for producing GPC
even under outdoor curing condition.
• The geopolymer concrete dry density varied from 2400 to 2500 kg/m3 for (100%
GGBS+0% fly ash) binder proportion, and for remaining proportions, it varied
from 2300 to 2400 kg/m3 which is likely to be conventional concrete.
• For higher density samples show the higher in the compressive strength when the
age of 7 days for all mixes.
• The significant increase in compressive strength was observed at alkaline/binder
ratio of 0.55 for a binder content of 300 kg/m3 .
• UPV performance is good for GPC at the age of 7 days, and it shows the similar
trend to that of conventional concrete noticed, i.e. 3.5–4.5 km/s. This range shows
the good sign of concrete quality and homogeneity.
References
1. Wang, S. D., Pu, X. C., Scrivener, K. L., & Pratt, P. L. (1995). Alkali-activated slag cement and
concrete: A review of properties and problems. Advances Cement Research, 27, 93–102.
2. Glukhovsky, V. D. (1959). Soil silicates. Gosstroiizdat, Kiev, (in Russian 1959).
3. Hardjito, D., Wallah, S. E., Sumajouw, D. M. J., & Rangan, B. V. (2004). On the development
of fly ash-based geopolymer concrete. ACI Materials Journal, 101(6), 467–472.
4. Pinto, A. (2004). Alkali-activated metakaolin based binders (PhD thesis). University of Minho.
5. Rangan, B. V. (2008). Mix design and production of fly ash based geopolymer concrete. Indian
Concrete Journal, 82, 7–15.
6. Anuradha, R., Sreevidya, V., Venkatasubramani, R., & Rangan, B. V. (2012). Modified guidelines
for geopolymer concrete mix design using Indian standard. Asian Journal Civil Engineering
(Build Hous),13(3), 353–364.
7. Black, J. R. (2012). Mix design process for alkaline-activated class f fly ash geopolymer concrete.
University of New South Wales at the Australian Defence Force Academy.
8. IS-383-1970, Specification for coarse and fine aggregates from natural sources for concrete.
Experimental and Analytical Behavior
of Recycled Aggregate Concrete Using
ANN
B. Suguna Rao, Vithal Joshi, and Srikanth M. Naik
1 Introduction
Concrete is being one of the most consumed material in the area of infrastructure.
Among the various ingredients present in the concrete, aggregates play an impor-
tant role. Aggregates play a vital role in concrete by providing volume stability to
the concrete. The increase in population and rapid urbanization has increased the
demand for increased infrastructure development. The construction activity is being
majorly using concrete as a material of construction. For satisfying the demand of
the increased construction activity, there is a huge requirement of aggregates. Avail-
ability of aggregates in large scale and at the time of need has become scarce and
is also causing environmental impacts. In order to facilitate the need of aggregates,
an effort has been made to utilize construction and demolition waste as aggregates
in concrete. A large quantity of C & D waste is being generated yearly and the
quantity measured is around 14.5 million tons as per the reports of Central Pollu-
tion Control Board (CPCB), Delhi. The quantity of waste being utilized in a useful
way is only around 3% and the rest of the waste goes into landfill areas. Increased
population explosion has also reduced the land availability. Utilizing C & D waste
is one of the solutions for environmental sustainability and economical benefit. C
& D waste being concrete waste has properties of concrete finding its suitability in
concrete production. According to the literature, recycled aggregates have specific
gravity lesser than natural aggregates and have higher water absorption than natural
B. Suguna Rao (B)

Department of Civil Engineering, M.S. Ramaiah Institute of Technology, Visvesvaraya
Technological University, Bangalore, Karnataka, India
V. Joshi · S. M. Naik
Department of Civil Engineering, M.S. Ramaiah Institute of Technology, Bangalore, Karnataka,
India

184 B. Suguna Rao et al.
aggregates which may be due to adhered mortar to RAC. While utilizing recycled
aggregates for structural concrete, the characterization of the material properties is
required to understand the water absorption and specific gravity of recycled aggre-
gate concrete. The compressive strength of concrete made from recycled aggregate
concrete with respect to controlled mix with same water–cement ratio was studied
and variations are observed with respect to elastic modulus, shrinkage and creep [1].
Hence, an effort is made to understand recycled aggregate concrete by characterizing
them initially and checking the fresh properties to be suitable for hardened concrete.
Concrete made from recycled aggregates are tested for their strength and results are
utilized for predicting their behavior using artificial neural networks (ANN). ANN is
a powerful modeling tool specially useful tool when high reliable results are expected
from varying variables [2]. ANN develops a multilayer perceptron made up of several
neurons grouped in three layers such as input layer, hidden layer and output layer
imitating the functioning of biological brain being utilized in wide range of applica-
tion from structural concrete to timber panels [3]. This study aims at understanding
the behavior of recycled aggregate concrete by conducting various experiments and
validating it through ANN.
2 Materials and Methodology
2.1 Materials
The materials used to carry out the experimental investigation of recycled aggregate
concrete are as follows:
1. Cement: Cement used here is OPC of 43 grades.
2. Fine Aggregate: Manufactured sand.
3. Coarse Aggregate:
(1) Natural aggregate of size 20 mm downsize for M40 grade of concrete and
12.5 mm downsize for M60 grade of concrete.
(2) Recycled aggregates are procured from rock well crystal recycling plant of
size 20 mm downsize for M40 grade and 12.5 mm downsize for M60 grade.
4. Admixtures:
(1) Mineral admixture: 10% silica fume and 10% fly ash.
(2) Chemical admixture: Superplasticizer Aura mix 400 as per workability
requirement.
5. Water: Portable water.
Experimental and Analytical Behavior of Recycled Aggregate … 185
2.2 Characterization of Materials
Initially, the basic characterization of the material used is carried out and is as shown
in Table 1 for cement and Table 2 for fine aggregate and Table 3 for coarse aggregate.
Results of the characterization of materials were within the acceptable limits of codes.
These materials are considered for preparing the mix design of recycled aggregate
concrete.
Table 1 Properties of cement

Properties of cement Test results
Setting time Initial setting time 40 min
Final setting time 345 min
Fineness (%) 6
Standard consistency (%) 33
Table 2 Properties of fine aggregate

Properties Fine aggregate
Fineness modulus 3.943
Moisture content (%) 4
Table 3 Properties of natural and recycled coarse aggregates

Properties Natural coarse aggregate Recycled coarse aggregate
12.5 mm 20 mm 12.5 mm 20 mm
downsize downsize downsize downsize
Water 1.3% 1.8% 3.2% 3.4%
absorption
Crushing value 23.67% 15.69% 32.39% 20.94%
Abrasion value – 26.75% – 30.72%
Impact value 25.49% – 34.39% –
Specific gravity 2.56 2.61 2.30 2.34
Fineness 8.182 6.716 8.298 6.752
modulus
Table 4 Mix proportion used

Materials M40-per cubic meter M60-per cubic meter
for M40 and M60 grade of
(Kg/m3 ) (Kg/m3 )
concrete
Cement 414.766 444.4
Water 139.984 150
FA 603.309 551.4
NA 1136.468 1000
RCA – –
SP 3.111 3.333
SF 51.846 55.56
FA 51.846 55.56
2.3 Mix Design and Testing
After testing the basic characterization of the materials, mix design procedure pro-
posed by Perumal [4] for high strength concrete is being adopted for M60 grade of
concrete and IS 10262:2000 is adopted for M40 grade of concrete. Fresh proper-
ties tests were conducted to fix the dosage of superplasticizer by considering 100%
replacement ratio as the worst combination of mix. Table 4 shows the mix design
proportions adopted for M60 and M40 grade of concrete for normal aggregate con-
crete. The mix design is adopted to achieve M60 and M40 grade of concrete for
replacement ratio of 0, 50 and 100% for W/B (water–binder) ratio of 0.27 and 0.36.
2.4 Experiment
Concrete cubes of sizes 150 mm × 150 mm × 150 mm were casted for W /B (water–
binder) ratio of 0.27 and 0.36 for M40 and M60 grade of concrete. The concrete
cubes were tested for 1, 3, 7, 14 and 28 days strength. Three specimens were casted
for each of the mix proportion and for each testing day considered. Compressive
strength of 180 specimens was observed. The achieved 28 day compressive strength
is plotted as shown in Figs. 1 and 2.
From the above experimental analysis, we can say that for replacement ratio of
50% there is a reduction in compressive strength by 15–25% and for 100% replace-
ment ratio of recycled aggregate the compressive strength reduces by 28–38%. When
compared with the grade of concrete, high strength concrete performed better than
concrete of grade M40 for W /B ratio of 0.27 than for 0.36. One of the reasons for
the variation in the strength of the concrete may be due to the sizes of the aggregates
of M40 and M60 grade of concrete.
Fig. 1 Compressive strength of M40 grade of concrete for different W/B ratio
Fig. 2 Compressive strength of M60 grade of concrete for different W/B ratio
3 Predictive Analysis Using ANN
The need of the hour is to develop sustainable concrete by utilizing the C & D waste
as aggregates. Several constrains exist while using secondary products like recy-
cled aggregates. The major concern is quality and consistency of the products. The
challenge in using recycled aggregates dependents on source, age of old concrete,
crushing technology adopted, etc. Hence, to understand the strength parameter of
recycled aggregate concrete, a powerful tool like ANN is used where the aggregates
used are consisting of varying properties and source. This varying property leads
to variation in the result of concrete. Predicting the compressive strength of con-
crete with its constituents is a nonlinear function [5–11]. The strength parameter of
concrete not only depends upon water-to-cement ratio, but also depends upon the
additive constituents. The lack of empirical relationships to judge the compressive
strength of concrete has created the need for using soft computing tools by various
researchers. Soft computing tool uses the optimization tools for learning and predict-
ing and creating new patterns based on the previous data. ANN consists of networks
of simple nonlinear computing elements called neurons. ANN is a useful tool used
for processing the data through training, testing and arriving at suitable results for
a specific nature of problem and is easily transportable between the computing sys-
tems. The capability of ANN to arrive at a model in which relationship between
dependent variable and independent variable is popular machine learning method.
This unconventional method of arriving at a solution has created an immense interest
in the field of neural networks. ANN has the capability of constructing a relation-
ship between the data provided for training. While using ANN, there is no need
to provide any functional relationship among the variables. ANNs modeling ability
to derive meaning from unknown and nonlinear interrelationships among variables
have been harnessed to aid the prediction of behavior of engineering and natural
systems. Compressive strength of concrete is unstructured in nature and involves a
high nonlinear relationship among its constituents. Therefore, an attempt is made to
predict the compressive strength of concrete considering 14 input parameters and
compressive strength as output parameter. Using machine learning technique like
ANN, to predict the compressive strength of concrete significantly reduces the time
and money for the industry.
3.1 Experiment Design was Designed Using ANN
A total of 14 input parameters were chosen to design the model to give one output
value which is compressive strength of concrete at 28 days.
Selection of Input Parameters Factors affecting the strength of concrete are con-
sidered as input parameters. Concrete constituents: Quantity of cement, water, sand,
natural aggregates and recycled aggregate are the most important factors affecting
the strength of concrete. While understanding the nature of aggregates influencing
the compressive strength of concrete fineness modulus of sand is overlooked in many
studies, which influences the workability, permeability and strength of concrete. The
properties of coarse aggregates like maximum particle size, SSD specific gravity and
water absorption are some of the factors that affect the strength greatly [12]. The data
collected for the study varied from low to high strength as the variation in strength
also depends upon crushing process involved, source of these aggregates and age
of the aggregates. Source of the aggregates is an important factor that needs to be
considered while proposing the concrete for structural purpose [13]. Table 5 shows
the list of input parameter used in this study.
Selection of Output Parameter The 28 days compressive strength of concrete was

the output parameter considered in this study.
Table 5 List of
Parameter Min. Max.
input parameters
Cement (kg/m3 ) 220 450
Mineral admixture (kg/m3 ) 0 102.5
Water (kg/m3 ) 120 271
Sand (kg/m3 ) 540 1020
N.A. (kg/m3 ) 0 1186.18
R.A. (kg/m3 ) 0 1070.9
W /B 0.3 0.76
F.M. 2.11 3.88
Max. size of CA (mm) 12.5 25
Water absorption (%) 2.24 10.6
SSD specific gravity 2.27 2.76
Rv (%) 0 100
Chemical admixture (%) 0 3
Impurity (%) 0 5
Compressive strength MPa 17 85
Data Collection A total of 136 sets of experimental data from different literature
sources were collected to train and check the reliability of the strength model [14–
32]. The available data is divided into three parts consisting of training set which
consists of 75% of the data, and the second part is 12.5% of data utilized in validation
while the remaining data is used for testing, to test the accuracy of the prediction.
3.2 Construction of ANN Model
Feed-forward back-propagation network was used in this study with 14 input vari-
ables and one unit in output. The network parameter included three hidden layer =
0, 1 and 2; and number of hidden neurons = 2–50; learning rate varied 0.1 variation
up to 2.0; the momentum factor used are 0, 0.3, 0.5, 0.7, 0.9 and 1; the variation
of learning cycle = 500, 1000, 5000, 10,000, 20,000 and 50,000. These momentum
factors cover the entire database available for training.
Functions used as solver are Quasi-Newton, Levenberg–Marquardt and conjugate
gradient decent. Activation functions used here are hyperbolic tangent, logistic and
linear. To maximize the R2 value and to ascertain the error of integral testing set after
conducting a series of trials, the best network architecture and parameters were used.
From the 14 input parameter included for the study, the best network architecture
was developed for 30 hidden units within one hidden layer and one output layer. The
rate of momentum is 0.1 with 1000 learning cycle. Quasi-Newton was the solver
function that resulted in hyperbolic tangent as activation function.
Table 6 Performance of
Test set Training set All sets
network
R2 0.95 0.94 0.94
MAE 2.19 2.36 2.47
Correlation 0.98 0.97 0.97
3.3 Analysis Results and Discussion
The performance of the network is judged based on R2 (coefficient of determination),

mean absolute error (MAE) and correlation coefficient values. Table 6 shows the
results for the best network model.
Coefficient of determination value was 0.95 for test set and 0.94 for training set
and for all sets considered together. MAE value was 2.19 for test set, 2.36 for training
set and 2.94 for all sets. Correlation value of 0.98 was obtained for testing set and the
ANN model provided with 28 day compressive strength closer to the experimental
determined values.
The performance of ANN predicted a relationship between different input param-
eters and output parameter. The prediction is depicted in Figs. 3, 4 and 5. The results
arrived from ANN are listed in Table 7.
From Table 7, we can observe that maximum and minimum absolute errors in the
testing sets were 5.78 MPa and 0.02 MPa, respectively. The range of errors is from
0.06 to 11.35%. The relative error was about 5.07% which is considered good and
acceptable. Predicted values of 28 day compressive strength matched very well for
target strength of concrete.
The ANN model predicted good performance of compressive strength which is
closer to the target strength. Some of the test data did not fit very well, and this might
be due to several reasons including:
Fig. 3 Training set

Fig. 4 Testing set used in

ANN
Fig. 5 Validation of ANN

model using 28 days
compressive strength data
from literature
(1) Conversion of cylindrical strength to cube strength of concrete leads to error

due to approximation of strength data.
(2) The missing information in various literatures regarding type of cement is one
of the setbacks to streamline the results.
(3) The processing method adopted to procure recycled aggregates, source of recy-
cled aggregate and strength index is not mentioned in many literatures which
will help to improve the accuracy of prediction.
Test results showed that the constructed ANN model was able to predict the
compressive strength values of RAC accurately. When using ANN to predict the
compressive strength value of RAC, the maximum particle size of aggregate, water
absorption and SSD specific density can generally reflect the properties of RA, how-
ever, the performance of ANN model can still be improved if more parameters can
be considered, such as cement type and quality of the recycled aggregate.
Table 7 Comparison of actual and predicted values

Sr. No. Actual value Predicted value Absolute error Absolute relative error (%)
1 43.8 43.45 0.35 0.79
2 38.9 41.2 2.3 5.91
3 20 22.27 2.27 11.35
4 36.16 33.94 2.22 6.13
5 35.8 32.78 3.02 8.43
6 32.04 32.06 0.02 0.06
7 68.82 69.85 1.03 1.49
8 60.72 60.91 0.19 0.31
9 64.2 61.34 2.86 4.45
10 65.12 59.34 5.78 8.87
11 60.72 58.58 2.14 3.52
12 44.9 43.01 1.89 4.20
13 44 43.08 0.92 2.09
14 43 39.36 3.64 8.46
15 36 39.84 3.84 10.66
16 50.41 46.14 4.27 8.47
17 53.34 53.88 0.54 1.01
4 Conclusion
In this paper, an effort is made to understand the basic characteristics of recycled

aggregate concrete with the inclusion of various ingredients. The parameters studied
were within the acceptable limits of the codes. The experimental analysis conducted
to study compressive strength of concrete gave us an infer that high strength concrete
performed better than normal strength concrete and at lower W/B ratio of 0.27. The
ANN model created, predicted the compressive strength of concrete with minimal
error. The model can be further improved by considering parameters like source
of recycled aggregate and method of processing. Along with compressive strength,
properties of concrete such as creep, shrinkage, modulus of elasticity and durability
can be studied using ANN.
References
1. Sabaa, B., & Ravindrarajah, R. S. (1997). Engineering properties of lightweight concrete

containing crushed expanded polystyrene waste. Sydney, Australia: Faculty of Engineering,
University of Technology.
2. De Veaux, R. D., & Ungar, L. G. (1996). A brief introduction to neural networks. http://www.
cis.upenn.edu/Datamining/publications/nnet-intro.pdf. Last accessed 26 May 2014.
3. Yeh, I. C. (1998). Modelling of strength of high-performance concrete using artificial neural

networks. Cement and Concrete research, 28(1), 797–808.
4. Perumal, K., & Sundararajan, R. (2004). Effect of partial replacement of cement with silica fume
on the strength and durability characteristics of high performance concrete. In 29th conference
on our world in concrete and structures, 25–26 Aug 2004. Singapore.
5. Duan, Z. H., Kou, S. C., & Poon, C. S. (2013). Prediction of compressive strength of recycled
aggregate concrete using artificial neural networks. Construction and Building Materials, 40,
1200–1206.
6. Chopra, P., Sharma, R. K., & Kumar, M. (2015). Artificial neural networks for the prediction of
compressive strength of concrete. International Journal of Applied Science and Engineering,
13(3), 187–204.
7. Shao, J., Ji, X., & Li, R. (2015). Application of BP neural network model in the recycled concrete
performance prediction. In International Conference on Advances in Energy, Environment and
Chemical Engineering (AEECE).
8. Chandwani, V., Agrawal, V., & Nagar, R. (2014). Applications of artificial neural networks
in modeling compressive strength of concrete: a state of the art review. International Jour-
nal of Current Engineering and Technology, E-ISSN 2277–4106, P-ISSN 2347–5161, ©2014
INPRESSCO® . http://inpressco.com/category/ijcet.
9. Yousif, S. T., & Abdullah, S. M. (2009). Artificial neural network model for predicting
compressive strength of concrete. Tikrit Journal of Engineering Sciences, 16(3), 55–66.
10. Chopra, P., Sharma, R. K., & Kumar, M. (2016). Prediction of compressive strength of concrete
using artificial neural network and genetic programming, Hindawi Publishing Corporation.
Advances in Materials Science and Engineering, 2016(Article ID 7648467), 10. http://dx.doi.
org/10.1155/2016/7648467.
11. Nikoo, M., Torabian Moghadam, F., & Sadowski, A. (2015). Prediction of concrete compressive
strength by evolutionary artificial neural networks, Hindawi Publishing Corporation. Advances
in Materials Science and Engineering, 2015(Article ID 849126), 8. http://dx.doi.org/10.1155/
2015/849126.
12. Malesev, M., Radonjanin, V., & Marinkovic, S. (2010). Recycled concrete as aggregate
for structural concrete production. Sustainability, 2, 1204–1225. https://doi.org/10.3390/
su2051204. http://www.mdpi.com/journal/sustainability.
13. Mirza, F. A., & Saif, M. A. (2010). Mechanical properties of recycled aggregate concrete
incorporating silica fume. In Second International Conference on Sustainable Construction
materials and Technology. http://www.claisse.info/Proceedings.htm.
14. Vyas, C. M., & Bhatt, D. R. (2013). Destructive strength properties of recycled coarse aggregate.
International Journal of Innovative Technology and Exploring Engineering, 2(3).
15. James, M. N., Choi, W., & Abu-Lebdeh, T. (2011). Use of recycled aggregate and fly ash in
concrete pavement. American Journal of Engineering and Applied Sciences, 4(2), 201–208.
Department of Civil, Architecture and Environmental Engineering, North Carolina A and T
State University, USA.
16. Sivakumar, N., Muthukumar, S., Sivakumar, V., Gowtham, D., & Muthuraj, V. (2014). Exper-
imental studies on high strength concrete by using recycled coarse aggregate. Research
Inventy: International Journal of Engineering and Science, 4(01), 27–36. Department of Civil
Engineering, Jay Shriram Group of Institutions, Tirupur, Tamil Nadu, India.
17. Sonawane, T. R., & Pimplikar, S. S. (2013). Use of recycled aggregate concrete. IOSR, Jour-
nal of Mechanical and Civil Engineering (IOSR-JMCE), 52–59, ISSN: 2278–1684. www.
iosrjournals.org.
18. Parekh, D. N., & Modhera, C. D. (2011). Assessment of recycled aggregate concrete. Journal
of Engineering Research and Studies, JERS, II(I), 1–9. Applied Mechanics Department; Sardar
Vallbhbhai National Institute of Technology; Surat; Gujarat.
19. Kou, S. C., & Poon, C. S. (2012). Enhancing the durability properties of concrete prepared
with coarse recycled aggregate. Construction and Building Materials, 35, 69–76.
20. Kou, S., Poon, C., & Agrela, F. (2011). Comparisons of natural and recycled aggregate concretes
prepared with the addition of different mineral admixtures. Cement and Concrete Composites,
33, 788–795.
21. Babu, V. S., Mullick, A. K., Jain, K. K., & Singh, P. K. (2013). Mechanical properties of
high strength concrete with recycled aggregate-influence of processing. The Indian Concrete
Journal, 88, 10–26.
22. Qasrawi, H. (2014). The use of steel slag aggregate to enhance the mechanical properties of
recycled aggregate concrete and retain the environment. Construction and Building Materials,
54, 298–304.
23. Mukharjee, B. B., & Barai, S. V. (2014). Influence of nano-silica on the properties of recycled
aggregate concrete. Construction and Building Materials, 55, 29–37.
24. Yadav, S. R., & Pathak, S. R. (2009). Use of recycled concrete aggregate in making concrete-
an overview. In 34th Conference, Our World In Concrete and Structures, 16–18 Aug 2009.
Singapore.
25. Kalpavalli, A., & Naik, S. M. (2015). Use of demolished concrete wastes as coarse aggre-
gates in high strength concrete production. International Journal of Engineering Research and
Technology (IJERT), 4(7), ISSN: 0181–2278, IJERTV4IS070935. www.ijert.org.
26. Poon, C. S., Shui, Z. H., Lam, L., Fok, H., & Kou, S. C. (2004). Influence of moisture states
of natural and recycled aggregates on the slump and compressive strength of concrete. Cement
and concrete research, 34(1), 31–36.
27. Poon, C. S., Kou, S. C., & Lam, L. (2007). Influence of recycled aggregate on slump and
bleeding of fresh concrete. Materials and Structures, 40(9), 981–988.
28. Suguna Rao, B., Suresh, A., & Naik, S. M. (2019). Shrinkage behaviour of high strength
concrete using recycled concrete aggregate. As proceeding of ICSCBM, pp 829–837. Springer
book chapter https://doi.org/10.1007/978-981-13-3317-0_74.
29. Yong, H. L., Yaw, T. Y., Ta, C. P., & Ching, Y. C. (2004). An assessment of optimal mixture for
concrete made with recycled concrete aggregates. Cement Concrete Research, 34(8), 1373–
1380.
30. Cabo, A. D., Lázaro, C., Gayarre, F. L., López, M. A. S., Serna, P., & Tabares, J. O. C. (2009).
Creep and shrinkage of recycled aggregate concrete. Construction Building Material, 23(7),
2545–2553.
31. Kou, S., & Poon, C. (2012). Effect of quality of parent concrete on the mechanical properties
of high performance recycled aggregate concrete. In Submitted to the 9th symposium on high
performance concrete design, verification and utilization, 9–11 Aug 2011. Christchurch, New
Zealand.
32. Mill Valley. (1989). CA: University Science.
Effect of Vehicle Impact on Reinforced
Concrete Structures
Mohammed Mujeeb, Venkata Dilip Kumar Pasupuleti,

and Archanaa Dongre
1 Introduction
There are number of reasons for the increase of vehicular density on city roads,
one being comfortable travel from home to any other place in city or outside the
city. Second reason would be traveling at any time with capacity of four or more.
Third reason is population density of the considered city. Over and above one of
the major factor is limited availability of public transportation. These reasons have
led huge traffic in cities. With improper planning of urban development or lack of
vision in expansion of cities, construction of apartments, flyovers have led to sharp
turns, columns of houses on roads. These abrupt structures are leading to the fatal
accidents. Even though these kinds of studies are not new to American and European
researchers, very less research has been done for Indian roads.
Studies have proven that most of these accidents have occurred during the night.
Similar kinds of accidents have also been reported on the Indian roads. And if proper
care is not taken in coming years, these types of accidents would be more, as many
abrupt structures have been spotted by the authors in the existing city, Hyderabad,
as shown in Fig. 1.
This study mainly concentrates in building the methodology to understand the
structural behavior due to vehicular impact. It is assumed that masonry structures
would most probably have minor to major damage based on vehicle weight and
M. Mujeeb (B) · A. Dongre

The Northcap University, Gurgaon, Haryana, India
A. Dongre
V. D. K. Pasupuleti
Mahindra Ecole Centrale, Hyderabad, Telangana, India

196 M. Mujeeb et al.
Fig. 1 Probable locations of vehicle and structural impact
speed of collision. As framed structures are more resistant towards sudden impacts
by vehicles, the same has been dealt in brief. Apart from on road collision; there are
situations where vehicles have ramped into columns while they are being parked or
while moving out of the structure. So, based on the previous cases, case studies have
been formed.
2 Literature Review
The ever-increasing speed of vehicles and the growing infrastructure has led to the
collision of vehicles with structures, apart from terrorist attacks. Cormic and coau-
thors have raised the point of terrorists attacks were changed from static blasting
to dynamic blasting, also said to be transformed from parking to hitting of criti-
cal elements of structure. In this regard, it is necessary to understand the structural
collapse behavior under such vehicle-structure collisions [1]. Tay et al. [2] carried
out vehicular crash test of a security bollard, and compared the results with those
of numerical simulations using two different loading approaches in LS-DYNA [2].
Sharma et al. [3] developed a basic framework for the estimation of dynamic shear
force capacity of an RC column subject to vehicle impact [3].
Kang et al. [4] have developed numerical simulations for vehicular crash with
two dimensional frames for different speeds of vehicle; results obtained from their
research has been shown in Fig. 2 [4]. Apart from the above studies there have been
some of the experimental studies for understanding the crash of vehicle. But there has
been very less research on experimental for structural behavior due to its setup and
costly process. That is one major reason for the development of analytical models
for understanding the structural behavior due to vehicular impact.
Effect of Vehicle Impact on Reinforced Concrete Structures 197
Fig. 2 Two-dimensional frame impact solutions for 40, 80 and 120 km/hr [4]
3 Vehicle Impact Load
Several codes have demonstrated the impact load calculations apart from fundamental
mechanics. For this study, calculations have been taken from both Euro code and
fundamental mechanics. But as the study is more focused on behavior, an approximate
load of 150 kN is considered for all cases.
3.1 Formulation of Impact Load
In Annex B of Eurocode EN1991-1-1:2002 [5, 6], which is related to the vehicle

barriers and parapets for car parks, the horizontal force of vehicle impact, in kN,
normal and equally distributed at any length of 1.5 m car park barrier, is given by
the expression:
0.5 mv2
F= (1)
δc + δb
where: m is the gross mass of the vehicle in (kg); v is the velocity of the vehicle (m/s),
normal to the barrier; δ c —deformation of the vehicle (mm); δ b —deformation of the
barrier (mm). Equivalent static force of 150 kN is applied for a vehicle mass less
than 2500 kg and velocity at impact upto 16.2 km/h as given in Table 1.
3.2 Application of Impact Load
The impact load is applied using time history as shown in the Fig. 3.
Table 1 Values of equivalent static load obtained by different codes [7]

Source Expression Vehicle mass Impact velocity Equivalent static
(kg) (km/h) load (kN)
CTE (Spanish … 3000 No shown 50
Code [7])
EUROCODE 1 F= <2500 16.2 150
PART 1.1 0.5 mv2 /(δc + 3000 20 308
ANNEX B δb )
EUROCODE 1 … No shown <20 50–100
PART 1.7
√
EUROCODE 1 F = vr · k·m 1500 10 59
PART 1.7 3000 20 166
ANNEX C
3000 10 83
Fig. 3 Impact load time history applied at the various heights of the building in Sap2000
4 Case Studies
In this study, four cases have been considered. Case (i) simple 2D reinforced concrete
frame with single bay, case (ii) 2D two bay by two bay reinforced concrete frame
structure, case (iii) 3D two bays by two bays with single storey structure, case (iv)
actual existing 3D reinforced concrete structure. All the developed numerical models
have been assumed with basic dimensions and their designs have been checked
according to IS 456 Indian standard code. For all the numerical models, dead loads
and live loads are considered to be same and mentioned briefly at the respective
sections. As impact its self is dynamic in behavior, linear dynamic analysis is carried
out for all the case studies. The duration of impact is assumed to be 0.02 s as shown
in the Fig. 3.
4.1 Case Study: 2D Single Frame Structure
The dimensions of two columns and beam are considered to be 230 mm × 230 mm.
Grade of the concrete for beam is taken as M25 and for the column is considered to
be M30. Grade of the steel used for rebar is taken as Fe415.
Figure 4a shows the Schematic diagram of two-dimensional frame and column
considered for the application of impact load. Figure 4b shows locations where
impact loads are applied on column ‘C1’. The combination of dead load and live
load applied on to the frame is equal to 6.35 kN/m and are calculated using IS 875
Indian standard code.
The linear dynamic results obtained for case-(i) are presented in the Fig. 4. For
an impact load of 150 kN acting at 0.25 h, maximum displacement is observed to be
4.94 mm as shown in Fig. 5. Similarly, for 0.5 and 1 h, the maximum displacement are
observed to be 16.28 mm, 29.16 mm and 38.71 mm respectively. It is also observed
Fig. 4 Two-dimensional frame and locations of impact load applied
Fig. 5 Time history response for impact

the maximum displacement at joint 2 and joint 4 are same for all the cases. This
clearly indicates, as the location of impact height increases, damage to the structure
would be more. Impact load considered for this study is 150 kN. And related to
the vibrations, all the sub cases have shown similar behavior. There is an increment
of 0.93 mm horizontal displacement between 0.25 and 0.5 h, where as for other
two subcases the increment of displacement is 0.65 and 0.62 mm. This shows as
the height of the impact increases there would be increase in maximum deflection,
but comparatively the difference amount of horizontal displacement remains almost
same. For this sub case and all other subcases, only horizontal displacements are
considered.
4.2 Case Study: 2D Two Bay by Two Bay Frame Structure
Similar study has been done for two storey with two bays to understand the same
phenomenon of impact effect for a horizontal load of 150 kN, applied suddenly at
the locations shown in the Fig. 6. Geometry of all the columns are taken as 0.3 m
× 0.3 m with a grade of M30 and all the beam of 0.23 m × 0.23 m with a grade of
M25 are considered. Dead load and live load acting on the structure are calculated
to be 6.35 kN/m and the same has been applied on the beam as shown in the Fig. 7.
Complete structure is assumed to be fixed at the bottom. Column and beam numbers
are given based on the material and geometrical properties. Length of each bay is
considered to be 3 m and height of each floor is considered to be 3 m.
The results obtained for case-(ii) are showed in the Fig. 7. This study shows
the effect of impact on internal column and on external column also to observe the
difference in behavior if the impact occurs at two different floors. In this study, it is
assumed their impact would happen at only one place at any given time. Obtained
results indicate there is a minute difference if the impact has occurred in the ground
Fig. 6 Two-dimensional frame and considered locations for impact load

floor at location one and at location two. And horizontal deflections are more for
the impact occurred at location two. Which indicates impact at the internal column
would be more dangerous than external column. But for the same column impact
on the first floor have shown the different behavior, that is for external column and
internal column, the displacements are same. Which means as the impact occurs on
the top floors; mostly the behavior of the structure would remain constant.
And for all the impacts, it has been seen there is almost negligible effect on the
deformation of the beams. Figure 7 shows the displacement obtained for number of
steps, instead of time. Joints 3, 6 and 9 have shown same displacement profile and
joints 2, 5 and 8 have shown same displacements profiles. As they are in the same
floors, and two crystal curves can be seen, that is between first floor and second
floor. Damping of the structure is assumed to be 5%, same is seen for all the results.
Consistent logarithmic decrement of the displacement along the steps considered.
4.3 Case Study: 3D Two Bay by Two Bay Single Story

Structure
After analyzing two 2D frames, three-dimensional frame is also considered to under-

stand the effect of the third dimension as shown in Fig. 8. And effect of third dimen-
sion is only seen when structure is unsymmetrical or the load applied is unsymmetri-
cal. As for the subcase considered, the load acting on the structure is unsymmetrical
and displacements profiles obtained are not similar to the subcase-1. Geometry and
material properties are similar to case study-2 and thickness of slab is considered to
be 0.12 m. Results for both the cases have been found to be same and this is due to
Fig. 8 Three-dimensional
frame and considered
locations for impact load
rotational behavior. Detailed time history response for given impact load is seen in
Fig. 9.
4.4 Case Study: 3D Two Story Existing Structure
Existing structure G + 2 has been considered to understand its response due to vehicle
impact loads. Plan of the considered structure and developed 3D numerical modal
are shown in Fig. 10. Three-dimensional existing structure with different column
sizes of grade M30 and different beam sizes with a grade of M25 along with slab of
0.12 m thick have been considered. Similar to earlier cases a load of 3 kN is assumed
Fig. 10 G + 2 structure and locations of impact load applied
to be acting on the slab. Results obtained are shown in the Fig. 11, unlike previous
cases, the displacements response obtained varied at different levels as the structure
is unsymmetrical.
Fig. 11 Time history response for impact at the relative height of 0.5 h at locations 1(a) 2(b) and
3(c)
5 Conclusions
Structures are prone to vehicular impacts for various reasons be at a soft storey, struc-
ture on mid of a road, sudden diversions, temporary constructions, poor banking or
over speeds, whatever the reason, impacts are increasing and structures are under-
going minor to major damage. In this work, a fundamental study is carried out to
understand the structural behavior to vehicular impacts at the different heights of the
structure. And to understand more clearly on the geometry of the structure, four cases
have been considered and linear dynamic analysis are presented. Numerical models
developed indicate maximum response is seen if the impact is occurred at the top of
the column. As slabs are considered as rigid diaphragm, internal and external impact
have led to similar results for symmetric structures. If structure is asymmetric, then
there is large difference in structure behavior and corner column impact would lead
to higher displacements.
References
1. Cormie, D., Mays, G., & Smith, P. (2009). Vehicle-borne threats and the principles of hostile
vehicle mitigation. In Blast effects on buildings (2nd ed.). Thomas Telford.
2. Tay, S. K., Lim, B., & Ng, S. H. (2012). Crash impact modelling of security bollard. In 12th
International LS-DYNA Users Conference, vol. 13, pp. 1–10, Detroit.
3. Sharma, H., Hurlebaus, S., & Gardoni, P. (2012). Performance-based response evaluation of rein-
forced concrete columns subject to vehicle impact. International Journal of Impact Engineering,
43(5), 52–62.
4. Kang, H., Shin, J., & Kim, J. (2013). Analysis of steel moment frames subjected to vehicle
impact. APCOM and ISCM, 11–14 Dec (2013).
5. European Standard. (1998). Road restraint systems, EN 1317. London: British Standards
Institution.
6. Eurocode 1: Action on structures—part 1–1: general actions—Densities, self-weight, imposed
loads for buildings, Final draft prEN 1991-1-1, comite europeo denormalization, julio (2001).
7. Ferrer, B., Ivorra, S., Segovia, E., Irles, R. (2009) Impact load in parking steel column: code
review and numerical approach. In ECCOMAS Thematic Conference on Computational Methods
in Structural Dynamics and Earthquake Engineering, 22–24 June 2009. Rhodes, Greece.
Seismic Behaviour of G+7 RC Open
Ground Storey Buildings with Fluid
Viscous Dampers
Shaik Shayza and Bodige Narender
1 Introduction
Earthquakes lead to ground shaking due to the sudden slip on a fault. From past
many decades, earthquakes have been occurring all over the world in a large number.
The damage severity depends on the intensity of the earthquake. Some severe earth-
quakes cause death and also destroy the basic livelihood of many people. Almost all
the earthquakes directly damage the structures built leading to the loss of life and
property. In the case of developing countries, the constructions can be seen almost
everywhere. Nowadays, as the population is increasing, the need for construction is
going high with limited availability of land. Many structures are being built without
following proper guidelines and safety measures. Many high-rise buildings are built
with lot of investments. In order to protect the structures from devastating earth-
quakes, the structures need to be earthquake-resistant. In many metropolitan areas,
multi-storey buildings such as office buildings and residential apartment buildings
are designed to have a parking facility in the bottom storey. Due to the lack of masonry
infill walls in the bottom storey, it has very low strength and stiffness compared to
the above storeys. Such stories are called “Soft Storeys” which are more prone to
damage and collapse during strong ground motions. Many researchers have tried
and implemented many techniques to reduce the severity caused by an earthquake.
Over the past decades, many research studies were carried out for seismic protection
of structures. The studies concentrated on traditional techniques such as increasing
stiffness of structures, post tensioning methods and enlarging the sectional dimen-
sions which resulted in high construction costs. Later, civil engineering structures
S. Shayza (B) · B. Narender

Department of Civil Engineering, Anurag Group of Institutions, Ghatkesar, Telangana, India
B. Narender

206 S. Shayza and B. Narender
are upgraded with modern seismic reduction methodologies for better performance
during seismic events. Damping comes into this category which helps in designing
earthquake-resistant structures in earthquake-prone areas. This can be achieved by
implementing different techniques in structures such as base isolation, shear walls,
bracings and dampers. These methods can be applied to both newly constructed as
well as existing structures. Fluid viscous dampers (FVD) are passive energy dissi-
pating devices widely applied to many civil engineering structures in recent times
to minimize the structural responses caused during an earthquake or wind. They are
known for easy installation, low construction cost and zero maintenance. They dissi-
pate energy via orifices. In the year (2008), Hwang and Huang [1] had modified the
existing design formulas of supplemental FVDs in different configurations which
were provided by FEMA 273 [2]. In the year (2014), Bisht and Setia [3] explained
the seismic behaviour of soft storey building with and without viscous dampers. In
the year (2018), Maurya and Singh [4] analysed a multi-storeyed RC building by
adding FVDs at corner and middle positions and compared the building responses.
The main objective of the present study is to analyse the behaviour of G+7 reinforced
(RC) open ground storey buildings using fluid viscous dampers placed in different
positions. The responses of open ground storey buildings without FVDs are com-
pared with responses of open ground storey buildings with FVDs placed at the corner,
middle and periphery. The responses are compared in seismic zones IV and V. From
the results, it is concluded that FVDs achieved a great attention in recent times by
offering massive benefits in the field of earthquake engineering.
1.1 Fluid Viscous Dampers
FVD consists of a piston rod made up of stainless steel with a bronze orifice head
and an accumulator. It consists of a cylinder filled with highly viscous fluid usually
silicon oil which is inert, non-toxic, non-flammable, non-ageing and thermally stable.
A piston head with orifices separates the cylinder into chamber 1 and chamber 2.
FVD works mainly on the principle of fluid flow via orifices. When the structure
is subjected to external ground motions due to the earthquake, the piston inside the
cylinder moves and pushes the viscous fluid via orifices, and the fluid passes from
chamber 2 to chamber 1 with large velocity and gets transferred into kinetic energy.
Later when the fluid movement slows down, the kinetic energy is lost due to the
turbulences. The pressure differences between two chambers result in generating a
large force known as the viscous force which provides resistance against structural
movements. FVDs can be operated on temperatures ranging from −40 to +70 °C.
Equation (1) shows the damping force of a FVD which is given by [5].
F = C · Vα (1)
Seismic Behaviour of G+7 RC Open Ground Storey … 207
Fig. 1 Typical FVD manufactured by Taylor Devices, Inc.
where,
F is the damping force, C is the damping coefficient, V is the velocity of the
piston relative to the cylinder, and α is the damping exponent. The typical FVD
manufactured by Taylor Devices, Inc., is shown in Fig. 1 [6].
2 Methodology
In this paper, linear dynamic analysis (response spectrum analysis) is performed using
ETABS software. This analysis considers dynamic forces which are applied to the
structures as per code-based design spectrum. It helps to determine the effect of the
higher modes of vibration and distribution of forces. In response spectrum analysis,
multiple mode shapes are taken into consideration. Depending upon the modal mass
and modal frequency, a response is read from the design spectrum for each mode.
Later, these responses are combined to determine the total response of the structure
by modal combination methods. This method is performed for the structures whose
modes, except the fundamental one, influence the response of a structure. Response
spectrum is the estimation of maximum response such as maximum absolute accel-
eration, maximum relative velocity and displacement of single degree of freedom
system for different values of damping for a particular ground motion.
Fig. 2 Plan of the building
3 Modelling Details
3.1 Basic Building Details
In this study, G+7 RC buildings are modelled using ETABS software. The plan of
the building is symmetric with dimensions 24 m × 24 m. The building has total five
numbers of bays in both x- and y-directions, with each bay having a width of 4.8 m.
Total height of the building is 24 m with each storey height of 3 m. Figure 2 shows the
plan view of the building. Beams with 300 mm × 400 mm and columns with 400 mm
× 500 mm dimensions are considered. Thickness of slab is 150 mm; thickness of
exterior and interior walls is taken as 230 mm and 155 mm, respectively. Materials
used are of M25 grade concrete and Fe415 rebar. Loads and load combinations are
considered based on IS 456-2000 [7] and IS 875-1987 [8] codes. Earthquake loads
are taken as per IS 1893-2002, clause 7.3 [9]. In this work, live load is taken as
2 kN/m2 , floor finish is taken as 1 kN/m2 , and wall load is calculated and taken as
13.66 kN/m.
3.2 Fluid Viscous Damper Details
FVDs considered in the present study are manufactured by Taylor Devices. The link
property data is taken from the standard dimensions given by the Taylor Devices,
Inc., 1956. In this study, FVDs of 250 kN force are considered. In link property
data, enter link type as damper, and total mass and weight are taken as 44 kg and
250 kN, respectively. In directional properties, U 1 direction is selected with fixed

end properties, and the rotating inertia R1 , R2 , R3 are considered to be zero. After
defining the link properties, the links defined are drawn in the building models in
diagonal configuration using draw links icon in ETABS software.
3.3 Structural Modal Details
In this work, five buildings are modelled in zone IV and zone V to understand the
seismic behaviour of buildings individually in different zones. The FVDs are added
to the building models in different positions in the bottom storey. The following are
the building models considered in the present study:
Model 1: Bare frame building (BF).
Model 2: Bare frame building with infill walls in all storeys except the bottom
storey (BFI).
storey and fluid viscous dampers in all exterior corners bays of bottom storey
(BFIFDC).
storey and fluid viscous dampers in all exterior middle bays of bottom storey
(BFIFDM).
storey and fluid viscous dampers in all exterior bays (corner and middle) of bottom
storey (BFIFDP).
The elevation view of the building models is shown in Fig. 3.
In this study, G+7 RC buildings are modelled and analysed by response spectrum
analysis using ETABS software. The buildings are modelled in seismic zones IV
and V. The parameters considered in the response spectrum method are base shears,
storey displacements, storey drifts, storey shears and storey stiffness. To study the
seismic responses after using FVDs at particular zone, the results are compared with
respect to zone IV and zone V. In order to find out the best position for placing the
FVDs in the buildings, the dampers are placed at corners bays, middle bays and
periphery of bottom storey, and the building responses are compared.
From modal analysis, the mode shapes, time periods and frequencies for all the
building models are obtained. Mode shapes present the configurations into which
the building will naturally displace. They help us to identify the point where the
building will amplify the effect of a load. For the building models considered in this
study, i.e. bare frame (BF), bare frame with infill walls (BFI), bare frame with infill
Fig. 3 Elevation views of (a) BF, (b) BFI, (c) BFIFDC, (d) BFIFDM, (e) BFIFDP
Fig. 4 First mode shapes of: (a) BF, (b) BFI, (c) BFIFDC, (d) BFIFDM, (e) BFIFDP
walls and FVDs at corner bays (BFIFDC), bare frame with infill walls and FVDs
at middle bays (BFIFDM) and bare frame with infill walls and FVDs at periphery
(BFIFDP), 12 modes are showing 90% of total mass participation. The BF model
has shown four mode shapes; BFI, BFIFDC and BFIFDM models have shown three
mode shapes; BFIFDP model has shown only two mode shapes. The first mode
shapes of all five building models considered in the study are shown in Fig. 4. From
Fig. 4, it can be observed that the modal deformations are reducing in the buildings
modelled with FVDs, i.e. BFIFDC, BFIFDM and BFIFDP. The positions where the
FVDs are added had shown rigid action unlike the positions without FVDs in the
building and because of this reason the deformations reduced in buildings modelled
with FVDs. On comparing the three FVDs positions, BFIFDP model has shown very
less deformations.
Table 1, shows the time periods and frequencies for 12 numbers of modes for
all five building models. As the stiffness of the structure increases, the time period
decreases. Time period is inversely proportional to frequency so as the time period
increases, the frequency decreases. From Table 1, it can be observed that the modal
time periods using FVDs decreased up to 58%, 67.35% and 69.6% in BFIFDC,
BFIFDM and BFIFDP models, respectively, when compared with BFI model. This
shows the buildings modelled with FVDs have lesser time periods than the buildings
modelled without FVDs.
212
Table 1 Time periods and frequencies for all five building models
Mode BF BFI BFIFDC BFIFDM BFIFDP
Period (s) Freq (Hz) Period (s) Freq (Hz) Period (s) Freq (Hz) Period (s) Freq (Hz) Period (s) Freq (Hz)
1 1.205 0.83 0.487 2.054 0.201 4.977 0.159 6.297 0.148 6.751
2 1.131 0.884 0.403 2.479 0.198 5.04 0.157 6.374 0.144 6.946
3 1.042 0.959 0.403 2.481 0.1 9.975 0.085 11.757 0.08 12.564
4 0.388 2.578 0.069 14.518 0.052 19.252 0.049 20.613 0.047 21.324
5 0.358 2.791 0.068 14.724 0.052 19.391 0.047 21.27 0.044 22.487
6 0.334 2.998 0.038 26.417 0.033 29.87 0.033 30.175 0.033 30.414
7 0.219 4.571 0.034 29.498 0.032 30.89 0.029 33.917 0.028 35.83
8 0.197 5.07 0.03 33.56 0.027 37.069 0.027 36.783 0.027 37.378
9 0.187 5.359 0.028 35.93 0.027 37.195 0.027 37.511 0.026 37.857
10 0.146 6.865 0.028 35.946 0.027 37.206 0.026 38.432 0.025 39.432
11 0.128 7.822 0.024 41.663 0.024 42.26 0.024 42.319 0.023 42.637
12 0.123 8.143 0.024 41.765 0.024 42.359 0.024 42.413 0.023 42.759
S. Shayza and B. Narender
Fig. 5 Time periods for different building models
Figure 5 shows the base shear values of different building models in seismic
zones IV and V. The buildings modelled with FVDs have lesser base shear values
when compared to models without FVDs in both the zones. The base shear values
in BFIFDC, BFIFDM and BFIFDP models decreased up to 85.91%, 94.28% and
96.31%, respectively, when compared with BFI model in zone IV whereas in zone
V the base shears reduced to 85.96%, 94.28% and 96% in BFIFDC, BFIFDM and
BFIFDP models compared to BFI model. Figure 6(a & b) shows the storey dis-
placements for different building models, i.e. bare frame building (BF), bare frame
building with infill walls in all storeys except the bottom storey (BFI), bare frame
building with infill walls in all storeys except the bottom storey and FVDs in all exte-
rior corners bays of bottom storey (BFIFDC), bare frame building with infill walls in
all storeys except the bottom storey and FVDs in all exterior middle bays of bottom
storey (BFIFDM) and bare frame building with infill walls in all storeys except the
bottom storey and FVDs in all exterior bays (corner and middle) of bottom storey
(BFIFDP) in seismic zones IV and V. The maximum storey displacements in bottom
storey after using FVDs decreased up to 84.21%, 89.47% and 94.73% in BFIFDC,
BFIFDM and BFIFDP models, respectively, when compared to BFI model in zone
IV and decreased up to 82.14%, 92% and 92.85% in BFIFDC, BFIFDM and BFIFDP
Fig. 6 Storey displacements for different building models a zone IV, b zone V
Fig. 7 Storey drifts for different building models a zone IV, b zone V
Fig. 8 Storey shears for different building models a zone IV, b zone V
models compared to BFI model in zone V. Figure 7(a & b) shows the storey drifts for
all five building models in zones IV and V. The maximum storey drifts in BFIFDC,
BFIFDM and BFIFDP models decreased up to 81.87%, 86.8% and 90.14%, respec-
tively, in zone IV and 81.97%, 91.19% and 93.42%, respectively, in zone V when
compared to BFI model in bottom storey. Figure 8(a & b) shows the storey shears for
all five building models in zones IV and V. The maximum storey shears in BFIFDC,
BFIFDM and BFIFDP models decreased up to 85%, 94.2% and 96.3%, respectively,
in zone IV and 86%, 94.27% and 96.3%, respectively, in zone V when compared
to BFI model in bottom storey. It can be observed that the damper force is resisting
the earthquake forces and significantly reducing the storey shears. Figure 9(a & b)
shows the storey stiffness for all five building models in zone IV and V. The storey
stiffness in BFIFDC, BFIFDM and BFIFDP models approximately increased up to
17%, 12.8% and 2.4%, respectively, in both the zones when compared to BFI model
in bottom storey.
Fig. 9 Storey stiffness for different building models a zone IV, b zone V
5 Conclusions
In this paper, the study on seismic behaviour of G+7 RC open ground storey buildings
located in seismic zones IV and V is carried out using ETABS software. The open
ground storey buildings with and without fluid viscous dampers in the bottom storey
are compared. Following are the conclusions of present work:
• The time periods in open ground storey buildings with FVDs decreased up to 65%
when compared to the open ground storey buildings without FVDs.
• The values of base shears, storey displacements, storey drifts and storey shears in
buildings with FVDs reduced up to 90% when compared to the buildings without
FVDs in both seismic zones IV and V which shows the reduction of structural
responses to a great extent.
• The storey stiffness increased significantly in bottom storey after using FVDs in
the open ground storey buildings, which are said to have very low stiffness due to
the absence of infill walls in the bottom storey. Increase of stiffness adds strength
to the building to withstand the vibrations during earthquakes.
• From the results, it can be observed that the building responses in seismic zone V
are higher when compared to seismic zone IV, and the application of fluid viscous
dampers in the buildings reduced the responses in both the zones effectively by
dissipating the energy during an earthquake.
• From the observations, out of all the positions compared, FVDs placed in all exte-
rior bays (BFIFDP) have shown better and effective results due to more number
of dampers and more amount of energy dissipation.
• Thus, it can be concluded that the addition of fluid viscous dampers to the buildings
reduces the structural responses which helps in minimizing the damage occurred
during an earthquake and allowing faster reoccupation after an earthquake.
References
1. Hwang, J.-S., & Huang, Y.-N. (2008). Design formulations for supplemental viscous dampers to
building structures. Journal of Structural Engineering, 134(1), 22–31. https://doi.org/10.1061/
(ASCE)0733-9445(2008)134:1(22).
2. FEMA 273/274—Federal Emergency Management Agency. (1997). NEHRP guidelines for the
seismic rehabilitation of buildings (Report No. 273/274). Washington, D.C.: Building Seismic
Safety Council.
3. Bisht, Y., & Setia, S. (2014). Seismic behaviour of a soft storey building with & without viscous
dampers. International Journal of Engineering Research and Applications (IJERA), 10–15 ISSN:
2248-9622.
4. Maurya, A. K., & Singh, V. K. (2018). Analysis of building using viscous dampers in seismic
zone-V. International Journal of Advances in Mechanical and Civil Engineering, 5(3). ISSN:
2394-2827.
5. Mathew, L., & Prabha, C. (2014). Effect of fluid viscous dampers in multistoried buildings.
International Journal of Research in Engineering & Technology, 2(9), 59–64. ISSN(E): 2321-
8843; ISSN(P): 2347-4599.
6. Taylor Devices Company. (1956). North Tonawanda, NY, US, Site: http://taylordevices.com/.
7. IS 456:2000, Code of practice for plain and reinforced concrete (fourth revision).
8. IS 875-1987, Code of practice for design loads (other than earthquake) for buildings and
structures.
9. IS 1893 (Part 1). (2002). Criteria for earthquake resistant design of structures Part 1 General
provisions and buildings (Fifth Revision). New Delhi: Bureau of Indian Standards.
Mathematical Model for the Compressive
Strength and Elastic Properties
of Triple-Blended Steel Fiber
Self-Compacting Concrete Based
on the Experimental Investigation
S. Vijaya Kumar, B. Dean Kumar, and B. L. P. Swami
1 Introduction
The present study clearly indicates that the self-compacting concrete is essential in
the concrete field to prevent external compaction. It can be filled under its own weight
at all the corners of the formwork and get completely compacted even in dense steel.
SCC is homogeneous and dense and has similar structural parameters of the vibrated
cement concrete (VCC). Investigations for required design and self-compatibility
methods have been carried out to make it as standard concrete. SCC can cast in a
way that there is no additional vibration needed.
1.1 Need for Design Mix
For obtaining SCC, the main criterion involves in design of a mix is the determina-
tion of the physical properties of the concrete. The fresh SCC shows better fluidity,
self-compacting and resistance against segregation. These contribute to reduce the
problem of honeycombing in the concrete. Hence, SCC can significantly raise reli-
ability along with durability. SCC shows better contribution in axial compressive
strength test and can satisfy the requirements.
S. Vijaya Kumar (B)

Department of Civil Engineering, Vasavi College of Engineering, Hyderabad 500031, India
B. Dean Kumar
Department of Civil Engineering, Jawaharlal Nehru Technological University Hyderabad,
Hyderabad, Telangana 500085, India
B. L. P. Swami
Department of Civil Engineering, MCET, Hyderabad, India

218 S. Vijaya Kumar et al.
The materials for SCC are same as vibrated cement concrete. It has cement, fine
aggregate, coarse aggregate and water along with admixtures. Presently, to the best
of authors’ knowledge, there is no standard approach for determining mix propor-
tions for SCC. Some more modifications will be required to gain strength and other
rheological properties like flow requirements. The fine aggregate content will balance
other constituents by volume for this type of special concrete.
1.2 Applications of SCC
Applications of SCC are minimization in manpower, more comfortable while placing,

smooth finishing, low sound level which are some of the benefits of this type of
concrete. SCC usually reaches the slump flow value of more than 600 mm without
segregation. SCC shall remain flowable and self-compacting for 540 s. This special
SCC is able to withstand a slope 3° in case of a free horizontal surface.
1.3 Objectives of the Current Study
The objective of this experimentis to study the presence of steel fiber percentage
and aspect ratio on the strength and elastic properties of the triple-blended self-
compacting concrete. The M40 grade is considered as base concrete. The base mix
is as per the IS 10262 specification; then, these material proportions are rearranged
in such a way that fine aggregate to total aggregate is around 0.58–0.64 to get the
requirements of SCC as per provisions. The triple blending includes the replacement
of cement with fly ash at 15% and condensed silica fume at 10% in each mix by
the weight of cement. The superplasticizer and viscosity-modifying agent are used
for good rheological properties. The workability tests like slump flow, V-funnel and
L-box tests are performed as per the directions of EFNARC-2005 [1] to fulfill the
SCC requirements. In this work, the percentage of steel fiber is varied from 0.20 to
0.80, and aspect ratios of 20, 30 and 40 were adopted for each percentage of steel
fiber. A comparison is extracted for the results of the compressive strength along
the Young’s modulus and Poisson’s ratio. These properties are evaluated, and the
necessary conclusions are extracted.
2 Literature Review
The ideology of self-compacting concrete was developed in the late 90s by Professor
Hajime Okamura. The prototype beams were examined in Japan by Professor Ozawa
in Tokyo University. The self-compacting concrete was designed at that time for
better strength and durability of concrete structural elements without the need of
Mathematical Model for the Compressive Strength … 219
external compaction. Researchers have been carried out in this special concrete by
incorporating supplementary cementitious materials (SCM).
Professor Okamura [2] developed mix design method for SCC, and the main
intention was to do the test on paste and mortar to examine the compatibility and
properties of the superplasticizer (SP). Su et al. [3] reported a mix design procedure
for SCC, and the main intention in doing so was to fill gaps in unstructured aggregates
with paste that was a binder in nature.
Okamura and Ouchi [4] investigated the current and future use of the SCC and
developed design mix approach by rearranging the fine aggregate-to-total aggregate
ratio, until flowability, passing ability and resistance against the segregation were
achieved.
Petersson et al. [5] did the experimental study to develop guidelines for self-
compacting concrete and discussed various methods for the workability of concrete.
Pajak et al. [6] reported the flexural variation of SCC by using straight, hooked-
end steel fibers at various percentages of 0.5, 1.0 and 1.5; conclusions were drawn
with vibrated cement concrete (VCC).
Corinaldesi et al. [7] organized self-compacting concrete mixes which use three
varied fibers consisting of steel, polyvinyl alcohol (PVA), polypropylene along with
different types of minerals (limestone and recycled concrete generated powder). The
results showed that the utilization of recycled concrete powder is better than limestone
powder for producing SCC.
Iyer et al. [8], fiber-reinforced concrete (FRC) is used in different elements such as
concrete pavements, floors and airport runways. The basalt chopped fibers in filament
shape are used to develop basalt fiber-reinforced concrete (BFRC). They studied the
possible enhancemnet in the compressive strength along with the modulus of rupture.
Sahmaran et al. [9] investigated on straight- and hooked-end steel fibers in fresh
and hardened state of the concrete. The authors reported that it is possible to attain
self-compaction up to 60 kg/m3 of fiber content.
Cunha et al. [10] reported the influence of steel fiber percentage on concrete for
compressive behavior of SCC. The compressive strength and modulus of elasticity
increased with age but decreased when steel fiber increased from 30 to 45 kg/m3 .
Kosmatka et al. [11], reported that SCMs (supplementary cement materials) gen-
erally enhance the workability of concrete. The fly ash cement and slag cement have
frequently been reported to enhance concrete rheological properties.
SCC have been developed by Vijaya et al. [12] to get strength and durabil-
ity in structures. The reference mix proportion of SCC was arrived from the NAN
SU method of mix for M40 grade plastic fiber-reinforced self-compacting concrete
(PFRSCC). It was developed by varying percentages of fibers from 0.0, 0.25, 0.5,
0.75, 1.00, 1.1, 1.2, 1.3 and 1.4%. EFNARC [1] gave some guidelines for mix design
for SCC. The specifications about the original compositions, the volume of paste,
fine particles, the dosage of superplasticizers and viscosity-modifying agents had
been given. Details of various workability tests along with the limiting values are
presented.
Ganeshan et al. [13] examined, the influence of the steel fiber on the structural
behaviour of fiber reinforced self compacting concrete (FRSCC) subjected to flex-
ure. In this investigation, optimum volume portion of fibers was observed as 0.5%,
from the experimental work considering 0,0.25,0.5 and 0.75% of steel fiber with the
variable aspect ratio 15, 25 and 35.
3 Materials Used
3.1 Cement 53 Grades
53 Grade cement was adopted for this experiment confirming to the Indian Standard
code IS: 10269-1987 [14].
3.2 Fine Aggregate (FA)
In this experimentation, river sand was used as fine aggregate. The physical tests on
the fine aggregates like specific gravity, gradation curve, bulk density, and fineness
modulus were conducted as per IS-2386 [15]. Here, 30% of FA was replaced by
artificial sand (Robo sand).
3.3 Coarse Aggregate (CA)
The coarse aggregate of size 10–16 mm was acquired from aggregate crushing unit.
The physical properties of aggregate specify its specific gravity, gradation curve,
bulk density and fineness modulus which were examined as per the code provisions
of IS-2386 [15]. The coarse aggregate was chosen by shape as per IS 2386 (Part I).
3.4 Fly ash
Fly ash of F type was used as one of the substitution materials for cement. It was
collected from thermal power plant in Telangana State. The specific surface area of
fly ash is found as 4750 cm2 /g by Blaine’s permeability apparatus.
3.5 Condensed Silica Fume (CSF)
CSF was collected from VBC. Ferro Alloys Ltd. (VBCFAL), Rudraram in Hyder-
abad. Its fineness was found to be greater than 10,000 cm2 /g.
3.6 Superplasticizer
Superplasticizer used in this experimentation is B233 (M/S BASF INDIA LTD).
3.7 Viscosity Modifying Agent (VMA)
VMA ensuring the consistency and minimize the fluid mix segregation, in this work
Glenium2 is used as VMA, this is brought from M/S BASF INDIA LTD, the Perfor-
mance variations are due to different in the material quality and the moisture in the
aggregate were adjusted for an effective quality control.
3.8 Steel Fibers
Mild steel fibers of 1 mm diameter was used at various aspect ratios of 20, 30 and
40. Fiber is varied from 0.2 to 0.8% in four stages.
3.9 Water
Water is the most important for concrete; for good concrete, it should be free from
impurities such as oil, alkali and acidic.
4 Concrete Mix Design and Testing the Specimens
The mix design for M40 grade of concrete as per specifications of IS 10262-2009
[16] by utilizing the materials selected. The details of the reference mix mentioned
in Table 1 and the design mix for SCC employed in this work are given in Table 2.
Table 1 Concrete mix proportions for M40 (VCC)a

Sl. no. Grade of Cement Fine aggregate Coarse Water/cement
concrete aggregate ratio
1 M40 1 1.38 2.4 0.45
a VCC Vibrated cement concrete
Table 2 Concrete mix proportions for M40 (SCC)a

Sl. no. Grade of Cement Fine aggregate Coarse Water/cement
concrete aggregate ratio
1 M40 1 2.33 1.37 0.45
a SCC Self-compacting concrete
Table 3 Quantities of materials required with mineral admixtures (20% fly ash and 10% CSF)
Sl. no. Materials required Quantities in kg. per m3 for M40
1 Cement 350
2 Fly ash 70
3 CSF 35
4 Fine aggregate 745
5 Robo sand 320
5 Coarse aggregate 624
6 Water 205
The mineral admixtures of fly ash and CSF were used as a replacement to OPC at
optimum percentages of 15 and 10 to achieve triple blending. The water/cement ratio
is to be maintained at 0.45. The quantity of materials used is given in Table 3.
4.1 Optimum Dosages of Chemical Admixtures Used
By gradually increasing the dosages of superplasticizer and VMA by trial, the final
dosages of 1.0 and 0.15% were arrived to achieve the requirements for the percentage
of steel fibers up to 0.4; later, for percentage of steel fibers of 0.6 and 0.8, the super-
plasticizer is 1.2 and VMA 0.2 percentage which is used to fulfill the requirements
of SCC.
4.2 Workability Tests
Workability tests as per EFNARC-2015 [1] specifications consisting of slump flow,

L-box, V-funnel tests were examined; the results of different trials are presented in
Table 4.
Table 4 Trial tests for workability
Sl. no Method Units Trial mix1 Trial mix2 Trial mix3 Acceptance
criteria
1 Slump flow mm 460 550 680 650–800
2 T50 cm slump s – 7 4 2–5
flow
3 V-funnel s 52 22 8 6–12
4 V-funnel at s 63 26 10 6–15
T5 min
5 L-box test H2 /H1 2.2 0.8 0.95 0.8–1
Fig. 1 Arrangement of the dial gauges for the elastic properties of SFSCC and testing setup
4.3 Tests on Hardened Concrete (Fibrous Triple-Blended

SCC)
4.3.1 Number of Mixes
A total of 13 mixes were tried in this experimental investigation. The steel fiber
percentages of 0.2, 0.4, 0.6 and 0.8 by volume of the concrete and three aspect ratios
of 20, 30 and 40 were adopted (Fig. 1).
5 Results
5.1 Compressive Strength
Three cube specimens of size 150 × 150 × 150 mm were casted for each combi-
nation; a total of 39 samples (36 + 3) for three combinations for aspect ratio and
four combinations of each percentage of steel fibers were casted. The normal cured
approach is adopted and tested at the age of 28 days for compressive strength. The
values are presented in Table 5. Its variation is plotted in Fig. 2. The compressive
strength of the reference concrete which did not include the blending of admixtures
and steel fibers is greater than 40 N/mm2 . This fulfills the requirements; SCC mix
with triple blending (fly ash 20% and CSF 10%) and without fibers indicated strength
that was moderately higher than the reference mix. The mineral admixtures consist-
ing of fly ash and CSF were added to increase the strength besides giving more
workability properties. It clearly indicates from Table 5, the compressive strength of
Table 5 Compressive strength of the SFSCC

Sl. no. Code no. Steel fiber percentage Aspect ratio Average cube strength
(N/mm2 )
1 M0 0 – 44.50
2 M1 0.20 20 47.25
3 M2 0.20 30 48.15
4 M3 0.20 40 49.82
5 M4 0.40 20 48.25
6 M5 0.40 30 51.25
7 M6 0.40 40 54.15
8 M7 0.60 20 53.25
9 M8 0.60 30 58.30
10 M9 0.60 40 60.74
11 M10 0.80 20 59.90
12 M11 0.80 30 63.72
13 M12 0.80 40 68.50
70
y = 0.0158x4 - 0.0531x3 + 0.0842x2 + 4.6403x + 40
65
The compressive strength in Mpa
60
55
50
45
20
40
30
35
40
30
0 0.2 0.4 0.6 0.8
percentage of steel fibres
Fig. 2 Variation of the compressive strength for steel fiber triple-blended SCC
SCC of M40 grade is increasing with an increase in steel fiber percentages. There is
a significant increase in the compressive strength due to fibers and its aspect ratio.
In the present investigation, the maximum percentage of fiber is kept at 0.8 and the
maximum aspect ratio is 40. The strength is higher with an aspect ratio of 40. Hence,
it is clear that up to a specific favorable percentage and a significantly proportionate
aspect ratio, steel fibers played a part in increasing the strength. Further increase in
the fiber percentage or aspect ratio may interfere with the flow of SCC. The mathe-
matical models for the compressive strength of steel fiber for different percentages of
steel fiber are given by most fit fourth-degree polynomial equation for typical aspect
ratio of 40.
y = 0.0158x 4 − 0.0531x 3 + 0.0842x 2 + 4.6403x + 40
y is the compressive strength in MPa, and

x indicates the percentage of fiber.
The above equations can be derived for other aspect ratios also. The addition of
steel fibers makes concrete more and more compact and denser. The enhancement
in aspect ratio (up to optimum value) contributes toward better strength as it makes
concrete become stiffer.
5.2 Modulus of Elasticity
For the Young’s Modulus and Poisson’s ratio, total of 39 numbers of standard cylin-
ders of size 150 mm diameter and 300 mm height were tested. The Young’s modulus
is determined on cylinder specimen by direct compression and measuring the strains
using dial gauges in two orthogonal directions. The Young’s modulus of concrete
depends upon the modulus of elasticity of the aggregates, cement and their relative
proportions. Average Young’s modulus is computed from the measurements in the
elastic range following the code procedure IS: 516 [17]. In the present study, the
Young’s modulus is increasing with increase in fiber percentage. Table 6 shows the
Young’s modulus values. SFSCC combines with triple mixing of fly ash 20% and CSF
10% with fiber 0.8% showing the highest value which is about 35% more compared
to the base mix. The Young’s modulus of triple-blended fibrous SCC of M40 grade
enhanced while increasing fiber percentage and the aspect ratio. The mathematical
equation for the Young’s modulus of the triple-blended steel fiber self-compacting
concrete had been drawn as exponential which is the fitting curve for the experi-
mental values for typical aspect ratio of 40. The arrangement of the dial gauges for
experimental setup is shown in Fig. 1. And variation of the Young’s modulus is shown
in Fig. 3. Equations can be derived for other aspect ratios. The strength increase with
the introduction of fiber causes higher stress and lowers the strain; hence, Young
modulus increases with higher percentage of fibers.
Table 6 Elastic properties of SFSCC

Sl. no. Code no. Steel fiber Aspect ratio Young’ modulus Poison’s ratio
percentage MPa
1 M0 0 0 32,669.00 0.152
2 M1 0.2 20 34,618.25 0.150
3 M2 0.2 30 34,930.50 0.142
4 M3 0.2 40 35,500.00 0.133
5 M4 0.4 20 36,000.00 0.147
6 M5 0.4 30 36,800.75 0.135
7 M6 0.4 40 37,200.50 0.130
8 M7 0.6 20 39,250.00 0.129
9 M8 0.6 30 40,010.25 0.128
10 M9 0.6 40 42,030.00 0.117
11 M10 0.8 20 43,900.25 0.124
12 M11 0.8 30 44,000.50 0.110
13 M12 0.8 40 34,931.12 0.107
48000
E = 32669e0.0801X
46000
20
44000
Young's Modulus in Mpa
30
42000
40
40000
38000
36000
34000
32000
30000
0.2 0.4 0.6 0.8
percentage of steel fibre
Fig. 3 Variation of the Young’s modulus for steel fiber triple-blended SCC
E = 32669e0.0801(X )
E is the Young’s modulus in MPa, and

x indicates the percentage of fibers.
5.3 Poisson’s Ratio
In the present study, Poisson’s ratio is decreasing while increasing steel fiber per-
centage. Table 6 shows Poisson’s ratio values. The concrete mix with triple blending
and with steel fiber shows lower value than the reference mix. It is noted that Pois-
son’s ratio of triple-blended fibrous SCC of M40 grade of the concrete reduces with
enhancing fiber percentage and aspect ratio. The variation of Poisson’s ratio with
respect to steel fiber percentage in this triple-blended steel fiber self-compacting
concrete is observed as linear. For aspect ratio of 40, best-fitted curves are shown
in Fig. 4. It is recorded that Poisson’s ratio for the optimum dosage of steel fiber
and maximum aspect ratio is 30% lesser than the reference mix. An increase in the
fiber percentage causes higher stress capacity resulting in lower strain; hence, the
Poisson’s ratio gets reduced with higher percentage of fibers.
y = −0.0116x + 0.152
• y is the Poison’s ratio, and

• x indicates the percentage of fibers
As regression analysis was conducted for best fitting curves; hence, separate
statistical distribution was not required.
0.16
y = -0.0116x + 0.152
0.15
20
0.14
The poissons's ratio
30
0.13
40
0.12
0.11
0.1
0.09
0.08
0.2 0.4 0.6 0.8
percentage of steel fibres
Fig. 4 Variation of the Poisson’s ratio for steel fiber triple-blended SCC
5.4 Influence of the Aspect Ratio
As the aspect ratio of the fiber increases, there is an increase in strength, as the aspect
ratio is high; it will interfere with the flow of concrete because of its weight. With
higher aspect ratio, there may be balling effect also.
5.5 Use of Triple Blending
The triple blending of cement using admixtures like fly ash and condensed silica
fume renders the concrete mix to flow smoothly and contributes better strength. The
use of the mineral admixtures in specified percentages in the concrete matrix will
imporove the strength and durability properties. In this investigation, M40 concrete
mix was adopted as the reference mix, and triple blending was carried out with
fly ash and silica fume at percentages of 20 and 10, respectively. In this case, the
benefit received out of triple blending may be marginal in strength, but it has more
contributed to better workability. Hence, for today’s concrete constructions in which
SCC is used, a triple-blended cement with mineral admixtures has advantageous for
strength, durability along with the rheological requirements.
6 Conclusions
The following are some of the conclusions which were noted in the experimental
investigation.
1. The optimum percentage and aspect ratio are found to be 0.8% and 40, respec-
tively, for steel fiber self-compacting concrete for fulfilling the requirements of
the SCC.
2. The compressive strength of fibrous SCC increases with fiber percentage and
aspect ratio; hence, the concrete becomes stiffer because of steel fibers.
3. It is nearly 15% increase in the compressive strength with an aspect ratio of 40
compared to 20 for all the steel percentages except 0.2%, and lower percentage
of steel fibers will not contribute much on the strength parameter.
4. The Young’s modulus of triple-blended fibrous SCC of M40 grade was enhanced
with increase in steel fiber percentage and also with aspect ratio; there is nearly
35% enhanced in the value of the Young’s modulus with 0.8% of steel fiber at
an aspect ratio of 40 compared to the basic mix, it clearly indicates due to steel
fibers in the SCC cause higher stress capacity and lower the strain.
5. The Poisson’s ratio of triple-blended fibrous SCC of M40 grade is reducing with
increase in fiber percentage and the increase in aspect ratio. It is noted that the
Poisson’s ratio for optimum dosage of steel fiber and aspect ratio has 30% lesser
than the reference mix, due to longitudinal strain more than the lateral strain in
this case.
6. By triple blending SCC with 20% fly ash and 10% condensed silica fume (CSF),
strength increase is marginal; due to these mineral admixtures, flowability of the
mix has increased.
7. For the compressive strength and elastic properties for the triple-blended steel
fiber self-compacting concrete, the mathematical models are developed and the
equations are almost agreeing with the experimental values.
Acknowledgements The authors wish to express their sincere thanks to the authorities of Vasavi
College of Engineering, Hyderabad, given required laboratory facilities in conducting the present
investigation. Special thanks to University Grant Commission for sanctioning the minor research
project for fiber-reinforced self-compacting concrete.
References
1. EFNARC. (2005). Specifications and guidelines for self-consolidating concrete.

2. Okamura, H. (1997). Self compacting high-performance concrete. In Concrete international
(pp. 50–54).
3. Su, N. (2003). A new method for the mix design of medium strength flowing concrete with
low cement content. Cement & Concrete Composites, 25, 215–222.
4. Okamura, H., & Ouchi, M. Self compacting concrete development, present use, and future. In
1st International RILEM Symposia on Self Compacting Concrete.
5. Petersson, O, Billiberg, P., et al. (1996). A model for self-compacting concrete. In P. J. M.
Bartos, D. L. Marrs & D. J. Cleand (Eds.), Production methods & workability of concrete.
Londres: E&FN Spon.
6. Pajak, M., & Ponikiewski, T. (2013). The flexural behavior of self-compacting concrete
reinforced with different types of steel fibres. Construction and Building Materials, 47,
397–408.
7. Corinaldesi, V., & Moriconi, G. (2011). Characterisation of self-compacting concrete prepared
with different fibers along with mineral additions. Cement & Concrete Composites, 33, 596–
601.
8. Iyer, P., Kenno, S. Y., & Das, S. (2015). Mechanical properties of fiber-reinforced concrete
made with basalt filament fibers. Journal of Materials for Civil Engineering.
9. Sahmaran, M., Yutseven, A., et al. (2005). Workability of hybrid fiber reinforced self-
compacting concrete. Building and Environment, 40(12), 1672–1677.
10. Cunha, V. M. C. F, J. A. O. Barros, et al. (2006). Compression behavior of steel fiber reinforced
self-compacting concrete: age influence and modeling.
11. Kosmatka, S. H., & Wilson, M. L. (2011). Design and control of concrete mixtures, engineering
bulletin 001, 15th ed. Skokie, IL: Portland Cement Association.
12. Vijaya, G., Vaishali, V, et al. (2011). The behavior of self-compacting concrete with waste
plastic fibers when subjected to acid attack. ISSN:0975-4024.
13. Ganeshan, N. (2005), Strength behavior of steel fiber reinforced self-compacting concrete in
flexure. In International Conference on Advances in Concrete composites and Structures, 2005
(pp. 2691–2698).
14. IS-10269- Specification of ordinary Portland cement properties.
15. IS-2386-Method of testing the aggregate for concrete.
16. IS 10262-2009—Guidelines for design mix for the concrete.
17. IS-516 reaffirmed 2004-Methods of testing for strength concrete.
Prediction of Concrete Compressive
Strength Using Fuzzy Logic and Model
Tree
Vardhan Nagarkar, Preeti Kulkarni, and Shreenivas Londhe
1 Introduction
Concrete is an essential material in Civil Engineering field. Ascertaining the com-

pressive strength of concrete has great importance as it offers an option to do the
essential modification on the mix proportion to avoid circumstances where concrete
does not attain the design strength and also for more economic use of raw material
and fewer construction failures, hence reducing construction cost. The property of
concrete strength differs depending on the factors like proportion of its constituents,
the construction methods, the loading and environmental conditions, etc. Tradition-
ally, ascertaining the strength of concrete includes actual testing which demands
material and time (minimum 28 days) [1]. Moreover, concrete compressive strength
is a nonlinear problem for Construction Engineering field. It is difficult to predict the
concrete strength due to its nonlinearity. Many researchers lately have been address-
ing this issue by using data-driven techniques like artificial neural network, support
vector regression, fuzzy logic, etc. Of these data-driven tools, fuzzy logic tool is
been widely used to predict strength of concrete. Fuzzy logic can efficiently predict
compressive strength of high strength concrete with silica fume than artificial neural
network with three inputs: cement content, age of concrete and silica fume content
with 24 fuzzy sets and centroid method for a defuzzification [2]. In a study, 7, 28
and 90 days strength values of fly ash-based concrete were predicted using fuzzy
logic which yielded better results than artificial neural networks [3]. Fuzzy logic has
V. Nagarkar (B) · P. Kulkarni · S. Londhe

Department of Civil Engineering, Vishwakarma Institute of Information Technology, Pune, India
P. Kulkarni
S. Londhe

232 V. Nagarkar et al.
been used for: self compacting concrete strength prediction [4], predicting concrete
properties subjected to sulfate attack [5], strength prediction of concrete and cement,
evaluation of concrete durability, crack diagnosis, prediction of shear strength of
FRP-reinforced concrete beams and retrofit and prediction of lateral confinement
coefficient for FRP-strengthened RC columns which was reviewed and evaluated in
[6]. Matrix formulation of fuzzy rule-based systems was studied in which a very gen-
eral and commonly used structure with p inputs, q outputs and n rules was done. The
training algorithms for adaptive membership functions are also provided in matrix
formulations [7]. Model tree was also used to predict strength of high performance
concrete which resulted in effective use of tree-based modeling approach. It was also
seen that M5P tree works well for the dataset in comparison with MLP and linear
regression [8]. Trend line analysis also gives an approving result for M5P. Model tree
can be efficiently used to predict recycled aggregate concrete and are characterized
with series of equations which can be readily used [9]. Literature discusses the use
of fuzzy logic and model tree for strength prediction. However, the development of
membership functions in fuzzy logic mostly seen is triangular or trapezoidal, and its
reasoning is seldom done. The rule development is majorly based on experience and
domain knowledge and seldom mention about any stasticial method. This method
of rule development can be a tedious activity and is not based totally on statistical
approach. Thus, a need to develop rules using a statistical approach is necessary
which can decrease the tedious traditional method. On the other hand, model tree
technique which is efficiently used for strength prediction develops series of equa-
tions to predict an output. This characteristic of model tree can be utilized in fuzzy
logic, and a hybrid model with the use of fuzzy logic and rule development using
the series of equations developed by model tree can be explored. Thus, the aim of
current project is to predict the 28-day strength of fly ash-based concrete using fuzzy
logic (FL) with the rule development using model tree. The current method can help
in developing the rules in a systematic approach. This model termed as hybrid model
is developed, and its performance is judged according to the error measures utilized.
The further part of the paper is organized as follows: Sect. 2 gives an overview of
fuzzy logic and model tree followed by Sect. 3 which will discuss the information
about data adopted, and Sect. 4 will discuss the methodology adopted for model
development in the current study. Section 5 further discusses the results. The paper
ends in Sect. 6 with the conclusions followed by the references.
2 Techniques Utilized
Fuzzy Logic (FL) technique is used in the current study with model tree.
Prediction of Concrete Compressive Strength Using Fuzzy Logic … 233
2.1 Fuzzy Logic
Fuzzy logic is a superset of conventional (Boolean) logic that has been extended to
handle the concept of partial truth values between ‘completely true’ and ‘completely
false.’ Fuzzy logic basically contains of three parts: membership function, fuzzy
inference system and defuzzification. Fuzzification: The component forms fuzzy sets
for input–output variables using membership functions. Fuzzy membership functions
may take many forms, but in practical applications, there are two kinds of membership
functions: (1) triangular and (2) bell-shaped (Gaussian). Triangular functions are
mostly used membership functions. Intuition, rank ordering and inductive reasoning
can be among the many ways to assign membership functions to fuzzy variables.
The intuitive approach is commonly used because it is simple and derived from
the innate intelligence and understanding of human beings. Readers are referred to
Tayfur for further details [10]. Fuzzy Inference System: The fuzzy inference system
is a popular computing framework based on the concepts of fuzzy set theory, fuzzy
if-then rules and fuzzy reasoning. There are two types of inference systems in fuzzy
logic Mamdani inference system and Sugeno inference system. A typical fuzzy rule
in a Sugeno fuzzy model has the form. If x is A and y is B, then z = f (A, B), where
A and B are fuzzy sets in the antecedent, while z = f (A, B) is a crisp function in
the consequent. Usually, f (A, B) is a polynomial in the input variables A and B,
but it can be any function as long as it can appropriately describe the output of the
model within the fuzzy region specified by the antecedent of the rule. Defuzzification:
Defuzzification converts the resulting fuzzy outputs from the fuzzy inference engine
to a number. There are many defuzzification methods, such as center of gravity
(COG), bisector of area (BOA), mean of maxima (MOM), leftmost maximum (LM)
and rightmost maximum (RM). Readers are referred to Tayfur for further details
[10].
2.2 Model Tree
MT utilizes divide-and-conquer approach, partitions the data and provides rules for
reaching the models at the leaf nodes. The linear models are then used to quantify the
contribution of each attribute to the overall predicted value. M5P, a reconstruction of
Quinlan’s M5 algorithm, is used for inducing trees of regression models and combines
a conventional decision tree with the possibility of linear regression functions at the
nodes. First, a decision tree induction algorithm is used to build a tree, and a splitting
criterion is used that minimizes the intra-subset variation in the class values down
each branch. The splitting procedure in M5 stops if the class values of all instances
that reach a node vary very slightly, or only a few instances remain. Second, the tree
is pruned back from each leaf. When pruning, an inner node is turned into a leaf with
a regression plane. In comparison with classical regression trees, model trees deliver
better compactness and prediction accuracy [8, 11].
Table 1 Concrete compressive strength real-time data set

Parameter Min–max values Average Std-dev
Cement (kg/m3 ) 130–460 302.38 73.12
Fly ash (kg/m3 ) 0–120 25.38 31.05
Fine aggregate (kg/m3 ) 398–1011 837.47 97.57
Coarse aggregate 10 mm (kg/m3 ) 0–958 390.67 143.93
Coarse aggregate 20 mm (kg/m3 ) 482–1242 790.45 135.25
Water (kg/m3 ) 127–202 167.12 15.46
Admixture (kg/m3 ) 0–5.2 2.68 1.31
Strength (MPa) 12–65 34.2 10.63
3 Data Utilized
Two hundred and forty-one instances of fly ash-based concrete are collected from
Ready Mix concrete plant situated in Pune. The concrete mix parameters present
are proportions in kg/m3 of fly ash, cement, fine aggregate, coarse aggregate 10 mm,
coarse aggregate 20 mm, water and admixture in concrete. 28-day strength of concrete
in MPa is the output in the models developed. The characteristics of the data are shown
in Table 1. The table shows the presence of variety or pattern of data which helps the
models to learn pattern and give meaningful results with unseen data.
4 Methodology Adopted
Concrete is a complex material, and its strength characteristics depend on mix pro-
portions of materials used and their properties [1]. To predict 28-day compressive
strength of fly ash-based concrete, fuzzy logic model with rules defined using model
tree was developed. The fuzzy logic model was developed with input parameters as
proportions in kg/m3 of fly ash, cement, fine aggregate, coarse aggregate 10 mm,
coarse aggregate 20 mm, water and admixture in concrete and 28-day strength of
concrete in MPa as output parameter. The membership function developed for the
model is Gaussian type and is used for each input as the data is discrete, and there is
no fixed pattern in data. Each of the input parameter is divided into six membership
functions having different ranges in which whole data set is included. Membership
function for one of the parameter, i.e., water is seen in Fig. 1.
Sugeno type of fuzzy interference system is used in the model with centroid
method of defuzzification which is shown in Fig. 2.
Further, the rules of the fuzzy system were designed using model tree. The rules
are designed as follows: For model tree technique, M5P algorithm was used for
calibrating the model, and WEKA was used for the same [8, 11]. The input parameters
as discussed in the current section and the output are presented to WEKA with M5
Fig. 1 Membership function for water parameter
Fig. 2 Fuzzy interference system for concrete strength

Fig. 3 Model tree developed
algorithm. The tree developed for the same is as shown in Fig. 3. Total eight linear
model equations (LM) developed as shown below.
LM num: 1: strength = 0.0925 ∗ cement + 0.0288 ∗ fly ash − 0.0112 ∗ fine Aggt
− 0.0085 ∗ 10 mm − 0.0012 ∗ 20 mm − 0.0294 ∗ water
+ 0.1888 ∗ admixture + 20.5771
+ 0.0069 ∗ 10 mm − 0.0011 ∗ 20 mm − 0.1368 ∗ water
+ 0.9255 ∗ admixture + 42.8644
+ 0.0023 ∗ 10 mm − 0.0011 ∗ 20 mm − 0.426 ∗ water
+ 1.2815 ∗ admixture + 88.6687
− 0.0191 ∗ 10 mm − 0.0139 ∗ 20 mm − 0.1013 ∗ water
+ 0.4494 ∗ admixture + 76.2583
− 0.0142 ∗ 10 mm − 0.0007 ∗ 20 mm − 0.1013 ∗ water
+ 0.4494 ∗ admixture + 41.8827
− 0.0142 ∗ 10 mm − 0.0063 ∗ 20 mm − 0.1013 ∗ water
+ 0.4494 ∗ admixture + 46.5108
− 0.0136 ∗ 10 mm − 0.0063 ∗ 20 mm − 0.1013 ∗ water
+ 0.4494 ∗ admixture + 42.7007
− 0.0342 ∗ 10 mm − 0.0195 ∗ 20 mm − 0.1152 ∗ water
+ 0.4494 ∗ admixture + 73.8673
The fuzzy logic model with the rules is developed using MATLAB 2017. Thus, this
fuzzy model using model tree is termed as a hybrid model. To check the accuracies and
robustness of the model, the experimental dataset is divided for training and testing
purposes. From the available data, 70% is randomly selected to be used for training
purposes, and the remaining 30% is used for model validation. The performance of
the model is compared with experimental data reported by the author and assessed by
statistical measures like correlation coefficient (R), root mean squared error (RMSE),
mean absolute error (MAE) and Nash–Sutcliffe Efficiency (E) [12].
A fuzzy logic model with Gaussian membership function and rules derived through
model tree is developed to predict 28-day strength of fly ash-based concrete. Gaus-
sian membership function was selected in the current work as the data was discrete,
and no fixed pattern in the data. The model tree equations developed show a high
influence toward cement content (as per the coefficient) followed by fly ash content.
This is in tune with the fundamental knowledge of concrete technology [1]. Negative
coefficient toward water parameter indicates its increase in the mix after a certain
limit can decrease the strength of concrete which agrees with the domain knowl-
edge. Thus, it is seen that model tree develops LM models which are in tune with
domain knowledge and thus have a statistical approach toward strength prediction.
The concrete strength hybrid model for one of unseen output is seen in Fig. 4.
Table 2 presents the performance of developed model in terms of error measures.
As Table 2 shows correlation coefficient ‘r’ of 0.96, the model performs in an
excellent way. This is also endorsed by the other error measures shown in Table 2.
The crux of the model is the rules developed using model tree. Model tree with
M5P algorithm, the basic tree is formed based on a splitting criterion. It uses the
standard deviation of the class values for each node as a measure of the error at
Fig. 4 Hybrid model output
Table 2 Performance of
Correlation Root mean Mean Nash
hybrid model
coefficient square error absolute efficiency
‘r’ ‘RMSE’ error ‘MAE’ ‘E’
0.96 2.48 1.88 0.95
that node and then calculates the expected error reduction as a result of testing each
attribute at that node. Thus, in spite of linear equations at the end of nodes, due
to the splitting criteria, model trees can address the nonlinearity in the problem. In
comparison with classical regression trees, model trees with smoothing and pruning
deliver better compactness and prediction accuracy. Thus, the equations developed
can be applied to fuzzy model as rules, which can also reduce the tedious activity
of forming rules through experience and observation. This concept also allows a
systematic method of forming rules. Figure 5 shows the observed and predicted
values of 28-day compressive strength of concrete. The graph clearly speaks that
results of hybrid model are much closer to the observed values and follow the same
trend.
Fig. 5 Observed and predicted values of concrete strength
6 Conclusion
In the current work, a hybrid model, i.e., fuzzy logic model with rules developed
using model tree is developed to predict 28-day strength of fly ash-based concrete in
MPa with input parameters as kg/m3 proportions of materials used in concrete. The
developed hybrid model displays an excellent performance with a correlation coeffi-
cient of 0.96. The excellent performance is also endorsed by other error measures as
RMSE, MAE and CE. The equations developed using model tree show a statistical
approach and also can avoid the tedious activity of rule formation the traditional way.
References
1. Shetty, M. S. (2005). Concrete technology (17th ed.). New Delhi: S. Chand and Company.
2. Tayfur, G., Erdem, K. T., & Kırca, O. (2014). Strength prediction of high strength concrete by
fuzzy logic and artificial neural networks. Journal of Materials in Civil Engineering, 26(11),
401–407.
3. Topcu, B. L., & Sarıdemir, M. (2008). Prediction of compressive strength of concrete containing
fly ash using artificial neural networks and fuzzy logic. Computational Materials Science, 41(3),
305–311.
4. Vakhshouri, B., & Nejadi, S. (2018). Prediction of compressive strength of self-compacting
concrete by ANFIS models. Neurocomputing, 280(6), 13–22.
5. Diab, A. M., Elyamany, H. E., Elmoaty, A. E. M. A., & Shalan, A. H. (2015) Comparison
between neural network and fuzzy logic on assessment of long term concrete compressive
strength and expansion due to sulfate attack. International Journal for Research in Applied
Science and Engineering Technology, 3, 175–192.
6. Seyed, A. A., & Naderpour, H. (2015). Application of fuzzy logic in reinforced concrete
structures. https://doi.org/10.4203/ccp.109.10.
7. Lotfi, A., Andersen, H. C., & Tsoi, A. C. (1996). Matrix formulation of fuzzy rule-based
systems. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics), 26(2),
332–340.
8. Deepa, C., SathiyaKumari, K., & Sudha, K. P. (2010). Prediction of the compressive strength of
high performance concrete mix using tree based modeling. International Journal of Computers
and Applications, 6(5), 18–24.
9. Deshpande, N., Londhe, S., & Kulkarni, S. (2014). Modeling compressive strength of recy-
cled aggregate concrete by artificial neural network, model tree and non-linear regression.
International Journal of Sustainable Built Environment, 3, 187–198.
10. Tayfur, G. (2012). Soft computing in water resources engineering. WIT Press.
11. Quinlan, J. R. (1992). Learning with continuous classes. In A. Adams & L. Sterling (Eds.),
Proceedings of 5th Australian Joint Conference on Artificial Intelligence. Singapore: World
Scientific.
12. David, R. L., Jr., & Gregory, J. (1999). Evaluating the use of “goodness of fit” measures in
hydrological and hydro climatic model validation. Water Resources Research, 35(1), 233–241.

Advances in Structural Engineering: K. V. L. Subramaniam Mohd Ataullah Khan Editors

Uploaded by

Copyright:

Available Formats

Advances in Structural Engineering: K. V. L. Subramaniam Mohd Ataullah Khan Editors

Uploaded by

Document Information

Original Title

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

Advances in Structural Engineering: K. V. L. Subramaniam Mohd Ataullah Khan Editors

Uploaded by

Copyright:

Available Formats

What topics are covered in the Lecture Notes in Civil Engineering series?

What topics are covered in the Lecture Notes in Civil Engineering series?

What are the submission guidelines for publishing in the Lecture Notes in Civil Engineering series?

What are the submission guidelines for publishing in the Lecture Notes in Civil Engineering series?

Lecture Notes in Civil Engineering

More information about this series at http://www.springer.com/series/15087

ISSN 2366-2557 ISSN 2366-2565 (electronic)

production. Techniques such as fuzzy logic and artiﬁcial neural networks to

Sangareddy, India Dr. K. V. L. Subramaniam

Seismic Fragility Assessment of Unreinforced Masonry

Performance of Self-ﬂowing Concrete Incorporated with Alumina

Prof. K. V. L. Subramaniam is a Professor in the Department of Civil

Dr. Mohd Ataullah Khan is an Assistant Professor in the Department of Civil

Jayaprakash Vemuri and K. V. L. Subramaniam

2 Material and Geometric Properties of URM Walls

Vd = Dtτu ; τu = minimum(τcs ; τws ; τb ) (2)

τb is the shear stress based on brick cracking and is given by:

-60 Experiment -150 Experiment

3 Nonlinear Time-History Analyses Using Synthetic

Acceleration (g) 0.50 Time (sec)

0.10 Time (sec)

Sites in Zone 4 (IS1893:2016)—PSA(DBE) = 0.30 g

Fig. 3 a Displacement versus PGA (wall-1), b displacement versus PGV (wall-1)

Fig. 4 a Displacement versus PGA (wall-2), b displacement versus PGV (wall-2)

Table 3 Fragility curves: median and dispersion

5 Summary and Conclusions

S. Vijaya Bhaskar Reddy and S. Sivasankar

S. V. B. Reddy · S. Sivasankar (B)

2.1 Steel Tube

2.2 Concrete Mix

Concrete strength of M25 was achieved by using commercially available materials

2.3 Glass Fibre-Reinforced Polymer (GFRP) Sheet

Fig. 1 Samples before

3.1 Test Specimen Details

Fig. 2 Samples after

3.2 Application of GFRP Sheets on CCFST Columns

4 Experimental Results and Discussion

4.1 Failure of Specimens

Specimen with no wrapping (control specimen) failed by local buckling located at

Table 1 Corrosion rate analysis

Fig. 3 GFRP wrapping

Fig. 4 Air removal using

Fig. 5 Failure of control

Table 2 Ultimate compressive load and its increment comparison

Fig. 6 Failure of CCFST-1L

Fig. 7 Failure of CCFST-2L

Fig. 8 Failure of CCFST-3L

4.2 Ultimate Compressive Load

4.3 Load–Deformation Relationship

Figure 10 presented the load–deformation curve of all tested specimens. It was

800 734 746

Fig. 9 Ultimate load comparison

Fig. 10 Load versus deformation

Fig. 11 Deformation curve for all specimens

4.4 Stress–Strain Relationship

Fig. 12 Stress–strain 140

Allan L. Marbaniang, Subhrajit Dutta, and Siddhartha Ghosh

1 Introduction and History

A. L. Marbaniang (B) · S. Ghosh