A Vital Source of Sustainable Energy For Pakistan

International Conference On
A VITAL SOURCE OF SUSTAINABLE ENERGY FOR PAKISTAN
ICHP-2017
Lahore-Pakistan, December 19-20, 2017
Conference Proceeding
Organized By:
Center Of Excellence in Water Resources Engineering
University of Engineering and Technology, Lahore-Pakistan
International Conference on
HYDROPOWER
A VITAL SOURCE OF SUSTAINABLE ENERGY FOR PAKISTAN
ICHP-2017
Lahore-Pakistan, December 19-20, 2017
Conference Proceeding
CENTRE OF EXCELLENCE IN WATER RESOURCES ENGINEERING

University of Engineering and Technology, Lahore-Pakistan
Co-Sponsored by:
Suggested citation:
AUTHOR, A. (2017), Title of the Paper. In: International Conference on
HYDROPOWER-A Vital Source of Sustainable Energy for Pakistan 2017. Centre of
Excellence in Water Resources Engineering: UET Lahore-Pakistan, pp. xx-xx. ISBN:
978-969-8670-06-01
© 2017 Centre of Excellence in Water Resources Engineering, UET Lahore-Pakistan.
All rights reserved. Published 2018.

Printed in Pakistan.
ISBN: 978-969-8670-06-01
The views expressed in this publication are those of the authors and do not necessarily reﬂect the views and
policies of the Centre of Excellence in Water Resources Engineering (CEWRE), its Board of Governors, or any
official.
CEWRE does not guarantee the accuracy of the data included in this publication and accepts no responsibility for
any consequence of their use.
CEWRE encourages printing or copying information exclusively for personal and noncommercial use with proper
acknowledgment of CEWRE. Users are restricted from reselling, redistributing, or creating derivative works for
commercial purposes without the express, written consent of CEWRE.
Centre of Excellence on Water Resources

Conference Secretary Engineering
UET, G.T. Rd, Lahore- Pakistan. 54890
▪ Engr. M. Kaleem Sarwar Tel: +92 42 99250257
Fax: +92-42-99250259
Editors Email: [email protected]
www.cewre.edu.pk
▪ Dr. Ijaz Ahmad
▪ Engr. M. Kaleem Sarwar For orders, please contact:
Conference Organizing Committee
Tel: +92 42 99250257
Compiled By Email: [email protected]
▪ Dr. Ijaz Ahmad
▪ Engr. Rana Zain Nabi Khan
Printed on recycled paper.
FOREWORD
“ The CEWRE was established with the objectives of high level goal-oriented capacity building of
professionals, teaching and research in water resources sector in the country. These objectives are
being achieved by imparting post graduate degrees, conducting specialized research, dissemination of
the knowledge through short courses, seminars, training workshops and International Conferences etc.
Centre provides the healthy research facilities to the students as well as faculty members. Many national
and international organizations collaborate with Centre to carryout research activities in term of
financing, data collection, analysis and finalization of research reports.
Hydropower is clean renewable energy source that doesn’t pollute environment. It is very efficient
energy source because some turbines can achieve efficiency as high as of 95% and more. It provides
about 20 % of the world's electricity and is the main energy source for more than 30 countries.
Hydropower's advantage over other renewable energy sources is the fact that average rainfall is highly
predictable and therefore output is reliable and river flow doesn't fluctuate from minute to minute like is
the case with wind energy.
Hydropower development has been recognized as one of the key drovers in optimally utilizing the
available water resources towards achieving the objective of energy security and green growth. Despite
Pakistan having a hydropower potential of 60,000 MW, it could develop so far only 7,320 MW. Further,
due to the recent incidences of hydro-meteorological disasters and concerns raised about social and
environmental aspects from many quarters, the development of hydropower potential has not made
much of headway. Therefore, there is a huge opportunity to further optimally harness the hydropower
potential towards acquiring energy security and green growth. There is a need of serious relooking of
the existing status, policy, environmental, financial, social and technical barriers in implementation.
To provide forum for exchange of experiences on technological innovations, best practices and scope
of hydropower development as a clean, green and sustainable source of development, Centre of
Excellence in Water Resources Engineering (CEWRE), University of Engineering and Technology,
Lahore organized an International Conference on “Hydropower – A Vital Source of Sustainable Energy
for Pakistan” at CEWRE, UET, Lahore during December 19-20, 2017.
”
Director
iii
Chief Guest
Prof. Dr. Fazal Ahmad Khalid S.I

Vice Chancellor
University of Engineering & Technology, Lahore
I take this opportunity to welcome you all at this 2 days International Conference
on Hydropower - A Vital Source of Sustainable Energy for Pakistan being held in
CEWRE UET Lahore. Hydropower offers a wide range of benefits-especially for
developing countries. The resource is environmentally responsible and has substantial
economic advantages.
There is an urgent need to develop this promising sector of renewable energy in all
developing countries as only a fraction of available potential has been harnessed so
far. To further boost the developing of this sector, a need was felt to share the views in
this field through organizing International Conference covering the advances made in
the planning, technologies selection and implementation of hydropower projects. With
all the varied efforts underway to promote the development of small hydropower, a need
has developed to share information on what has worked, and what has not.
I wish a meaningful discussion and a great success for the intended purpose in this
conference with pragmatic and positive approach that incorporates sustainability
requirements in the decision-making process for growth and prosperity of all.
Thank you!
iv
Guest
Guest of Honor
of Honor
Prof. Dr. Niaz Ahmad Akhtar

Vice Chancellor
University of Engineering & Technology,
Taxila
First of all, I feel honored and humble for being invited in this conference on
Hydropower- A vital Source of Sustainable Energy for Pakistan.
It is my great pleasure to congratulate the team of CEWRE which effectively organized

this conference and provided an opportunity for the likeminded and concerned
individuals to gather and discuss the issues which Pakistan is facing regarding water
and power generation. A well planned and well managed hydropower plant is a real
long asset of any country which results in spin off benefits in many areas like agriculture,
industry, drinking water, tourism and cultural.
I would also like to pay my gratitude to all the honorable speakers and participants.
Moreover, it is my utmost desire that the decisions and the discussion taken place today
among the professionals in this field find its rightful place in the policy circles of our
society and bring about meaningful change in how we look and work on these issues.
Thank you!
v
Acknowledgement
The Centre of Excellence in Water Resources Engineering would like to acknowledge the excellent work
of our conference organizing committee which comprised of the following:
Gratitude is also extended to all of our distinguished research presenters and to our keynote speakers. We
would also like to acknowledge the important contributions of all of our conference participants and in
particular those who use their services for sharing with us their personal insights and aspirations.
Prof. Dr. Habib Ur Rehman

Conference Chairman
Centre of Excellence in Water Resources Engineering
UET Lahore
Prof. Dr. Noor Muhammad Khan

Conference Coordinator
Department of Civil Engineering, UET Lahore
Engr. M. Kaleem Sarwar

Conference Secretary
UET Lahore
vi
Organizing Committee
Engr. M. Kaleem Sarwar

UET Lahore
Dr. Ghulam Nabi

UET Lahore
Engr. Mohammad Masood

UET Lahore
Dr. Ijaz Ahmad

UET Lahore
Dr. Muhammad Waseem

UET Lahore
Engr. Rana Zain Nabi Khan

UET Lahore
vii
Program
December 19, 2017 (Tuesday)

09:15 10:00 Registration of Delegates
Inaugural Session
10:00 10:15 Guests to be seated & arrival of Chief Guest
10:15 10:25 Recitation from the Holy Quran & National Anthem
10:25 10:35 Welcome Address: Director CEWRE, UET Lahore
10:35 10:55 Opening Address: Vice Chancellor, UET Lahore
10:55 11:15 Address by the Chief Guest
11:15 11:35 Presentation of Shield to Chief Guest and Group Photo
11:35 12:00 Coffee/Tea Break
December 19, 2017 Technical Session-I (12:00 to 14:30) Venue: Seminar Hall CEWRE
Session Chairman (PROF. DR. ABDUL SATTAR SHAKIR, DEAN FACULTY OF CIVIL ENGINEERING)
Engr. Haji M. Farooq
12:00 12:20 Keynote Lecture
Former GM Wapda
Engr. Munwar Iqbal
Director (Hydel) PPIB
Hydraulic Transient Analysis for Proposed Penstock Engr. Javed Munir
12:40 13:00
of Jari Tunnel GM Nespak
Application of Computational Flow Dynamics (CFD)
Engr. Mohsin Munir
13:00 13:15 Analysis for Surge Inception and Propagation in Low
Head Hydropower Projects Nespak
13:15 13:25 Q & A Secession and Concluding Remarks by The Session Chairman
13:25 13:30 Distribution of Shields/Certificates
13:30 14:30 Lunch/Prayer Break
December 19, 2017 Technical Session-II (14:30 to 16:30) Venue: Seminar Hall CEWRE
Session Chairman (PROF. DR. HABIB-UR-REHMAN DIRECTOR CEWRE)
Prof. Dr. Khamaruzaman
UTP Malaysia
Statistical Analysis of Inflows of River Indus At Dr. Khurram Shahzad
14:50 15:10
Tarbela In the Wake of Climate Change Inst. Of Southern Punjab
Scientific Evaluation of Water Footprint Methods Dr. Atiq-Ur-Rehman Tariq
15:10 15:25 for Hydropower Generation
CUST, Islamabad
Evaluation of Suitable Design Flood Frequency
Dr. M. Yasin
15:25 15:40 Approaches for the Mountainous Watershed (A
Case Study of Upper Indus Basin) PU, Lahore
Impacts of Hydro Climatic Variables Trends on
Dr. M. Saifullah
15:40 15:55 Water Resources of Yihe River Basin During the
Past 50 Years China
viii
Program
December 20,2017 Technical Session-III (09:00 to 11:30) Venue: Seminar Hall CEWRE
Session Chairman: Prof. Dr. Khamaruzman, UTP Malaysia
9:00 9:10 RECITATION FROM THE HOLY QURAN & National Anthem
Javed Rashid
Hydropower Economist
Dr. Raza-Ul-Mustafa
9:30 9:50 Keynote Lecture UTP, Malaysia
Raheel Ahmad Rana

9:50 10:05 Risk Involved on Hydropower Projects in Pakistan
CEWRE UET Lahore
Muhammad Yasar
Sensitivity Analysis of Hydraulic Parameters for
10:05 10:20 Dr. Ghulam Nabi
Water Hammer in Penstock Design
CEWRE UET Lahore
Physical Modeling to Control Scouring Under Dr. Atiq-Ur-Rehman Tariq
10:20 10:35
Natural Flow Conditions CUST, Islamabad.
Assessment of Defects, Remedial Measures and
Major Qaiser Karim
10:35 10:50 Development Prospects of Sick Jaglot Hydropower
Projects to Meet Future Demands. MES, Rawalpindi
December 20,2017 Technical Session-IV (11:30 to 14:30) Venue: Seminar Hall CEWRE
Session Chairman (Engr. Tariq Altaf V.P. NESPAK)
Dr. Shams-Ul-Mulk
Former Chairman Wapda
Dr. M. Nadeem
Leads University
Engr. Ubaid Ullah
12:10 12:25 Modeling of Double Stilling Basin-A Case Study Eric J Lesleighter
EGC (Pvt.) Pakistan
Engr. Wajid Ijaz
Modeling of Spillway Breach for Attabad
12:25 12:40 Dr. Ata-Ur-Rehman Tariq
Landslide Dam
Muet, Jamshroo
Design Of U/S Overflow Cofferdam of Patrind Engr. Rizwan Fareed
12:40 12:55
Hydropower Project Pes (Pvt.) Lahore
Enhancing the Hydropower Potential of Runoff
River Type Project Operating in Chitral Pakistan: A Engr. Javed Zulfiqar
12:55 13:10 Case Study of Chitral Hydel Station on Lutkho
WAPDA
River
13:30 14:30 Lunch/Prayer Break
December 20,2017 Concluding Session (14:30 to 16:30) Venue: Seminar Hall CEWRE
14:30 14:40 Recitation from the Holy Quran
14:40 15:00 Presentation of Conclusions and Recommendations of Conference
15:00 15:20 Address by the Chief Guest of Concluding Session
15:20 15:40 Distribution of Certificates to Participants, Organizers & Presentation of Shield to
Chief Guest
15:40 15:55 Vote of Thanks by Director CEWRE & Group Photo
ix
Table of Contents
Sr.
Title of Paper Pages
No.
1 Detection and estimation of sediment transport trends in the upper Indus
River during the last 50 years 1-6
Authors: Sardar Ateeq-Ur-Rehman, Minh Duc Bui, Peter Rutschmann
2 Hydraulic Transient Analysis for Proposed Penstock of Khari Power
House at Jari Dam Downstream 7-12
Authors: Javed Munir, Syed Abbas Ali, Irfan Mahmood
3 Application of Computational Flow Dynamics (CFD) Analysis for Surge
Inception and Propagation in Low Head Hydropower Projects 13-18
Authors: M. Mohsin Munir, Taimoor Ahmad, Dr. Javed Munir
4 Statistical Analysis of inflows of River Indus at Tarbela in the Wake of
Climate Change 19-27
Author: Dr. Muhammad Khurram Shahzad
5 Scientific Evaluation of Water Footprint Methods for Hydropower
Generation
28-33
Authors: F. Seemab, W. Sherani, M. Abbas, M.N. Alam, R. Farooq, Z.
Ashfaq, M.A.U.R. Tariq
6 Evaluation of Suitable Design Flood Frequency Approaches for
Hydropower Structures on the Mountainous Rivers (A Case Study of
Upper Indus Basin) 34-41
Authors: Muhammad Yaseen, Ijaz Ahmad, Bilal Nasir, Muhammad
Imran Azam, Muhammad Hassan Rehman, Muhammad Afzel
7 Impacts of hydro climatic variables trends on water resources of Yihe
River Basin during the past 50 years
42-46
Authors: Muhammad Saifullah, Ijaz Ahmad, Muhammad Zaman, Zhijia
Li, Abdul Nasir
8 Sensitivity Analysis of Hydraulic Parameters for Water Hammer in
Penstock Design 47-55
Authors: Muhammad Yasar, Muhammad Waseem, Ghulam Nabi
9 Physical modeling to control scouring under natural flow conditions
Authors: Farooq. R., Ghumman A.R., AlSaleem, S.S., Seemab, F., Tariq. 56-63
M.A.U.R.
10 Assessment of Defects, Remedial Measures and Development Prospects
of Sick Jaglot Hydropower Project
64-70
Authors: Muhammad Waseem, Qaiser Karim, Muhammad Kaleem
Sarwar
11 Double-Stilling Basin Modelling, Pakistan – Case Study
Authors: Ubaid Ullah, Eric J Lesleighter, Muhammad Iqbal, Umair 71-76
Mannan
12 Modeling of spillway breach for Attabad landslide dam
77-85
Authors: Muhammad Wajid Ijaz, Ata-ur-Rehman Tariq
13 Design of Upstream Overflow Cofferdam of Patrind Hydropower Project
86-91
Authors: Rizwan Farid, Ijaz Ahmad, Rana Zain Nabi Khan
x
Table of Contents
Evaluation of Small Hydropower Scheme Operating in Chitral, Pakistan; a

14 Case Study of Chitral Hydel Station on Lutkho River 92-97
Authors: Engr. Javed Zulfiqar, Engr. Nazakat Hussain
Impact of Baglihar Hydroelectric Power Plant on Design Flood of
Pakistan
15 98-100
Authors: Adnan Majeed, Rana Zain Nabi Khan, Neelam Pari, Ahsan
Naseem
Risk of Indus Basin Water Transfer in Violation of Indus Basin Treaty
16 101-106
Authors: Ahmad Raza, Muhammad Waseem, Navid Tahir, Ijaz Ahmad
Subsurface Flow Analysis of Hydraulic Structures using Bligh’s Theory,
Lane’s Weighted Creep Theory and Modern Khosla’s Theory
17 107-113
Authors: Hasnain Jillani, Muhammad Haris Asghar Khan, Hammad ur
Rehman
Assessing Hydropower Resilience under Changing Climate: Murum-
Bakun Cascade in East Malaysia
18 114-119
Authors: Khamaruzaman B. Wan Yusof, Mubasher Hussain, Muhammad
Raza-ul-Mustafa
Rainfall-Runoff Relationship for Small Watershed in Potohar Area
19 120-126
Authors: Moien Ahsan, Sonia Zafar, Ghulam Nabi
Selection of Best Diversion Facility for the Construction of Kohala
Hydropower Dam
20 127-134
Authors: Muhammad Yaseen, Muhammad Naveed, Muhammad Kaleem
Sarwar, Muhammad Hassan Rehman, Muhammad Imran Azam
Hydropower Optimization Station using PSO and GA Techniques
21 Authors: Ijaz Ahmad, Muhammad Zaman, Shouqi Yuan, Liu Junping, 135-140
Muhammad Saifullah
One Dimensional Numerical Simulation of Scour and Deposition in a
Channel using Finite Difference Method
22 141-147
Authors: Muhammad Zain Bin Riaz, Muhammad Masood, Rana Zain
Nabi Khan
Risks Involved on Hydropower Projects in Pakistan
23 148-159
Author: Raheel Ahmed Rana
Analysis of regime behavior of lower Gugera branch canal after
24 rehabilitation/remodeling of the system. 160-168
Author: Sajid Mehmood, Ghulam Nabi
Impact Analysis of Floods on Fixation of Road Profile Grade Level for
25 Roads in Hilly Terrain 169-176
Authors: Zafar Iqbal, Kaleem Sarwar, Sajid Mahmood, Muhammad Afzal
Hydraulic Performance Assessment of an Orifice Spillways using CFD
26 Modeling 177-184
Authors: Zohaib Nisar, Muhammad Kaleem Sarwar, Ghulam Nabi
xi
Online available at www.cewre.edu.pk CEWRE/ICHP/2017/001/pp: 1-6
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Detection and estimation of sediment transport trends in the upper Indus

River during the last 50 years
Sardar Ateeq-Ur-Rehman1*, Minh Duc Bui1, Peter Rutschmann1
1
Department of Hydraulic and Water Resources Engineering
Technical University of Munich
*
[email protected]
Abstract: The water availability during the spring and summer has changed significantly in the Indus River over
the last 50 years. The change in water availability has also affected the sediment load transport capabilities of
the river at the upper Indus basin (UIB). To assess the changes in sediment transport patterns, neural network
models coupled with discrete wavelet transforms (WA-ANN) were developed. The modelling process used daily
discharges and distinct suspended sediment concentration (SSC) samples from the last 50 years. Sediment load
trends were obtained using daily sediment loads calculated by WA-ANN. The results showed that sediment load
transport patterns in UIB have been shifting from the summer to the spring months due to an increase in effective
discharge. The results also showed, at Partab Bridge (an upstream gauge station draining the glacier zone),
sediment load and flow volume have been increasing, except in August. Surprisingly, the flow volume at Besham
Qila gauge station (downstream of Partab Bridge in precipitation zone) has been decreasing during the summer
months and causing an overall reduction of sediment load transport capability of the Indus river in-between
Partab Bridge and Besham Qila. As a result of the change in water availability in different zones, the sediment
load transport capacity of the river has been increasing in glacial melt zones and decreasing in snowy and
precipitation zones. Consequently, it was concluded that sediments are being deposited in-between glacial melt
(Partab Bridge) and precipitation (Besham Qila) zones of UIB. The current findings are important to develop the
reservoir operational rules and sediment management strategies for the under construction and existing dams
under sediment load transport changing scenarios.
Keywords: Sediment load estimation, Sediment pattern, Besham Qila, Partab Bridge
Introduction 675 Milliman et al. (1984)

Detection and precise estimation of sediment 300 Summerfield and Hulton (1994)
load trends are important for designers and 200 Collins (1996)
engineers to plan and operate dam and reservoir 197 Ali (2009)
structures. One of the most popular methods for 200 Dasu Hydropower Consultants
estimating sediment load is to develop sediment (2013)
rating curves (SRCs) based on the observed data
of suspended sediment concentration (SSC) and In contrast to many studies of climate
water discharges. The SRC method has a limited degradation impact on river flow patterns in UIB,
accuracy for the Indus River due to hysteresis very few investigations concerning the impact of
phenomenon in sediment transport in the domain. flow pattern changes on the sediment load
As it can be seen in Tab. 1, the estimated annual carrying capacity of the river have been
sediment loads for the Indus River vary in a wide conducted (Azim et al. 2016).
range from 200 Mt yr-1 to 675 Mt yr-1 over the During recent years, the Artificial Neural
last 50 years. Therefore, applying SRC methods Networks (ANNs) have gained warm reception
to detect and estimate trends of sediment load as a new perspective in the fields with non-linear
may cause significant flaws in the results. nature. In general, the advantages of ANNs over
other conceptual models are: the application of
Table 1 Estimates published for the Indus river ANNs does not require a prior knowledge of the
suspended sediment load (SSL) at Besham Qila process. ANNs have the inherent property of
and Tarbela dam. nonlinearity since neurons activate a nonlinear
SSL References filter called an activation function. ANNs can
480 Holeman (1968) have multiple input with different characteristics,
400 {Peshawar University} (1970) which can make ANNs able to represent the time-
475 Meybeck (1976) space variability. In spite of suitable flexibility of
200 Lowe and Fox (1982) ANNs in modelling time series, sometimes
Proceeding of International Conference on Hydropower – A Vital Soruce of Sustainable Energy for Pakistan 1
ANNs have a weakness when signal fluctuations
are highly nonstationary and physical
hydrological processes operate under a large
range of scales, varying from one day to several
years. In such a situation, different methods have
been proposed among which are wavelet
transforms. They have become a useful method
for analysing such variations and trends in
hydrological time series. A wavelet has been
defined as a small wave whose energy is
restricted into a short period of time and is an
efficient method for signals that are non-
stationary, have short-lived transient
components, features at different scales, or Fig. 1 Location of gauge stations in the study
singularities. A non-stationary signal can be area, developed by Dasu Hydropower
decomposed into a certain number of stationary Consultants (2013).
signals by wavelet transform. Then ANN is
combined with wavelet transform (WA-ANN). It The time lengths of the data sets are from 1969 to
is thought that the WA–ANN models are more 2008 at Besham Qila and from 1962 to 2008 at
accurate than the conventional methods since Partab Bridge. Downstream of Besham Qila, the
wavelet transforms provide useful Tarbela dam is located which is the only dam in
decompositions of the original time series, and UIB.
the wavelet-transformed data improves the The specific sediment yield from the drainage
performance of the conventional ANN model by area at Besham Qila was estimated
capturing useful information on a various approximately 1,197 Mt km-2 yr-1. The high
resolution levels. sediment yield in this area is due to the large
In this paper, we develop WA-ANN models for number of small and relatively steep catchments
suspended sediment load (SSL) estimation based discharging straight into the Indus river. An
on daily observed flow discharges and possible amount of 80 to 85% of the annual sediment load
data of SSC samples in the upper Indus River. was observed at the Tarbela dam in July and
The designed WA-ANN models can be used to August. The contribution of rainfall, glacier and
reconstruct the missing daily SSL. Further, a non- snow melt to the total flow at the Tarbela dam in
parametric Mann-Kendall (MK) test can provide monsoon is about 33% to 55%. Partab Bridge
the detailed changes in the trends of SSL without gauge station is located 300 km upstream from
requiring the data to be normally distributed Besham Qila.
(Zhang et al. 2006). The MK seasonality test The gauge station is located at a river elevation
shows the trends by removing the cycles from the of 1,250 masl with a drainage area of 142,800
time series. Given a statistical significance of the km2 measures the flow of glacierized areas of the
trend, the Sen's slope estimator defines the Karakoram and Himalaya (Ali 2009). The
magnitude. The calculated results describe the contribution in flow at Besham Qila from Partab
effect of change in flow patterns on sediment Bridge is approximately 74%. The SSC sampling
load transport capacity of the Indus River in UIB frequency at both gauge stations represents an
over the last 50 years. average 22% of a daily sampling in a year at
Besham Qila and 14% at Partab Bridge. The
Study area and data description sampling is done in such a way that it represents
Data of river flow and observed SSC at two- all events in the year. More details of data quality,
gauge stations, Besham Qila and Partab Bridge, data collection, period of records for the Indus
are used in this study (Fig. 1). river can be found in Ali (2009).
Methods
First, based on the information of the observed
data in the region, we built a suitable relationship
between inputs and SSC (output) by applying a
trial-and-error method to optimise the
architecture of the conventional ANN
approaches. To improve the performance of
ANN, we decomposed the raw input data time
series (daily flows) into approximation (an) and
details (d1,…,dn) using wavelets (Fig. 2). The and mean absolute error (MAE). While the
approximation consists of high scale and low relative statistic parameters are correlation
frequency components of the signal. While the coefficient (R) and Nash-Sutcliffe coefficient of
details consist of low scale and high frequency efficiency (NSE).
components of the signal which are obtained After getting the daily SSL time series, we
from low-pass and high-pass filters respectively. analysed statistical significance of the trends
These decompositions were fed to input neurons, using the MK tests, which have been proposed
which in turn pass them on to the hidden layer first by Mann (1945) and later by Kendall (1975)
neurons after multiplying by a given weight. A to test statistic distribution. This test is found to
hidden layer neuron adds up the weighted input be an excellent tool for trend detection without
received from each input neuron, associates it being affected by outliers. Eq. 1 gives MK test
with a bias, and then passes the result on through for Z statistic.
a nonlinear activation function. The output 𝑆−1
neurons do the same operation as that of a hidden , 𝑆>0
√𝑉𝑎𝑟(𝑆)
neuron. Therefore, a neuron output in a layer
𝑍 = 0, 𝑆=0
depends on the signal received from the previous 𝑆+1
layer, its defined weight and activation function , 𝑆<0
type. Applying the trial-and-error method again, {√𝑉𝑎𝑟(𝑆)
we chose an optimal architecture of WA-ANN Whereas S is the sum of positive or negative
models. More details about ANN and wavelets signs, 𝑉𝑎𝑟 is variance of S. A positive value of Z
can be found in Haykin (1999) and Shoaib et al. indicates an upward trend and vice versa. If there
(2016). is no trend in the time series, the statistic Z has a
standard normal distribution. In case of trend,
Sen's slope estimator was used to get the
magnitude of the trends. The Sen’s slope test was
originally developed by (Sen 1968) to estimate
the true slope using simple a non-parametric
procedure. The Sen’s test is not affected by
outlier and is closely related to MK test. More
information about the MK and Sen's slope tests
can be found in the books written by Gilbert
(1987) and Gibbons et al. (2009).
Fig. 2 Schematic diagram of WA-ANN model.
Results and discussions
In the present study, we trained the network using The best performed WA-ANN model provided
70% of the data for training, 15% for testing and values for correlation coefficient, R=0.90 at
15% for validation on a random basis. The daily Besham Qila and R=0.81 at Partab Bridge. The
discharges were decomposed up to eight levels NSE was also at 0.81 and 0.66, respectively. The
using Daubechies (db1) wavelet. The output of estimated mean annual SSLs at Besham Qila and
WA-ANN was un-decomposed SSC samples. Partab Bridge were at 160 and 188 Mt yr-1. The
The permutation of logsig, tansig, radbas and value estimated by WA-ANN for Besham Qila is
purelin transfer functions was used in the hidden less than the previous estimates ranging from 200
and output layers. The Levenberg-Marquardt to 675 Mt yr-1 (Tab. 1). Additionally, the SSL
algorithm, due to its simplicity, was used to trend at Besham Qila shows a decreasing
calibrate the network. The neurons in the hidden tendency (Fig. 3a), while it has been increasing at
layer were selected based on the criteria Partab Bridge (Fig. 3b).
described by Fletcher and Goss (1993). The The Mann-Kendall (MK) test has also shown a
stopping criteria of the model was a maximum of decreasing trend in SLL at Besham Qila with a
1000 epoch. The final network was saved for true Sen's slope of 0.634 Mt yr-1 and an
later use to generate the daily missing SSL. Due increasing trend at Partab Bridge with a slope of
to different data characteristics of both gauge 0.20 Mt yr-1 (Tab. 2). At both gauge stations, the
stations, we developed two separate WA-ANN sediment load is significantly decreasing in
models. August, which is due to less snowmelt and the
The performance of WA-ANN models was cooling effect. At Besham Qila, the sediment
assessed using absolute (i.e. represented in the load is also decreasing in April, June, July, and
units of output) and relative (i.e. dimensionless) August.
statistic parameters. The absolute statistic
parameters are root mean square error (RMSE)
regimes of UIB , which require certain changes
in the operating curve of the existing reservoirs
operational rules for future. These changes may
be the use of additional water for power
generation (Hussain et al. 2017) during low flows
in winter months or for irrigation or flushing
operation in May when more water with less SSC
is available. Flushing in May will not only erode
Fig. 3 Annual SSL shows decreasing trend at (a)
the deposits from the dam due to less SSC in
Besham Qila and increasing trend at (b) Partab
incoming flow but will also provide the
Bridge.
opportunity to re-fill the reservoirs in the
proceeding high flow months, e.g. June or July.
The flow in April at Besham Qila is contributed
with base flow and rainfall. The decreasing trend
Conclusion
of sediment transport in April is due to dryness
The results predicted by the designed WA-ANN
and less river flow. The increase in SSL in May
models along-with the Mann-Kendall test, Sen's
at Besham Qila is caused by early snowmelt at
slope estimator and linear regression show a
low altitude, which is causing increase in flow
shifting trend of sediment load from summer to
and effective discharge for sediment transport.
spring in the upper Indus River over the last 50
On the other hand, decreasing trend in discharge
years. The annual sediment load share in May has
during monsoon season in-between Partab
increased by more than 100%, which is very
Bridge and Besham Qila (Böhner and Lucarini
significant. In contrast, it has decreased by 13%
2015) is causing to decrease the sediment
in August. While annual sediment load has been
transport capacity of the river. For examples in
increasing in glacial melt zones, i.e. at Partab
June, Besham Qila is showing decreasing trend
Bridge gauge station, the contribution of early
in sediment load, however, it is opposite at Partab
snowmelt has caused a particularly pronounced
Bridge (Tab. 2). Consequently, the increase in
increase in the spring. As a result of early snow
sediment load noticed at Partab Bridge is not
melt, the sediment load transport capacity of the
being transported downstream until Besham
Indus River has been decreasing during the
Qila, which indicates that sediment load is being
summer at Besham Qila. The decrease in
deposited in the channel in-between Partab
transport capacity led to a long-term decrease in
Bridge and Besham Qila. The deposition effect is
sediment load, in contrast to a long-term increase
more prominent in 2006 when 390 Mt SSL was
in flow volume in snow and precipitation zones
estimated at Partab Bridge, but it did not pass
of the Indus River. For future water and food
through Besham Qila, even in the subsequent
security along the Indus River command area, it
years (Fig. 3).
is necessary to estimate the impact of long-term
Nevertheless, changes in both snow and glacier
sediment load variations on water resources
melt regimes all together can result a
capacities located on the Indus River.
sophisticated alteration of the hydrological
Table 2 Annual and monthly sediment load trends at Besham Qila and Partab Bridge.
Annual/Month Mann-Kendall test Sen’s slope (Mt yr-1)
statistic trend type confidence true slope lower limit upper
Z interval limit
Besham Qila
Annual -1.21 downward 89% -0.634 -1.397 0
January 2.74 upward 95% 0.00111 0.00062 0.00212
February 2.76 upward 95% 0.00158 0.00060 0.00237
March 1.08 upward 84% 0.00200 0.00000 0.00431
April -1.14 downward 87% -0.00280 -0.00528 0.00000
May 2.66 upward 95% 0.07692 0.00000 0.13333
June -0.35 downward 64% -0.05263 -0.05013 -0.05013
July -0.82 downward 79% -0.18182 -0.32738 0.00000
August -1 downward 84% -0.24138 -0.46863 0.00000
September 0.61 upward 73% 0.03704 0.03704 0.03704
October 0.51 upward 69% 0.00161 0.00168 0.00168
November 2.16 upward 95% 0.00125 0.00015 0.00227
December 1.05 upward 85% 0.00040 0.00000 0.00104
Partab Bridge
Annual 0.38 upward 65% 0.20 -0.42 0.82
January 0.2.17 upward 95% 0.00040 0.00000 0.00090
February 0.71 upward 76% 0.00000 0.00000 0.00023
March 1.51 upward 93% 0.00034 0.00000 0.00060
April 3.09 upward 95% 0.00300 0.00125 0.00455
May 3.98 upward 95% 0.05882 0.03078 0.08452
June 2.14 upward 95% 0.33333 0.07418 0.50000
July 0.41 upward 66% 0.05556 -0.06559 0.25000
August -1.28 downward 90% -0.26829 -0.40000 -0.11538
September 1.35 upward 91% 0.11111 0.01250 0.18750
October 2.87 upward 95% 0.00816 0.00438 0.01348
November 3.58 upward 95% 0.00222 0.00125 0.00311
December 2.13 upward 95% 0.00080 0.00034 0.00118
Acknowledgements Gibbons, R. D., Bhaumik, D. K. and Aryal, S.,

This work was jointly funded by the German 2009. Statistical Methods for Groundwater
Academic Exchange Service (DAAD) and Monitoring: Second Edition.
Higher Education Commission (HEC) of Gilbert, R. O., 1987. Statistical methods for
Pakistan. Hydrological data was provided by the environmental pollution monitoring, John
Water and Power Development Authority, Wiley & Sons.
Pakistan and Dasu Hydropower Consultants. Haykin, S., 1999. "Multilayer perceptrons."
Their help is greatly appreciated. Neural networks: a comprehensive
foundation, 2, 156-255.
References Holeman, J. N., 1968. "The sediment yield of
Peshawar University, 1970. "The sediment load major rivers of the world." Water Resour.
and measurements for their control in rivers Res., 4(4), 737-.
of West Pakistan." Department of Water Hussain, M., Nadya, S., Yusof, K. and Mustafa,
Resources, Peshawar. M., 2017. "Potential impact of climate
Ali, K. F., 2009. "Construction of Sediment change on river inflows to the Batang Ai
Budgets in Large Scale Drainage Basins: The hydro plant,." The International Journal on
case of the upper Indus River." PhD thesis, Hydropower & Dams(01), 44-48.
Department of Geography and Planning, Kendall, M., 1975. "Rank Correlation Methods
University of Saskatchewan, Saskatoon. (4th edn.) Charles Griffin." San Francisco,
Azim, F., Shakir, A. S. and Kanwal, A., 2016. CA, 8.
"Impact of climate change on sediment yield Lowe, J. and Fox, I., "Sedimentation in Tarbela
for Naran watershed, Pakistan." Int. J. reservoir." Proc., Commission Internationale
Sediment. Res. des Grandes Barrages. Quatorzieme
Böhner, J. and Lucarini, V., 2015. "Prevailing Congres des Grands Barrages, Rio de
climatic trends and runoff response from Janeiro.
Hindukush-Karakoram-Himalaya, upper Mann, H., 1945. "Nonparametric tests against
Indus basin." arXiv preprint trend." Econmetrica, 13(3), 245-259.
arXiv:1503.06708. Meybeck, M., 1976. "Total mineral dissolved
Collins, D. N., 1996. "Sediment transport from transport by world major rivers/Transport en
glacierized basins in the Karakoram sels dissous des plus grands fleuves
mountains, Erosion and sediment yield: mondiaux." Hydrolog. Sci. J., 21(2), 265-
Global and regional perspectives " 284.
IAHS(236). Milliman, J., Quraishee, G. and Beg, M., 1984.
Dasu Hydropower Consultants, 2013. "Detailed "Sediment discharge from the Indus River to
Engineering Design Report, Part A; the ocean: past, present and future." Marine
Engineering Design." 7. Geology and Oceanography of Arabian Sea
Fletcher, D. and Goss, E., 1993. "Forecasting and Coastal Pakistan, 65-70.
with neural networks: an application using Sen, P. K., 1968. "Estimates of the regression
bankruptcy data." Information & coefficient based on Kendall's tau." J. Am.
Management, 24(3), 159-167. Stat. Assoc., 63(324), 1379-1389.
Shoaib, M., Shamseldin, A. Y., Melville, B. W. in major world drainage basins." J. Geophys.
and Khan, M. M., 2016. "Hybrid Wavelet Res-Sol. EA., 99(B7), 13871-13883.
Neural Network Approach." Artificial Zhang, Q., Xu, C.-y., Becker, S. and Jiang, T.,
Neural Network Modelling, Springer, 127- 2006. "Sediment and runoff changes in the
143. Yangtze River basin during past 50 years." J.
Summerfield, M. A. and Hulton, N. J., 1994. Hydrol., 331(3), 511-523.
"Natural controls of fluvial denudation rates
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Hydraulic Transient Analysis for Proposed Penstock of Khari Power

House at Jari Dam Downstream
Javed Munir1*, Syed Abbas Ali2, Irfan Mahmood3

1*
General Manager, Water Resources Division, NESPAK
2
Senior Engineer, Water Resources Division, NESPAK
3
Senior Engineer, Water Resources Division, NESPAK
*
[email protected]
Abstract: Jari pond separates from main Mangla reservoir below EL. 1102 ft asl, creating trapped water
storage. The technical feasibility of enhancing power generation capacity of existing Khari powerhouse has been
evaluated in the study. Additionally, the option of a new powerhouse capable of generating 3.5MW power has
also been analyzed and penstock proposed accordingly. The total head loss in the penstock for all options have
been considered. The most feasible option has been assessed for safety against the development of water hammer
or hydraulic transient pressures in case of sudden gate/valve closure. Conclusions to the study have been
proposed at the end.
Keywords: Transient analysis; Penstock; Khari powerhouse
Introduction
Jari dam is constructed across a saddle along the
rim of Mangla reservoir away from the main dam
as shown in Fig. 1. Jari dam is 4495.8 m long,
with a height of 92.66m above river bed. Another
ridge in the reservoir separates the Jari pocket
from the main Mangla reservoir. This ridge is cut
to El. 335.89. Jari pocket separates from main
reservoir when water level in the reservoir drops
below El. 335.89 m. Jari pocket becomes a part
of the main reservoir when water level in the
reservoir rises above El. 335.89 m. A 2.13m (7
ft) diameter tunnel was constructed to drain
trapped water in Jari pocket with tunnel invert at
El. 320.04m. An intake gate 2.74m (9 ft) wide x
1.52m (5 ft) high has been provided for flow
regulation.
Fig. 1 Location Map
Fig. 2 Longitudinal Section of Jari Tunnel
Scope of Hydraulic Studies: g = generator efficiency (taken as
The scope under this study is as below: 97.5%)
i) to check if 3.5MW can be generated by ttr = transformer efficiency (taken
passing 4.25m3/s through existing Khari as 98%)
powerhouse tapping Q = discharge
ii) to propose a new penstock for 3.5MW Hn = net head = total head – head losses
power generation with for a new
powerhouse at different location by Water Hammer Analysis
bifurcating and enlarging the existing The closure of a wicket gate/valve at the end of a
tapping, keeping the existing powerhouse as conduit results in head rise at u/s side of
it is gate/valve called water hammer. As a result of
iii) to perform hydraulic transient analysis for this head rise, a pressure wave is generated that
the technically feasible option travels in the upstream direction and gets
reflected from the free surface reservoir at the
Hydraulic Analyses upstream end of the conduit travelling back
The performed hydraulic analyses involved towards the gate/valve. The wave motion results
calculation of head losses, power generation, in the development of negative pressures in the
water hammer analysis, etc. Head loss conduit. This to and fro movement of positive
determination at main tunnel and respective and negative pressure wave is gradually
tappings under consideration included dissipated to zero by friction of the conduit. The
calculation of major loss due to friction and other pressure wave velocity is calculated by using
minor losses. Friction losses were calculated following equation,
using Darcy Weisbach equation as given below: 1
λL V 2 a=
hf = 1 D C1
√ρ (
D 2g K + tE )
w
where, where,
hf = head loss a = pressure wave velocity
L = conduit length r = density of water
D = conduit diameter Kw = volume modulus of water
V = flow velocity in conduit D = diameter of conduit
g = acceleration due to gravity C1 = factor for anchorage and support of conduit
 = Darcy Weisbach friction factor; calculated t = thickness of lining of Tunnel or thickness of
using Colebrook White formula pipe/penstock
E = Young’s modulus of elasticity of the lining
1 e 2.51 or pipe material
0.5
= −2log [ + ]
λ 3.7 D R e λ0.5 The maximum head rise just at the upstream side
where, of the gate/valve is calculated from the Allievi’s
Re = Reynolds number chart. To determine the maximum pressure rise
e = absolute roughness in mm from Allievi’s chart, two constants given below
Trash rack, intake, transition, bend, gradual are to be calculated.
aV
expansion, gradual contraction, bifurcation, exit Pipeline Constant K = 2gH
o
and wicket gate/valve losses collectively fall aT
under minor loss category and each has been Time Constant N = 2L
calculated using the following equation: where,
V2 Ho = total head = reservoir level – elevation
hl = K at centerline of wicket gate/valve
2g
where, T = closure time of gate/valve
K = coefficient of minor loss (taken from Design L = total length of the conduit up to
of Small Dams, USBR) gate/valve
Power has been calculated using the following For conduit with varying cross sections, different
equation, type of lining or pipe material and
P = ρ g η Q Hn anchorage/support, the weighted average of
where, pressure wave and flow velocities is calculated
P = power in Watts for each section of the conduit. The weighted
 =overall efficiency= t g tr average values are used to calculate the pipeline
t = turbine efficiency and time constants for determining maximum
(taken as 92.0%) pressure rise from Allievi’s chart.
Analysis of Existing 1m Dia Tapping for
1MW Khari Powerhouse
The description of length and alignment for
concrete and steel lined section of Jari tunnel and
existing Khari powerhouse tapping are given
below:
Jari Tunnel
Six trash rack units 1.02m x 2.0m
Rectangular entrance 1.52m x 2.74m Fig. 4 Power v/s Reservoir Level for 4.25
Transition from rectangular to circular tunnel m3/sec through Existing Tapping
2.13m diameter tunnel
Length of 2.13m dia concrete lined tunnel In order to achieve the target of 3.5MW through
332m the existing Khari powerhouse tapping,
Vertical bend in concrete lined tunnel 53o calculations were carried out by passing
Horizontal bend in concrete lined tunnel 5.76m3/sec through the existing tapping. The
36o power versus reservoir level is shown in Fig. 5.
Length of 2.13m dia steel lined tunnel 445m
Free standing 2.13m dia steel lined tunnel
28.5m
Existing Steel Penstock

Diameters of conical off-take 1120
mm/800mm
Length of 800mm pipe 2.33m
Vertical bend after cone 90o
Butterfly valve diameter 800mm Fig. 5 Power v/s Reservoir Level for 5.76m3/sec
Transition after butterfly valve 800mm to through Existing Tapping
1000mm
Horizontal bend after butterfly valve 15o The calculation results show that 3.5MW can be
Length of 1.0m penstock 11m achieved by passing 5.76m3/sec through the
The maximum tail water level in the stilling basin existing tapping at reservoir level of El. 378.56m.
considered was at El. 298.86m. The raising of The flow velocities through the 1000mm
Mangla Dam was taken into account and the diameter penstock and 800mm diameter pipe at
maximum conservation level of El. 378.56m was start of penstock are 7.34m/s and 11.46m/s
considered. The power calculations were respectively. These velocities greatly exceed the
performed by varying the reservoir level from limiting velocity of 5.0m/s. The velocity in
maximum conservation level of El. 378.56m 800mm diameter pipe exceeds limiting velocity
down to El. 323.09m. for abrasion i.e. 9.14m/s (30.0ft/s).
The power calculations were carried out by Therefore, it was concluded that the existing
passing 4.25m3/sec discharge through the Khari powerhouse tapping was not technically
existing penstock with aforementioned setting. viable for the generation of 3.5 MW power.
The calculated power versus reservoir level curve
is shown in Fig. 4. Proposed Enlargement of Tapping (1219mm)
The results of the analysis show that maximum for 3.5 MW Powerhouse
power of 2.64MW can be achieved at reservoir As discussed previously it was not possible to
level of El. 378.56m. Therefore, it is not possible generate 3.5MW from the existing Khari
to produce the desired power of 3.5MW by powerhouse by increasing the discharge to
passing 4.25m3/sec through the existing Khari 4.25m3/sec. The increase in discharge from
Powerhouse tapping. 4.25m3/sec to 5.76m3/sec to generate 3.5MW
The hydraulic analysis also shows that flow power resulted in increased flow velocities
velocity in the 1000mm diameter penstock is exceeding the maximum permissible limit. This
5.41m/s which marginally exceeds the generally induced the possibility of abrasion in 800mm
acceptable velocity limit of 5.0m/s, but the pipe of the penstock.
velocity in 800mm diameter pipe at the start of Therefore, it was proposed to provide the 3.5MW
penstock is 8.45m/s which greatly exceed the powerhouse at separate location while keeping
acceptable velocity limit of 5.0m/s. 1MW powerhouse intact. The layout plan of
proposed scheme is shown in Fig. 6 and
longitudinal section of proposed penstock in Fig. while keeping the existing 1MW penstock intact.
7. It was also proposed to use the tapping of The tapping proposed will pass 5.66m3/sec to
existing Khari powerhouse for the proposed produce 3.5MW power at reservoir level of El.
scheme and enlarge an entire length to 1.2m 378.56m. The flow velocity in entire penstock
diameter. The tapping will bifurcate after 90o will remain within 5.0m/s not exceeding the
bend to branch towards the proposed powerhouse permissible velocity limit.
Fig. 6 Layout Plan of Proposed Powerhouse
Fig. 7 Longitudinal Section of Proposed Penstock for 3.5 MW Powerhouse
The description of tapping is given below: Third vertical bend after bifurcation 7.5
First vertical bend after tapping 90 degrees degrees
Power tunnel length between tapping and 1.2m diameter penstock from bifurcation to
bifurcation 7.74 m powerhouse 57.64 m
Butterfly valve diameter 1.2 m (48 inch) The power versus reservoir level developed after
Bifurcation diameter 1.2 m hydraulic analyses for the proposed scheme is
First horizontal bend 35 degrees shown in Fig. 8. The results show that 3.5MW
Second horizontal bend 37 degrees power is achieved for the proposed configuration
Second vertical bend after bifurcation 7.5 with reservoir level at El. 378.56m by passing
degrees 5.66m3/sec discharge through the penstock. The
power will reduce as the reservoir level is
lowered. Power generation equal to 3.0MW can Water Hammer Analysis for 3.5 MW
be achieved at reservoir level of El 367.5m. So Powerhouse Tapping
far the reservoir level of El. 378.56m has not been Water hammer analysis was carried out to
achieved. Normally the reservoir is filled up to determine the maximum pressures in Jari tunnel
the normal pond level by the end of the monsoon and penstock for the case when full load rejection
season if adequate flows are available. After the takes place and wicket gate/valve at the turbine
monsoon season low flow season starts and the are closed in a short period of time. The
reservoir starts depleting. maximum pressures have been calculated for six
The hydraulic analysis shows that the proposed (6) and ten (10) second wicket gate closure times
scheme is technically viable and 3.5MW power separately.
can be produced from the proposed powerhouse. As explained above, the weighted average of
However, when reservoir level is low and it is not pressure wave velocity and flow velocity has
possible to produce power from this facility, the been used for the determination of maximum
existing 1MW unit can be operated. pressure at wicket gates from Allievi’s chart. The
pressure at other locations in the tunnel has been
determined by considering linear decrease of
head rise up to the reservoir.
Head losses in the penstock and Jari tunnel have
been neglected for the determination of water
hammer pressure. The summary of the results of
water hammer analysis is given in Table 5.
Fig. 8 Power v/s Res. Level for 5.66m3/sec
through Proposed Penstock of 3.5MW Power
Unit
Table 5 Summary of Water Hammer Calculations for Proposed 3.5MW Powerhouse
Stat Tota
Flow Distance Tunnel Hea
Diamet Res. ic l
Veloci from Centerline d
Location er Elv. Hea Hea
ty Intake Elevation Rise
d d
(m) (m/s) (m) (m) (m) (m) (m) (m)
Wicket Gates Closure Time
= 6 sec
At start of steel 378. 73.3 39.5 112.
2.13 4.44 332.00 305.26
lining 56 0 9 89
At tunnel start in 378. 75.1 92.7 167.
2.13 4.44 778.00 303.43
culvert 56 3 7 91
At tapping for 378. 80.0 96.1 176.
2.13 4.44 806.50 298.55
Khari PH 56 1 7 18
378. 80.0 104. 184.
At Turbine 1.20 5.01 873.25 298.52
56 4 13 17
Wicket Gates Closure Time = 10 sec
At start of steel 378. 73.3 20.5 93.8
2.13 4.44 332.00 305.26
lining 56 0 6 6
At tunnel start in 378. 75.1 48.1 123.
2.13 4.44 778.00 303.43
culvert 56 3 7 30
At tapping for 378. 80.0 49.9 129.
2.13 4.44 806.50 298.55
Khari PH 56 1 4 95
378. 80.0 54.0 134.
At Turbine 1.20 5.01 873.25 298.52
56 4 7 11
The calculations show a total head rise of 134.11m at the turbine and 129.95m in free
184.17m at the turbine and 176.18m in the free- standing tunnel in the culvert.
standing tunnel in the culvert for six (6) second The tunnel must be safe against the total head
wicket gates closure time. For ten (10) second developed due to full load rejection in both cases
wicket gates closure time, the total head rise is i.e., wicket gate closure time of 6 and 10 seconds.
The penstock for additional power has been hydraulic transients. It is suggested that a relief
proposed accordingly. It is suggested that a relief valve may be provided at the turbine which
valve may be provided at the turbine which automatically opens when load rejection takes
automatically opens when load rejection takes place to protect Jari tunnel from water hammer
place to protect Jari tunnel from water hammer pressures.
pressures.
References
Conclusions [1] Creager W.P. and Justin J.D., 1950,
The carried-out water hammer analysis shows Hydroelectric Handbook, 2nd Edition,
that the existing Khari powerhouse penstock John Wiley and Sons, London
cannot be used for generation of 3.5MW and [2] Gulliver J.S. and Arndt R., 1991,
therefore, a new penstock has been proposed for Hydropower Engineering Handbook,
3.5MW generation. The proposed penstock must McGraw Hill, USA
be safe against the total head developed due to [3] Mosonyi E., 1987, Low Head Hydropower
full load rejection at wicket gate closure time of Plants,
6 and 10 seconds; the penstock for additional [4] US Bureau of Reclamation (USBR), 1987,
power has been proposed accordingly. Design of Small Dams, 3rd Edition
The proposed penstock is hydraulically safe [5] Warnick C.C., 1984, Hydropower
against developed water hammer pressures or Engineering
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Application of Computational Flow Dynamics (CFD) Analysis for Surge

Inception and Propagation in Low Head Hydropower Projects
M. Mohsin Munir1*, Taimoor Ahmad2, Dr. Javed Munir3
1*
National Engineering Services, Pakistan, Pvt, Ltd
2
3
*
[email protected]
Abstract: Determination of maximum elevation of a flowing fluid due to sudden rejection of load in a
hydropower facility is of great interest to hydraulic engineers to ensure safety of the hydraulic structures. Several
mathematical models exist that employ one-dimensional modeling for the determination of surge but none of these
perfectly simulate real-time circumstances. The paper envisages investigation of surge inception and propagation
for a Low-Head Hydropower project using Computational Fluid Dynamics (CFD) analysis in FLOW-3D
software. The fluid dynamic model utilizes its analysis for surge by employing Reynolds’ Averaged Navier-Stokes
Equations (RANSE). The CFD model is designed for a case study at Taunsa Hydropower Project in Pakistan.
Various scenarios have run through the model keeping in view upstream boundary conditions. The prototype
results were compared with the physical model testing results and proved quite accurate coherence with the
physical model testing and offers insight into phenomenon which are not apparent in physical model and shall be
adopted in future for the similar low head projects limiting delays and cost incurred in the physical model testing.
Keywords: Surge, FLOW-3D, Numerical Model, Taunsa, RANSE
Introduction of immense importance especially during the

Maximum elevation that the flow will achieve design phase of the hydropower project. The
due to a sudden rejection of load in a hydropower rising wave can easily overtop the banks of the
facility is an important parameter to hydraulic fore-bay or the headrace channel and damage
engineers. The information is required in setting appurtenant structures of the hydropower facility.
the maximum height of sidewalls to prevent Moreover, studying the propagation of wave
overflow in the headrace channel as well as to through the headrace channel and into the
understand the surge propagation upstream in reservoir is required to assess the effect of rapidly
headrace channel to schedule the opening of varying flow in the headrace on the reservoir
gates of the main barrage and balance the water levels.
discharge through the barrage and power Several mathematical methods exist for one-
channel. dimensional determination of surge for different
In Pakistan, physical model studies are the only conditions i.e. for straight channels, for
practical medium available to understand and propagation of surge on a gradient, analysis of
analyze the three dimensional and time- reflected surges and the Johnson method. The
dependent complexities of the fluid flow first three methods are simplifications of the
phenomenon. Physical models can only be setup original problem and are based on several
at the final design stage and their execution is assumptions. These methods do not truly
expensive. However, the avant-garde represent the original problem and these
computational flow dynamics has emerged not conditions are rarely met in practice. Similarly,
only as an alternative analysis and design tool but the Johnson method can be arduous as the
can also be employed to analyze phenomenon computation proceeds since numerous surges
that are not possible to assess with physical will be produced and propagated. Hence it
testing. becomes inaccurate and difficult to assess the
surge using these methods.
Literature Review Other time dependent analysis increases the
A monoclinal wave exhibiting a rapidly varying complexities of the problem by introducing
flow may result in a hydropower project due to additional independent variable of time since the
sudden closure or opening of the powerhouse resulting equations become partial differential
control structure such as the sluice gates or equations instead of ordinary differential
wicket gates. This hydraulic transient analysis is equations. Method of Characteristics and Implicit
and Explicit Finite-Difference Methods have in Pakistan. It is a run of the river, low head
better accuracy and rigor but are time-consuming hydropower scheme envisaged to provide power
and cannot be applied fully to original conditions. to the national grid. Table 1 summarizes the
In special cases, such as analysis of hydropower salient features of the project.
projects comprising of open headrace channel
with tailrace tunnel (Closed conduit), implicit Table 1 Salient Features of 135MW Taunsa
methods are preferred for the closed conduit. HPP
Such complexities can only be modelled using a
powerful software package like FLOW-3D, Type of Turbine Horizontal Bulb
based on numerical solution schemes that can Mode of Operation Run of the River
accurately predict fluid flow using the concept of Installed Capacity 135 MW
fluid volume tracking.
Gross Head 6.0m
FLOW-3D was developed at the Los Alamos
National Laboratory in the 1960s and 1970s as a Rated Head 5.8m
general purpose computational fluid dynamics Design Discharge 3,155.5 m3/sec
simulation package. FLOW-3D uses an Eulerian Turbine Units 9 Units
framework in which volume tracking technique Headrace Width 203m
models the free surface. This method, based on Headrace Length 1100m
fluid volume fraction analyses the amount of
Vertical Gate Height 16m
fluid in each cell and is robust enough to handle
the breakup and coalescence of fluid masses.
FLOW-3D analysis are often limited by
FLOW-3D uses several models to numerically
computation power available. For numerical
simulate turbulent flows. For this case study, the
simplicity in this case, the three-dimensional
Renormalization group (RNG) k-epsilon
model was setup based on only one hydropower
turbulence model was used with a no slip or
unit and not the complete power-station. FLOW-
partial slip wall shear boundary condition. The
3D can effectively use the symmetry conditions
RNG turbulence model uses statistical models to
of the project layout without affecting the
solve the turbulent kinetic energy (k) and the
accuracy of the surge analysis. A 3D model was
turbulent kinetic energy dissipation rate (Є),
used since flow has three-dimensional
renormalizing the Navier Stokes Equations to use
characteristics.
cater for the effects caused by smaller scale
motion.
Model Preparation
FLOW-3D employs the Volume-of-Fluid (VOF)
The model for Taunsa hydropower project was
technique developed by Hirt and Nichols in 1981
prepared in AutoCAD and imported in stereo-
to deal with interfaces between two fluids and to
lithographic format which is used for rapid
model. VOF depends on the volume fraction
prototyping, 3D printing and computer-aided
which is assigned value of either 1 or 0 depending
manufacturing. Three dimensional models were
the cell is occupied by the fluid or not. It is
prepared for the headrace, transitions,
governed by the following convection transport
powerhouse and tailrace. Fig. 1 illustrates the 9-
equation:
unit powerhouse with gates open. Fig. 2 shows
𝜕𝐹 𝜕𝐹𝑢𝑗
+ =0 the detail of transitions of a single unit with bulb
𝜕𝑡 𝜕𝑥𝑗 turbine housed inside.
A detail on VOF method related issues can be

found in Bombardelli et al. (2001) while its
development can be referred to C.W. Hirt et al
(1981). The free surface is not only unknown in
3-D computations, but it also acts a boundary for
the problem. A sharp representation of interface
is necessary to be maintained apart from locating
the free surface in a 3-D Eulerian grid, doing it
effectively in terms of computational time.
Case Study of 135MW Taunsa Hydropower

Project
Taunsa Hydropower Project has been sited along Fig. 1 Isometric View of Taunsa HPP Model
right bank of Taunsa Barrage across Indus River
memory and computation time. Increasing the
number of cells requires greater computation
power and longer simulation time. Average
simulation time in for Taunsa HPP was 15 hours.
Keeping in view the two factors, minimum cell
size was selected so that full computation power
was being employed.
Boundary Conditions
Boundary conditions have a huge impact on the
final results of the simulation. It is necessary to
assess the boundary conditions that best replicate
the real-time conditions and actual simulation.
Boundary conditions applied for the problem are
specified below:
X minimum Specified Pressure Boundary
X maximum Outflow Boundary
Y minimum Symmetry Boundary
Y maximum Symmetry Boundary
Fig. 2 Wireframe of a single in-take and draft Z minimum Wall Boundary
tube with bulb turbine Z maximum Symmetry Boundary
The upstream boundary condition consisted of a
Meshing specified pressure to maintain a prescribed
Meshing is a consequential part of the analysis reservoir elevation. The downstream boundary
process which not only determines the numerical utilized the FLOW-3D outflow boundary
accuracy of the model but also the memory and condition. A symmetry boundary condition was
time required for the simulation. Meshing was applied along right and left side of the mesh
done and refined in FLOW-3D model setup section to take advantage of the inherent
options. FLOW-3D has an advantage of symmetry in the problem and thereby decrease
FAVOR; Fractional Area/Volume Obstacle computational time while maximizing spatial
Representation Method which allows modeling resolution. Boundary conditions used have been
of complex geometries based on equations explained below:
formulated as functions of the area and volume
porosity. Symmetry Boundary: There is no mass flux
As specified before, for simplicity as well as (flow) through a symmetry boundary. There is
increased accuracy, meshing was done for one also no shear stress or heat transfer applied at this
unit. This is a common practice since for large boundary type. It is useful for reducing the size
models, no of total active cells are limited by of a simulation when symmetry exists by cutting
computer memory. For Taunsa Hydropower the simulation at the symmetry plane.
Project, the extents, no of cells and cell sizes are
tabulated in Table 2. Wall Boundary: The wall boundary is similar to
the symmetry boundary in that mass flux across
Table 2 Summary of meshing in FLOW-3D the boundary is not allowed. However, with a
wall boundary, heating and viscous stresses can
Total active
2320000 be applied.
cells
Extent No of Cell
Dimension Specified Pressure Boundary: This boundary
(m) cells size (m)
type sets a pressure condition. The pressure can
X 400 400 1.00 be constant by setting a value in the dialog box,
Y 18.5 50 0.37 or time dependent by selecting the pressure
Z 29 116 0.25 button.
Cell sizes for the mesh were selected based on Outflow Boundary: This boundary type is useful
two factors. First consideration in this regard is for surface waves because they are able to leave
the model accuracy. In case of surge analysis, the flow region without reflecting back into the
height of surge (z-direction) is the most domain. This boundary type looks at flow
important. Hence, smallest cell size has been conditions just inside the mesh and matches them
selected for z axis. Second consideration is CPU
to allow fluid to freely dissipate through the mesh
extent.
Case Study Scenarios

Two flow scenarios have been analyzed for the
case study as tabulated in Table 3. During normal
flow operation, initial water level elevation of
135.94m will be maintained in the headrace
channel and 130.25m in the tailrace channel
during full operation of 9 units. The model is Fig. 4 Surge inception at time of load rejection
allowed to operate freely for 30 seconds after
which the gates are shut at 1.4m/s to close the
orifice in 5 seconds in order to replicate sudden
closure conditions. Type B surge or the rejection
surge occurs as a result of sudden decrease in
power output. Similarly, during flood flow
operation, initial water level elevation of
136.25m will be maintained in the headrace
channel and 133.26m in the tailrace channel.
Other conditions are kept same as in normal flow
conditions. Fig. 5 Flow reversal after 10 seconds of load
Table 3 rejection
Flow U/S water D/S water In Fig. 3, flow vectors at full flow conditions are
Condition level (m) level (m) illustrated by arrows. This shows an
Normal Flow 135.94 130.25 uninterrupted movement of fluid. After load
Flood Flow 136.25 133.26 rejection, a region of fluid immobility starts to
Normal Flow 135.94 130.25 develop as observed from Fig. 4. This region acts
Flood Flow 136.25 133.26 a cushion against incoming discharge and
reverses its direction which results in surge
Results inception. Fig. 5 shows the region of fluid
The rejection surge occurs as a result of sudden immobility develops upstream with time as surge
decrease in power output. Fig. 3, 4 and 5 illustrate wave begins to achieve greater elevation. The
the surge inception and propagation at different phenomenon holds true for both normal flow
times after load rejection in the extent just conditions and flood flow conditions.
upstream of the gates. The points of flow
direction reversal can easily be tracked for the
upstream advancing surge. This is a feature that
is only possible with computational flow
dynamics analysis since it is not possible to
observe these phenomenon in physical model
testing.
Fig. 6 Change in water elevation vs time after
load rejection for Normal Flow Conditions
Taunsa
Normal Flow Conditions

As the surge moves upstream, change in flow
depth will ensue. FLOW-3D can track the change
in elevation of free surface efficiently for any
spatial location of the model. Changes in surface
elevation with time were observed for three
Fig. 3 Flow through the power unit at full
locations i.e. at turbine unit entrance, 85m
operation
upstream of the entrance and at headrace
entrance.
The temporal and spatial change in free surface
elevation at 3 locations have been shown in Fig.
6. Table 4 shows the summary of results which
depict a maximum surge of 2.25m at unit Conclusion
entrance. Time of arrival of first wave has also Observations of the surge inception and
been tabulated. The surge is expected to reach propagation are in line with the theory and
the headrace entrance in 260 seconds after gate assumptions specified by VenTe Chow and other
closure with an average velocity of 4.84m/s. As authors. For upstream advancing surge, when the
the surge wave moves upstream, the height of surge wave reaches any point in the headrace, the
wave dampens. water elevation behind the wave approximately
equals the maximum elevation of the surge wave.
Table 3 Summary for Normal Conditions The simulation results were compared with the
Max Surge Time mathematical formulae mentioned above as well
Location as the physical testing model carried out at
Surge Elevation (sec)
Unit Irrigation Research Institute, Nandipur, Punjab.
2.25m 138.19m 7 The results of the physical model testing were
Entrance
85m U/S 2.21m 138.15m 14 within ±2% of the results achieved with
Headrace simulations in FLOW-3D. This has built greater
0.53m 136.47m 260 confidence in modelling using computational
Entrance
flow dynamics especially for modelling in field
of power generation.
Applications of computational flow dynamics
have been increasing in engineering applications
over recent times. The computer numerical
models are a cost-effective alternative to physical
modeling techniques offering more flexibility
during design and analysis and can offer insights
Fig. 7 Change in water elevation vs time after into phenomenon that are not apparent in
load reject Flow Conditions Taunsa Headrace physical model testing. However, CFD analyses
limited by computational power. The present 3D
In case of flood flows in Taunsa hydropower, an detailed analysis of a single unit required more
upstream water level of 136.25m is maintained than 24 hours of computation per simulation to
in the headrace with a corresponding tail-water examine a 70-second time-history. A longer
level of 133.26m in tail-race. Other conditions time-history is often more desirable.
are kept same as in normal flow conditions. The The analysis of Taunsa hydropower headrace
surge is expected to reach the headrace entrance channel using computational flow dynamics is a
in 206 seconds with an average surge velocity step-forward in supplementing the results from
of 6.7 m/s. Maximum surge elevation of 138.4m mathematical modelling and conventional
(2.21m) is observed at the power unit entrance. physical model testing in Pakistan. It is not only
The results have been tabulated in Table 5. Fig. a robust analysis software solution but can be
7 shows the change in free surface elevation at employed as efficacious tool in the design
the three selected locations vs time after load process of hydraulic structures where it can offer
rejection from the power unit. extensive flexibility to assess and compare
different proposed designs and their efficiencies.
Table 4 Summary for Flood Conditions This design support can save a lot of time and
Max Surge Time money by the optimization process before final
Location designs are physically tested at Irrigation
Surge Elevation (sec)
Unit Research Institute, Nandipur.
2.21m 138.39m 6
Entrance
85m U/S 2.00m 138.20m 12 References
Headrace 1. VenTe Chow (1959). Open Channel
0.32m 136.52m 206 Hydraulics. United States of America:
Entrance
McGraw-Hill Book Company Inc. 561-565
Flood Flow Conditions 2. Rich, G.R. (1951) Hydraulic Transients, 2nd
It is observed that the surge travels faster and ed., Dover Publications, New York, NY,
achieves a lower maximum free surface U.S.A
elevation for the flood flow conditions 3. M. Hanif Chaudhry (2014). Applied
compared to the Normal Flow Conditions. Hydraulic Transients. 3rd ed. London:
Springer. 450-474
4. Yakhot, V.; Orszag, S. A.; Thangam, S.; 11. Rodi, W.: 1984, Turbulence Models and
Gatski, T. B.; Speziale, C. G. (1992). their Application in Hydraulics, IAHR
Development of Turbulence Models for Shear Monograph, Delft.
Flows by a Double Expansion 12. Vošnjak, S., “Verification of the Flow-3D
Technique. Physics of Fluids A. 4 (7), 1510- mathematical model by a physical hydraulic
1520 model of a small hydropower plant”,
5. Flow Science Inc. Los Alamos New Mexico, presentation, European FLOW-3D User
USA. 1997. FLOW-3D® User’s Manual Meeting 2006, CFD Consultants, Tübingen
Version 9.3
6. Flow Science Inc. Los Alamos New Mexico, DECLARATION
USA. 1997. FLOW-3D® User’s Manual I, the corresponding author, confirm that the
Version 10 Chapter 2: Tutorial authors listed on the manuscript are aware of
7. C.W. Hirt and B.D Nichols. 1981. Volume of their authorship status and qualify to be authors
Fluid (VOF) Method for the Dynamics of Free on the manuscript according to the guidelines
Boundaries, J. Computational Phys. above. I confirm that all of the content, Fig.s
8. Teklemariam, E., Korbaylo, B. W., (drawings, charts, photographs, etc.), and tables
Groeneveld J.L., Fuchs, D.M. 2002. in the submitted work are either original work
Computational Fluid Dynamics: Diverse created by the authors listed on the manuscript or
Applications in Hydropower Project's Design work for which permission to reuse has been
and Analysis, CWRA 55th Annual Conference obtained from the creator. The submitted work
June 11-14. has not been published elsewhere. In case the
9. D.T. Souders and C.W. Hirt, Modeling article is accepted for publication in ‘Proceedings
Roughness Effects in Open Channel Flows of the Pakistan Academy of Sciences’, its
10. Bombardelli, F. A., Hirt, C. W. and García, copyright will be assigned to Pakistan Academy
M. H.: 2001, ‘Discussion on ‘Computations of Sciences. Authors must obtain permission to
of Curve Free Surface Water Flow on Spiral reproduce copyrighted material from other
Concentrators’ ’ by B. W. Matthews, C. A. J. sources and ensure that no copyrights are
Fletcher, A. C. Partridge and S. Vasquez, J. infringed upon.
Hydr. Eng. ASCE 127(7), 629–630.
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Statistical Analysis of inflows of River Indus at Tarbela in the Wake of

Climate Change
Dr. Muhammad Khurram Shahzad1*

1*
Department of Civil Engineering, Institute of Southern Punjab, Multan; Email:
*
[email protected]
Abstract: Pakistan, after signing Indus Basin Treaty with India in 1970, mainly relies upon the discharges from
the three western rivers of Indus Basin i.e. Indus, Jhelum and Chenab along with the inflow of River Kabul that
enters Pakistan from Afghanistan near Nowshera. The major portion of these Western rivers’ inflow is contributed
by River Indus. The effect of climate change on the glacial melt of Himalayan-Karakorum-Hindu Kush (HKH)
region and the regional precipitation of Upper Indus Basin (UIB) is widely reported. However, the impacts of
changing climate could be only verified if its foot prints are reflected in the discharges of the river Indus that
drains the basin. Pakistan holds restricted capability to withstand the impacts of climate change on pushing the
flows towards extreme, excess and otherwise, due to its limited water storage capacity. Similarly, the limited
capacity of conveyance system, natural passages and the regulatory structures, inhibits the high flows to be routed
through the system without flooding. Therefore, the climate change induced scenarios of long term positive /
negative trend and / or increased fluctuations in intra-year and inter-year discharges, both adversely affect the
water regulation of the country. Consequently, this study is focused on the investigation of temporal variation of
flows of Indus at Tarbela. The seasonal and yearly trend is investigated by applying statistical techniques
including Mann-Kendall Test. The 10-daily discharge data of Indus at Tarbela from 1961 to 2010 is used to
conduct the analysis. A less significant decreasing trend is observed in the Kharif season and yearly discharges
of Indus, whereas, significant increasing trend is observed in the Rabi flows.
Keywords: Statistical analysis, River inflows, Tarbela dam
Introduction some of the contributing glaciers will completely

The phenomenon of global climate change is an disappear.
outcome of natural processes and anthropogenic The upstream snow and ice reserves of South and
activities. The natural processes include natural South East Asian River basins; Indus, Ganges,
variability, solar activity, volcano eruption, land Brahmaputra, Yellow, and Yangtze, fed the
& ocean change, where the major anthropogenic water demand of over 1.4 billion people. The
activities affecting the climate are urbanization, people are likely to be affected under different
land use change, aerosols and the emission of scenarios of climate change (Imerzeel, W.W. et
greenhouse gases (Rasul, G. et. al., 2011). The al., 2010). However, there is a conflict of expert
effect of climate change on various opinion in the case of glacial melt of Karakorum-
meteorological parameters like temperature, Hindu Kush (HKH) region. Initially, more runoff
precipitation, humidity etc. has been widely yield was predicted in this region because of
reported. In general, there is a consensus of the increase in global temperature under different
experts, on global warming mainly due to the climate change scenarios (Akhtar, M, et. al.,
emission of GHG; however, its impact on the 2008). However, many studies that have
hydro-meteorological processes is not uniform in analyzed the meteorological processors of this
different geographical regions. Initially, the 3rd region came to different conclusions (Gosain,
assessment report of Intergovernmental Panel on A.K., 2006 et. al.) from the pattern of retreat and
Climate Change (IPCC) has captured the impacts thinning of the glaciers of Eastern Himalaya. In
of climate change on water resources. The fact, the anomaly of thickening of Karakorum
variations in magnitude and frequency of glaciers was reported (Julie, G. et. al., 2012). It is
precipitation have been predicted. It was predicted that Indus and Brahmaputra are more
predicted that global warming will result in the prone to the reduction of discharges under the
diminishing of glaciers due to additional melt, influence of climate change that may affect
that may cause increased discharges around 60 million people (Imerzeel, W.W. et al.,
consequently flooding in few future years but 2010). In a study by Fowler, H.J. and Archer,
after that the flows will sharply decrease because D.R. (2005) the historical analysis of temperature
trends indicated that the winter and summer
temperatures of UIB are increasing and glaciered area is located within Pakistan (Mercer,
decreasing respectively. Where, the verification 1975). More than 80% of total flow of the UIB is
of trend in meteorological parameters could be contributed from less than 20% of its total area.
verified by analyzing the catchment response that This area mainly includes heavy snowfall zone
occurs in the form of runoff. This study focused and glaciered basins above 3500 m elevation
on the investigation of historical discharges of (Hewit, K. et al., 2017). There are thousands of
main stem of Indus River that drains the major glaciers in the Karakorum region with 23 bigger
part of Upper Indus Basin. Descriptive statistical glaciers accounting for about 60% of total glacier
checks are applied to estimate the mean and areas (Young, G.J. and Hewitt, K., 1988). It is
deviation from the mean on seasons of different one of the largest glaciated areas outside the polar
time intervals. A long-term trend in the inflow ice belt. The glaciers of Himalayan-Karakoram
data is explored by applying various statistical region contain huge amount of frozen water
techniques. storage. The temperature rises in the late spring
to early summer initialized the melting of snow
Characteristics of Upper Indus Basin (UIB) in the mountains that reaches to the maximum in
The Indus River emerges from the Tibetan mid-summer months. The salient features of
Plateau and travels initially east to west and then distinct hydrological regimes of UIB has been
north to south descending from above 5500 m to investigated by Archer (2003) and presented
the mean sea level before emptying into Arabian three kinds of controls that are responsible for
Sea. The Indus basin mainly fall in the category Indus flow; summer temperatures, winter &
of semi-arid to arid region, where, the excess spring precipitation, and summer monsoon
moisture is available in the UIB located in the rainfall. Young, G.J. and Hewitt, K. (1988)
region of northern Pakistan ranging up to China reported that snow and ice melt is responsible for
in the east and Afghanistan in the West. The about 80% of flow for the main stem of Indus
HKH range runs from west to east in an arc with above Tarbela. Where, in the years of heavy
Hindu Kush in the west and the main Himalayan winter snowfall followed by cool summer, the
chain in the south east. Where, the central glaciers grow in size and vice versa.
Karakorum region, hosting much of the glaciers, The Tarbela dam, commissioned in 1976 with
is contributing the major portion of Indus flows gross storage capacity of 11.62 MAF, is situated
(Young, G.J. and Hewitt, K., 1988). In UIB, the at the main stem of Indus River at a distance of
Indus River is mainly contributed by the flows 1126 km from its source. The historical record
from two set of sub-basins; 1) the northern sub- indicates that it received the annual runoff of 45
basins includes Hunza River that travel south- MAF to 82 MAF in dry to wet year with an
west, whereas, the Gilgit River moves in from the average of 61 MAF from the drainage area of
North-Western side and merge into Hunza River 169645 km2. The installed hydro-electric power
near Gilgit city, 2) the eastern sub-basins generation capacity of 3478 MW will be
comprises of Shyok River which combines with increased to 6928 MW after the completion of
River Indus, in which Shigar River add its share fourth and fifth extension projects. Where, the
of flow before Katchura gaging site, after which one third of live storage capacity of the Tarbela
this confluence move towards north-west and is reservoir has been lost due to siltation. The
joined by the consolidated flow coming from the operational rule curves of releases from the
northern sub-basins between the gaging sites of Tarbela are based on future water availability
Partab Bridge and Alam Bridge. This combined estimates and the crop water demand of the
discharge in Indus River from the northern and Irrigation system in combination with auxiliary
eastern sub-basins is later joined by the demand for hydroelectric power generation.
contributions from the Astore River and Gorband
River from its left and right side respectively, Descriptive Statistics and Trend Analysis of
before entering into the Tarbela storage. Where, Indus flows at Tarbela
the flow from the Siran River directly enters into The inflows of Indus River at Tarbela are
the Siran Pocket in the Tarbela Reservoir. normally calculated on the basis of volumetric
Indus River accounts for approximately 90% mass balance of inflow, outflow & change in
food production and 13 GW of hydroelectricity storage with the help of rating curve of Tarbela
generation in Pakistan (Cook, E.R., et al., 2013). that is updated from time to time. Where, the data
Due to high altitude, most of the precipitation before the Tarbela dam may have been obtained
occurs as a snow in this region. The outcome is from the nearby gage to compile one monolithic
extensive glaciation in the Indus basin with total data series. The historical daily discharge data of
covered area of 16,300 km2 (Faran Ali and De Indus at Tarbela from 1961-62 to 2010-11 from
Boer, 2003), where, around 13000 km2 of this Water and Power Development Authority
(WAPDA) is used in the descriptive statistical positive and negative, are the results of each
analysis of univariate time series. The 50-year difference of pair that are summed separately.
data is organized into volumetric discharges of The process is repeated in a special array as per
following time intervals: 10-daily, seasonally sequence given in Eq-4. Refer to Eq-4 & 5; the
(Rabi Season: Oct-Mar and Kharif: Apr-Sep), net result of all the positive and negative numbers
annually (hydrologic year: Apr-Mar) and 10- counts indicates whether there is negative or
yearly in order to investigate the characteristics positive trend in the data. The positive S and Z
of flow for shorter, mid-term and longer duration. value indicate positive trend and vice versa.
In the statistical analysis of the discharges the The S value in MK can be computed as:
center (mean and median), terminal limits
(maxima, minima) variability from the mean S = ∑𝑛−1 𝑛
𝑖=1 ∑𝑗=𝑖+1 𝑠𝑖𝑔 (𝑞𝑗 − 𝑞𝑖 ) (4)
(standard deviation and Coefficient of Variance)
and the trend is determined. Conventional non- +1 𝑖𝑓 (𝑞𝑗 − 𝑞𝑖 ) > 0
parametric Mann Kendall (MK) Test is used to
check the trend in data. Before applying the test, Sgn (𝑞𝑗 − 𝑞𝑖 ) = { 0 𝑖𝑓 (𝑞𝑗 − 𝑞𝑖 ) = 0
the data is checked for the presence of serial −1 𝑖𝑓 (𝑞𝑗 − 𝑞𝑖 ) < 0
correlation by computing sample Pearson (5)
correlation coefficient (r) by using Eq-1:
∑(𝑞𝑖 −𝑞̅𝑖 )(𝑞𝑖+1 −𝑞̅𝑖+1 )

𝑟= (1) Where, qi and qj are the discharge data values in
√∑(𝑞𝑖 −𝑞̅𝑖 )2 (𝑞𝑖+1 −𝑞̅𝑖+1 )2
year i and j, and j > I. And n shows the total
number of values in the dataset. As the n values
The presence of serial correlation can be tested in this case were greater than 10, therefore
for any confidence interval, say 95% in this case normal approximation test is used. But prior to
by using two tailed tests as under: that the Variance of S is computed to take into
account the effect of tied (equal) values in the
−1∓1.96√(𝑛−2)
𝑟 (95%) = (2) data series as they reduce validity of normal
(𝑛−1)
approximation
The variance S of data series is computed by
If the serial correlation coefficient for the
Eq-6
discharge data series fall within the interval
computed by Eq-2, than the data is serially 1
correlated, hence, the pre-whitening has to be Var (S) = [𝑛 (𝑛 − 1)(2𝑛 + 5) −
18
conducted before applying the MK test. ∑𝑚
𝑝=1 𝑡𝑝 (𝑡𝑝 − 1)(2𝑡𝑝 + 5)] , (6)
Where, the Mann Kendall Test assumes that
variable under consideration is function of time Where, m is number of tied groups and tp is the
as under: number of ties for pth values. The Z statistics can
be computed as:
𝑞𝑖 = 𝑓(𝑡)𝑖 +∈𝑖 (3)
𝑠−1
𝑖𝑓 𝑆 > 0
In the Eq-3, q is any variable, discharge in this √𝑉𝑎𝑟 (𝑆)
case, of ith time that is dependent upon time with Z= 0 𝑖𝑓 𝑆 = 0 (7)
a residual, ∈𝑖 . 𝑠+1
𝑖𝑓 𝑆 < 0 ,
As per basic concept of statistics, two hypotheses {√𝑉𝑎𝑟 (𝑆)
are considered, Null (H0) and Alternate (H1).
Where, the Null hypothesis states that there is no In order to test the trend, the following two
trend in data and alternate hypothesis indicates tailed tests is applied:
otherwise. The S and Z statistics of MK Test
indicate the trend and significance level of the H0 is rejected if |𝑧| >|𝑍1−𝛼⁄2 | (8)
trend in the data.
The MK test is applied by using the procedure H0 is accepted if |𝑧|<|𝑍1−𝛼⁄2 | (9)
illustrated in various studies (Kendall, M.G.,
1975; Ahmad, I. et. al., 2014; Salmi, T. et. al. Where, 𝛼 indicates the significance level. The
2002). The S value computations are based on trends in the discharge data series of different
pairing the first data point in the series with all time intervals is tested with 𝛼 ranging from 0.1
the subsequent data points and the process is to 0.001.
repeated for the next data point pairing with all After trend analysis the best fit linear line is
subsequent points in the series. The signs, plotted with least squared error optimization in
the scatter plot of discharge time series and their that the flows start rising in the month of March,
respective time dependent equations are possibly due to early snow melt, and reach the
developed. The equation of straight line is used peak by end of July by which time the summer
to plot the trend line: rainfall add into snow melt from the glaciers of
UIB. After hitting the peak, the discharges recede
𝑞𝑖 = 𝛽(𝑡𝑖 ) + 𝐶+ ∈𝑖 (10) gradually till end of august. From September
onwards, the receding in flows increases sharply,
The coefficient of determination (R2) is used as because the summer monsoonal rainfall dries up
an objective function to reach the best fit and temperature in the high altitude start
regression line as under: decreasing, till they hit the base flows in
November. The base flow period remains steady
2
𝑆𝑆𝑟𝑒𝑠 (𝑞𝑖 −𝑞̂𝑓 ) till the end of February. The mean and median,
𝑅2 = 1 − 𝑆𝑆𝑡𝑜𝑡
=1− (𝑞𝑖 −𝑞̅ )2
(11) both are the indicator of data center, the perfect
overlapping of these values reflect un-biasness in
Where, the data, however, the outlier or occurrence of
qi = observed discharges extreme values lead mean-median mismatch as it
qf = discharge estimates by the best fit linear can be seen in peak flow months of July and
regression line august.
Where, the residual ∈𝑖 is not further treated and In general, the discharges of snow-fed rivers
the deterministic part of Eq-10 is used to estimate show strong seasonality with high flows in hot
the discharges. summer months and low flows to no flows in cold
Further, the Weibul distribution (Eq. 12) is used winter months. The seasonal flow variations
to calculate the exceedance probability of from year to year are the outcome of variations in
discharges against the percent of time for plotting inter-annual meteorological forcing especially
the flow duration curves of 1st and the last decade the degree day temperature or radiation etc.
of available data series. Whereas, in the mixed fed rivers, the occurrences
of rainfall result in additional variability in the
𝑀 flow pattern. The plot of minimum, average and
𝑃 = [(𝑛+1)] ∗ 100 (12)
maximum flows on 10-daily basis from the
Where, available record of 50 years starting from year
P = the probability that a discharge will be 1961-62 in Fig-2 shows that there is quite narrow
equaled or exceeded (% of time) range between minimum and maximum flows in
M = the ranked position the Rabi season.
n = the number of events
Results and Discussion

The mean and median plot of 10-daily discharges
of Indus at Tarbela, as shown in Fig-1, indicates
Fig. 1 Mean and Median of 10-daily flows of Fig. 2 Minimum and maximum 10-daily flows of
Indus at Tarbela Indus at Tarbela
However, the flow range widened in the Kharif October. The huge fluctuations of flow in June,
months and reaches to maximum width in the late July and August could be attributed to several
June and remains wide till 2nd 10-daily of August reasons, albedo affect in the late snow and glacial
but after that it starts shrinking till end of melt coupled with the variability in rainfall
received in the monsoon months. The low snow of the mean value that occurred in September.
melt together with scarce rainfalls ends up in low The peak Cv in May, September and February
flows and vice versa. However, the top and may be attributed to the fluctuations in late snow
bottom range are based on historical minimum and glacial melt, occurrences of summer rainfall
and maximum 10-dailies and that does not and early snowmelt respectively.
necessarily mean that they occur in the same
year. The flows in any one 10-daily can be
strongly affected by short term weather
conditions. Consequently, it is quite possible that
one 10-daily receive less than average inflow and
the subsequent one more than average flows and
vice versa.
Fig-4: Characteristics of Seasonal flows of Indus

at Tarbela
Fig 4. volumetric discharges and its fluctuations
Fig. 4 Normally, the volumetric discharges of

long time durations show less fluctuations,
however, it is different in the case of Indus.
Historical Rabi, Kharif and annual flow
characteristics of Indus at Tarbela are plotted in
The maximum recorded Kharif Season and
Fig. 3 Standard deviation and coefficient of annual flows are double than the corresponding
variation of 10-daily Indus flow at Tarbela minimum flows. Whereas, the minimum
recorded flows in Rabi season are 2/3rd of the
The variation of discharges around the mean is maximum recorded flows of Rabi. However, the
best captured by the standard deviation. coefficient of variation is less in the annual data
However, since the standard deviation is series in comparison to Kharif and Rabi flows
dependent upon the magnitude of the mean, that is in according to the general principle of less
therefore, the data with high mean values ends up variation in the averages or sums of longer time
in large standard deviation magnitudes and vice intervals.
versa. Whereas, the coefficient of variation is a The MK tests are applied separately for
unit-less quantity which indicate the deviation in individual 10-dailies and on seasonal and annual
relation to the magnitude of mean values. As basis. The results are given in Table-1. The Z-
anticipated, the shape of standard deviation curve statistics indicate the direction of trend. The
of mean 10-daily discharges of Tarbela, as significance of trend is mentioned in the separate
presented in Fig-3, is similar to the shape of mean column, where the notations like +, *, **, ***
discharges curve as shown in Fig. 1. However, indicates significance level at 90%, 95%, 99%
the coefficient of variation is better representing and 99.9% respectively. In the Kharif season, a
the variations in flows with respect to mean significant positive (rising) trend in noted in the
discharges. Unlike, the curve of mean and month of May, whereas, a non-significant
standard deviation, the plot of Cv is multimodal negative (falling) trend is observed from June to
with at-least three defined peaks occurring in the September. MK test of overall Kharif season
month of May, September and February-March. resulted in non-significant negative trend.
Where, the highest coefficient of variation is 40%
Table 1 Results of Mann Kendall (MK) Test

Time series Test Z Significance Time series Test Z Significance
APR I 2.5 * OCT I 0.4

APR II 0.8 OCT II 1.0
APR III 1.0 OCT III 0.8
MAY I 2.7 ** NOV I 2.2 *
MAY II 3.8 *** NOV II 3.6 ***
MAY III 2.7 ** NOV III 4.2 ***
JUN I -0.6 DEC I 3.0 **
JUN II -1.2 DEC II 2.4 *
JUN III -1.1 DEC III 3.2 **
JUL I -1.0 JAN I 2.0 *
JUL II -1.5 JAN II 2.0 *
JUL III -1.0 JAN III 2.0 *
AUG I -0.5 FEB I 3.4 ***
AUG II -1.0 FEB II 2.5 *
AUG III -1.3 FEB III 3.4 ***
SEP I -0.7 MAR I 2.9 **
SEP II -1.0 MAR II 1.7 +
SEP III 0.8 MAR III 2.6 **
Kharif Season -0.8 Rabi 3.6 ***
Season
Annual -0.2
Fig. 5 Fitting best fit trend line in seasonal flows Indus at Tarbela
The analysis revealed a positive trend in all the Fig-5 presents the Rabi, Kharif and annual
10-dailies of Rabi season. A very significant volumetric discharge of Indus from year 1961-
positive trend is noted in the months of 62 to year 2010-11. The trend line and
November, December and February. Where, respective straight-line equation show slight
the months of January and March also show negative trend in annual and Kharif season
increasing trend with reasonable significance flows, whereas, a visible positive trend can be
level. The outcome of MK test indicates overall observed in the Rabi season flows.
significantly positive trend is in the bulk Rabi
season flows. Where, the annual series has
shown slightly negative trend.
Fig. 6a Trend of decadal accumulative Fig. 6b Trend of decadal accumulative volumetric
volumetric discharges of Indus at Tarbela discharges of Indus for Rabi season
for annual values and Kharif season.
In order to further investigate the seasonal and snow melt, which is the finding of this study.
annual trends, the seasonal discharge volumes These results further agree with the findings of
are added per decade in chronological order to study conducted for swat basin, a sub-basin of
smoothen the yearly variations in the seasonal UIB, by Ahamd, I. et. al. (2014).
volumes. Plot of 10-yearly sum of Kharif and The impact of decade wise reduction in flows is
annual discharge volumes of 5 decades is given further explored by comparing the flow duration
in Fig-6a. The trend line indicates strong curves of yearly discharges of 1st (1961 to 1970)
downwards trend in Kharif flows with coefficient and the last decade (2001- to 2010) in the
of determination of 0.3. However, the coefficient available data series as shown in Fig-7. The log-
of determination of trend line for 10-yearly scale of y-axis indicates the sharp reduction of
accumulative discharge volumes is relatively discharge from 1st to last decade against almost
small, i.e. 0.01. The plot of 10-yearly all the exceedence probabilities.
accumulative volumetric discharges for Rabi
season, as presented in Fig-6b, shows very strong Conclusion
rising trend with a significant high R2 value, i.e. The descriptive statistical analysis of 10-daily
0.85. This rising trend of winter flows of Indus at discharge data series of 50 years from 1961 to
Tarbela endorses the findings of the study by 2011 revealed that on the average discharges
Fowler, H.J. and Archer, D.R. (2005) in which it varies from 09 to 52 MAF in yearly cycle from
is stated that winter temperatures of UIB are Rabi to Kharif season respectively. The means
rising which lead to increase flows in winter. are slightly higher than median discharges,
Similarly, the sinking of mean summer especially, during the peak season which
temperature in UIB as reported Fowler, H.J. and indicates the occurrence of high magnitude
Archer, D.R. (2005) are in harmony with the peaks.
results of current study as the estimated
decreasing summer temperature consequently
results in reduced flows in Kharif due to less
Fig. 7 Flow duration curve for the decade: a) 1961 to 1970, b) 2001 to 2010
The coefficient of variations of 10-daily data of accumulated Kharif and annual flows are
indicated, three peaks in the yearly cycle, that comparatively small (0.3 & 0.01 respectively).
occurred in the months of May (0.39), September Extrapolating the best fit trend line for the
(0.28) and Feb-Mar (0.31). High Cv in these estimation of future discharges indicated that the
months owes to the variations in the performance flows of Rabi increase by 3 MAF (4%), whereas,
of local meteorological processors responsible the Kharif and annual flows decrease by 7 MAF
for early snowmelt, late glacial melt and end (1.3%) & 4 MAF (0.6%) per decade respectively.
summer season rainfalls in UIB. The interesting fact is that around 50% reduction
High inter-annual variations are apparent from in the flows of summer is compensated by the
the fact that maximum recorded flows are twice increased winter flows. Since, no major
as much the minimum flows. Same pattern is anthropogenic activity has been seen in UIB that
observed in the Kharif flows. However, there is could affect the flow pattern; therefore, the trend
slightly less variation in Rabi flows with the in discharges could be attributed to the impacts
minimum flows are 30 % lesser than the of climate change.
maximum flows. The implications of occurrence The analysis of flow duration curves for 1st and
of dry year with total annual flows reducing last decade of discharges in the available data
down to 45 MAF may strictly constraints the verify the reduction of discharge availability
ability of the system to meet the annual water against various exceedance probabilities of time.
demand especially in the absence of carry over The implication of these curves are important
dam. with respect to future operational rule curves of
The test results of MK verified the upward trend Tarbela, especially considering the matter in the
in Rabi flows and downward trend in Kharif and context of continuous reduction of live storage of
consequently annual discharges which may the reservoir and increased water demand from
results in continuous reduction in the capacity of the reservoir for hydroelectric power generation
the system to meet annual water demand. Where, at the completion of 4th and 5th extensions of
the magnitude of reduction has been estimated by Tarbela.
running the trend line in the seasonal and annual
discharge series. However, low coefficient of Acknowledgement
determination (R2) caused by the yearly scatter The Water and Power Development Authority,
deter the author to use the trend lines equations Pakistan is hereby acknowledged for the
for predicting the seasonal and annual flows. discharge data
Therefore, the analysis was repeated on 10- References
yearly accumulated discharges. Evidently, the
scatter reduced in the plot of accumulated Archer, D., 2003: Contrasting hydrological
discharges in large temporal increments. High R2 regimes in the upper Indus basin, Journal of
(0.85) is observed in the equation developed for hydrology, 274, 198-210
the linear trend line of Rabi flows, however, the Akhtar, M., Ahmad, N. and Booij, M.J., 2008:
coefficients of determinations for the equations The impact of climate change on the water
resources of Hindukush-Karakorum- Immerzeel, W.W., Beek, L.P.H. V and Bierkens,
Himalaya region under different glacier M.F.P., 2010: Climate change will affect the
coverage scenarios, Journal of Hydrology, Asian towers, Science, 328, 1382-1385
355, 148-163 Julie, G., Etienne, B. and Yves, A., 2012: Slight
Ali, K.F. and Boer, D.H.D., 2003: Construction mass gain of karakoram glaciers in the early
of sediment budget in large-scale drainage twenty-first century, Nature geoscience,5,
basins: the case of upper Indus River, 322-325
Erosion prediction in un-gauged basins: Khattak, M.S., Babel, M.S and Sharif, M., 2011:
Integrated methods and techniques Hydro-meteorological trend in Upper Indus
(proceedings of symposium), IAHS Publ. Nr. River basin in Pakistan, 46, 103-119
279, 206-215 Kendall, M. G., 1975: Rank Correlation
Ahmad, I., Tang, D., Wang, T., Wang, M. and Methods, Griffin, London, UK
Wagan, B., 2014: Precipitation trend over Mercer, J.H., 1975: Glaciers of the Karakorum,
time using Mann-Kendall and Spearmann’s Mountain glaciers of the northern
rho tests in Swat River basin, Pakistan, hemisphere, 1, 371-409
Advances in Meteorology, 2015, 1-15 Rasul, G., Chaudhry, Q. Z., Mahmood, A, Hyder,
Cook, E.R., Palmer, J.G., Ahmad, M., K. W., Dahe Q., 2011: Glaciers and Glacial
Woodhouse, C.A., Fenwick, P., Zafar, M.U., Lakes under Changing Climate in Pakistan,
Wahab and M. Khan, N., 2013: Five Pakistan Journal of Meteorology, 8, 1-8
centuries of upper Indus river flow from tree Salmi, T., Maata, A., Anttila, P. Rouho-Airoloa,
rings, Journal of hydrology, 486, 365-375 T. and Amnell, T., 2002: Detecting trends of
Fowler, H.J. and Archer, D.R, 2005: Conflicting annual values of atmospheric pollutants by
signals of climate change in the Upper Indus the Mann-Kendall Test and Sen’s Slope
Basin, Journal of Climate, 19, 4276-4292 estimates – The excel template application
Gosain, A.K., Rao, S. and Basuray, D., 2006: Makesens, Finnish Meteorological Institute,
Climate change impact assessment on 31, 1-35
hydrology of Indian river basins, Current Young, G.J. and Hewitt, K., 1988: Hydrology
Science, 90, 346-353 research in the upper Indus basin, Karakorum
Hewitt, K., Wake, C.P., Young, G.J. and David, Himalaya, Pakistan, Hydrology of
C., 2017: Hydrological investigations at Mountainous Areas, Proceedings of the
Biafo glaciers, Karakorum range, Himalaya; Strbske Pleso workshop, IAHS Nr. 190, 139-
an important source of water for the Indus 152.
river, Annals of glaciology, 13, 103-108
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Scientific Evaluation of Water Footprint Methods for Hydropower
Generation
F. Seemab1, W. Sherani1, M. Abbas1, M.N. Alam1, R. Farooq1, Z. Ashfaq1, M.A.U.R. Tariq1*

1*
Capital University of Science and Technology, Islamabad
*
[email protected]
Abstract: The water footprint has attracted interest as a standard that indicates the use of freshwater resources
and its impacts. Although hydropower is most economical and environmental friendly, it also has a large water
footprint per unit energy as compared to other sources. The methodology to quantify the water footprints of
hydropower is in its preliminary stages. Currently, there are three developed methods namely WF1, WF2, and
WF3. These methods have their own limitations but are widely used to calculate water footprint for different
hydropower plants. The aim of this paper is to evaluate and perform a detailed systematic assessment of these
water footprint methods for hydropower generation. The task includes the assessment of complexity and sensitivity
of the involved factors, practical implications, as well as the accuracy and reliability along with the limitations
that the method inherits. First two methods are based on the consumptive use, while the third considers input and
output of the reservoirs. WF1 is a single parameter simple, and easy to measure method with reliable and
consistent results. However, it grossly ignores the water losses due to evapotranspiration that might occur in the
absence of reservoir, which is considered in WF2. The practical implications in measurement of
evapotranspiration and the demarcation of the system boundary are the two biggest challenges for the WF2 that
may arises the issues of accuracy, reliability and consistency of the results. WF3 accounts the net water balance
by considering the water leaving and entering the surface of reservoir, namely the evaporation as output and
precipitation as input. WF1 and WF2 results have always a positive value whereas WF3 can give negative value
if the input is greater than the output which means the net water is surplus. The negative value of WF3 still
requires physical based interpretation of WF of a hydropower project.
Keywords: Water footprint methods, hydropower generation

of so-called green water. Green water indicates
Introduction the precipitation on land that does not run off or
Water footprint is an indicator of use of water that recharge the groundwater but is stored in the soil
measures the volume of water used to produce a or temporarily stays on top of the soil or
product. It is mainly divided into two vegetation. This footprint is the volume of
components i.e. water consumed during rainwater consumed during the production
production and water polluted during production process. It is mostly used in agriculture and
process. Water consumption refers to the water forestry. The grey water footprint refers to the
that is lost from the reservoir or from the volume of water that is required to assimilate
available catchment area due to evaporation and waste, quantified as the volume of water needed
evapotranspiration processes whereas water to dilute pollutants to such an extent that the
footprint refers to the water which is used for quality of the ambient waste remains above
production of goods which are used by people. agreed water quality standards. (Herath, Deurer,
Water footprint is divided into three parts i.e. Horne, Singh, & Clothier, 2011) WFs can be
blue, green, and grey. The blue water footprint is assessed for different entities, for example,
an indicator of consumptive use of fresh surface products, consumers, businesses, nations, or
water or groundwater. The term ‘consumptive humanity.
water use’ refers to one these four cases: 1) Water Water is vital and essential for energy
evaporation, 2) Water incorporation into the production. It is a hidden input for almost all
product, 3) Water not returning to the same types of power plants. It is used to grow biomass
catchment area e.g. it is returned to sea 4) water for bioenergy and in the extraction, processing,
not returning in the same period i.e. withdrawn in and transport of fuels for energy production.
a scarce period and returned in a wet period. The Energy and water are two critical and important
blue water footprint measures the amount of necessities for modern world. Thermal power
water available in a certain period that is plants require a huge amount of cooling water
consumed (in other words, not immediately and hydropower plants results in evaporative loss
returned within the same catchment). The green from large surface areas of the reservoirs (Lee,
water footprint is an indicator of the human use Han, Elgowainy, & Wang, 2017). The WF of
electricity refers to the water volumes consumed
and polluted in the different stages of the supply Where RF is the annual volume of rainfall falling
chain of electricity. It is distinguished in three on the reservoir (m3) (Herath et al., 2011)
major stages in the production of electricity i.e.
fuel supply, construction, and operation. The first Problem statement
stage is relevant only for fuel-based electricity There is no well documented and accepted
(when electricity is based on coal, lignite, oil, methodology to quantify the WF of
gas, uranium, or biomass). In the other cases hydroelectricity (Herath et al., 2011). However,
(hydro, solar, wind and geo-electricity), we only there are three developed formulae to measure
consider two production stages: construction and the water footprint. There is no consensus as to
operation. There are three methods adopted for which method should be adopted as a standard
the calculation of water footprint of hydropower (Coelho et al., 2017).These methods have their
generation i.e. WF1, WF2, and WF3. (Herath et own limitations but are widely used to calculate
al., 2011) water footprint for different hydropower plants.
The aim of this paper is to investigate the detailed
WF-1: (consumptive water use) systematic assessment of these water footprint
In this method the water footprint can be methods for hydropower generation.
calculated as the evaporative water loss from the
surface of the reservoir divided by the energy Literature review
produced by that hydropower plant (Herath et al., Hoekstra introduced the concept of water
2011) footprint in 2007. The three adopted
𝐸𝑜 methodologies of WF calculation have been used
𝑊𝐹1 = widely throughout the world for WF estimation.
𝑃
Here, Eo is the annual open-water evaporative Gleick evaluated the gross consumption by plants
loss from the reservoir (m3) and P is the energy in USA in the 90’s. This method was
production of the power plant (GJ). conceptualized by Hoekstra in 2007 and is now
known as WF1. WF2 was introduced because
WF-2 (net consumptive approach) WF1 does not consider the situation before the
The second approach also considers consumptive dam. (Herath et al., 2011) worked on the case
water use, but it compares the consequences of study of New Zealand and proposed WF3 stating
land use changes created by the dam. Building of that regional climatic conditions are not
a dam results in the replacement of vegetation by considered in the first two methods. Since, the
a free-water surface. Thus, evapotranspiration introduction of WF3 it has been rarely used and
from the vegetation is replaced by open-water thus has been used so far for only three model
evaporation from the reservoir. Taking this into cases i.e. Brazil, Norway and New Zealand. WF1
account, the WF-2 (m3/GJ) considers the net and WF2 have been used widely by all the
evaporative water loss from the area occupied by scientists. Hoekstra has used these two methods
the reservoir (Herath et al., 2011). on 35 plants out of which 7 are in Brazil and 12
are in remaining Latin America. Furthermore,
𝐸𝑜 − 𝐸𝑇𝑐 many others have used these two methods in
WF2 = ( )
𝑃 different case studies of Austria, Egypt, China,
Norway, New Zealand and many others (Coelho
ETc is the amount of water lost by et al., 2017).
evapotranspiration. The parameters of evaporation,
evapotranspiration, and rainfall will be taken
WF-3 (net water balance) from the nearest meteorological station to the
This moves beyond the simply consumptive use reservoir. WF1 yields high values as compared to
definition of the water footprint. A simple water WF2 whereas WF3 can yield negative values if
balance was used to estimate the water footprint the amount of rainfall occurred on a reservoir is
considering supply side of hydrology. WF3 greater than the evaporation rate.
accounts the net water balance by considering the
water leaving and entering the surface of Methodology
reservoir. The net water balance is calculated as: The three methods normally used in WF
𝑂𝑢𝑡𝑝𝑢𝑡 − 𝐼𝑛𝑝𝑢𝑡 = 𝐸𝑣𝑎𝑝𝑜𝑟𝑎𝑡𝑖𝑜𝑛 − 𝑅𝑎𝑖𝑛𝑓𝑎𝑙𝑙 calculation yields three different results. The aim
So, WF3 is calculated as: is to evaluate scientifically that which method out
of these is the most accurate, reliable, and easy to
𝐸𝑜 − 𝑅𝐹 measure based on scientific parameters.
WF3 = ( )
𝑃
Theoretical legitimacy and power production is minimum. WF1 will
WF1 follows the definition of water footprint yield lower values for the opposite scenario. WF2
given by Hoekstra and Chapagain (2007). values can come close to WF1 if a period occurs
Furthermore, it also follows the definition used where evapotranspiration before the construction
by (Gerbens-Leenes, Hoekstra, & van der Meer, of the dam was not significant. WF3 can yield
2009) to estimate the WF of hydropower on negative values for the month in which there is
global average basis. This method was suggested large amount of rainfall (Coelho et al., 2017).
by (Mekonnen & Hoekstra, 2012) who used this
method to estimate water consumption of Sensitivity to Area
hydropower plants. The method accounts the WF1 is most sensitive to reservoir area. The
water consumed in the process under surface area of reservoir accounts for 38.6%
consideration. The second approach also percent variability of WF1. WF2 is not
considers consumptive water use, but it compares significantly sensitive to the reservoir area but
the consequences of land use changes created by WF3 is also sensitive to this parameter because it
the dam. A similar approach has recently been relates to the effectiveness of dam to collect
discussed in WF assessment of beer, and this was rainfall. The major part of variability of WF1 and
termed “net green water” loss (SABMiller, 2009) WF3 is due to local climate and reservoir area.
. In the third method, the focus moved beyond the But still there is an unexplained variability which
consumptive use. WF3 was developed to assess may be due to the difference in efficiency of
the hydrologically rational water footprint to power generation (Herath et al., 2011).
quantify the influences of regional climatic
conditions (Herath et al., 2011). WF1 and WF2 Comparison between dams for Water
measure the consumptive use but pay no attention efficiency
to supply side of hydrology which is accounted In the case of comparison between multiple
in WF3. dams, if the WF1 values of a dam results in lower
WF1 is controversial as it appears to be too values as compared to other dams than it
simple by not considering the water lost by demonstrates that it has higher water efficiency
evapotranspiration before reservoir as compares to others. Similarly, in case of WF2
impoundment. WF3 was also criticized stating if the values are lower a dam in comparison to
that the method simply indicates if the others than it has high water efficiency as
evaporation is greater than the precipitation on compared to the others.
the reservoir surface divided by the power
production. WF3 was further elaborated that this Sensitivity based on Region
method is important for weighing the advantages For tropical regions, WF1 is not suitable for
and disadvantages of building a dam in a certain explaining the consumption because
region (Bakken, Killingtveit, Engeland, evapotranspiration values are high even before
Alfredsen, & Harby, 2013). Based on theory the construction of the dam. Thus, this method
WF2 seems the most legitimate as it estimates the allocates excessive water consumption to
WF based on the condition before and after the hydropower generation that already existed
construction of the reservoir. before its implementation. This effect is
minimized when applied to temperate regions
Results comparison where evaporation rates are low before
WF1 results in larger values as compared to WF2 construction of the dam. WF2 is more accurate
whereas WF3 also yields negative values when than WF1 because it considers
input is greater than the output. The negative evapotranspiration into consideration which has
value suggests that a reservoir has received more a significant value for tropical regions as
water as rainfall than they lost through compared to temperate regions. The accuracy of
evaporation. With reference to the results WF1 WF2 is more than WF3 because the latter method
and WF2 are reliable and WF3 results seems does not consider the real consumption of water
unrealistic because negative value is unreliable and also does not considers the
and unacceptable (Herath et al., 2011). evapotranspiration before the construction of the
reservoir (Coelho et al., 2017).
Variability based on change in temperature
throughout the year Scientific Innovation
The values of WF1 vary with respect to the The study focuses on the criticality of each
evaporation and power production. The highest parameter involved in the evaluation of water
value of WF1 will be observed for the month/ footprint. Furthermore, complexity, the ease of
period in which evaporative loss is maximum, measure, accuracy and reliability of results
obtained by these methods lies in the scope of this negative for those months in which the rainfall is
paper. greater than evaporation while it yields positive
values during summer time as evaporation rate is
Limitation very high in contrast to the precipitation.
This study ignores economic perspective and is WF1 and WF3 are sensitive to the reservoir area.
focused only on water consumption. Among all If reservoir area increases than the rate of
methods, the deleterious impacts that the dam evaporation also rises. Hence evaporation
might have on the local ecosystem i.e. water increases which results in increase in water
scarcity is ignored because all three methods are scarcity. This scarcity is found out by evaluating
focused purely on water quantity. the water scarcity index. Water efficiency is the
most important factor for evaluating the
Result/Output hydropower capacity of a reservoir. By
A scientific evaluation of the results obtained evaluating the hydropower capacity of multiple
from the past studies are considered and dams, if it is found out that WF1 values of one
compared to check the complexity, ease of dam is lower as compared to the remaining dams
measure, accuracy, and reliability of the methods than this dam has greater water efficiency which
and to devise the most appropriate method to be means that evaporation rate is lower, or the power
used for future calculations of WF. WF1 is too generation capacity of this particular dam is high.
simplistic and follows the definition of Hoekstra The power generation capacity is also dependent
but yields large values that appear to be on the speed of spinning of the turbine.
unrealistic and unreliable whereas WF3 has not The water footprint methods are region sensitive
been backed by scientists and is found out to be which means that in a region where evaporation
region sensitive, because of these factors WF3 is lower, and rainfall is higher, than in that region
has not been widely used till now. WF2 appears the WF3 values will appear negative but for
to be the most accurate method because it tropical region the WF3 values will be unreliable
compares the situation before and after as the evapotranspiration plays a significant role
construction of reservoir. This method is being in such region. Similarly, for the case of
theoretically and scientifically backed by many temperate region where evapotranspiration is not
scientists like Bakken. Moreover, the method is that high as compared to other region, then in this
not sensitive to the area and gives reliable results case WF3 will yield reliable results but in this
based on region. case results obtained by WF2 are also reliable.
Discussion Conclusion
From water footprint perspective, it is WF1 and WF2 only considers the consumptive
advantageous to locate hydroelectric dams in wet water use but not the supply side of hydrology
regions where evaporation rate is low. which is accounted in WF3. WF1 and WF2 are
Furthermore, the WF is also an indicator of water inadequate to understand the impact of
efficiency which will indicate the power hydroelectric power generation on freshwater
generation capacity of a dam. Theoretically, WF2 resource availability. Water related issues are
is the soundest method because it compares the highly local and WF3 shows sensitivity to it. The
situation before and after construction. The study also shows the effect of local climate
values obtained by WF2 are reasonable and differences and the surface area of the reservoirs
justifiable as both evaporation and on the water footprint calculated using different
evapotranspiration are obtained from methods. The methods are still under
meteorological stations and can be incorporated development and needs further research and
to evaluate water footprint. WF1 on the other understanding of the concept to refine the
hand yields large values which shows over evaluation techniques. All methods have their
estimation whereas WF3 can yield negative value pros and cons, but detailed systematic assessment
which altogether seems absurd and inappropriate. will help in proposing the most efficient method
The values of these three water footprints vary for hydropower evaluation. Based on theoretical
with temperature and change of climate understanding, complexity evaluation and
throughout the year. In the case of dry climate, reliability and ease consideration, WF2 seems to
the values of WF1 are highest as compared to be the most reliable for the time being but still
WF2 because the gross evaporation is greater this method needs to be studied further to
than the net evaporation. The values of WF2 and incorporate the precipitation in this method.
WF1 can sometimes be approximately same for
the months in which the evapotranspiration is Recommendation
almost negligible. For WF3 the values can appear WF2 method does not include rainfall on the
reservoir and it needs research in future to Lee, U., Han, J., Elgowainy, A., & Wang, M.
incorporate precipitation in this formulation. Still (2017). Regional water consumption for
there is no study to incorporate seepage losses in hydro and thermal electricity generation in
these methods. Detailed analysis is needed in the United States. Applied Energy.
future to incorporate seepage and ecosystem Mekonnen, M., & Hoekstra, A. (2012). The blue
effects in all three methods. water footprint of electricity from
hydropower. Hydrology and Earth System
References Sciences, 16, 179-187.
Bakken, T. H., Killingtveit, Å., Engeland, K., Oel, P., & Hoekstra, A. (2010). The green and
Alfredsen, K., & Harby, A. (2013). Water blue water footprint of paper products:
consumption from hydropower plants– Methodological considerations and
review of published estimates and an quantification.
assessment of the concept. Hydrology and Pahlow, M., Snowball, J., & Fraser, G. (2015).
Earth System Sciences, 17(10), 3983-4000. Water footprint assessment to inform water
Bakken, T. H., Modahl, I. S., Engeland, K., management and policy making in South
Raadal, H. L., & Arnøy, S. (2016). The life- Africa. Water SA, 41(3), 300-313.
cycle water footprint of two hydropower Papadopoulou, M., Charchousi, D., Tsoukala, V.,
projects in Norway. Journal of Cleaner Giannakopoulos, C., & Petrakis, M. (2016).
Production, 113(Supplement C), 241-250. Water footprint assessment considering
doi: climate change effects on future agricultural
https://doi.org/10.1016/j.jclepro.2015.12.03 production in Mediterranean region.
6 Desalination and Water Treatment, 57(5),
Bakken, T. H., Killingtveit, Å., & Alfredsen, K. 2232-2242.
(2017). The Water Footprint of Hydropower Scherer, L., & Pfister, S. (2015). Water scarcity
Production—State of the Art and footprint of selected hydropower reservoirs.
Methodological Challenges. Global Scherer, L., & Pfister, S. (2016). Global water
Challenges. footprint assessment of hydropower.
Bueno, E. d. O., Mello, C. R. d., & Alves, G. J. Renewable Energy, 99(Supplement C), 711-
(2016). Evaporation from Camargos 720. doi:
hydropower plant reservoir: water footprint https://doi.org/10.1016/j.renene.2016.07.021
characterization. RBRH, 21(3), 570-575. SABMiller, W. (2009). Water footprinting:
Coelho, C. D., da Silva, D. D., Sediyama, G. C., identifying & addressing water risks in the
Moreira, M. C., Pereira, S. B., & Lana, Â. M. value chain. SABMiller, Woking, UK, and
Q. (2017). Comparison of the water footprint WWF-UK, Goldalming, UK.
of two hydropower plants in the Tocantins Chavez-Rodiguez, Mauro F., Silvia, A.N., 2010.
River Basin of Brazil. Journal of Cleaner Assessing GHQ emissions, ecological
Production, 153, 164-175. footprint, and water linkage for different
Costa, L., & Neto, J. A. (2017). Proposal for a fuels. Environmental Science and
water resource management strategy model Technology 44, 9252-9257.
using the water footprint concept. Brazilian Zhou, Y., Hejazi, M., Smith, S., Edmonds, J., Li,
Journal of Operations & Production H., Clarke, L., Calvin, K. and Thomson, A.,
Management, 14(3), 371-380. 2015. A comprehensive view of global
Egan, M. (2011). The Water Footprint potential for hydro-generated
Assessment Manual. Setting the Global electricity. Energy & Environmental
Standard. Social and Environmental Science, 8(9), pp.2622-2633.
Accountability Journal, 31(2), 181-182. doi: Abbasi, S.A. and Abbasi, N., 2000. The likely
10.1080/0969160X.2011.593864 adverse environmental impacts of renewable
Gerbens-Leenes, W., Hoekstra, A. Y., & van der energy sources. Applied Energy, 65(1),
Meer, T. H. (2009). The water footprint of pp.121-144.
bioenergy. Proceedings of the National Yu, B. and Xu, L., 2016. Review of ecological
Academy of Sciences, 106(25), 10219-10223. compensation in hydropower
Herath, I., Deurer, M., Horne, D., Singh, R., & development. Renewable and Sustainable
Clothier, B. (2011). The water footprint of Energy Reviews, 55, pp.729-738.
hydroelectricity: a methodological Van der Ent, R.J., Savenije, H.H., Schaefli, B.
comparison from a case study in New and Steele‐Dunne, S.C., 2010. Origin and
Zealand. Journal of Cleaner Production, fate of atmospheric moisture over
19(14), 1582-1589. continents. Water Resources
Research, 46(9).
Fthenakis, V. and Kim, H.C., 2010. Life-cycle Döll, P., Fiedler, K. and Zhang, J., 2009. Global-
uses of water in US electricity scale analysis of river flow alterations due to
generation. Renewable and Sustainable water withdrawals and reservoirs. Hydrology
Energy Reviews, 14(7), pp.2039-2048. and Earth System Sciences, 13(12), pp.2413-
Hammond, G.P., 2006, February. ‘People, planet 2432.
and prosperity’: The determinants of Postel, S.L., 2000. Entering an era of water
humanity's environmental footprint. scarcity: the challenges ahead. Ecological
In Natural Resources Forum (Vol. 30, No. 1, applications, 10(4), pp.941-948.
pp. 27-36). Blackwell Publishing Ltd. Matthews, H.S., Hendrickson, C.T. and Weber,
Hertwich, E.G. and Peters, G.P., 2009. Carbon C.L., 2008. The importance of carbon
footprint of nations: A global, trade-linked footprint estimation boundaries.
analysis. Environmental science & Druckman, A. and Jackson, T., 2009. The carbon
technology, 43(16), pp.6414-6420. footprint of UK households 1990–2004: a
Torcellini, P.A., Long, N. and Judkoff, R., 2003. socio-economically disaggregated, quasi-
Consumptive water use for US power multi-regional input–output
production. model. Ecological economics, 68(7),
pp.2066-2077.
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Evaluation of Suitable Design Flood Frequency Approaches for

Hydropower Structures on the Mountainous Rivers (A Case Study of
Upper Indus Basin)
Muhammad Yaseen1*, Ijaz Ahmad2, Bilal Nasir1, Muhammad Imran Azam3, Muhammad Hassan
Rehman2, Muhammad Afzel2
1
Centre for Integrated Mountain Research, University of the Punjab, Lahore, Pakistan
3
Centre of Excellence in Water Resources Engineering, University of Engineering and Technology,
Lahore 54890, Pakistan
2
College of Hydraulic and Environmental Engineering, China, Three Gorges University, Yichang
443002, China
*
[email protected]
Abstract: The proper assessment of design flood is a major concern for many hydrological and hydraulic
applications in mountainous watersheds. A number of statistical approaches can be used; however, each method
has its own limitations and assumptions being applied to the real world. The aim of this study is to evaluate the
most suitable design flood frequency approaches in the mountainous watersheds of Upper Indus Bain (UIB),
Pakistan. Different flood frequency approaches i.e. Gumbel, Weibull, Normal, Lognormal, Lognormal Type 3,
Pearson Type 3 and Log Pearson Type 3 were applied at 33 stream gauges of 21 watersheds in UIB. These
approaches were evaluated by the chi-square test. Log Pearson-III frequency distribution showed good results
with minimum chi-suare test value on 30 stream gauging stations as compare to Gumbel distribution.
Keywords: Flood frequency analysis, Indus River basin, distributions
Introduction The flood discharge adopted for the design of

Reliable estimates of extreme flood events are hydraulic structures while putting economic and
required for the design and operation of vital hydrological factor into consideration is known
infrastructures and also for more general flood as design flood and as it (design flood increases),
risk management and planning. This information the cost of the structure decreases but there will
is generally obtained through the use of flood be a reduction in the probability of yearly
frequency estimation techniques. These damage. It is of great importance to select a
techniques are based upon the principle of design flood, which is unlikely to occur during
statistical analyses of series of observed flood the design life of a hydraulic structure and
events, providing estimates of the likely difference between the design return period and
magnitude of future extreme events through the estimated life of the structure should be quite
extrapolation. large. Large return period of hydraulic structure
The Indus River is the main river tributary in (spillway of dams) is taken to reduce hazard of
which most of the smaller streams originating failure (Izinyon and Igbinoba, 2011).
from mountainous areas of Himalayas of Frequency analysis is a procedure for estimating
Baltistan, Pakistan. The Indus River provides a the frequency of occurrence or probability of
major source of water for agriculture and occurrence of the past or future events. A number
hydropower. Due to lot of human activities in the of approaches can be used including statistical
inundation plain, it becomes important for the approach and the continuous simulation and
protection of floods. Some of the structures are design storm methods. However, each method
constructed by human to control floods are has its own limitations and assumptions being
leaves, reservoir and channel improvement. For applied to the real world. Statistical approaches
an economic and efficient design of these (Gumbel, Weibull, Normal, Lognormal,
measures, flood has been estimated with Lognormal Type 3, Pearson Type 3 and Log
particular level of accuracy. Maximum or peak Pearson Type 3) were used to determine the
flood can be estimated which may occur in a inaccuracy of data as well as design flood for
particular site; an ideal solution can then be particular site. Reliable flood frequency
proffered by a hydraulic Engineer (Haan, 1977). estimates are essential for flood plain
management, protection of public infrastructure,
cost reduction of food related matters to 286,000 km2 (see Figure 2). The study area
government and private enterprises, for assessing contained three major basins namely Jhelum,
hazards related to the development of flood Indus and Kabul. There were 21 sub-basins in
plains and epidemic control (Tumbere, 2000). In all these three basins. The stream flow gauges are
order to ensure safety and economic hydrologic installed at all the sub-basins of Upper Indus
design in the catchment area, the Gumbel Basin on different locations are shown in Fig.1.
distribution, a stochastic generating structure that The flow data of these sites will be collected from
produce random outcomes was used to model the Surface water Hydrology Project (SWHP),
annual peak discharge data of indus basin. Not WAPDA for the period 1961-2012. The
only the river has its prime importance in the geographical distribution of these stations is
river basins of Pakistan but also has the gratitude shown in Fig. 1. Characteristics of these selected
among the biggest rivers in the world. In gauging sites are given in Table 1.
addition, the country has an agro based economy Determination of the frequency of occurrence of
with arid or semi-arid climatic conditions. extreme hydrologic events is very important in
Therefore, accurate estimates of flood water resources planning and management.
magnitudes become an important part of the Frequency analysis was done by using
design and operation of water resource systems, HyfranPlus tool with applying the Gumbel,
land use planning and management, flood Weibull, Normal, Lognormal, Lognormal Type
insurance assessment and protection of inhabited 3, Pearson Type 3 and Log Pearson Type 3
areas, etc. distribution. One of the most commonly used
tests for goodness of fit of empirical data to
The paper presents results of flood frequency specified theoretical frequency distributions is
analysis of annual maximum peak (AMP) flows the Chi-Square Test. This test makes a
observed at 33 stream gauges of 21 watersheds in comparison between the actual number of
UIB. Present study estimates the occurrence observations and the expected number of
exploration of Indus basin gauging stations. On observations (expected according to the
the basis of our analysis, we find out flood distribution under test) that fall in the class
quantity on any return period. intervals. Expected numbers of observations are
calculated by multiplying the expected relative
Description of study area frequency by the total number of observations.
Present study was carried in the Upper Indus The chi-square test value was calculated by the
Basin (UIB). The catchment of this basin falls in following equation.
range 33º, 40ʹto 37º, 12ʹ N latitude and 70º, 30ʹ to 𝐗 𝐂 = (O − E) 2/E
77º, 30ʹ E, longitude. The Upper Indus basin Where
boundary was derived from Digital Elevation O=observed Values
Model (DEM) just upstream of Massan as shown E=Expected Values
in Fig.1. Most of area of this catchment is lies in The expected frequencies are computed on the
China and India. Due to unavailability of data basis of a hypothesis. If under the hypothesis the
from China and India, the study area was computed value of X C is greater than some
confined to the catchment falling within Pakistan critical value. This test is called Chi-square test
boundary as shown in Fig.1. The elevation varies of hypothesis. If chi-square results well enough
from 254 m to 8570 m above mean sea level. for all distribution than we select distribution
There are so many rivers which contribute water having least chi-square value.
to main Indus River.
The important sub basins are Chitral, Swat, Results and discussions
Kabul, Hunza, Gilgit, Astore, Shigar, Shyok, Table 2 presents the summary of chi-square test
Kunhar, Neelum, Kanshi, Poonch, Soan, Siran, results values for different distributions applied
Sil, Haro etc. Indus River originates from the at different gauging stations of Upper Indus
north side of the Himalayas at Kaillas Parbat in Basin. The chi-square test value helps in
Tibet having altitude of 18000 feet. selection of suitable frequency distribution, the
chi-square test value should be less than 12
Material and Methods (C.T.Haan). However, if the values of Chi-square
Stream flows measurement of UIB is carried out test for more distributions remain below or equal
by the Water and Power Development to 12 then the distribution with relatively lowest
Authority—Surface Water Hydrology Project value of test is selected. Table representing the
(WAPDA-SWHP) with the earliest records chi-square Test results for Gumbel and Log
commencing in 1960. The stream gauges have a Pearson-III
wide range of drainage area from 262 km2 to
Fig. 1: The Upper Indus Basin confined in Pakistan boundary showing rivers, stream gauges and
elevation.
Fig. 3: (a) Number of gauges by drainage basin area. (b) Number of gauges with different
length of discharge records.
distributions of different stream gauging stations the point of Yogo stream gauging station of
of Upper Indus Basin. At Garhi Habibullah Shyok River both distributions representing the
stream gauging station of Mangla (Sub-Basin) data adequately, but the value of Chi-Square Test
Log-Pearson Type-III representing the good for Log-Pearson III is 3.15 as compare to Gumbel
results with lowest value of Chi-Square as Method. Similarly, at the point of Khairabad both
compare to Gumbel Distribution. In Mangla distributions representing the data adequate with
River basin, Log-Pearson Type-III for six stream low values of chi-square test, but Gumbel is more
gauges (out of 8) has the most suitable results. At fit on the data as compare to Log-pearson III with
lowest value of chi-square test. Generally, Log lowest value of Chi-Square Test on Upper Indus
Pearson III is suitable as compare to Gumbel with Basin.
Table 1 List of stream gauges used in the present study and their characteristics.
S. N. Station Lat Lon River Basin Area No of Mean Annual
(dd) (dd) (Km2) Years Streamflows
(cumec)
1 Naran 34.9 73.7 Kunhar Jhelum 1036 43 227.84
2 G. 34.4 73.4 Kunhar Jhelum 2355 43 432.18
Habibullah
3 Muzaffarabad 34.4 73.5 Neelum Jhelum 7275 43 1373.69
4 Chinari 34.2 73.8 Jhelum Jhelum 13598 43 922.91
5 Domel 34.4 73.5 Jhelum Jhelum 14504 38 1106.11
6 Kohala 34.1 73.5 Jhelum Jhelum 24890 43 2706.49
7 Azad Pattan 33.7 73.6 Jhelum Jhelum 26485 43 2946.63
8 Kotli 33.5 73.9 Poonch Jhelum 3238 43 1724.24
9 Palote 33.2 73.4 Kanshi Jhelum 1111 43 295.87
10 Kharmong 35.2 75.9 Indus Indus 67858 31 1871.07
11 Yogo 35.2 76.1 Shyok Indus 33670 41 2252.5
12 Shigar 35.4 75.7 Shigar Indus 6610 18 1107.72
13 Kachura 35.5 75.4 Indus Indus 112665 43 4970.11
14 Gilgit 35.9 74.3 Gilgit Indus 12095 43 1350.42
15 Dainyor Br. 35.9 74.4 Hunza Indus 13157 43 1571.25
16 Alam Br. 35.8 74.6 Gilgit Indus 26159 43 2760.36
17 Bunji 35.7 74.6 Indus Indus 142709 12 2053.94
18 Doyain 35.5 74.7 Astore Indus 4040 40 626.12
19 Shatial Br. 35.5 73.6 Indus Indus 150220 30 9425.71
20 Karora 34.9 72.8 Gorband Indus 635 36 186.19
21 BeshamQila 34.9 72.9 Indus Indus 162393 43 10605.37
22 Daggar 34.5 72.5 Brandu Indus 598 43 90.13
23 Phulra 34.3 73.1 Siran Indus 1057 43 242.74
24 Kalam 35.5 72.6 Swat Kabul 2020 43 361.33
25 Chakdara 34.6 72 Swat Kabul 5776 43 757.32
26 Chitral 35.9 71.8 Chitral Kabul 11396 43 1067.31
27 Jhansi Post 33.9 71.4 Bara Kabul 1847 43 64.87
28 Nowshera 34 72 Kabul Kabul 88578 43 3303.74
29 Gurriala 33.7 72.3 Haro Indus 3056 43 611.26
30 Khairabad 33.9 72.2 Indus Indus 252525 26 11702.71
31 Thal 33.4 71.5 Kurram Indus 5543 43 213.92
32 Chirah 33.7 73.3 Soan Indus 326 43 224.12
33 Chahan 33.4 72.9 Sil Indus 241 43 74.8
34 DhokPathan 33.1 72.3 Soan Indus 6475 43 1302.7
35 Massan 33 71.7 Indus Indus 286000 42 14320.91
Fig. 3 Comparison of most suitable flood frequency distributions.
Table 2 Chi-Square Test Results of Upper Indus Basin(UIB)

Sr. Station Chi Square Test Value Most Suitable
No. Method
Gumb Weibu Norm Lognorm Lognorm Pearso Log
el ll al al al Type 3 n Pearso
Type 3 n
Type 3
1 Naran 7.9 7.0 7.9 7.9 7.9 6.2 4.9 Log Pearson Type 3
2 Garhi 12.5 26.7 47.6 15.8 10.8 230.6 7.9 Log Pearson Type 3
Habibullah
3 Muzaffarabad 10.0 21.3 23.4 12.5 14.1 105.8 10.0 Log Pearson Type 3
4 Domel 11.3 12.5 15.1 7.9 7.1 73.2 7.1 Log Pearson Type 3
5 Kohala 3.3 24.6 27.5 4.5 4.1 153.5 6.2 Gumbel
6 Azad Pattan 11.6 35.5 30.5 13.3 10.4 166.1 12.1 Lognormal Type 3
7 Kotli 10.0 10.8 22.5 6.2 6.6 6.6 6.6 Lognormal
8 Palote 12.1 6.2 28.0 9.1 9.5 17.9 9.5 Weibull
9 Kharmong 4.0 5.0 3.0 3.0 2.0 3.0 2.5 Log Pearson Type 3
10 Yogo 6.0 8.8 3.7 4.5 6.4 5.2 8.0 Lognormal
11 Shigar 5.3 5.3 6.0 7.3 6.0 90.0 8.7 Gumbel
12 Kachura 4.9 4.9 7.9 5.8 3.7 7.4 4.9 Lognormal Type 3
13 Gilgit 14.6 30.5 22.1 12.5 10.4 219.7 6.6 Log Pearson Type 3
14 Dainyor Br. 6.6 13.7 12.9 7.9 6.6 6.6 7.4 Gumbel
15 Alam Br. 7.9 12.1 6.6 5.8 4.9 5.8 4.1 Log Pearson Type 3
16 Bunji 1.3 0.5 0.5 1.3 0.5 0.5 0.5 Log Pearson Type 3
17 Doyain 1.2 4.4 2.0 1.2 3.2 4.4 3.2 Gumbel
18 Shatial Br. 8.7 12.0 8.3 9.7 8.3 5.0 6.4 Pearson Type 3
19 Karora 5.3 6.2 16.9 8.9 8.9 7.6 14.7 Gumbel
20 Besham Qila 5.4 16.7 12.9 9.1 5.4 5.8 5.8 Gumbel
21 Daggar 6.2 6.2 7.9 3.3 3.7 4.5 6.6 Lognormal
22 Phulra 7.4 10.0 10.4 7.0 5.8 5.8 5.4 Log Pearson Type 3
23 Chakdara 6.2 15.4 20.8 7.0 7.9 230.6 7.9 Gumbel
24 Chitral 4.1 13.7 10.4 3.3 2.4 2.0 2.4 Pearson Type 3
25 Jhansi Post 26.7 11.2 63.1 6.6 5.4 8.3 7.0 Lognormal Type 3
26 Nowshera 2.8 12.5 8.3 3.3 2.0 153.5 3.3 Lognormal Type 3
27 Gurriala 9.5 20.8 52.2 10.4 10.8 4.1 11.6 Pearson Type 3
28 Khairabad 2.5 0.9 0.4 2.5 4.2 0.9 2.5 Pearson Type 3
29 Thal 7.0 7.4 10.4 7.0 4.9 1.2 3.3 Pearson Type 3
30 Chirah 4.9 10.4 23.4 13.3 10.8 8.7 13.3 Gumbel
31 Chahan 17.9 5.8 54.3 5.4 4.1 7.4 7.4 Lognormal Type 3
32 Dhok Pathan 10.0 6.6 16.7 16.7 8.7 15.0 5.4 Log Pearson Type 3
33 Massan 3.0 4.5 9.1 3.0 3.0 2.2 2.6 Pearson Type 3
Fig. 3 Comparison of design flood magnitude from different flood frequency method in different
rivers of Kabul River Basin.
rivers of Jhelum River Basin.
rivers of Indus River Basin.
Conclusions Kosi river breach: an evaluation”

• Identification of hydrological similarities CURRENT SCIENCE, VOL. 95, NO. 12,
was carried out for un-gauged and gauged 25 DECEMBER 2008.
sites for flood frequency analysis. There Haan, C.T. 1977 “Statistical Hydrology”
are the following conclusions. McGraw Hill Book Company, Singapore.
• Flood Frequency analysis showed that Log Solomon, O. and Prince, O 2013 “Flood
Pearson Type-III distribution providing Frequency Analysis of Osse River Using
suitable flood magnitudes on mostly areas Gumbel’s Distribution” Civil and
of Upper Indus Basin with minimum values Environmental Research ISSN 2224-5790,
of Chi-square test. Vol.3, No.10.
From the Previous discussion and Analysis of Pegram,G. and Parak,M. 2004 “A review of the
results it may be worthwhile recommending the regional maximum flood and rational
following. The study area was within formula using geomorphological
geographical region. Consequently, there are information and observed floods” ISSN
limits on the accuracy with which flow 0378-4738 = Water SA Vol. 30,No. 3.
characteristics were estimated from the Tumbare, M. J. (2000), “Mitigating Flood in
rationalization techniques. South Africa”, paper presented at the 1st
WARSFA/water Net Symposium: sustainable
References use of water Resources, 1-2 November,
Reddy, D. V. Kumar, D. Saha, Dipankar Maputo.
Mandal, M. K. 2008 “The 18 August 2008
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Impacts of hydro climatic variables trends on water resources of Yihe

River Basin during the past 50 years
Muhammad Saifullah1,4*, Ijaz Ahmad2, Muhammad Zaman3, Zhijia Li4, Abdul Nasir1
1*
Department of Structure and Environmental Engineering, University of Agriculture, Faisalabad,
Pakistan
2
Center of Excellence in Water Resources Engineering, University of Engineering and
Technology, Lahore 54890, Pakistan
3
Department of Irrigation and Drainage, University of Agriculture, Faisalabad, Pakistan
4
College of Hydrology and Water Resources, Hohai University, Nanjing 210098, China
*
[email protected]
Abstract: For sustainable water resources, the hydro climatic variables are important and influence on the
availability of water resources. In this study, Sen innovative method, Spearman’s rho test was used to analyze
the trends of hydro climatic variables of Yihe river basin during 1961-2011. Results indicated that precipitation,
streamflow and evapotranspiration showed decreasing trend in the Yihe river basin. The trend patterns of hydro
climatic variables matched by spearman’s rho test which showed good agreement in the most cases. The findings
of this study can be used as a reference for the planning and management of water resources of the river system
Keywords: climate change, Innovative trend pattern, Spearman’s rho test, Streamflow
Introduction numerous scientists investigated the impact of

A lack of water resources has become one of the climate change on water resources of different
critical factors forcing the development of the rivers of the world at various spatiotemporal
societies and economics of many countries of the scales (Zhang et al. 2014; Kormann et al. 2015).
world (Jia et al. 2015). Water resources are Similar studies have also investigated for
influenced by climate change. Recent hydrologic Chinese river (Amo-Boateng et al. 2014; He et
regime changes attributable to climate variability al 2013; Zhao et al. 2015). Zhang et al. (2011)
impacts which brought into focus the search for estimated water resources variations across
trend analysis. A lot of scientist investigated China and identified that mean annual
systematic trends of key climate variables streamflow reduced in arid and semi-arid region
attribute to climate change. It is well known that of North and increased in South of China.
climate change is expected to have significant Although previous studies have found the
impacts on the water resources attributable to decline trend in water resources of the Yihe
positive or negative trends in hydro climatic time River, it is still unclear whether the water
series (flood, drought, etc.). The variation in resources will continue to decline in future, and
rainfall runoff process may be alter the to what extent hydro climate variables are
availability and quality of water resources and responsible for the decline. Thus, the objective
change the spatiotemporal characteristics of of the study analyzed these variables through
watershed, such as timing of flood events, and spearman’s rho test and auto corelation. Also,
the frequency and severity of flood and droughts. Sen’s innovative method (Sen 2012) is used to
These variations cannot be considered to be analyze the trend pattern of hydro climatic
constant, which leads that future predictions are variables at different time scale. Hence this study
no more statistically indistinguishable from his will be helpful to identify the regime of water
historical records. resources of Yihe river basin.
Nowadays, there are different statistical methods
for analyzing hydro climatic variables such as t- Materials and Methods
test; regression analysis, Pearson correlation Study Area Description
coefficient, the spearman’s Rho, Sen’s slope, One of the large tributaries of Yellow River is
auto correlation and most commonly used Yihe River and major upper tributary sub-basin
Mann-Kendall method. Trend analyses have of the Yihe River is Dongwan basin as (Fig. 1),
been extensively used for detecting variability in which exists in the semi-arid climate zone. The
climate and hydrologic works. Consequently, the Dongwan watershed is situated in Henan
Province, China, between longitudes 111o to The time series data come from random process,
112o E and latitudes 33.5o to 34.5o N, with a the expected value and variance of r1 are:
drainage area of 2856 km2. It drains the area of
rising mountains Fuinu located in the Qinling
Mountain. There is a massive area of forest land
in the upstream region and soil erosion is not a
serious problem.
Innovative trend analysis technique

Sen (2012) divided the time series into
subsection and plotted the time series on where z is test static of test, n is number of
Cartesian coordinate system which is observations.
independent of the trend in time series, pre-
assigned significance level, magnitude of trend, Spearman’s rho Test.
sample size and the amount of variation within Spearman’s rho (Lehmen 1975; Sneyers 1990)
time series. This plot draws triangle and trend test is another rank-based nonparametric method
free line draws at 45o. The upper and lower part used for trend analysis and the innovative trend
exhibited as increasing and decreasing trend technique validated through this test. In this
respectively. If all the observations lie on 1:1 method, the time series data assumes
line, it means that there is trend free time series. independent and identically distributed, the null
The scatter points above or below the 1:1 line hypothesis identify no trend over time; the
indicate increasing or decreasing monotonic alternate hypothesis (H1) estimate that a trend
trends, respectively in innovative trend analysis exists and that data increase or decrease with .
technique. The trends of hydro climatic The test statistics sp and standardized statistics
variables can be analyzed through this simple
sp are defined as:
but still powerful innovative trend analysis
methodology. The Monte Carlo simulation was
applied to validate this method and then applied
for two annual stream flow data sets and one
precipitation time series from Turkey (Sen
2012).
Where n is number of observations, Di is rank of
observation, Rsp is rho coefficient and Zsp test
statics respectively.
Results and Discussion Innovative pattern of

hydro climatic variables
A simple qualitative method was introduced by
Sen in 2012 which estimated the trends of total
time series. This trend pattern was used for
annual time series of hydro climatic variables in
Fig. 1 Location of the study area
the present research in which the findings
confirm from other methods. As other researcher
Autocorrelation
also investigated the trend of hydro climatic
All the hydro climatic variables were first tested
variables using this innovative pattern (Sen
through autocorrelation. The lag one
2012; Nourani et al. 2015).
autocorrelation coefficient is calculated as:
Fig. 2 Result of Innovative trend pattern of evapotranspiration
Fig. 3 Result of Innovative trend pattern of precipitation
For this purpose, the time series of hydro climatic variables show decreasing trend and
climatic variables were divided into two parts high value of precipitation show increasing
with respect to time and arranged in ascending trend. The streamflow and evapotranspiration
order. Based on scatter diagram plot by time identified as decreasing trend as previous studies
series of hydro climatic variables from 1961- (He et al. 2013; Xing-Jie et al. 2012). The time
2011, the trend analysis via Sen’s method series possess serial correlation, but proposed
justified good agreement with obtained results of idea estimate the trend of hydro climatic
Spearman ‘s rho test in the most cases. The result variables as (Table.1). The annual time series of
of Sen’s scatter diagram justifies the same result hydro climatic variables plot as scatter diagram
as determine from the rho’s test. The scatter as (Fig. 2-4).
diagram divided into low, medium and high
cluster. The low and medium clusters of hydro
Table 1: Result of auto correlation and Spearman’s rho test

Precipitation Evapotranspiration Streamflow
Months Correlation rho static Correlation rho static Correlation rho static
Jan 0.07 -0.65 0.9 -2.39** 0.70 -1.68***

Feb 2.03** 1.36 -0.10 -2.01** 1.16 -2.08**
Mar 2.62* -1.33 3.09* -2.48** 2.28* -1.60
April 1.54 -3.70* 0.76 -1.39 0.70 -3.89
May -0.55 0.04 1.38 -2.50** 2.19** -3.83*
June -0.02 -0.18 2.69* -4.45* -0.91 -1.70***
July 1.36 2.00** 4.56* -5.24* -0.64 -2.73*
Aug -1.51 -1.07 3.42* -4.52* -0.69 -1.00
Sept 2.14** -0.64 3.44* -3.44* 1.59 -2.40**
Oct -0.90 -2.23** 3.30* -3.76* 1.64 -3.26*
Nov -0.20 -1.97** -0.41 -1.27*** 0.98 -3.10*
Dec 0.52 0.27 -0.05 -2.22** 2.54** -2.18**
Annual 0.98 -0.67 4.96* -5.26* 1.00 -0.71
Fig. 4 Result of Innovative trend pattern of runoff
Note: ***, Significant trend at 90% significance precipitation and streamflow are highly
level, ** Significant trend at 95 % and * dependent and sensitive to each other. Finally,
significant trend at 99% respectively. the result showed the changes in the water
The trend of hydro climatic variables found to be resources of Yihe river basin. Overall, we
similar as innovative trend pattern. The potential believe that this study provides an elaborate view
evapotranspiration showed decreasing pattern of hydro climatic variables trends in Yihe River
during low, medium and high cluster values. The basin, using nonparametric tests in conjunction
high cluster of precipitation show slighter with accumulated deviation curve and Parde
increasing trend which show good match with coefficient which should be useful for further
spearman’s rho test as in Table 1. The investigation. This study provides an idea
streamflow shows decreasing trend as innovative regarding the effect of natural forces on water
trend patter. The result of all month of hydro resources. More importantly, this study opens
climatic variables and innovative pattern found new horizon for additional research that can
to be similar. address human activities on the water resources
of the region.
Conclusion
In this study, Sen’s innovative method was used References
to analyze the hydro climatic variables in which Amo-Boateng M., Z. Li. and Yiging G. 2014
findings were in accordance with Spearman’s Inter-annual variation of streamflow,
rho test results and confirmed the existence of precipitation and evaporation in a small
decreasing trend in precipitation, humid watershed (Chengcun Basin, China).
evapotranspiration and streamflow. The PET Chinese Journal of Oceanology and
show decreasing trend using innovative trend Limnology 32(2): 455-468.
pattern and spearman’s rho test. The
He B., C. Miao. and Shi W. 2013 Trend, abrupt Sen Z. 2012 Innovative trend analysis
change, and periodicity of streamflow in the methodology. Hydrologic Engineering,
mainstream of Yellow River." Environ Technical note, 17(9), 1042-1046.
Monit Assess 185(7): 6187-6199. Sneyers, R. 1990 On the statistical analysis of
Jia X., C. Li. Cai Y. Wang X. and Sun L. 2015 series of observations. World
An improved method for integrated water Meteorological Organization, Technical
security assessment in the Yellow River Note 143, WMO no. 415.
basin, China. Stochastic Environmental Xing-Jie J., Ji-Jun W., Wan-Long G., Ye-Yu Z.
Research and and Feng-Xiu L. 2012 Trends in Annual and
Risk Assessment (Online Published). doi Seasonal Pan Evaporation in the Lower
10.1007/s00477-014-1012-2 Yellow River Basin from 1961 to 2010.
Kormann C., Francke T. and Bronstert A. 2015 Advances in Climate Change Research,
Detection of regional climate change effects 3(4), 195-204. doi:
on alpine hydrology by daily resolution 10.3724/sp.j.1248.2012.00195
trend analysis in Tyrol, Austria. Journal of Zhang Q., Peng J., Singh V. P., Li J. and Chen,
Water and Y. D. 2014 Spatio-temporal variations of
Climate Change, 6(1), 124-143. doi: precipitation in arid and semiarid regions of
10.2166/wcc China: The Yellow River basin as a case
Lehmann E. L. 1975 Nonparametrics, Statistical study.
Methods Based on Ranks. Holden-Day, San Global and Planetary Change, 114, 38-49. doi:
Francisco, Calif, USA. 10.1016/j.gloplacha.2014.01.005
Nourani V., Nezamdoost N., Samadi M. and Zhang Z.X., Chen X., Xu C.Y., Yuan L.F., Yong
Vousoughi F.D. 2015 Wavelet-based trend B. and Yan S.F. 2011 Evaluating the
analysis of hydrological processes at nonstationary relationship between
different timescales. Water and Climate precipitation and streamflow in nine major
Change (in press) basins of China during the past 50 years. J
Hydrology, 409:81–93.
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Sensitivity Analysis of Hydraulic Parameters for Water Hammer in

Penstock Design
Muhammad Yasar1*, Muhammad Waseem1, Ghulam Nabi1
1
*
[email protected]
Abstract: Water hammer is a transient flow which is produced as a result of sudden velocity changes in penstock
of hydropower plant. This phenomenon may lead to initiate strong pressure wave in penstock. if the pressure wave
exceeds the design pressure of penstock, the result would be collapse of pipe. Sensitivity analysis is an analysis
that finds out how sensitive an output is to any change in specific input while keeping other inputs constant. In
hydropower system, it refers to an analysis of how each of the input variables (such as diameter, length, wave
velocity, etc.) affects the fluctuation of pressure wave in penstock. In this study, governing equations of water
hammer i.e. continuity and momentum equations are numerically solved and simulated by developing a computer
program in “Visual Basic” environment. The fluctuations in positive and negative pressures were calculated by
varying hydraulic design parameters one by one considering the rest of parameters as constant. The program
employs method of characteristic (MOC) for numerical solution. Depending upon the sensitivity of the water
hammer phenomenon, it is obligatory to prepare adequate and cost-effective design of penstock to avoid any
cavitation. In addition, it shows, by reducing the penstock pipe diameter and wave velocity and increasing pipe
length, the maximum pressure head increased, and the minimum pressure head decreased. The software
developed in this study was also applied to hydropower project in northern area of Pakistan i.e. Kyal Khwar
hydropower plant. The developed program showed good agreement with the original data prepared by the
commercial software used by the consultant of Kyal Khwar hydropower plant.
Keywords: Water hammer, Transient flow, Valve, Surge Wave, Penstock pipe, VISUAL BASIC
Introduction found to reduce the water hammer effect and to

Water hammer is a transient flow phenomenon prevent water column separations and fatigue
which initiates due to suddenly change in the failures, surge tank recommended as good
flow pattern the velocity and pressure varies engineering practices. Furthermore, water
abruptly. This phenomenon may lead to hammer effects were also reduced by increase the
propagate severe wave in the pressurized pipe; valve closure time. Araki and Kuwabara (1975)
the intensity of such wave is so high which proposed a differential equation for elastic theory
oscillate very high to very low value, it poses to of water hammer. Moreover, they introduced a
penstock to danger at higher pressure and may relation for considering head loss simplification
collapse the penstock and other conveyance by using finite difference method, with the help
facilitates if the system is not equipped with of a special computer program. A special
proper protection devices. Hydraulic transient computer program known as (WHAMO),
phenomenon occurs due to flow changes rapidly developed by US Amy Corps of Engineers, was
in pressurized conduits for several reasons such being used to solve the equation, by inserting
as, sudden valve closure, pump stop, and load appropriate boundary conditions.Wylie and
failure and changing elevation in reservoir. Streeter (1979) presented similar equation for
Wylie and Streeter (1978); White (1979). The analyzing water hammer phenomenon in a
peak magnitude of the transient pressure depends simple pipeline, with an open reservoir on the
on pipe length, profile, wave velocity, friction upstream side and a valve on the downstream
coefficient and diameter. Wylie and Streeter side. They have used the method of
(1978). Due to the strong pressure wave, the characteristics to develop a basic computer
pressure head changes rapidly which may cause program scripted in Fortran Computer Language.
severe damages to the pipe (Parmakian, 1963). Nemet (1974) presented a theory on water
Abuaziah et al. (2013) showed that the hydraulic hammer and developed a computer program to
transient analysis is an effective and reliable tool analyse the pressure increases due to water
to determine the needs for surge protection hammer effect. In addition, he also provided the
systems against transient phenomenon. They format for the computer system data input for
developed program. Mansuri et al. (2014a, b) partially closed valve were inconclusive, because
studied, fundamental equation of water hammer of limitations related to the design of the
solved numerically by method of characteristics experiment. Their research demonstrates the
by using MATLAS software, the study showed assumption that a water hammer wave interacts
the sensitivity analysis in the pressure variations with a boundary and no losses is valid for
by changing some parameters of pipe line i.e. situations with no flow in the pipeline system. In
diameter, length and wave velocity, the study addition, it provides a foundation on which
also shows by increasing these parameters, the further research can be undertaken. Various
pressure changes decrease & vice versa. They methods of analysis were developed for the
also found that by increasing pipe roughness, problem of transient flow in pipes such as: (a)
negative and positive pressure range decrease, approximate analytical approaches where are
while increase in reservoir water level cause nonlinear friction part in the momentum equation
severe change in the pressure wave in the pipe. is either neglected or linearized (b) numerical
Bernard (2013) showed that the thickness of solutions of the nonlinear system or method of
penstock, Length, initial velocity and bend angle Characteristics (Wylie and Streeter, 1978).
has a significant impact on the fluctuation of Among these two methods, method of
pressure wave and structural response. He also characteristics (MOC) is the most popular
observed, in the presence of soil have a approach for handling hydraulic transient flow.
substantial benefit for decreasing the von-mises Its importance lies in its ability to convert the two
stresses. Norazline and Amin (2015) studied and partial differential equations (PDEs) of
showed that profile of pressure wave and continuity and momentum into four ordinary
amplitude depends on the valve operation, time differential equations that are solved numerically
of closure of valve and type of closing laws and using finite difference techniques (Gray 1953,
showed that the pressure wave profile varies Streeter and Lai 1962).
frame square to triangular and trapezoidal shape In this study, the numerical modeling method for
depending on the closing laws and the frequency method of characteristics (MOC) has been
of the pressure decreases as the closing time is presented for sensitivity analysis of water
decrease and number of segment increases. Jalil hammer problems by varying pipe parameters
and Solemani (2011) showed that the most such as: diameter length and velocity. For this
sensitive parameters for which maximum purpose; a program in Visual Basic environment
hammer developed are the (a) moment of inertia has been developed. To test the software,
of pump and electrometer, (b) diameter of the fluctuations of surge wave by changing of pipe
pipe, (c) material of the pipe (d) thickness of pipe diameter, length and velocity, are investigated.
(e) water temperature and by increasing these Engineers and researches working in this area can
parameters, the maximum water hammer benefit from the software detailed. The study
decrease and vice noses. Wuyi and Hnang (2013) sensitivity analysis of numerical model by
showed that the compensation strainer should do changing parameters helps to understand well
effect on pressure wave either it is positive about hammer phenomenon and design aspects
pressure wave or negative pressure wave. The for improved design of penstock.
study also proves if suitable strainer was selected
for pressure vessel, due this water hammer effect Material and Method
and severe pressure fluctuation should be
decreases up to some extent. Algirdas et al.
(2009) studied that the flow energy loss
coefficient in pipe just d/s of the valve is the most
important parameter and play vital role on the
variance estimate pressure peak. It also showed
that the pipe wall roughness and pressure at the The study consists of sensitivity analysis of
pumping head were also play a role on pressure hydraulic parameters for water hammer problems
wave fluctuation. Cooper and Johnston (2010) in penstock design by changing penstock pipe
investigated the system boundaries on transient diameter, length and wave velocity. The full
wave, when the system contains a fully closed elastic water hammer in one-dimensional
valve, a partially closed valve and a reservoir. unsteady pressure flow equations given by
They found negligible effect on the nature and (Wylie and Streeter, 1978)
magnitude of the pressure wave after interaction Where, H is the head of reservoir (m), Q is the
with these boundaries i.e. fully closed valve and discharge (m3/sec), D is internal diameter of pipe
reservoir with no flow. The experimental results (m) and “a” is the celerity of wave (m/sec). The
obtained from the reservoir with flow and
fundamental equation of water hammer is Applying some mathematical operations in Eqs.
depending upon the momentum and continuity (3 and 4), the two-normal differential equations
equation of flow. Eq (3) shows the dynamic Are deduced as shown in Eqs. (5 and 6)
equation for non-compressible fluids.
H a Q af Q
+ + V V − sin  = 0 (5)
V V 1 P  V t gA t 2 gD A
+V + . + + g. sin  = 0 (3)
t x p x 2 D
H a Q af Q
The continuity equation for a segment of pipe − + VV − sin  = 0 (6)
t gA t 2 gD A
length as shown in Eq. (4) (Mansuri et al. 2014)
In order to solve the Eqs. (5 and 6) by
V 2 1 P 1 P
a + + V =0 (4) characteristic line method, a grid of
x p t p x characteristic, now established, the two
characteristics C+ and C- initiate from node A &
To the analysis of water hammer phenomenon B respectively and intersect at nodel point “P” (as
Eqs. (3 and 4) are solved simultaneously. shown in fig. 2), at nodes A & B, here value of
(Q , H), where the conditions are known, and at
Method for Numerical Solution node “P”. The conditions are unknown where the
Method of Characteristics (MOC): value of (Q , H) are calculated. Using liner finite
There are many methods and technologies are difference scheme between node A & P and
available for the solution of the governing between node B & P, deduced the two linear
equations of water hammer (momentum & equations as shown in Eqs.(7 and 8) (Strecter,
continuity) among them method of characteristic 1967).
(MOC) is one of the best method to be considered
𝑎 𝑓 ∆𝑥
with some minor losses and modify for different 𝐻𝑝 − 𝐻𝐴 + 𝑔𝐴 (𝑄𝑝 − 𝑄𝐴 ) + 2𝑔𝐷𝐴2 𝑄𝐴 |𝑄𝐴 | = 0
boundary conditions. This method has ability to
(7)
convert the two- partial differential equation
(momentum & continuity) into normal 𝑎 𝑓 ∆𝑥
differential equation, which were solved 𝐻𝑝 − 𝐻𝐵 − 𝑔𝐴 (𝑄𝑝 − 𝑄𝐵 ) − 2𝑔𝐷𝐴2 𝑄𝐵 |𝑄𝐵 | = 0
numerically using finite difference techniques (8)
(Streeter 1967).
𝐶𝑝 = 𝐻𝑖−1 + 𝐵𝑄𝑖−1 − 𝑅𝑄𝑖−1 |𝑄𝑖−1 | = 0 (9)
𝐶𝑀 = 𝐻𝑖+1 + 𝐵𝑄𝑖+1 + 𝑅𝑄𝑖+1 |𝑄𝑖+1 | = 0 (10)
𝐻𝑝𝑖 = (𝐶𝑃 + 𝐶𝑀 )/2
𝑄𝑝𝑖 = (𝐻𝑃𝑖 − 𝐶𝑀 )/𝐵
𝑎 𝑓 ∆𝑥
𝑊ℎ𝑒𝑟𝑒 𝐵 = 𝑎𝑛𝑑 𝑅 =
𝑔𝐴 2𝑔𝐷𝐴2
Fig. 1 (x, t) grid for solving simple pipe problem
by Characteristic lines Wylie and Streeter In Eqs. (9 and 10) CM and Cp are known
(1967). coefficients based on value of H and Q in known
time step. By solving these two linear equations,
two unknown values 𝐻𝑝𝑖 and 𝑄𝑝𝑖 in the next time
step has been found.
Boundary Conditions
The H and Q values at the ends of the pipe were
determined by using boundary conditions. These
conditions are, Chaudhry (1979).
U/S Boundary Condition (Constant Head

Fig. 3 Schematic diagram of pipe connecting Reservoir)
reservoir U/S and D/S a valve. At the U/S end there is constant head reservoir,
the H-value remains constant for all time. Thus,
for ὶ = 1 geometrical axis of the pipe. The program in each
time steps calculates the value of CDA using
linear interpolation. Simultaneously, the value of
Hpi and Qpi in valve was calculated.
D /S Boundary Condition (Valve at end)
At the D/S end there is valve, the  – t Results and discussion
relationship for valve closure is given by General
In the current study, Table 1 shows the
t Ef parameters, to test the model and perform
 = (1 − ) sensitivity analysis to understand the behavior of
tc
water hammer phenomenon.
Where, t is current time of analysis (sec), tc is
time of valve closure (sec) and Ef is modulus of Table 1 General hydraulic data of pipe for
elasticity of fluid flow (Gpa). For slow closure of testing the model and analysis of water Hammer
valve, closure time (tc) is normally greater than (Wylie and Streeter, 1978)
reflection time (tr) and for rapid closure of valve, Description of data Data
Length of pipe (L) 600 m
closure time (tc) is less than reflection time (tr).
Head of reservoir (Hp) 150 m
Diameter of pipe (D) 0.5 m
Celerity of wave (a) 1200 m/sec
Friction coefficient (f) 0.018
(CD A)0 0.009
Fluctuation of Pressure Wave Due to Valve

Closure
The hydraulic transient analysis performed for
the pipe parameters listed in table 1. In this
analysis, slow valve closure condition has been
assumed. Moreover, analysis has been performed
for method of characteristics (MOC) (as shown
Fig. 4  – t, relationship due to Slow valve in Fig. 5) and maximum (Hmax) and minimum
closure (Hmin) pressure head values are computed. For
example, Hmax and Hmin values are 268.53m
The governing equations of water hammer and 116.66m, respectively. Moreover, the
(momentum and continuity of flow), are solved maximum and minimum pressure head values are
numerically by method of characteristics (MOC), occurring at normalized length (X/L) of 1.
a computer program in visual basic (VB)
environment has prepared. The prepared module Sensitivity Analysis of Hydraulic Parameters
solves the water hammer equations numerically In Penstock
by considering constant head reservoir at u/s end Penstock Diameter (D) as Variable
and a valve at d/s end of the penstock. Through Fluctuations of pressure head for the change of
value operation and determine the value of CDA penstock diameter calculated using computer
for each time step as shown in Eq. (13) software. Both MOC and MOI have been used to
study hydraulic transient analysis due to valve
Qp = C D A 2 gHp (13) closure (Tc>Tr).
CD is discharge coefficient, Hp is Head of

reservoir (m) and A is the area of orifice (m2).
The behaviour of water hammer phenomenon on
a system including a pipe within variable
diameter, length and wave velocity within
constant head of reservoir is assessed. In order to
analyse the transient phenomenon a penstock
pipe length (L) is divided into N = 5, segments
Fig. 5 Pressure fluctuation for slow valve
(pipe's full length, pipe's 4/5 length, pipe's 3/ 5
closure by method of
length, pipe's 2/ 5 length and pipe's 1/ 5 length)
characteristics(MOC)
for the calculation of pressure wave. The datum
level for hydraulic are considered for the
Standard pipe diameter of penstock pipe has been reduced (0.5m to 0.3m), the number of water
selected as 0.5 meters. The diameter of the thrusts on valve increases for the same time
penstock pipe has been increased and decreased period and magnitude of theses thrusts also
to study its effect on pressure head, while keeping increases (268.53m to 427.97m) as shown in
other parameters (listed in table 1) as constant. Fig (7). Therefore, valve life decreases due to
Fig. (7) Shows maximum and minimum pressure water hammer effect. Thus, number of cycles
heads vary rapidly for diameters below standard should be considered in sensitivity analyses of
diameter, while the pressure variation for water hammer effect in pressure conduits. For
diameters above the standard diameter start diminishing water hammer effect, theoretically
converging and the change in values of pressure pipes of exorbitant diameters should be used.
head is very small. For minimum pressure head However, considering economic aspects this is
values (-64.99m & -365.32m), which were below not a viable solution therefore optimum pipe
the vapour pressure of water (<10.3 m), diameter should be determined after performing
cavitation may occur. several iterations. In the prevailing scenarios, a
good criterion to establish a systematic design
FLUCTUATION IN PRESSURE HEAD(Tc>Tr) and/or develop maintenance-task should include
the readiness to employ optimum internal
penstock-diameter (and accessories) and lesser
rigid pipe materials.
Different scenario analysis presented in this
PRESSURE HEAD (m)
section conclude that the sensitive amplitude of

pressure wave occurs at the end of pipe (X/L =
1), hence end of pipe must be considered as
critical zone and should be designed accordingly.
0 0.2 0.4 0.6 0.8H(max) 1
DIAMETER (m) H(min)
Hydraulic Transient Analysis for Penstock
Length (L) as variable
Fig. 6 pressure variation with changing diameter In the next phase fluctuations of pressure head for
the change of penstock length calculated using
Those results reached above indicate that the prepared computer module. the penstock pipe
water hammer magnitude is sensitive with length increases from 600 m to 1200 m, the
respect to the penstock diameter. The reason for corresponding increase in the values of
this sensitiveness was the variation in the maximum pressure head is 8.61% (291.66m) to
transient-pressure amplitude due to change in 52.77% (410.23m); and the corresponding
standard diameter (0.5m). Study shows that the decrease in the values of minimum pressure head
decreasing exponential trends of water hammer is 28.89% (82.96m) to 176.97% (-89.79m) as
pressure with respect to diameter as shown in Fig shown in Fig (11). On the other hand, with the
(6). Hence, it was evident that when pressure- decrease of penstock pipe length below the
amplitude increases (268.53m to 688.93m), it standard length (600m), the maximum pressure
poses greater vulnerability to the fluid- structure. head value decreases while the minimum
It has also been concluded that the frequency of pressure head value increases. Present study
pressure wave is dependent upon the shifts in indicates, as the penstock pipe length decreases
pipe-diameter. from 600m to 100m, the corresponding decrease
in the values of maximum pressure head is 8.54%
(245.61m) to 38.26% (165.79m); and the
corresponding increase in the values of minimum
pressure head is 15.29% (134.5m) to
17.89%(137.53m) as shown in Fig (8)
Fig. 7 pressure variation w.r.t. time by changing

diameter
Study depicted, when the penstock-diameter was
Fig. 8 pressure variation with changing pipe
length
Present study showed that the Pipe length has Fig. 9 pressure variation with varying wave
control the reflection time (Tr=2L/a= 1sec) and velocity
the inertia of water inside the pipe. The longer the
pipe is (1200m), the longer the reflection time (2 The study showed, wave velocity also controls
sec), in addition, the longer the pipe, the larger the reflection time (Tr = 2L/a = 1sec), higher
the mass of water that has been affected on the wave velocity lesser the reflection time and vice
moment of inertia of the water column, so that the versa. To avoid cavitations and pipe rupture, used
pipe length is greater than 600 m in length higher wave velocity which was less than
(1200m), the risk of sub pressures exists (- velocity of sound (1420 m/sec) i.e. 1000> a <=
89.79m < 10.30m) which may lead to produced 1200m/sec, the reason is that in higher wave
cavitations as shown in Fig (8). velocity lesser the reflection time so due this
wave's sweep occurs more and rapid and this
Hydraulic transient analysis by Wave could encounter waves in opposite direction. It
velocity in penstock (a) as variable could be seen that increasing wave velocity
In this study celerity of wave (a) is varied, while would results in more increasing transiency.
keeping other parameters (listed in table 1) as
constant for the analysis of water hammer Model Application
phenomenon in penstock, first assumed pipe Keyal Khwar Hydropower Project:
wave velocity (a = 1200 m/sec) as standard one The power conduit system of the Keyal Khwar
and thereby incremental in celerity of wave by HPP consists of between power intake and
same proportions to the standard one, which Underground powerhouse of a headrace tunnel of
results increased or decreased in the pressure about 7.2 km length, a surge tunnel and a pressure
head. by increasing the celerity of wave by 8% shaft/tunnel of about 900 m length as the major
(1300 m/sec), to 33% (1500 m/sec); resulted in components. The powerhouse accommodates
the maximum value of pressure head is decreased two 5-nozzle Pelton turbines with vertical axis
up to 4.01% (257.75m) and also the minimum with a spherical valve as turbine inlet valve. The
value has increased up to 4.81% (122.27m). In elevation of the needles of the Pelton turbine is at
second case when pressure wave is reduced by 685.0 m. At the rated turbine discharge of10 m3/s
8% (1100 m/sec) to 58% (600 m/sec) resulted the the rated power is 63 MW. With the objective to
minimum value was decreased up to 75.89% assure a high degree of flexibility of plant
(28.13m) and maximum value was increased operation, and to protect the headrace tunnel
6.19% (285.16m) as shown in Fig 9. from excessive internal pressure fluctuations, a
surge shaft has been provided by means of an
inclined surge tunnel. Transient analyses were
performed in order to assess the maximum and
minimum heads in the pressure shaft/tunnel. The
schematic system as applied to Keyal Khwar
HPP is shown in Fig. 14.(WAPDA 2011)
Table 2 Input Hydraulic data for Keyal Khwar
Hydropoer Project
Description of data Data
Water level of reservoir 1422.5 m
Inlet level of valve 685.0 m
Diameter of Pressure Shaft(D) 2.2 m
Length of Pressure shaft (L) 900 m
Length of Surge Shaft 900 m
Discharge (Q0) 20 m3/sec
Friction coefficient conc. lined 0.002
in situ steel
Closure time of valve 30 sec
Maximum Pressure Head 1460.50 m
Fig. 10 Schematic Layout of Keyal Khwar HPP The corresponding maximum heads for the
closure of the needles of the Pelton turbines in 30
At first a transient analysis was performed for s at FSL = 1422.5 m are: (WAPDA 2011)
regular load rejection subsequent to full load Maximum head at Power House =1479.00 m
operation, i.e. with closure of the needles of the Maximum head at beginning of
Pelton turbines. The results of the corresponding pressure shaft = 1460.50 m
transient analysis are shown in the Fig. 15.
Fig. 11 pressure fluctuation due to slow valve

closure by consultant computer program
Fig. 12 Fluctuation of Pressure Head of Keyal Khwar HPP due to slow valve closure
Transient analysis in penstock with (1521.61m) and minimum value is decreased

NMSAWHP.exe up to 4% (1362.50m) respectively.
Penstock Diameter (D) as a Sensitive
Variable
In this scenario increased the pipe diameter
from 2.2m to 3.0m, the maximum value of
pressure head (Hmax) was decreased from
0.19% (1458.97m) to 0.92% (1448.37m) and
the minimum value (Hmin) was increased up to
0.08% (1421.46m), and decreased the pipe
diameter by same proportion the maximum
value of pressure head is increased up to 4%
further study.
I. No valve should be closed in less than 10
seconds, 30 seconds or more is preferable
i.e. tc (closure time) greater than or equal to
tr (reflection time = 2L/a), to avoid any
damage
II. Use optimum diameter (i.e. 2.2m & 1.5 m)
of penstock pipe in Keyal Khawar and
Satpara Hydropower system respectively,
to avoid any misshape due critical
amplitude of transient wave.
III. Study the pipe line profile along with
pressure head line, if the distance between
Fig. 13 pressure fluctuation by chaining them is small, and then provide protective
diameter due to slow valve closure by measure to minimize the effect of water
MOC hammer phenomenon.
IV. The critical amplitude of transient wave
Fig (13) also shows at 2.2m diameter of pipe, occurred at along the normalized length
the pressure head, Hmax, and Hmin are (X/L= 1) of penstock pipe, so end of pipe
1461.79 m and 1420.31m respectively which considered being the most critical zone. It
were nearly equal to the value computed with is required to provide protective remedial
CHARNETS software(1460.5m), which measure at normalized length (X/L= 1) of
validate the developed program. Moreover, the penstock pipe.
maximum and minimum pressure head values References
are occurring at normalized length (X/L) of 1. 1. Abuaziah, A. Oulhaj, K. Sebari , A. A.
This must be equipped with proper protection Saber , D. Ouazar And N. Shakameh
devices and design accordingly. (2013). Modeling and controlling flow
transient in pipeline systems. Mar. Sci.
Conclusions Agron. 3(1). PP. 12-18
Those results reached above indicate that the 2. Akan A. O. (2006). Open channel
water hammer magnitude is sensitive with Hydraulics, Elsevier ltd, Jordan Hill,
respect to the penstock diameter. Study shows Oxford. PP. 315-330
the decreasing exponential trends of water 3. Algirdas K., V. Kopustinskas and M.
hammer pressure with respect to penstock vaisnoras (2009). Water hammer model
diameter. Hence, it is evident that when sensitivity study by the FAST method.
pressure-amplitude increases, it poses greater ENERGETIKA, 55(1): PP. 13-19
vulnerability to the fluid-structure. Study also 4. Allievi L. (1925). Theory of Water
depicted, when the penstock-diameter is Hammer, E. Halmes, trans., Proceedings,
reduced, the number of water thrusts on valve American Society of mechanical Engineers.
increases for the same time period and 5. Araki. M. and T. Kuwabara (1975),
magnitude of theses thrusts also increases. Analysis of Pump-Turbine System
Therefore, valve life decreases due to water Including Pipelines, Hitachi Review, 24(5).
hammer effect. Thus, number of cycles should 6. Bernard. D (2013). Numerical modeling of
be considered in sensitivity analyses of water water hammer effects in penstocks. M.Sc.
hammer effect in pressure conduits. Study also Thesis, Department of Civil Engineering,
depicted, that the Pipe length has control the University of Ottawa, Canada. PP. 1-226.
reflection time and the inertia of water inside 7. Chaudhry M.H. (1968). Boundary
the penstock pipe. Generally, the pipe length is Conditions for Analysis of Water hammer
greater than 600 m in length, the risk of sub in Pipe System. M.Sc. Thesis, Civil
pressures exists which may lead to produced Engineering Department, the University of
cavitation. Results indicated by high velocities British Columbia. PP. 1-149.
i.e. above 1000 m/se, the pressure fluctuation 8. Chaudhry M.H. (1979). Applied Hydraulic
range has decreased by multiple of 8% that may Transient. Van Nostrand Reinhold
avoid cavitation and high-pressure problems. Company, Litton Educational, New York.
PP. 45-64
Recommendations 9. Chao W. and J. D. Yang (2015). Water
The followings recommendations are made for hammer simulation using explicit-implicit
coupling methods. Journal of Hydraulic
Engineering, ASCE, 4(141): PP. 15. Mosonyi .E and H.B.S. Seth (1975). The
04014086-96 surge tank a device for controlling water
10. Cooper and H.R. Johnston (2010). The hammers. Water power and Dam
effect of boundaries on water hammer construction. 27(2,3)
waves. Final Year project, Civil and 16. Norazlina S. and N. Amin (2015). Analysis
Natural Resource Engineering (NRE), of water hammer with different closing
University of Canterbury. PP. 1-10 valve laws on transient flow of hydrogen-
11. Fiszera ul. J. (2010). Mathematical model natural gas mixture. Department of
and numerical computations of transient mathematical science, University
pipe flows with fluid-structure interaction. Technology Malaysia (UMT). PP. 1-12.
Hydraulic machinery Department, the 17. Nemet A. (1974). Mathematical Models of
Szewalski institute of fluid flow, Gdansk, Hydraulic Plants. Escher Wyss News. No. 1
Poland. (14): PP. 77-94 18. Parmakian J. (1955, 1963). Water hammers
12. Jalil E. and A. Solemani (2011). Maximum Analysis. Englewood Cliffs, N.J.: Prentice-
water hammer sensitivity analysis. Hall inc.
International scholarly and scientific 19. WAPDA 2011. Detailed design report of
research and innovation, 5(1): PP. 1-04 Keyal Khawar hydropower project,
13. Mansuri B., F. Salmasi and B. Oghati Lahore.
(2014). Sensitivity analysis of water 20. White F.M. (1979). Fluid Mechanics,
hammer problem in pipelines. International McGraw-Hill inc., New York: various
Journal of energy and environment, pages
University of Tebraiz, Tabriaz. 5(2): PP. 21. Wylie E. B. and V.L. Streeter (1967, 1978).
124-131 Fluid Transient, McGraw-Hill inc., USA:.
14. Mansuri B., F. Salmasi and B. Oghati 2(1), PP. 31-63
(2014). Effects of pipe’s roughness and 22. Wuyi W., W. Huang and C. Li (2013).
reservoir head levels on pressure waves in Sensitivity analysis for the resistance on the
water hammer. Journal of Civil performance of a pressure vessel for water
Engineering and urbanization, Tabraiz hammer protection, 136(1)
university, Tabraiz. 4(1): PP. 36-40.
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Physical modeling to control scouring under natural flow conditions

Farooq. R.1, Ghumman A.R.2,3, AlSaleem, S.S.3, Seemab, F.1., Tariq. M.A.U.R.1*
1
Capital University of Science and Technology, Islamabad
2
University of Engineering and Technology Taxila
3
Qassim University Saudia Arabia
*
[email protected]
Abstract: The most common reason for the failure of hydraulic structures is the scour phenomena at their
supports. If the scouring is reduced/controlled 90% of failures can be avoided. Numerous mathematical models
have been developed for sediment transport. However, scouring around the pier in case of cohesive soil has not
been addressed satisfactorily. This lead to perform physical modeling under various recommended flow alteration
techniques found in the literature. Many remedial techniques have been applied to control local scour. Mainly,
there are two kinds of techniques to reduce erosion process near pier base i.e., bed armoring technique and flow
alteration technique. Applying modifications in a pier is an effective technique to control scour. A number of
methods have been employed to emasculate scouring process in order to assist the bridge designs. Present study
aims to reduce the stagnation of the flow and vortex formation in front of the bridge pier by providing a collar, a
cable and openings in combinations around a pier. The scour depth around octagonal pier was tested for various
combinations of above modifications. Combination of a collar with cable around the octagonal pier reduced more
scour depth than that for other two configurations. Similarly, the scour depth around the octagonal pier with
collar and openings was less than that of octagonal pier with cable and openings. Providing a collar, a cable and
openings together on the octagonal pier, found to be the best combination in reducing scour depth. It was found
49.4% less than that for the octagonal pier without any modifications. A plain octagonal and circular pier was
also tested for maximum erosion depth which was found 14% less for the former pier.
Keywords: Erosion, bridge pier, collar, cable, openings
Introduction
Bridge plays vital role in transportation. Incidence of bridge failures due to scour at their supports is a
common occurrence. The erosive process around pier starts due to the downward flow and resulting
vortices system. The pier blocks the approaching water as a result velocity at the front face of pier
reaches to zero causing an increase in pressure. Thus water level on upstream side of pier increases. As
the flow velocity decreases from surface to the bed, the dynamic pressure on the pier face also reduces
downwards. The downward flow creates a hole in front of the base of pier, rolls up and by interaction
with the coming flow forms a complex vortex system.
Local scour around a bridge pier is the primary result of the downwards flow effect especially towards
upstream face of pier. Actually, the downward flow initiates the formation of horseshoe vortex towards
pier base that would ultimately increase the scouring process (Kumar et al. 1999), as shown in Fig. 1.
Fig. 1 Demonstration of flow and scour process around pier
Literature review around the pier base in order to block the removal
Local scour at circular bridge piers has been of sediments at the pier. Collars of different
investigated by many researchers. In the river shapes and varying thickness are practiced
hydraulics, the local sediment scour at the bridge around the bridge piers as protective plates
piers has resulted to be one of main reasons for against the sediment removal (Melville and
the structure failure (Melville and Coleman, Coleman, 2000) and (Fotherby and Jones, 1993).
2000). Whitehouse (1998), and Sumer and The thickness of collar could be increased up to
Fredsoe (2002) described scour around the bridge certain limit otherwise it would cause high
pier or pile in a steady current. They studied turbulence in the flow resulting in excessive
comprehensively the process of scouring at scour (Whitehouse, 1998). Zarrati et al. (2004)
bridge piers from various aspects, including local studied the effect of width of a collar on the local
scour mechanism & flow characteristics and scour around pier and found that the optimum
anticipation of maximum probable scour depth. width of collar is thrice to that of pier in
Many studies are available in the literature controlling erosion. Thus, in present study the
encompassing various features of bridge pier width of collar equal to three times to that of pier,
scour in non-cohesive sediment beds including was kept for all the experiments. The provision
shape effects (Breusers and Raudkivi, 1991; of a collar initiates the scour on the downstream
Richardson et. al., 1993; Melville, 1997; and of the pier and forms a wake vortex there. This
many others). On the contrary, studies on local wake vortex progresses toward upstream to
scour around piers on cohesive sediment beds are create horseshoe vortex. This horseshoe vortex
rather inadequate (Ansari et al., 2002; Debnath enhances scouring thereby dissipating the effect
and Chaudhuri, 2010). Also, the above studies on of collar (Mashahir and Zarrati, 2002).
bridge pier scour in cohesive sediment beds were Entesar EL-Ghorab (2013) suggested a technique
carried out on circular piers. B.W. Melville to provide openings on upstream side of a pier.
(1975) explained that the downward flow These openings starting from upstream face of
behaves as vertical jet and moves the bed pier continue transversely to lateral sides and
material. He also pointed out that downward flow straight to the downstream side. In his
initially erodes to develop the horseshoe vortex experiments, he varied discharge, depth, size and
thus increasing the size of scour hole. vertical spacing of openings. He used opening
It was found in the literature that the shape of the size equal to 10%, 15% and 20% at vertical
pier can intensively affect the flow pattern around spacing of 0.5, 1.0 and 1.5 times the pier width
it. Tseng et al. (2000) and Richardson and Davis respectively. He found that the opening size
(2001) stated that the strength of the downward equal to 20% with vertical spacing 1.0 times the
flow, horseshoe and wake vortices are greater in pier width was the best in decreasing flow
the case of square piers as compared to circular pressure thereby reducing the scour.
piers in non-cohesive sediment bed. Dey et al. (2006) and Elham and Manouchehr
Many remedial techniques have been applied to (2012) examined the provision of threading or
control local scour around bridge piers. There are cable around a vertical pier at different angles,
two kinds of techniques to reduce erosion process i.e., 15o, 30o and 45o. They suggested that the
near pier base. cable wrapped at 15o with a cable-pier ratio of
1. Bed armoring technique 0.15, was the most effective to countermeasure
2. Flow alteration technique and weaken the downward flow effect causing
Zarrati et al. (2006) examined the application of horseshoe and wake vortices. The provision of
riprap at the base of collared pier and found a threading around a vertical pier is an economical
significant reduction in sediment movement technique to minimize local scour.
around the pier. They concluded that two collared Thus, in addition to a collar, other methods could
piers in a line parallel to the flow direction with also be adopted to countermeasure local scouring
riprap at the bed, decreased erosion up to 60%. around bridge piers. These procedures include
A number of other pier modification techniques combination of a cable and openings around the
were also employed to dissipate effect of pier. The main objective of this experimental
downward flow thus minimizing horseshoe investigation is to develop techniques whose
vortex consequently reducing local scour. The application would make bridge piers safer and
present research is aimed at investigating the more strengthened against scouring action by
erosion around bridge piers by using pier stream flow and flood. These measures are
modification techniques. One of the pier providing protections to bridge piers, in the form
modification techniques that often used to of different combinations of collar, cable and
minimize pier scour, is the provision of a collar. openings.
Fotherby et al. defined collar as a device built
Experimental procedure acts on the bed resulting in the formation of
Pier Alteration horseshoe and wake vortices. In all the
Pier alteration is very essential as this component experiments, bridge pier models made of wood
of bridge is directly affected by the scour action were used. The maximum width of both
during normal and flood seasons. When the octagonal and circular piers equal to 7 cm was
upstream flow collides with the pier, the water used. All other dimensions of pier and collar are
moves downward, all around the pier. This flow shown in the Fig. 2.
Fig. 2 Detail of pier
& collar dimensions 1. In all collar-pier experiments, the width

M.M. Abdel-Motaleb (1997) studied the of collar equal to thrice to that of pier was
provision of openings in the pier. In his maintained.
experiments he used openings, each having a 2. In all cable-pier experiments, the cable of
diameter equal to 10% of the pier width. R.J. thickness 10.5mm was wrapped around
Garde and U.C. Kothyari (1998) carried out the pier. The vertical spacing between
experiments on pier models by providing a slot in the cable loops was equal to width of the
the direction of flow. He found that the maximum pier; in fact, each cable is parallel to the
scour depth reduced up to 30%. The upstream to other.
downstream openings provided another path to 3. In all openings-pier experiments, the
divert the flow. This reduced the effect of diameter and vertical spacing of
downward flow velocity along the pier causing openings was 20% and 100% of the pier
less scour. width, respectively.
Following are the details of the pier
configurations as shown in Fig. 3.
Fig. 3 Details of pier modification
Fig. 4 (a) Cross section of Experimental channel
Fig. 4 (b) Top view of Experimental channel
Experimental Set-up and Materials

All the experiments were performed in a smooth
rectangular flume (2000 cm length, 100 cm width
and 75 cm depth) equipped with recirculating
water facility. The flume is located in the
Hydraulics Laboratory of Civil Engineering
Department at University of Engineering &
Technology, Taxila, Pakistan. The flume base is
built of concrete and side walls are made of 12
mm thick glass sheet. For all experiments same
flow conditions was maintained. Fig. 4a and Fig.
4b show a sketch of experimental setup.
The flume has a straight entrance with working
segment consists of 600 cm long and 30 cm deep Fig. 5 Gradation curve for fine sand.
layer of non-cohesive uniformly graded fine sand
material in which wooden pier is centered. The The experimental process was made out under
fine sand with grain size, i.e., D50 = 0.28 mm is clear water condition. At the end of the channel,
used. The sediment used in the experiments is there is a circular gated weir which control and
considered as uniform having specific gravity measure the discharge. To measure flow velocity
2.79 and geometric standard deviation (σg) = at different locations in the channel and also near
(d84/d16) = 1.21; where d16 = diameter for stream bed digital velocity sensor is used.
which 16% by mass of sediments are finer; and Similarly, due to scouring phenomenon, the
d84 = diameter for which 84% by mass of scouring depth around vertical pier is measured
sediments are finer. The standard deviation of by means of digital point gauge with an accuracy
grain sizes was σg = 1.21 < 1.30 to terminate the of ±0.01 mm.
impact of non-uniformity of the bed material on
scour depth (Zarrati et al., 2010). Particle size Testing Procedure
analysis of the fine sand is shown in Fig. 5. A series of twenty-one experiments on five
different pier configurations, all under clear-
water scour condition, were performed. For each
experiment, the initial plane bed level was
measured precisely. As the flow continues, the
scour progresses at the bed until the equilibrium
depth is reached. Same flow depth and discharge
was maintained for all the experiments. In order
to monitor the eroded profile of scour hole, a the equilibrium time, when the scour depth did
meter rod attached with a small flat circular tip not vary by more than 5% of the pier diameter in
was used. the succeeding 24 hours and developed the
At the end of each experiment, water was drained following equation for the prediction of the
off from the flume carefully to keep scour pattern equilibrium time for the clear-water condition.
ds 𝑢 𝑡
unaffected from the draw down flushing. The = exp{-0.03  ln 1/6} (1)
scour holes and scour pattern profiles around dse 𝑈𝑐 𝑡𝑒
octagonal pier were then accurately measured

Here ‘ds’ is the scour depth, ‘dse’ is the
with the help of depth gauge. Finally, temporal
equilibrium scour depth, ‘u’ represent local
development of scour for all pier configurations
velocity, ‘Uc’ is the critical velocity, ‘t’ is the
was plotted. Fig. 6 shows the description of
time and ‘te’ is the equilibrium time.
experimental tests.
In the current study, a long-time experiment was
performed for 12 hours according to the Melville
and Chiew (1999) criteria. The results suggested
that, depth of scour increases with time and there
is a good correspondence between experimental
results and equation (1) for the equilibrium time.
It was observed that 90% of the equilibrium scour
depth was achieved in the initial 3 to 4 hours.
Since the peak flood flow may not last long
enough to develop equilibrium scour depth, the
duration for further experiments was maintained
as 8 hours.
Results and discussion

In the current study, experiments were conducted
to determine the effectiveness of pier
modification on scouring action. In order to
Fig. 6 Detailed experimental setup of bridge minimize scour around pier, different pier
pier configurations were used. Initially, to examine
the response of pier geometry on scour dynamics,
Test Duration Effect two pier shapes, i.e., circular and octagonal, were
The equilibrium clear water scour depth is studied under same flow conditions. After getting
achieved asymptotically with time thereby it much satisfactory results from octagonal shaped
takes a very long time for the equilibrium scour pier, both in terms of overall spatial distribution
hole to evolve. Several researchers have looked and temporal development of scour, as shown in
temporal development of scour depth Fig. 7, further modifications and experiments
experimentally under both clear water and live were performed on octagonal pier.
bed conditions. Due to clear water conditions,
scour depth progresses slowly with time towards
the equilibrium state. However, under live bed
conditions, the scour depth reaches to
equilibrium state rapidly (Chabert and
Engeldinger 1956, Raudkivi 1976, Melville and
Chiew 1999). The equilibrium scour depth under
clear water is 10% more than that under live-bed
conditions (Graf 1996). Based on their
experimental work, a number of researchers have
Fig. 7 Temporal evolution of scour at the
suggested time duration for scour depth to reach
upstream of pier
equilibrium state. Kumar et al. (1999) halted their
experiments, as the scour depth did not vary over
Temporal evolution of scour was analyzed
a period of 3 hours by more than 1 mm. Similarly,
specifically at the upstream of the pier for all the
Mia and Nago (2003) stopped their experiments
experimental runs. From the experimental
as the scour depth did not increase by 1mm over
results, maximum scour depth was observed for
a period of 1hour. They also concluded that
plain octagonal pier. The erosive action of
almost major scouring takes place after the first 3
downward flow caused horseshow vortex
to 4 hours. Melville and Chiew (1999) specified
resulting in maximum scour of octagonal pier as octago 6.93
6) Y Y Y
it was unprotected and without any nal cm
modifications.
Three combinations of different configurations However, the combination of all three
were examined, under same flow conditions, modifications, i.e., provision of collar, threads
around octagonal pier, i.e., threaded pier with and openings in the pier resulted in the most
collar, threaded pier with openings and pier with effective configuration in reducing scour depth
openings and collar. The maximum scour depth produced by downward flow and the formation
for octagonal threaded pier with collar was found of horseshow vortex. Comparison of scour depth
to be less than that of other two cases. The of plain circular pier (PCP), plain octagonal pier
reduction in scour depth occurred due to (POP), octagonal pier with opening (OPWO),
provision of collar in combination with threads octagonal threaded pier with collar (OTPWC),
causing obstruction to downward flow and octagonal threaded pier with opening (OTPWO)
resulting horseshoe vortex. Initially scouring and octagonal threaded pier with collar and
action blocked at the upstream of pier. The flow opening (OTPWCO) were made as shown in Fig.
is shifted to downstream of pier forming wake 9.
vortex there resulting in small scour pits. These
pits expand and progress towards upstream and
initiate scour depth there. Due to this delayed
process, the equilibrium stage of scouring takes
more time. As the scouring action, for collared
pier was started with delay, addition of threads
makes pier more effective as it further delayed to
start scouring due to which equilibrium time is
longer than that of other two cases. Maximum
scour depth around threaded pier with openings
was observed to be less than collared pier with
openings. The threads of cable obstructed the Fig. 9. Temporal evolution of scour at the
downward flow more effectively than the upstream of pier
openings in the pier as shown in Fig. 8.
Conclusion
This experimental study aimed at assessing the
capability of various pier modifications to
minimize erosive power of flow acting on the
river bed around bridge pier. The main objective
of this research is to reduce maximum scour
depth using a collar, cables and openings
provided in the pier separately and in
combination in clear water flow over mobile sand
Fig. 8 Temporal evolution of scour at the bed. It was concluded,
upstream of pier 1. Under same conditions the maximum scour
depth around octagonal and circular pier was
Table 1 Summary of Experiments monitored and found to be 9.1% less around
the former pier.
Sr. pier cab openi coll scour 2. The scouring under combination of different
no. shape le ng ar depth configurations was also determined.
circula 13.02 Octagonal threaded pier along with collar
1) - - - was more effective in reducing scour as
r cm
octago 11.83 compared to the other two cases. i.e., collared
2) - - - pier with openings and threaded pier with
nal cm
octago 9.87 openings. For the first case the reduction of
3) Y Y scour was 7.5% and 12.8% than that for the
nal cm
second and third case respectively.
octago 8.61
4) Y Y 3. A combined provision of collar, cable and
nal cm
openings in the pier found to be the most
octago 9.31
5) Y Y effective configuration for reducing scour.
nal cm
The resulting scour was 41.4% less than that [9] Graf, W.H. 1996. Fluvial Hydraulics. John
for the plain octagonal pier. Wiley & Sons, N .Y.
4. In all experiments of collared pier with
different combinations, because scouring [10] Kumar V., Ranga Raju K.G. and Vittal N.
action started with delay, the equilibrium 1999, Reduction of local scour around
time was longer than those without collar. bridge piers using slot and collar. Journal of
Hydraulic Engineering, ASCE, Vol. 125,
References No. 12, pp. 1302-1305.
[1] Ansari, S.A., Kothyari, U.C., Ranga Raju,
K.G. 2002, Influence of cohesion on scour [11] M.M. Abdel-Motaleb, minimizing of scour
around bridge piers. Journal of Hydraulic around bridge piers using a new method for
Research, Vol. 40, N0. 6, pp. 717–729. vortex reduction, Ain Shams Engineering
Journal, (1997) Vol. 32, No.2, pp. 115–125.
[1] B.W. Melville 1975, local Scour at Bridge
Sites, Univ. of Auckland, Auckland, New [12] Melville, B.W. 1997, Pier and abutment
Zealand. scour: integrated approach. Journal of
Hydraulic Engineering, Vol. 123, No. 2, pp.
[2] Breusers, H.N.C., Raudkivi, A.J. 1991, 125–136.
Scouring. IAHR Hydraulic Structures
Design Manual No. 2. Balkema, Rotterdam, [13] Melville, B. W., and Chiew, Y. M. 1999,
The Netherlands. Time Scale for Local Scour at Bridge Piers,
Journal of Hydraulic Engineering, ASCE,
[3] Chabert, J., and Engeldinger, P. 1956, Etude Vol. 125 No. 1, pp. 125-136.
des affouillements autour des piles des
ponts, (scour around bridge piers). [14] Melville B. W. and Coleman S. E. 2000,
Laboratoire National d’Hydraulique, Bridge scour. Water Resources
Chatou, France. Publications, LLC, Colorado, U.S.A., pp.
550.
[4] Dey S., Sumer B. M., FredsØe J. 2006,
Control of scour at vertical circular piles [15] Mashair M. B. and Zarrati A. R. 2002,
under waves and current. Journal of Effect of collar on time development of
Hydraulic Engineering, ASCE, Vol. 132, scouring around rectangular bridge piers.
No. 3, pp. 270–279. 5th International Conference on
Hydroscience and Engineering, Warsaw,
[5] Debnath, K., Chaudhury, S. 2010, Cohesive Poland, pp. 9.
sediment erosion threshold: a review. ISH
Journal of Hydraulic Engineering, Vol. 16, [16] Mia, Md. F., and Nago, H. 2003, Design
No. 1, pp. 36–56. method of time-dependent local scour at a
circular bridge pier. Journal of Hydraulic
[6] Entesar A.S. EL-Ghorab. 2013, Reduction Engineering, Vol. 129, No. 6, pp. 420-427.
of scour around bridge piers using a
modified method for vortex reduction, [17] Raudkivi, A.J. 1986, Functional trends of
Alexandria Engineering Journal, Vol. 52, local scour at bridge piers. Journal of the
pp. 467-478. Hydraulics Division, ASCE, Vol. 112, No.
1, pp. 1-13.
[7] E. IZADINIA and Manouchehr H. 2012,
Simultaneous use of cable and collar to [18] Richardson, E.V., Harrison, L.J.,
prevent local scouring around bridge pier, Richardson, J.R., Davis, S.R. 1993,
International Journal of Sediment Research, Evaluating scour at bridges. Hydr. Engrg.
Vol. 27, pp. 394-401. Circular No. 18 (HEC-18), Fed. Hwy.
Admin., U.S. Dept. of Transp., Washington,
[8] Fotherby L. and Jones J. S. 1993, The D. C.
influence of exposed footings on pier scour
depths, Proc. Hydraulic Engineering, San [19] Richardson, E.V., Davis, S.R. 2001,
Francisco, July 25–30, ed. H.-W. Shen, S. Evaluating scour at bridges. HEC18 FHWA
T. Su, and F. Wen, pp. 922–927. NHI Federal Highway Administration, US
Department of Transportation, Washington,
D. C.
[20] R.J. Garde, U.C. Kothyari. 1998, Scour [24] Zarrati A. R, Gholami H, and Mashahir M.
around Bridge Piers, PINSA 64, A (4), pp. B. 2004, Application of collar to contro
569-580. scouring around rectangular bridge piers.
Journal of Hydraulic Research, Vol. 42, No.
[21] Sumer, B.M., Fredsøe, J. 2002, The 1, pp. 97–103.
Mechanics of Scour in the Marine
Environment. World Scientific, Singapore. [25] Zarrati A. M., Nazariha M., and Mashahir
M. B. 2006, Reduction of local scour in the
[22] Tseng, M.H., Yen, C.L., Song, C.S. 2000, vicinity of bridge pier groups using collars
Computation of three-dimensional flow and riprap. Journal of Hydraulic
around square and circular piers. Engineering, ASCE, Vol. 132, No. 2, pp.
International Journal of Numerical Methods 154-162.
in Fluids, Vol. 34, No. 3, 207–227.
[26] Zarrati, A.R., Chamani, M.R., Shafaie, A.,
[23] Whitehouse R. J. S. 1998, Scour at marine Latifi, M. 2010, Scour countermeasures for
structures: A manual for practical cylindrical piers using riprap and
applications. Thomas Telford publications, combination of collar and riprap. Int. J.
Thomas Telford Ltd, 1 Heron Quay, Sediment Res., Vol. 25, pp. 313–322.
London, United Kingdom. pp. 198.
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Assessment of Defects, Remedial Measures and Development Prospects of

Sick Jaglot Hydropower Project
Muhammad Waseem1, Qaiser Karim1*, Muhammad Kaleem Sarwar1
1
Centre of Excellence in Water Resources Engineering, University of Engineering and
Technology, Lahore, Pakistan
*
[email protected]
Abstract: This study was carried out to identify the major causes of failure of the project, suggest remedial
measures and to explore the power potential of Sai & Damot nullah. A field survey was carried out to get essential
data of civil works for the identification of major causes of failure and to propose the remedial measures.
Hydrological, topographical and geological data analysis was conducted to explore power potential of both
nallah’s (Sai and Danot). Defects in Jaglot hydropower scheme includes wrong design of intake structure,
unnecessary lengths of power channel, non-consideration of better/ideal power house site, channel construction
on large fills in critical areas and insufficient cross flow structures. Defects in Jaglot hydropower scheme can be
rectified by constructing tyrolean weir type intake, cross flow structure, spillway at RD 2.5, desilting base at RD
2.8, repairing of expansion Joint of Penstock and improving the cross section of power channel at same critical
locations. Data analysis shows that power potential of Sai nullah can be enhanced from 1 MW to 177 MW by
using a discharge of 0.77 m3/s from Damot nullah.
Keywords: Jaglot hydropower project; Assessment of defects; Remedial measures
Introduction firstly, to rectify the one-megawatt power plant

Due to its geo-strategic importance and increase after thorough assessment of its defects, and
in population, Jaglot was upgraded to tehsil secondly to explore the possible power potential
headquarters. Accordingly demand of electricity of the Sai nullah and its tributary to meet the
for the town increased manifolds. Previously the future demand of the area. This study was
energy requirements of the town were being met conducted to recognize and point out the main
by a 100 Kw hydropower station constructed near causes of failure of the one-megawatt Jaglot
Jaglot Town. The source of water for power hydropower project, to suggest remedial
generation was the Sai nullah, a right tributary of measures for rehabilitation of the sick project,
Indus River. In order to meet the increasing and to explore power potential of the Sai &
demand of the town, Public Works Department Damot nullah’s to meet the increasing power
planned to install a new one-Megawatt capacity demand of the area.
hydropower plant. A proper rectangular power
channel was constructed to feed the new one-
megawatt power plant (2 x 500 KW) through two
penstocks. The power plant worked only for six
and a half years during which it had a large
number of failures due to various reasons. The
last major failure occurred two and a half years
back because of land sliding in which 800 m long
section was washed away. It took almost two
years to repair the system in which the section
washed away was replaced by 800 m long 36 
steel penstock pipe. During this time period the
only permanent source of power supply to the
area was the above mentioned 100 KW
hydropower plant supplemented with a 250 KVA
diesel generator for emergency. Keeping in view
increased power demand of the people and
strategic importance of the area it is necessary,
Fig. 1 Layout of Jaglot Hydropower Plant
Channel Construction on Large Fills in
Defects in jaglot hydropower scheme Critical Areas
To identify the defects in sick Jaglot It has been observed that in gorges of the rocky
Hydropower, field survey was carried out and strata, the head race canal has been constructed
following objects were identified: on the well compacted fill. No cutting and
blasting of the rocks were observed in turn large
Inability to Utilize the Channel Capacity dry masonry retaining wall having depth upto 15
One of the major problems is wrong design of to 20 feet were observed. The detailed checking
intake structure. One megawatt plant was revealed that there were a large number of
installed without any proper intake structure, due breakages due to rock falling and washing away
to which the limitation in availability of adequate of channel by flash water current created by
volume of water for the power generation storm water.
remains a permanent feature. Although adequate Insufficient Cross Flow Structure
discharge in the Nullah is available but same The power channel has been constructed on the
cannot be tapped, due to inappropriate design of forward slopes the mountain, while constructing
intake. the channel due consideration had not been given
to the storm/rain drainage of the mountain. The
Illegitimate Extension of Power Channel channel linearly crosses all the natural courses of
beyond the Requirement water, but no cross-flow structures had been
On carryin out physical survey, it was found out constructed on the channel. However, later on
that the length of power channel was eight cross flow structures were added but were
unnecessarily increased which was not the site not properly designed as per the expected loading
requirement due to which the channel extended and flow conditions. As a result, all of them were
into a stratum having unstable formation with a damaged. Eight cross flow structures against the
very frequent record of land sliding during the requirement of twenty are grossly insufficient to
raining and snowing season. Because of this cater for the storm drainage of the project area.
reason there were extensive breakages of power
channel due to landslides. The worst example is Unstable Strata
that of due to a land slide, the power plant The power channel passes through a number of
remained closed for over one and a half year in unstable strata consisting of sandy gravel and
years 2004 and 2005 causing extensive loss to the morain. While passing through these stratas no
scheme and over expenditure of funds for the additional reinforcement in bed and beds
channel construction. foundation been provided to safeguard against
settlement and seepage.
Non-Consideration of Better/Ideal Power A number of failures were observed due to the
House Sites leakage/seepage of water through the
There are two power houses sites i.e. site A and foundations/bed. The seepage water saturated the
site B which are at a reduce distance of 1.775 km underneath strata causing it to flow under self-
and 2.500 km from the inlet gate having an weight resulting into failure of water channel.
available gross head of 95 and 85 meters
respectively. At the site ample space is available Reduction in Cross-Sectional Area of Power
for the construction of powerhouse, the proposed Channel
foundation of power house and supports of Reduction in cross-sectional area of power
penstock lie on the solid granite rock. The site B channel without any valid limitation of slope or
allows the construction of the penstock with depth was observed at a number of places.
supports foundation on solid rocks of graduraid
formation and the foundation of powerhouse can
be laid on solid firm strata. Had these options
been considered, the cost of the project would
have been reduced to a considerable extent and
the problem of land sliding would have been
minimized. The maintenance and the operating
cost would have been reduced to a mark
difference and the project would have been
financially conductive project.
the construction of same length of water channel
and utilizing the previous existing penstock
facility.
Splitting of one Penstock into Two

The first 750 m of penstock is a single pipe of 36-
inch dia, approximately 50 m short of
powerhouse this pipe through a thrust block was
split into two pipes of 30 inches dia each. The
angles substandard by the downstream and
upstream turbine feeding pipe is 90o and 130o
respectively.
Fig. 3 Reduction in cross-sectional area of

power canal.
Insufficient Side Wall Height

The height of wall of the head race canal is
grossly insufficient to stop any sliding
boulder/stone entering the power canal. During
rain a lot of water enters the power canal as the
shoulders/walls of the power canal is at level or
lower than the uphill surrounding area.
Head Loss
During the survey it was observed that without
any solid valid technical reasons/limitation a fall
of 9 m was created at RD 2.4 km causing a head
loss of 9 meters. On enquiring from local
villagers, it revealed that the construction of
channel started from both intake and powerhouse
ends when both ends being constructed closed
together a 9 meters elevation difference between
channels was observed to counter it a 9 m fall was
constructed.
Fig. 5 Splitting of one penstock into two.
Following were the affects:

i) As per the operators and own
observation: The turbine and
generator were giving rated output i.e.
500 KW but the upstream couple was
Fig. 4 Head loss due to water fall giving an output of 325 to 350 kilo watts
against its design capacity of 500 kw.
Penstocks ii) When only the upstream generator is
Length of Penstock operated its output does not go beyond
For utilizing 85 m head for the power generation. 350 kw.
The length of penstock laid is over 800 meters iii) The synconsisation of both generators
which is a very expensive affair as compared to have been disturbed due to which
upstream generator was burnt three times reason, there is excessive wear and tear of the
costing repairs bill more than 1 crore 10 turbines blades and other mechanical equipment
lacks. causing repeated break down of the power supply
causing unnecessary loss in revenue and
Failure of Electric Distribution System excessive maintenance expenditure which
The electric distribution system failed due to non- otherwise could have been avoided.
synchronisation of generators. Due to this reason
the electric distribution has been affected. The Jaglot hydro power plant rectification of
electricity produced by both the generators is defects/ remidiel measures
being supplied to the consumers through separate Water Rights
transformers and electric lines made a short fall The people should only be allowed to take water
in the supply and demand. from Irrigational channels. Use of water from
power channel should be forbidden.
Missing of Desilting/Sediment Control
Structures Weir
a) Desilting Sediment Control Structures A Retaining wall type, free over fall weir at Rd
In this Micro hydropower scheme besides the 17.4 Km of Sai Nullah should be constructed to
other problems previously mentioned, the maintain a constant water level to ensure that a
desilting basin is missing due to which, this smooth flow enters the power canal be
scheme is facing a lot of breakage of Mechanical throughout the year irrespective of low or high
equipment and furthermore the entire length of flow conditions.
the power canal has to be desilted manually. Tyroleon Weir construction
Although the running water is generally snow A proper Inlet Structure may be constructed for
melt clear water having negligible sediments getting adequate quantity of water during Low
(very less quantity of sand or silt), but as soon as flow conditions and for desilting of water. The
it drizzles, light or medium rain occurs. The construction of Tyrolion weir is most appropriate
quantity of sediments increases many times. As solution
the channel has not been designed on non-
scouring non-silting velocity, the bulk of the clay, Protective and Safety works
silt and sand settle down in the bed of the channel The construction of protective works as per the
due to the low velocity of water in various site requirement such as Raising of Channel side
stretches of power channel. The manual desilting walls, Breast Walls, retaining walls, RCC Chanel
of the entire channel is clossal and it is an in the unstable stratas (deep fill etc), Enroute
inefficient job due to the dimension of the Over Flow structures for disposal of Mountains
channel and the constant running water Run off water may be constructed as per site
conditions. This becomes a major reason of the requirements.
enhanced operating and maintenance cost of the
project/system. Cross Flow Structures
The Cross flow structure may be redesigned with
Spillways steeper slopes (>30*) having wing structures as
In the project no spillway or over flow structure per site requirement may be constructed .
has been incorporated in the design of head race
canal. During the rains the rain water of almost 4 Spillway
km linear length of mountain enter the water The Spillway may be made on solid rock next to
channel resulting into its spill over at various proposed cross flow structure at Rd 2.55.
sections causing extensive damage to it. The
damage is of a much higher degree where the Desilting Basin
natural course of water is disputed specially in Desilting basin may be made on solid rock at Rd
gorges and depression where the channel has 2.8.
been made on fill with the help of retaining
structures. Penstock
Redesigning of main penstock splitting from one
Forebay Intake 36-inch pipe to two 30-inch pipes be done as a
There is no properly designed forebay for smooth curve transition to avoid Eddies and
housing of penstock pipes and any desilting cavitation and for better Synconisation of
arrangements for the protection and safety of the Generators. The damaged joints should be
mechanical equipment is also missing. For this repaired and worn out rubber seals may be
replaced.
D = 2.69 (22 x 0.0122 x 170/80)0.1875
Proposed future development to enhance the = 0.765 m
power potential Say 1.1 m
Power Potential of Sai Nullah
At present only 1 MW hydropower station has Poer potential of damot nullah
been established on Sai Nullah whose mean Damot Nullah Power Canal
average flow is more than 10.4 cumecs. The A hydropower plant may be designed on winter
power channel discharge carrying capacity can available discharge of the Damot Nullah i.e. 0.77
be enhanced by removing its deflects and slight cumecs. The section recommended for proposed
modification in its design: Damot Gah Scheme is of stone masonry well
plastered in trouble free areas. RCC section for
Table 1 Power Canal Design Data the suspected area of land sliding and area
Sr. Design Parameter Design Value suspecting free fall debris and boulders. The
No. recommended channel section is as given in the
1 Depth of power 1.20 m following table:
canal Table 2 Power Canal Design
2 Width of power 1.50 m Sr. Design Parameter Design Value
canal No.
3 Depth of flow 0.95 m 1 Channel Depth 0.9 m
4 Bed slope So 0.002 2 Depth of flow 0.65 m
5 Stone masonry , 0.019 3 Channel Width 1.00 m
n 4 Channel Bed 0.002
Slope
As it can be easily observed that the best possible 5 Discharge, Q 0.77 cumecs
dimensions of rectified channel will be = 1.2 m x
1.50 m with depth of flow 0.85 m. This would
Power Potential of Damot Nullah
give a discharge of 1.86 cumecs. This discharge
By using design flow of 0.77 m3/sec from Damot
can be utilized for power generation. With
Nullah the data will be as follows:
available width of channel as 1.85 m, a discharge
Head available, H = 80 m
of 2.5 m3/sec can be passed through plastered
Discharge available, Q = 0.77 cumecs
stone masonry rectangular section.
By using the available discharge at new site, the P=g H Q 
gross available head is 85 m. The additional Power, P= (0-.85) x 9.8 x 0.77 x 80 =513 KW
power potential of Sai Nullah will be as under:
Forebay
A rectangular structure with the following
P=gHQ
dimensions can be made to house the pair of the
= (0.85) x 9.8 x 2.5 x 85 = 1.77 MW
penstocks.
Length: 15 m
Penstocks
Width 4.5 m
It is recommended that there should be one
Depth 3.5 m
penstock one having a capacity of 2 cumecs. The
dia of the penstocks can be calculated by using
The structure is located on semi-hard rock
the relations as mentioned in methodology for
formation which is granodiroid having ample
permissible loss of 5% will be as under. The
space to house the structure. The alignment and
proposed dimensions of the penstocks will be:
the location are shown in the Fig. 6.
D = 2.69 x (n2 Q2 L/H)0.1875
Fig. 6 Proposed site of forebay
Penstock /requirement resulting in a substantial

A pair of penstocks of 130 meters of length increase in Project and project’s O&M costs.
having 0.70 m internal dia are proposed which f) To avoid future land sliding /power canal
will be housed on the supports following the failure by employing 800m long 0.92 m dia
ground profile which is has an approximate slope pipe was not a technically and economically
of 75o. a sound solution.
g) Splitting one 0.92 m dia Penstock into two (Y
Powerhouse shape) of 0.769 m dia pipes in 15 m length
An external type powerhouse at an elevation through smooth transition to avid eddies and
1710 m.a.s.l. is proposed. It will be equipped hydraulic losses is hydraulically a sound
with 2 units having a input of 80 m heads and solution.
using a discharge of 0.38 cumecs each which
will give an output of 0.25 MW each. Recommendations
Tail Race a) All the defects as pointed in previously,
A tail race of 200 m length having a cross essential for the safe and trouble-free
sectional area of 1 m x 1 m can be proposed. The operation of scheme may be rectified on
Ist 100 meters will requires a digging of 4 m to priority.
enhance the head of the scheme.
b) Missing structure may be constructed on
Conclusions priority.
a) The one Mega Watt hydro Power Project on
Sai Nullah had neither been thoroughly c) At present the channel depth and width is 1
planned nor had it has been executed. m and 1.75 m respectively. The depth of flow
b) Standard Engineering practices such as site in the channel is 0.8 meter. The depth of the
Investigations, Proper Labourity testing, Site channel should be increased to 1.8 m and
Survey, Topographic Survey, Geological depth of flow should be kept as 1.5 m. The
Mapping, Project lay out optimization and site permits such a modification in
Design had not been done. dimensions. This increase in dimension
c) The Northern area PWD planned and should only be upto RD 1.775.
executed the project with least technical
Knowledge, Skills Qualifications and d) The existing stone masonry channel should
knowhow of construction of hydropower be well plastered to reduce the frictional
projects losses and enhance the velocity of water in
d) The planning and Execution had been done the channel. With this development the
by Northern area PWD. The Operation and enhanced channel capacity theoretically will
Maintenance is still not in their responsible enhance as per the manning equation:
hands. Any minor or major repairs still takes
considerably long time due to inter e) Site A at RD 1.543 km having a power
departmental tussle. It should be transferred generating potential of over 2 MW provides
to WAPDA for better Efficiency. good site conditions for the construction of a
e) The length of power canal had been powerhouse with a shortfall of steep gradient
increased without any technical justification
(angle of elevation of penstock will be regulation and distribution system in the
having a vertical angle of 70o to 75o). Chishtian subdivision at the main and
secondary canal levels. IIMI-Pakistan
f) Damot Nullah scheme having a power National Programme, Report No. R-599.
potential of over 0.5 MW may be developed Jarle Ravn. 1992. Planning and Implementation
to enhance the power production to meet of Hydropower Projects; published by
present and future power demand. Norwegian Institute of Technology Division
of Hydraulic Engineering.
g) The existing small HPP in Northern John J. Cassidy. 1984. Site Development and
Areas having problem in generating Hydraulic Analysis, published by McGraw-
continuous power may be checked in the Hills Book Company, London.
field to rectify the problem and to improve L. Monition M. Le Nir and J. Roux. 1984.
power and energy production. Micro Hydroelectric Power Stations;
translated by Joan McMullan; A Wiley
References Interscience Publication
American Society of Civil Engineers. 1989. Civil Layman’s guide. 1993. How to develop a small
Engineering guidelines for planning and hydro site. European Small Hydropower
designing hydroelectric developments. Small Association.Mosonyi, E. 1986. Low Head
scale hydro, Vol. 4. Power Plants.
C.C., Warnick in collaboration with Howard A McKinney et al. 1983. Micro hydropower
Mayo Jr. P.E/James L. Carson/ Lee H. Handbook, 2 vols. U.S. Department of
Sheldon, P.E. Hydropower Engineering. Energy. A basic manual for very small hydro
Published by Prentice-Hall Inc. Englewood applications.
Cliffs, NJ. Noyes, R. 1980. Small and Micro
COE, U.S. Army Corps of Engineers, 1979. Hydroelectric Power Plants, Technology
Hydropower Cost Estimating Manual. and Feasibility.
COE, 1979. Feasibility Studies for Small-Scale Peurifoy, R.L. and Ledbetter, W.B. 1985.
Hydropower Additions: A Guide Manual. Construction planning, equipment and
Daughery, R.. and J.B. Franzini. 1977. Fluid methods. McGraw-Hill Co., Fourth Edition.
mechanics with engineering applications. Punmia, B.C. and Pande B.B. Lal. 1983.
GTZ-WAPDA. 1999. Layout and Sizing; High Irrigation and water power engineering.
Head Hydropower. Sarwar, K. M. 2005. Application of HPC model
GTZ-WAPDA. 2000. Hydropower in General, for the optimum layout of Bunji Dam
High Head and Low Head Hydropower. Hydropower Project. M.Sc thesis of
GTZ-WAPDA. 2000. Power Plant Components; CEWRE, UET, Lahore.
High Head Hydropower. William P. Greager and Joel D. Justin.
Habib, Zaigham & Miarcel Kuper. 1998. Hydroelectric Handbook, published by John
Performance assessment of the water Wiley & Sons, Inc. New York.
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Double-Stilling Basin Modelling, Pakistan – Case Study

Ubaid Ullah1*, Eric J Lesleighter2, Muhammad Iqbal3, Umair Mannan4
1
Principal Hydraulics Engineer, Engineering General Consultants (Pvt. Ltd.), Pakistan
2
Lesleighter Consulting Pty Ltd, Sydney Australia,
3
Principal Hydraulics Engineer, Associated Consulting Engineers (Pvt. Ltd.), Pakistan
4
Senior Hydraulics Engineer, Associated Consulting Engineers (Pvt. Ltd.), Pakistan
*
[email protected]
Abstract: The Mohmand Dam is a 213m tall concrete-faced rockfill dam (CFRD) under design by a
consortium of consultants in Pakistan. The consortium comprises SMEC International (Australia), Nippon
Khoe (Japan), National Engineering Services (Pakistan), Associated Consulting Engineers (Pakistan),
Engineering General Consultants (Pakistan), and BAK Consulting Engineers (Pakistan). The dam is for
construction on the Swat River in Pakistan. A strategic component of the hydraulics studies has been large-
scale physical model studies carried out by the Irrigation Research Institute, Pakistan. The paper will
discuss in detail the use of, and the hydraulic behavior of a double hydraulic jump stilling basin facility
incorporated in a 600m long concrete chute. The upper basin was designed to operate with a maximum
head of about 100m, and the lower basin was designed to operate with a maximum head of approximately
120m with respect to tailwater level. The studies considered discharges up to approximately 25,500 m 3/s.
Detailed pressure transducer measurements of transients as part of the design of the basins, and the chutes
incorporated several aerators along the length of the chutes.
Keywords: Dams, spillways, hydraulic jump basins, energy dissipation, turbulence, pressure
transients.
Introduction Hydrology
The Mohmand Dam Hydropower Project In July 2010, the Swat River experienced
(MDHP) is a large power project to be built on extreme flooding with a discharge estimated as
the Swat River approximately 200 km northwest 9,909 m3/s and considered to have an Annual
of Islamabad. The location of the dam is Exceedance Probability (AEP) of 1 in 1,000. The
illustrated in Fig. 1. The project investigation and return periods for the Project went through a
design passed through several studies with a number of studies, updating and finally an
detailed feasibility study preceding the lengthy accepted series of magnitudes. It was confirmed
study that derived the double-stilling basin by WAPDA in January 2016 that the Project
spillway arrangement – the subject of the present design team should adopt a panel of expert’s
paper. The work, comprising site and dam recommendation for the Probable Maximum
selection, detailed hydrology, reservoir Flood (PMF) of 27,427 m3/s inflow discharge.
sedimentation aspects, power station sizing, The computed peak value for AEP 1 in 1,000 and
diversion detailing, and hydraulics progressed 1 in 10,000 are 10,669 m3/s and 18,640 m3/s,
the earlier feasibility studies and all has been respectively. Fig. 2 presents the inflow flood
carried out on behalf of the Water and Power hydrographs at Mohmand Dam site.
Development Authority (WAPDA) by a As will be described below, the energy
consortium of consultant companies. dissipation arrangement selected a two-stilling
The diversion works comprise two 15m diameter basin arrangement. The flood routing yielded a
tunnels. One is to be developed into a permanent, PMF outflow discharge from the reservoir of
low level outlet facility for necessary releases to 25,362 m3/s. For the spillway energy dissipation
the downstream and for drawdown purposes. The design the design discharge was selected as 90%
power intake will direct flows into a separate of the PMF for the upper stilling basin and the
tunnel leading to an 800MW power station on the AEP 1 in 10,000 discharges for the lower basin.
right bank of the river a short distance
downstream of the dam and near the spillway
discharge location.
Fig. 2 Inflow flood hydrographs
Fig. 1 Location of Mohmand Dam Project,

Fig. 3 Early Plunge Pool arrangement
Pakistan
The flip bucket-plunge pool arrangement was
Spillway type and key dimensions
abandoned, and consideration moved to the
The proposed spillway was to be located on the
application of a hydraulic jump stilling basin
left abutment of the dam, following the dam-type
alternative as the means of dissipating the energy
selection of a concrete-faced rockfill dam from the spillway. The dam with a crest level at
(CFRD). The type of spillway received detailed
EL 563, and a parapet wall level at EL 564.5,
consideration. Originally, the project would provide for a reservoir with a full supply
commenced with the plan to use a flip bucket and
level of EL555m, and a spillway crest level of EL
plunge pool energy dissipation arrangement, for
539m.
which the plunge pool pre-excavation would be a Several chutes and stilling basin arrangements
large-volume depression on the river’s left and
were contemplated. The headworks also passed
against a steep excavation of the hill on the through several alternatives for the number of
spillway’s left. Essential to the provision of an
gates, and whether part of the spillway would
acceptable plunge pool dissipater was the
remain ungated with the crest at FSL. The result
consideration of rock scour and its longitudinal was seven gates, each 15m wide, and piers 5.3m
and lateral extent. Even the pre-excavation of the
thick, all placed on a curved crest alignment on a
plunge pool would require a large slope 500m radius. The chute was converged from the
excavation on the left side. The site
total gross crest width of approximately 137m to
investigations revealed rock largely classified as
a width of 100m. This led to a unit outflow
a foliated schist. Based on the Consultant’s
discharge at PMF of approximately 255 m2/s.
experience it was considered very erodible under
With a reservoir level in the region of EL 560 and
the action of velocities around 45 m/s. the river bed in the dissipation area at EL360,
The main issue with the plunge pool erosion,
clearly the 200m head placed stringent conditions
apart from a likely depth to 60m below river bed
on the spillway design. Early considerations of a
level, was its lateral expansion and movement single stilling basin indicated basin inflow
with the result that the entire left hill excavation
velocities around 60m/s, and in due course it was
would be undermined and be subject to collapse. decided to investigate the use of a double stilling
Fig. 3 is a portion-plan of an early plunge pool
basin configuration, somewhat similar to the
possibility; it shows the large excavation on the
arrangement used some decades earlier on the
left side of the plunge pool. Such collapse in turn
Mangla Dam spillway, also in Pakistan. The
would produce a large volume of scoured and
investigation of the double basin configuration is
collapsed material to form a huge blockage in the the thrust of this paper. Detailed physical model
river, affecting the power station, the permanent
studies were carried out at the Irrigation Research
outlet works and the spillway itself.
Institute (IRI), Nandipur, Pakistan.
Fig. 4 shows a plan view of the spillway and Fig. and discharge into the lower basin with an invert
5 a profile, depicting the chute from the level at EL348.
headworks into the upper basin with an end weir
Fig. 4 Plan of the double basin spillway
Fig. 5 Profile of the spillway
The upper basin was designed to operate with a basins were provided with conventional chute
maximum head of about 100m, and the lower blocks, and baffle blocks were sized according to
basin was designed to operate with a maximum the jump characteristics and the USBR guidelines
head of approximately 120m. The studies on sizes and spacing. The model was constructed
considered discharges up to approximately in Perspex and instrumented with many
25,500 m3/s. Detailed pressure transducer piezometers and several locations for pressure
measurements of transients in the upper basin transducers. Fig. 7 is a view of the model in
were made as part of the design of the basins, and operation with a AEP 1 in 10,000 discharge, and
the chutes incorporated several aerators along the Fig. 8 shows the lower basin at AEP 1 in 1,000
length of the chutes, the geometry of which was discharge.
studied and varied on the hydraulic model.
Hydraulic model description

The model was built and tested with a scale of
1:60. Fig. 6 shows a general arrangement plan of
the model boundary. Each stilling basin was
designed initially with estimation of spillway
losses, and basin length and depths based on the
hydraulic jump characteristics, on the basis that a
USBR Type III arrangement would be used. The
the length of the aerated zone as well as
relocating the aerators to command the chute
length sufficiently to provide assurance that the
full chute flow would have adequate aeration.
There is sufficient experience – model and
prototype – to allow confidence in the designs,
both in their location and in the air duct areas to
meet the demands of the jets from the ramps. Fig.
9 shows the dimensions of the five aerators. Fig.
10 shows the flow profile at the two upstream
aerators.
Pressure transients
A key consideration in stilling basin design is the
amount of uplift forces due to the combination of
under pressures, pressure transients and
Fig. 6 Coverage of the spillway model transmission of pressures through joints. The
high-energy conditions in both stilling basins
The model studies dictated close consideration of the pressure
Many of the key and major aspects of the transients in the stilling basins for the slab and
hydraulics, leading to a final design arrangement, anchoring design. Pressure transducers were used
were studied. First, to obtain satisfactory on the floor of the basin both upstream and
retention of the hydraulic jumps, by making downstream of the baffle blocks, and on the
alterations to the basin lengths and floor levels, sidewall of the stilling basin. Records of
and numerous changes to the energy dissipation pressures were obtained at a sampling speed of
baffle blocks and aerators. A workable and safe 300Hz for generally up to 5 minutes (model).
design was achieved through the studies. The
second major stage of the studies were the
measurements of transient and static pressures in
key locations. Several aspects of the model
dimensions and details were subject to change:
• Upper basin length from 90m to 120m and
end weir height from 14m to 16m
• Aerators reduced from six to five – two on
the upper chute and three on the lower chute
• Aerator geometry
• Location, height and number of the baffle
blocks in the upper basin, and Fig. 7 View of the upper basin operation for
• Lower basin lowered from EL 355 to EL348 AEP 1 in 10,000
having regard to the tailwater rating based
headworks 5 km downstream.
The baffle block utilized the shape developed by
USBR studies of a “supercavitating” block
during testing for the Folsom Dam auxiliary
spillway (USBR, 2009). The purpose was to
“push the limits” for which baffle blocks could
be used in a cavitation environment, meanwhile
ensuring generous aeration of the lower flow
layers in the chute and into the stilling basins.
The aerators on the Mohmand model showed the
nappe profiles well. The performance led to the
lowering of ramp heights in some cases to reduce
Fig. 8 Lower chute and lower basin entry Q1,000
Fig. 9 Final aerator geometry
By way of illustration, Fig. 11 shows the

deployment of 8 transducers on the floor of the
model upper basin. Fig. 12 shows a small part-
sample of the 2,300s (prototype time) total
capture of the transients at two of eight
transducers in the upper basin. The information,
together with cross correlation analysis of signals
from pairs of transducers and spectra, provided
information for the design of the basin floor
thickness as well as anchors.
A sample of the spectral density plots for two
transducers for the PMF is shown in Fig. 13.
Clearly, the major fluctuations power is around 1
Hz or less, frequencies which are well within the
“capability” of the structure floor slabs to
respond and therefore relevant for any dynamic
analysis of the slab/anchor system.
Fig. 10 Nappe profiles at aerators 1 & 2
Fig. 11 Transducer locations in the upper basin for one test configuration
Fig. 12 Sample transient pressures at two transducers, upper basin
Fig. 13 Sample spectral density plots for

transducers 0 and 5 for the PMF
Conclusions
The paper, describing detailed studies on a
physical model, shows the value of the exercise
in improving significantly on desk-type
estimates. Many aspects of the hydraulic
structures needed to be addressed by making
modifications to the basins, the aerators, and the
stilling basin appurtenances. The conditions are
major by all comparisons, with high heads and
large potential discharges. The double stilling
basin presented a workable and desirable option
to fit within a narrow corridor with a high
mountain (and appreciable excavation) on one
side and the dam on the other side. The model
results allowed confidence in the detailed
design exercise which followed.
Acknowledgments
The authors express their appreciation to the
IRI, Nandipur for its skill in building and
testing the model, particularly research officers
Tariq Anwar and Sana Ullah. Thanks also to
Manly Hydraulics Laboratory’s (Sydney)
assistance in providing instrumentation and
advice for the model transient work, and
running the pressures analysis.
References
USBR, Denver, Colorado, 2009, Cavitation
Potential of the Folsom Auxiliary Stepped
Spillway, Hydraulic Laboratory Report HL
2009-07.
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Modeling of spillway breach for Attabad landslide dam

Muhammad Wajid Ijaz1,2*, Ata-ur-Rehman Tariq3
1
U.S.-Pakistan Centre for Advanced Studies in Water, Mehran University of Engineering and
Technology, Jamshoro, 76062, Sindh, Pakistan
2
Environmental Protection Agency, Lahore, 54000, Punjab, Pakistan
3
Pakistan Engineering Services, Lahore, 54000, Punjab, Pakistan.
*
[email protected]
Abstract: Landslides occur due to rock falling on account of slope failure caused by different hydrological,
geological and weathering perturbations. A drastic event of landslide occurred on 4th January 2010 at Attabad,
Gilgit-Baltistan, Pakistan and a terribly massive landmass chocked the Hunza river. It claimed huge losses of
living beings and infrastructural damages. There were great suspicions about overtopping of its lake water. In
order to avoid the foreseen failure of the dam and extraordinary flooding, an artificial spillway over the saddle
of the dam was excavated. After overtopping of water, the predictions did not come true. Keeping it in view, the
spillway was modeled in River Analysis System of Hydrologic Engineering Center (HEC-RAS) with synthesized
infill gradation and erosion of its bed was simulated in sediment transport module. The results showed that there
was very small erosion and the dam was stable against normal overtopping conditions. A hydraulics-oriented
scenario of planned breaching of the dam and safe drainage of its lake was tested in the calibrated model. The
results demonstrated that by introducing repeatedly check-dam of 15 m height at the mouth of the spillway and
its subsequent controlled breaching can effectively lower the spillway invert up to 50 m and remaining part may
be accepted or else has to be removed with mechanical or blasting source. For authentic erosion results,
additional particle gradation and further measured data should be used in modeling.
Keywords: Natural dam, sediment transport, erosion, Hunza River
Introduction Gilgit River in 1911 and in Hunza valley in 1977.

Landslide dams are created due to falling of In recent years there have also been some minor
landslide mass due to various geological, flood events due to the sudden drainage of supra
hydrological, weathering processes etc. across a glacial lakes. Dam burst events were relatively
flowing stream (Liu et al. 2017). Landslide dams common during the period from 1833 to 1933.
are always different in their geomorphology and The most critical glacial lake outburst floods
chemistry (Fort et al. 2014: Stefanelli et al. 2016; occurred in August 1929, June 1841, August
Ahmed et al, 2017). Composition and shape of 1885 (landslide), when the massive flood waves
landslide dams primarily depend upon valley resulted in a significant rise in water levels
characteristics. Landslide dams may be larger (Calligaris et al. 2010).
than man-made dams in the world. Water stored In pursuance of the developing situation, Shah et
behind such landslide dams ultimately seep al. (2013) and Iqbal et al. (2014) compiled the
through or over top the landslide dam mass for background of the incident and present efforts
attaining its natural parent channel path. Stability underway. Butt et al. 2013 conducted inundation
of these natural dams depends on the volume of analysis against different peak values using HEC
landslide mass, tectonic configuration, and rate RAS and geographical information system (GIS).
of flow of runoff from mountainous drainage NDMA, (2010) arranged a scenario based dam
network. Failure often occurs a short time after break analysis from NESPAK who using HEC
formation; 50% of breached landslide dams RAS reported subsequent inundation extents and
documented in the world failed within 10 days flooding depth in confluences of Gilgit, Pakistan.
from the formation (Costa and Schuster, 1988). Furthermore, Chen et al. (2017) have recently
Karakorum Range of Himalaya traverses the undertaken a dam stability analysis using an
Hunza valley. Sixty historical landslide dam empirical model in which likelihood of its failure
burst events have been reported in the northern was quantified and hydrological consequences of
part of Pakistan. This gives an average recurrence dam failure were assessed. All the studies
frequency of about one event every 3 year. Much pertaining to Attabad landslide dam were focused
smaller landslide dam failures took place on the majorly over flooding issue and likelihood of
stability, but no results were reported about the dam in order to mitigate the expected losses due
use of infill gradation in a scientific way. Thus to sudden breached flooding.
present study aimed to fill this literature gap and
breaching analysis of the spillway was by Methods
coupling the lake/river hydrology with the Breaching analysis was performed using River
morphological configuration of the landslide Analysis System developed by Hydrologic
dam. This study would help the disaster Engineering Centre (HEC RAS). Mobile bed
managers and hydrologists for effective planning analysis of sediment transport module was used
and mitigation of such natural calamities in a to model the lowering of spillway invert.
more rational manner. Spillway channel was reproduced in the
geometric editor with 15 cross sections
Methodology measuring reach length 2000 m. Channel shape
Study area was trapezoidal with bottom width 1 m, to width
In result of October 2005 earth quake in Pakistan, 6 m and average depth 16 m (NDMA).
the seismic disturbance generated cracks in Since 4th January 2010 water was pounding
Hunza valley near Attabad district Gilgit behind the dam and on 29th May 2010 lake water
(Hussain and Awan, 2009). This area has been overtopped the spillway invert. Available data up
marked as a fault and high hazard zone by to 14 June 2010 was used in quasi-unsteady flow
Geological Survey of Pakistan since August form in order to calibrate the model. Three soil
2009. Eventually, it resulted in a massive gradation curves were constructed on the basis of
landslide due to the disintegration of rock from results of mechanical sieve analysis obtained
mountain slopes. The landslide mass estimated as from Royal Material Testing Laboratory,
45 million m3 fell in Hunza River (36.307ºN, Rawalpindi, Pakistan, and value of Manning
74.816ºE) on 4th January 2010. The slide roughness coefficient (n) were computed using
occurred from the northern wall of Hunza Valley Strickler formula.
and a landslide dam was created across the Hunza Breaching analysis is carried with the assumption
River (Petley, 2010), shown in Fig. 1. Hunza that whole landslide mass could be eroded by the
River is considered as a major tributary of Indus erosive power of overtopping water. So,
River in Pakistan. maximum depth is put equal to the depth of
landslide mass from river bed level. Laursen-
Copeland sediment transport function was
selected for breaching analysis (Copeland &
Thomas, 1989). Because the spillway channel
and its whole mass were not uniformly settled, so
it is treated as nonhomogeneous and graded bed
channel. Exner 5 method was selected to simulate
the sorting and armoring process. Report 12 was
used as fall velocity calculation method.
The average depth of degradation along the
spillway channel on the crest of the landslide dam
Fig. 1 Attabad landslide site location in
was designated as “Measured Minimum Depth”
Northern Areas of Pakistan
and an average depth of degradation along
downstream shoulder as “Measured Maximum
The slide occurred on a steep cliff consisting
Depth” in a unit of length as shown in Fig. 2.
primarily of diorite and granodiorite of the
However, calibration was done by making
Cretaceous-Tertiary age Karakoram batholiths.
logical variations in the bed gradation during
These rocks contain pervasive veins of pegmatite
each trial. Measured minimum and maximum
and quartz and are heavily tectonized, with both
depths were used as a reference while the change
folding and large-to local-scale faulting being
in bed gradation made possible for closing the
evident. The slope was extensively mantled with
modeled results to measured results.
glacial tills and colluvium, which was integrated
into the slide (Petley et al. 2011).
The crest of the landslide dam is approximately
140 m above the river bed. A spillway as
triggering element was formed by massive
excavation on the 2 km long crest of landslide
unsteady flow module of HEC RAS was routed
through spillway channel in order to achieve the
high erosive power of flowing water.
Results & Discussion

Following particle size gradation curves were
prepared with results of sample 1, 2 and 3. These
samples represent finer part of the fill. It is most
likely that larger grains could as well as part of
the avalanche infill representing gravel, cobbles,
and boulders. Gradation of Sample 1, 2 and 3
varies from clay to gravel classes of particle size
those are shown in Fig. 3 (a to c).
Fig. 2 Schematic diagram of Maximum and The spillway erosion results on the basis of the
Minimum Depth of degradation three-sample gradation are shown in Fig. 4 (a to
c). There was a large difference among measured
Scheme of breaching scenario testing and modeled results on using these gradation
The inflow rate of Hunza River at the time of curves. There is a major difference of magnitudes
spillway overtopping was approaching towards of the depth of erosion among modeled
its peak so within 10 days of overtopping a maximum and a minimum depth of erosion. It
maximum depth of degradation achieved was was confirmed during sediment transport
38.7 m while the minimum depth of degradation analysis that particle gradation curves are not a
was 4.95 m. After some time, inflow rate of true representative of whole landslide mass.
Hunza River and outflow rate through spillway It is obvious that there is always a large variety
channel were equalized and finer material eroded of grain classes present in landslide dams.
had been transported. It was hypothesized that Keeping in view the natural gradation of
exposed material in spillway channel was landslides the theoretical gradation was
comparatively coarser and flow rate through formulated. Numerous trials were carried out by
spillway was not large enough` to continue altering the percentage of passing corresponding
further erosion with its weak stream power. That to each grain size class during calibration in order
is why it was assumed that for the time being an to bring modeled erosion results closer to the
armor layer had been developed. So the dam did observed spillway erosion. The adopted
not breach as it was expected at the time of gradation curve is shown in Fig. 3 (d).
spillway excavation because of low flow rates
and channel self-stabilization through armoring.
It was formulated in order to restart the erosion
from the spillway. Scheme of the solution was
arranged in such a way that erosive power of
overtopping water could be enhanced so that
expected armor layer breaks up and erosion is
restarted. If the mouth of the spillway is plugged
with earth plug / stop log/bulkhead, the water
surface elevation would automatically start
increasing. After a reasonable time, the plugged
mouth of spillway be opened by removing the Fig. 3(a) Infill gradation curves (Sample 1)
earth plug by controlled blasting or any other
mean. This stored water will be routed and rush
through armored spillway channel and its high
erosive power should restart the erosion which
would ultimately lead towards breaching of that
landslide dam; this process was repeated fifteen
times until sufficient lake water lowering was
achieved. This hypothesized option was
implemented in the model with an inline structure
as check dam and hydrograph resulted by the
simultaneous breaching of this check dam in Fig. 3(b) Infill gradation curves (Sample 2)
Fig. 4(b) Results of erosion of spillway bed and
downstream shoulder of the dam, based on
Sample 2
Fig. 3(c) Infill gradation curves (Sample 3)
Fig. 4(c) Results of erosion of spillway bed and

Sample 3
Fig. 3(d) Synthesized gradation
Modeled results were much closer to actual

measured results of erosion (Fig. 4 (d)) and thus
the theoretical gradation curve was used for
further which enables the gradation curve to be
used for further analysis. Theoretical gradation
encompasses the landslide mass containing all Fig. 4(d) Synthesized gradation
grain size classes (from clay to
cobbles/boulders). Best efforts were made to train the modeling tool
with theoretical gradation. Spillway discharge
rate data was available up to the end of June 2010
but data regarding the depth of erosion was
available only for first nine days of June 2010. So
HEC RAS was precisely calibrated for
simulation within a short span of time. Setting up
initial and boundary conditions of HEC RAS
similar to the calibration portion, analysis of
breaching was extended up to the end of June
2010 and erosion results are shown in Fig. 5.
Fig. 4(a) Results of erosion of spillway bed and
Sample 1
Fig. 5 Results of erosion of spillway bed and

downstream shoulder of the dam up to known
spillway discharges
this hydrograph routing a 7.29 m degradation of
Table 1 Actual state of the erosion of Attabad spillway bed was experienced as shown in Fig. 6
landslide mass (d).
Measured Depths of Measured Depths of
Erosion Along Erosion Along 3rd Breaching
Spillway Reach Downstream After routing of 2nd hydrograph the spillway
Shoulder mouth was once again plugged with 15 m high
First 3 Further 6 First 3 Further 6 check dam in order to expedite the erosion of
Days days Days days landslide mass. In result of sudden breach of
1.2 m 0.75 m 21.5 m 3.2 check dam the hydrograph was generated shown
in Fig. 6 (e). After routing this hydrograph, the
In first three days, 21.5 m depth of erosion was
spillway bed got 3.37 m depth of erosion which
measured because all fines were present on upper
is depicted in Fig. 6 (f).
layers of landslide deposit while in next six days
only 3.2 m depth of erosion was experienced
4th Breaching
which shows that coarser grain sizes are exposing
After routing of 3rd hydrograph the spillway
(Table 1).
mouth was once again plugged with 15 m high
check dam in order to expedite the erosion of
Successive Case Testing
landslide mass. In result of sudden breach of
With the help of trained software, breaching
check dam the hydrograph was generated shown
analysis was conducted for two months with
in Fig. 6 (g). After routing this hydrograph the
successive case testing. On the basis of
spillway bed got 6 m depth of erosion which is
repeatedly constructing / plugging the spillway
depicted in Fig. 6 (h).
mouth, allowing the river flows to store and pond
up behind the plug and then later removing the
5th Breaching
plug to suddenly release the water stored behind
After routing of 4th hydrograph the spillway
the plug. Simulation results are shown in Fig. 6.
The simulation has shown that more than 70 m
depth of erosion could be lowered from crest of
successive erosion with the proposed breaching
plan. In result of sediment transport from
in Fig. 6 (i). After routing this hydrograph the
landslide deposit, the eroded material will be
spillway bed got 4.3 m depth of erosion which is
deposited on river bed d/s of landslide dam as
depicted in Fig. 6 (j).
depicted in simulation.
6th Breaching
1st Breaching
In result of erosion up to period of known flow
rates the crest of spillway was lowered 7.7 m
under natural condition. Due to non significant
progress in spillway degradation the mouth of
spillway in u/s side was plugged with 5 m high
in Fig. 6 (k). After routing this hydrograph, the
check dam. For restarting the erosion by breaking
up the presumed armor layer check dam was
depicted in Fig. 6 (l).
suddenly breached with controlled blasting and
hydrograph generated by impounded water was
7th Breaching
routed through spillway channel as shown in Fig.
6 (a). With this poundage very nominal depth of
erosion of 0.6 m was experienced as shown in
Fig. 6 (b).
2nd Breaching
in Fig. (m). After routing this hydrograph the
Due to very nominal depth of erosion in result of
first poundage with 5 m check dam, the spillway
depicted in Fig. 6 (n).
bed would mechanically excavated 7.5 m so that
spillway mouth may be plugged with 15 m high
check dam. Hydrograph generated with such
huge poundage is shown in Fig. 6 (c). In result of
8th Breaching check dam in order to expedite the erosion of
After routing of 7th hydrograph the spillway landslide mass.
Fig. 6 (k to t)
Fig. 6 (u to dd) Fig. 6: Flood hydrograph with instantaneous peak flow (on the left side) generated
from removal of check dam and subsequent lowering of spillway invert and d/s shoulder (on the right
side) of Attabad landslide dam (a) to (dd) exhibiting breaching scenario 1 to 15, respectively.
In result of sudden breach of check dam the in Fig. 6 (y). After routing this hydrograph, the
hydrograph was generated shown in Fig. 6 (o). spillway bed got 5.29 m depth of erosion which
After routing this hydrograph, the spillway bed is depicted in Fig. 6 (z).
got 3.5 m depth of erosion which is depicted in
Fig. 6 (p). 14th Breaching
9th Breaching mouth was once again plugged with 15 m high
After routing of 8th hydrograph the spillway check dam in order to expedite the erosion of
mouth was once again plugged with 15 m high landslide mass. In result of sudden breach of
check dam in order to expedite the erosion of check dam the hydrograph was generated shown
landslide mass. In result of sudden breach of in Fig. 6 (aa). After routing this hydrograph, the
check dam the hydrograph was generated shown spillway bed got 5.5 m depth of erosion which is
in Fig. Fig. 6 (q). After routing this hydrograph, depicted in Fig. 6 (bb).
the spillway bed got 2.3 m depth of erosion which
is depicted in Fig. 6 (r). 15th Breaching
10th Breaching mouth was once again plugged with 15 m high
After routing of 9th hydrograph the spillway check dam in order to expedite the erosion of
mouth was once again plugged with 15 m high landslide mass. In result of sudden breach of
check dam in order to expedite the erosion of check dam the hydrograph was generated shown
landslide mass. In result of sudden breach of in Fig. 6 (cc). After routing this hydrograph, the
check dam the hydrograph was generated shown spillway bed got 0.17 m depth of erosion which
in Fig. 6 (s). After routing this hydrograph, the is depicted in Fig. 6 (dd).
spillway bed got 5 m depth of erosion which is By the results of successive breaching
depicted in Fig. 6 (t). simulation, a total depth of erosion got by
spillway bed is 70 m within 2 months.
11th Breaching
After routing of 10th hydrograph the spillway From results of the model in the above Fig.s, it
mouth was once again plugged with 15 m high has been shown that in the start of case testing the
check dam in order to expedite the erosion of crest of dam gets more eroded than in later
landslide mass. In result of sudden breach of months. The small decrease in dam crest level is
check dam the hydrograph was generated shown experienced because as dam losses its height
in Fig. 6 (u). After routing this hydrograph the similarly lake storage behind the dam gets shorter
spillway bed got 2.1 m depth of erosion which is and thus breaching rate retards consequently.
depicted in Fig. 6 (v). It may be noted that complete description of
landslide infill material (in terms of particle
12th Breaching gradation, uniformity of fill over various location
After routing of 11th hydrograph the spillway and depth, geotechnical parameters of the fill
mouth was once again plugged with 15 m high material) was far from being completely
check dam in order to expedite the erosion of accurate. In this context, the results presented
landslide mass. In result of sudden breach of herein may be deficient. Further that the
check dam the hydrograph was generated shown proposition of constructing a plug at the mouth
in Fig. 6 (w). After routing this hydrograph the spillway, letting river water to store and rise
spillway bed got 4.3 m depth of erosion which is behind the plug and controlled breaching of the
depicted in Fig. 6 (x). plug may have many technical, safety, logistic
and social challenges and this should be adopted
13 Breaching after all due analysis and considerations.
mouth was once again plugged with 15 m high Impacts of Dam Breaching
check dam in order to expedite the erosion of Due to planned breaching of this landslide dam
landslide mass. In result of sudden breach of following impacts are likely to be observed:
1. As the dam got eroded its crest started collected and used in modeling analysis for more
lowering. In result of crest lowering the lake reliable and authentic results.
water impoundment level will decrease
which ultimately leads towards a decrease in References
lake length according to Fig. 7. Ahmed, M. F., Rogers, J. D., & Ismail, E. H.
(2017). “Correlations between Fluvial
Knickpoints and Recurrent Landslide Dams
along Upper Indus River”. Environmental &
Engineering Geoscience, 1078-7275.
Calligaris, C., Comi, M., Tariq, S., Bashir, F.,
Karim, D., and Khan, H. (2010). “Attabad
landslide survey in Hunza, Ev-K2-CNR
Committee”.
(http://www.mtnforum.org/rs/ol/counter_docdo
wn.cfm?fID=6241.pdf dated 10-oct-2010)
Chen, X., Cui, P., You, Y., Cheng, Z., Khan, A.,
Ye, C., & Zhang, S. (2017). “Dam-break risk
analysis of the Attabad landslide dam in
Fig. 7 Change in lake length with respect to Pakistan and emergency
dam height after successive breaching countermeasures”. Landslides, 14(2), 675-
683.
2. Breaching of Attabad Landslide Dam in this Copeland, R.R. and Thomas, W.A. (1989).
manner drains the inundating lake water in a “Corte Madera Creek Sedimentation Study.
safer mode. Consequently, structures and Numerical Model Investigation. US Army
drowned portion of KKH will once again Engineer Waterways Experiment Station,
expose so the danger of catastrophic flooding Vicksburg” MS. TR-HL-89-6
on downstream side will be minimized. Costa, J. E. and Schuster, R. L. (1988).
“Formation and failure of natural dams”.
Summary & Conclusions Geological Society of America Bulletin, Vol.
In this study flows of Hunza River were used for 100(7), 1054-1068.
simulation of lake levels behind Attabad (http://www.fsl.orst.edu/wpg/events/S04/Costa_
landslide dam and the likelihood of its failure 1988.pdf dated 10-oct-2010)
were assessed based on a synthesized infill Fort, M., Braucher, R., Bourlès, D., Guillou, V.,
gradation curve. Following conclusions are Nath Rimal, L., Gribenski, N., & Cossart, E.
drawn: (2014). “Geomorphic impacts, age and
1. Attabad landslide dam is stable without any significance of two giant landslide dams in
external intervention and dam is unlikely to the Nepal Himalayas: Ringmo-Phoksundo
breach under natural flow conditions. (Dolpo District) and Dhampu-Chhoya
2. Introducing repeatedly check dam and (Mustang District)”. In EGU General
subsequent blasting can effectively lower the Assembly Conference Abstracts (Vol. 16).
dam height up to 70 m. This will considerably Iqbal, M. J., Shah, F. H., Chaudhry, A. U. H., &
rescue the area from inundation effect. Baig, M. N. (2014). “Impacts of Attabad lake
3. The landslide dam of shortened height and a (Pakistan) and its future outlook”. European
short lake cannot be eliminated by hydraulic Scientific Journal, ESJ, 10(8).
erosion and may have to be accepted. Else NDMA daily report. (2010). “Update on Atta
mechanical excavation or controlled blasting Abad Lake 30th July 2010”.
has to be resorted for dam’s complete http://ndma.gov.pk/Documents/Hunza_Land
removal. slide_2010/One%20Pager%2030%20July.d
oc dated 10-oct-2010)
It is, thus, recommended that for safe drainage of Liu, A., Zheng, L., Deng, J., & Huang, Y. (2017).
Attabad lake check dam should be introduced and “Landslide susceptibility of the Xiangjiaba
breached subsequently for spillway channel bed Reservoir area associated with the Yaziba
erosion by instantaneous high discharges. This Fault”. Bulletin of Engineering Geology and
should be repeated many times until maximum the Environment, 1-11.
erosion is achieved. Additional gradation and
geotechnical data of landslide mass should be
Petley, D. N. (2010). “The Landslide at Attabad Himalayan Arc”. International Landslide
in Hunza, Gilgit/Baltistan, NDMA Centre, Durham University, United
Pakistan”. Version 1.01. (Unpublished) Kingdom
http://ndma.gov.pk/Documents/Hunza_Landslid Shah, F. H., Ali, A., & Baig, M. N. (2013).
e_Report/Report%20by%20 “Taming the monster–Attabad landslide
Proff.%20Dr.%20David%20N.%20Petley% dam”. Journal of Environmental Treatment
20on%20Hunza%20 Landslide.pdf dated 10- Techniques, 1(1), 46-55.
oct-2010) Stefanelli, C. T., Segoni, S., Casagli, N., &
Petley, D. N., Rosser, N. J., Karim, D., Wali, S., Catani, F. (2016). “Geomorphological
Ali, N., Nasab, N., and Shaban, K. (2011). analysis for landslide dams”.
“Non-seismic landslide hazards along the
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Design of Upstream Overflow Cofferdam of Patrind Hydropower Project

Rizwan Farid1*, Ijaz Ahmad2 , Rana Zain Nabi Khan2
1*
Pakistan Engineering Services (Pvt) Ltd. Lahore
2
*
[email protected]
Abstract: An overflow type of cofferdam has been proposed at 150 MW Patrind hydropower project for cost
saving of diversion work. The objectives of the study included the design of an overflow type upstream cofferdam;
estimate the quantity of spill water over the cofferdam during high flood, evaluating adequacy of low level outlet
and the determination of the downstream required protection work for the cofferdam for safety against overspill
flood flows. HEC RAS model was used for flood routing. A 16-m high rock fill cofferdam is proposed with
upstream membrane of clay and downstream fine filter. The downstream slope of the embankment is proposed to
be protected by a stepped chute of gabion along the slope of the embankment. Upstream slope of the embankment
is steeper i.e. 2H: 1V while slope of the downstream face is relatively flatter 3H: 1V and 10 m wide crest of
cofferdam is proposed. 1 m thick concrete slab is also provided on the crest of the cofferdam to prevent the
damage from overflowing water. During high flow season 181 cumecs water is overtopped over the cofferdam
with 2.1 m water depth above the cofferdam crest. The pondage is created between cofferdam and diversion weir
upto the elevation of 740.8 m amsl, so the main weir should be completed upto the elevation of 741 m amsl before
the high flow season. Seepage quantity i.e. 0.001 cumecs is estimated through the cofferdam. The cost of the
diversion works with non-overflow type cofferdam is 53% more than the overflow type cofferdam.
Keywords: Overflow, Cofferdam, Hydropower
Introduction Thus, diversion scheme need to be capable of

Construction of the hydro-power plant is diverting the peak flow of selected frequency
normally done in the river stream. For creating flood. Any higher flood is liable to overtop the
dry working conditions, the river flow is cofferdam leading to its complete failure /
temporarily diverted by construction of a destruction alongwith large damage to the under
diversion weir or diversion dam across the main construction works and delays in project
river course. Size and type of the diversion completion.
structure depends upon the flood characteristics The choice of diversion through cofferdam,
of the river and topographical conditions at the diversion channel or diversion tunnel also
diversion site. It can be a diversion tunnel if the depend upon the geological conditions of the
topography is favorable. A planed diversion area, if the geological conditions are not
scheme can minimize potential flood damage to favorable for off channel diversion, part of the
the work in progress and overall project main stream/river can also be used as a diversion
completion period. channel for phased construction.
Stream flow records provide dependable
information about stream flow characteristics. Research Methodology
Depending upon size of the drainage area and its The information relevant to hydrology,
geographical location, floods in a stream may be hydraulics, topographic survey and the design
from snowmelt, rain falling on snow, seasonal parameters required for the analysis of the
rainfall of cloudbursts. Each one of the types of problem in hand were taken from Feasibility
runoff have peak flows and their low flow period Study of Patrind Hydropower Project. Data
dependent on the climatology of the catchment includes the low and high flow diversion flood,
that influences the selection of diversion diversion tunnel rating for 7m diameter, Patrind
scheme. layout plan, Kunhar river x-sections at every
The 10 or 25-year frequency flood is generally 100-m interval and weir site survey of 1 m
selected as the design flood for diversion scheme contour interval. As the nature of this study is
based on the previous analysis or past experience analytic, accordingly a compact methodology
for sizing the diversion works and risk factor. was adopted to achieve the specific objectives.
crest elevation of the main weir is 765 m amsl.
Design of Cofferdam Three low level bottom outlets are also provided
For the design of overflow type cofferdam two in the main weir having 3 m diameter each and at
seasons low flow and high flow seasons were an invert level of 735 m amsl.
selected against the 10 years return period. In the second scenario only, the downstream
Adequacy of the diversion tunnel was checked geometry of the cofferdam is defined like a
during the low flow condition and the quantity of stepped gabion chute with 1 m drop and 3 m tread
the spill water over the cofferdam was estimated to dissipate the energy and cross sections are
against the high flow seasons. Velocities at taken at every start and end of the step. The
different elevation of cofferdam were also model was run for Qs and the minimum flow to
estimated and the adequacy of the bottom outlets getting the maximum velocities against these
was also determined against the spill water. For discharges. The Manning’s roughness
these purposes, the HEC – RAS Model was used. coefficients are changed at the downstream of the
cofferdam for continuous and stepped gabions.
Formulation of HEC-RAS Model The Manning’s “n” for main channel is taken as
The geometric data was prepared by assigning the 0.055.
the 6-km long reach of Kunhar River. In this
reach the diversion weir, cofferdam (Inline Selection of Stone Sizes
structure) and diversion tunnel is provided. The For the stability point of view, it is necessary to
height of the inline weir is 16 m; crest elevation estimate the rock size which is stable against the
of the weir is 750 m amsl and the bed at 734 m maximum velocity along the d/s slope of the
amsl. The diversion tunnel flow was simulated as cofferdam. For this purpose, Isbash method and
flow over a lateral weir. The main weir crest Isbash curves was used as a guide for the
elevation of the main weir is 765 m amsl is selection of stable rock sizes / gabion crate size.
modeled as a multi barrel culvert. Three low level
bottom outlets are provided in the diversion weir Calculation of Scour Depth and Stone Apron
having invert level of 735 m amsl and 3 m The scour depth at the downstream toe of the
diameter of each is incorporated as culvert cofferdam back slope is calculated by using the
barrels. Unsteady flow conditions were prepared Lacey’s formula as given below:
for low and high flow condition. In low and high 1
flow condition, the inflow hydrographs are taken

f s = 1.76d 2
(2.1)
1
as upstream boundary condition and the normal Q  3
depth is taken as downstream boundary R = 0.47  (2.2)
condition.  fs 
The unsteady model was run to check the Where
adequacy of the diversion tunnel i.e the R = scour depth in meters
cofferdam cannot be overtopped during low flow Q = flow in cumecs
condition (QL = 308 cumecs). The model was d = diameter of particle in mm
run for the high flow QH (i.e 1,034 cumecs) fs = silt factor
through the cofferdam (inline structure) and The diameter of the particle is calculated by using
diversion tunnel (lateral weir) for the estimation the Isbash Method against the maximum velocity
of the quantity of spill water (Qs) and depth of at the downstream toe of the cofferdam and by
the spillage over the cofferdam. using this diameter of the particle (d) in relation
2.1 the silt factor is calculated. By using relation
Estimate the Velocities and Check the 2.2 the scour depth is calculated against the Qs
Adequacy of Bottom Outlets and scour depth is obtained by multiplying with
In order to the get the velocities at every 1 m concentration factor 1.27 for straight reach.
height of the cofferdam the model was run under
for steady mode for Qs without lateral and inline Against the calculated scour depth launching
weir. For this purpose, the model was run with or stone apron is provided at the toe of the
without stepped chute gabion. cofferdam. The length of the apron is taken as 1.5
In the first scenario the u/s and d/s slopes of the times of the depth of the maximum scour.
cofferdam were taken as 2H: 1V and 3H:1V The volume of the stone per foot run of an apron
respectively and x-sections are taken at 2 m would be calculated as;
interval. In this model the main weir/dam was
5 xRx2.0 = 4.5R
taken as the bridge and culvert at R.D 11+010 the
Where PATRIND HYDROPOWER PROJECT
Kunhar River Main
Plan: Plan 33 6/11/2008
Legend
D is the scour depth in ft. EG Max.WS
WS Max.W S
Thickness of the stone apron is calculated by 750
Crit Max.WS
Ground
using the following relation;
T = Vs /W
Elevation (m)
745
Where
Vs = volume of stone per unit length (ft2) 740
W = length of the stone apron (ft)

T = thickness of stone (ft) 735
1320 1340 1360 1380 1400 1420

Results and Discussion Main Channel Distance (m)
This section describes the results of Hydraulic Fig. 1 Water Surface Profile at Cofferdam (with
analysis and the design of the cofferdam. The Continuous Gabion)
hydraulic analysis includes the result of HEC
RAS Model for steady and unsteady flow 754 10.0
conditions, estimation of rock sizes which is 752 9.0
stable against the highest velocity and 750 8.0

Velocity Profile
downstream protection work against scouring. 748 Cofferdam 7.0
ELEVATION (M AMSL)
W/S Profile
VELOCITY (M/SEC)
HEC-RAS Model was used to check the 746 6.0
adequacy of the diversion tunnel during low flow 744 5.0
742 4.0
(QL) conditions and the simulation proves that 740 3.0
the water will not overtop the cofferdam and the 738 2.0
diversion tunnel is adequate to pass QL. 736 1.0
HEC–RAS Model is run against the QH to 734

100 90 80 70 60 50 40 30 20 10 0
0.0
estimate the quantity of spill water over the DISTANCE (M)
cofferdam having crest elevation 750 m amsl. Fig. 2 Velocity and Water Surface Profile at
The result indicates that the 181 cumecs flood Cofferdam (with Continuous Gabion)
water will spill over the cofferdam with
maximum water elevation of 752.1 m amsl In order to further reduce the velocities at
during the high flow season. The 853 m3/sec will downstream of the cofferdam the model is run
be diverted through the diversion tunnel due to with downstream stepped gabion. In this
increased water level. simulation only, the downstream geometry of the
For the design of the cofferdam velocities are cofferdam is replaced by the stepped chute
needed along u/s, top and d/s slopes. For this gabion to dissipate the energy. The water surface
purpose, two simulations are run. In first profile at the location of the cofferdam is shown
simulation, the river x-sections are modified to in the Fig. . The velocity and water surface
take the shape along the slope of the cofferdam profiles at the cofferdam are shown in Fig. .
with continuous gabion at the downstream. The The analysis shows that after the stepped chute
model was run for Qs and minimum flow (1 gabion the velocities are reduced at the
cumecs) without lateral and inline structure. The downstream of the cofferdam. The maximum
maximum velocities are obtained against Qs (181 7.95 m/sec velocity is reached at the river station
cumecs). The water surface profile at the location 15.023 on the downstream side of the cofferdam.
of the cofferdam is shown in the Fig. . Velocity The downstream slope of the cofferdam is
and water surface profiles are shown in the Fig. protected against this velocity.
at the upstream and downstream of the PATRIND HYDROPOWER PROJECT
Kunhar River Main
Plan: Plan 32 6/11/2008
Legend
cofferdam. The analysis shows the maximum EG Max.WS
750 WS Max.W S
velocity 8.53 m/sec is reached at the river station Crit Max.WS
Ground
15.023 on the downstream side of the cofferdam.

745
Elevation (m)
The water poundage between cofferdam and the

weir is upto the elevation of 740.8 m amsl which 740
is 7.8 m above the river bed. The water level is
2.8 m above the overt of outlet (738 m amsl) so 735
the main weir should be completed upto the

elevation of 741 m before the high flow season. 1320 1340 1360
Main Channel Distance (m)

1380 1400
Fig. 3 Water Surface Profile at Cofferdam (with

Stepped Gabion)
the crest level (3 m high) of the cofferdam to
754 9.0 prevent damage from overflowing water.
752 8.0 The construction of the cofferdam is set in three
750
Velocity Profile
Cofferdam
7.0
stages during low flow period. In first stage the
748
6.0
excavation of the foundation and filling will be
ELEVATION (M AMSL)
W/S Profile
VELOCITY (M/SEC)
746
744
5.0
completed. In second stage the fill material
4.0
742 (rocks) will be dumped and compacted along the
3.0
740
2.0
upstream and downstream slope of the cofferdam
738
1.0
and in the final stage the fine material (clay) will
736
734 0.0
be dumped over u/s face to reduce the seepage.
100 90 80 70 60 50 40 30 20 10 0
DISTANCE (M)
Cofferdam Design Consideration

Fig. 4 Velocity and Water Surface Profile at
Stepped chute gabions of minimum size 1 m
Cofferdam (with Stepped Gabion)
thick and 3 m long are proposed to be provided
as protection measure for the downstream slope.
Slope Protection
These slope is reinforced by a 1 m x 1 m grid of
As the velocities on the downstream side of the
1/2-inch (# 4) steel bars, anchored backed 3 m
cofferdam dam are very high and need to be
into the rockfill embankment. Gabions are also
protected by providing suitable sized
provided at the place where the hydraulic jump is
stone/gabion. Stone sizes against these velocities
formed.
are calculated by using the Isbash Method. The
maximum 2.5 m size stone is required against the
high velocity of 7.95 m/sec. The maximum size
of the stone is 2.5 m which is very large and it is
very difficult to obtain and place on the site. So,
the gabions of minimum 1 m high and 3 m wide
are suitable at the downstream side of the
cofferdam against this high velocity. The chute
type gabions tied with ½ inch (# 4) bars are
provided to dissipate the energy. The detail of the
gabion and the downstream chute are provided in
the design of the cofferdam.
Cofferdam Design Consideration

A rockfill dam is designed for overflow type of
cofferdam and utilization of the fill materials
available in the vicinity of the weir site. It is the
rock fill dam with upstream membrane of clay.
At the upstream side, a layer of clay is provided
to prevent from seepage and ease of construction
during flowing water. The downstream slope of
the embankment is proposed to be protected by a
stepped chute of gabion and 1 m thick fine filter
is provided beneath the gabion to protect against
washout of the cofferdam fill material. The Fig. 5 Proposed Cross Section of Cofferdam
foundation seepage is controlled by providing a
6-m deep positive cutoff wall beneath the upper Downstream Floor Scour Protection
membrane of the rockfill cofferdam. Proposed x- The downstream floor is also required to be
section of the cofferdam is shown in Fig. . protected against the scouring. The hydraulic
Upstream slope of the embankment is steeper i.e. jump is formed at the 10 m above the toe of the
2H:1V while slope of the downstream face is cofferdam as shown in the Fig. 4.10 where
relatively flatter 3H:1V as the stability gabion is already provided. So, a riprap in the
requirements against very high velocities and 10 form of the launching apron is provided against
m wide crest of cofferdam is proposed. A 1 m the scour depth and the layer of geo membrane is
thick concrete slab is proposed on the crest and also provided behind the launching apron to
on the upstream slope from elevation 747 m upto protect from seepage as shown in the Fig. . The
size of the stone was calculated by using the
Isbash method against the maximum velocity of Seepage Analysis
2.47 m/sec at the toe of the cofferdam. The water losses due to seepage through the
V2 cofferdam fill, foundation and abutments are to
D50 = be kept minimum from the point of view of
 −w 
2 gC 2  s  conservation of water and control of seepage
 w  gradients. Detail seepage analysis has been
V = 2.47 m/sec (V max) carried out by the Darcy’s law, in order to control
V = 8.10 ft/sec the exit gradients and limit the overall
γs = 165 lb/ft3 permissible seepage to a safe and acceptable
γw = 62.4 lb/ft3 level. Details of Seepage analysis is given below.
c = 0.86 (for high turbulence) Q = KIA
D50 = 0.84 ft say 1 ft I = Δh/L
Determine the silt factor by using Lacey’s L = 55 m
Formula against the 4.6 mm diameter of the stone Δh = 17.5 m
1 I = 17.5/55 = 0.318
f s = 1.76d 2
K = 1x10^-5 m/sec
Fs = 3.05 A = 8 m2
Scour depth is calculating against the silt factor Q = (1x 10-5) x (0.318) x (8) x (40)
3.05 and discharge 181 cumecs (6391 cusecs) by Q = 0.001 m3/sec
using the following relation. The seepage through the rockfill cofferdam is
1 very less as compared to the overspill water so
Q  3
R = 0.47  the quantity of the seepage is neglected.
 fs 
Cost Estimate
R = 6.1 ft say 2 m
Cost of the overflow type cofferdam and non-
Scour depth is calculated by multiplying R with
overflow type cofferdam is the important
the concentration factor 1.27 for straight reach
parameters of the study. For cost estimate the unit
D = 1.27 R
rates as given in the BOQ of the Feasibility Study
D = 2.6 m
were applied.
Stone apron is designed against the maximum
scour depth 2.6 m.
Overflow Type Cofferdam Cost
Length of apron (W) = 1.5 x D
The overflow type cofferdam is 16 m high with
= 3.9 m say 4 m
2;1 and 3;1 upstream and downstream slope
The downstream slope of the cofferdam is 3:1 so
respectively. The quantities and the cost of 16 m
the volume of the stone per foot run of an apron
high overflow type cofferdam is given in the
is calculated by using slope is given as;
Table .
10 xRx2.0 = 6.3R Table 1 Cost Estimate of Overflow Cofferdam
The maximum scour depth is 2.6 m and the Unit
Amou
Quanti Rate
volume of stone per running meter of an apron is Material Units
ty Rs/Un
nt
7.6 m2. (Rs)
it
Thickness of apron = Vs /W =2m Excavation of
CUM 1020 171.8 175236
Trench
Embankment fill, 755044

CUM 15995 472.05
Stone filling 0
199438
Clay CUM 3828 521
8
Fine Filter CUM 1495 345 515775
Gabion Fill,
CUM 1600 600 960000
Boulders
Welded Steel
TONNE
Mesh: Deformed 5.19 85,000 440725
Fig. 6 Detail of Launching Apron Bars G-60
S
127300
Concrete CUM 380 3,350
0
Stone Apron CUM 228 532.05 121307
• Launching apron 2 m deep and 4 m long is
provided at the toe of the cofferdam to protect
Total Amount (Million Rs.) 13.03 against scouring. The scour depth is
calculated against 181 cumecs flow and 2.47
Non-Overflow Type Cofferdam Cost m/sec velocity at the toe of the cofferdam.
From the rating curve of the diversion tunnel the • The poundage is created between cofferdam
QH (1034 cumecs) is passed through the and diversion weir upto the elevation of 740.8
diversion tunnel at an elevation of 753 m amsl. m amsl (7.8 m above the bed level). The water
So, the crest of the non-overflow type cofferdam level is 2.8 m above the overt of outlet (738 m
is 755 m amsl with 1 m freeboard (i.e 21 m high). amsl), so the main weir should be completed
The upstream and downstream slopes of the both upto the elevation of 741 m before the high
overflow and non-overflow cofferdams are same. flow season.
The quantities and cost of the non-overflow type • 0.001 m3/sec seepage quantity is estimated
cofferdam is given in the Table through the cofferdam, which is very less as
From the Table and Table it is clear that the cost compared to overspill quantity, so it is
of the cofferdam increases with the increase in neglected.
the height of the cofferdam. The non-overflow • The cost of the non-overflow type cofferdam
cofferdam is approximately 46 % expensive than is 46 % more than the cost of the overflow
the overflow cofferdam. type cofferdam
Table 2 Cost Estimate of Non-Overflow Cofferdam References

Unit
Amoun USBR. (1987), “Design of Small Dams”, Water
Uni Quanti Rate
Material
ts ty Rs/U
t Resources Technical Publication, Third Edition,
(Rs) SSOP, Washington, D.C, p 491-496
nit
Excavation of CU Pakistan Engineering Services (Pvt) Ltd. (2007),
1020 171.8 175236 “Feasibility Study of Patrind Hydropower
Trench M
Embankment CU 472.0 161629 Project”, Main Report, Vol -1, DHA, Lahore,
34240 Pakistan.
fill, Stone filling M 5 92
Placement of CU 269878 HEC. (2001), “Technical Reference Manual of
5180 521 HEC-RAS” Version 3.0, US Army Corps of
Clay M 0
Engineers, USA.
Total Amount (Million Rs.) 19.04 Chaudhry, M. Hanif. (1993), “Open Channel
Flow”, Washington state University,
Conclusions Washington, p 246-247.
• A 16 m high rockfill cofferdam with upstream Casagrande. A. (1973), “Embankment Dam
membrane of clay is proposed, the upstream Engineering” Jhon Wiley and Sons, Inc., USA, p
and downstream slopes of the cofferdam is 2:1 97-103.
and 3:1 respectively and 1 m thick reinforced Tariq, Ata-ur-Rehman, (2006), “Dam and
concrete slab provided on the upstream face Reservoir Engineering”, Class Notes, Center of
from elevation 747 to 750 m amsl (3 m) and Excellence in Water Resources Engineering,
on the crest of the cofferdam to prevent UET, Lahore, Pakistan, p. 21-22.
damage from overflowing water. Novak, P, A.I.B. Moffat and C. Nulluri, (1990),
• During high flow season the 181 cumecs (Qs) “Hydraulic structures” University of Newcastle
is overtopped over the cofferdam with upon Tyne, Chapman and hall, London, p. 301-
maximum water flow depth of 2.1 m above 302.
the cofferdam. Fell, et al. (1992), “Geotechnical Engineering of
• The maximum velocity of 7.95 m/sec velocity Embankment Dams”, A.A. Bakema, Rottedam,
is obtained against Qs. The stepped chute Netherlands, p 450-455.
gabions of 1 m high and 3 m long are provided Yanmaz, A. M., (2000), “Overtopping Risk
at the downstream slope of the cofferdam to Assessment in River Diversion Facility Design”,
protect against the high velocities Canadian Journal of Civil Engineering, Vol. 27,
No:2, PP. 319-326.
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Evaluation of Small Hydropower Scheme Operating in Chitral, Pakistan; a

Case Study of Chitral Hydel Station on Lutkho River
Engr. Javed Zulfiqar1*, Engr. Nazakat Hussain1
1*
Water and Power Development Authority, Lahore, Pakistan.
Abstract: This research presents the optimization with different alternatives of existing Chitral one megawatt
(1 MW) hydropower station at Lutkho River through the hydrological analysis. The study involves the estimation
of synthetic flows of study area by correlation with having same Climatological and hydrological characteristics,
maximum design flood using Log Pearson Type III for various return periods, proposed a permanent concrete
weir structure, layout alternate with existing power channel and with extended power channel. In order to
estimation of power generation flow duration curve was carried out through derived flows. The analyses revealed
that with the permanent concrete weir structure and existing channel and with extended channel the energy can
enhance up to 1.8 MW and 5 MW.
Keywords: Hydropower potential, Runoff river, Chitral
Introduction average river gradient is 3.1%. [HPO Publication

Nature has been very generous to the Northern No. 271, 2009]
Pakistan in its endowments of towering mountain The Existing scheme of power generation at
peaks, gigantic glaciers, majestic rivers and Chitral was designed to utilize the portion of
splendid green vales and meadows with abundant flows from Lutkho River. There is no permanent
fruit trees. The steep streams and rivers with weir for diversion of discharge from Lutkho
abundant flows offer cheap and environmentally River to power house is available. Manual
friendly power potential to be harnessed in these dumping of boulders as weir is done in the lean
mountainous areas. Chitral is situated in the water months and this process is repeated every
extreme Northern part of Pakistan, in Khyber year and the temporary weir flushed away
Pakhtun Khua province. The climate of the particularly in the summer session due to high
Chitral city is very cold snowy winter and flows in Lutkho river. The main objective of this
relatively pleasant summer. Most of the research is to optimize the hydro power potential
watershed remains covered with snow and of power station at Lutkho River.
glaciers in winter season. The flow in the Lutkho
River is mainly due to glacier and snow melting.
Chitral valley has a number of potential streams
where hydropower potential schemes were
identified. A few of them have been planned and
are being considered for implementation.
The proposed Chitral weir site has been located
on the river Lutkho about 1 km downstream of
Chitral village at a latitude of 35o 56’ 14” and
longitude of 71o 48’ 37”. The proposed weir site
was about 10 km from Chitral town towards
North. Lutkho River is a right tributary of Chitral
River. Its proposed powerhouse site is located on
the right bank of the river Lutkho, about 5 km
upstream of Chitral town. The drainage area of
the Chitral weir site on the Lutkho River is about
2340 km2. The mean elevation of the catchment
area is 4009.5 m.a.s.l. The drainage area of Fig. 1 Existing view of temporary stone weir
Chitral River at Chitral is about 11,400 km2. The
length of Lutkho River is about 72 km and the
Hydropower Overview of Chitral Valley with some long term gauging station to prepare a
In Chitral and surrounding areas, most of the flow series of at least 30 years for reliable
power generation, transmission and distribution estimation of power and energy. The discharge
system is being maintained and operated by data of the nearby hydrological stations Chitral,
WAPDA. There is one existing hydel power having long term record (1964 to 2010) was
station 1 MW in addition to power supply from collected and the synthetic flows have been
National grid. The Chitral city was connected to derived. Ten daily flows at the proposed weir site
national grid through a 33-kV transmission line. have been estimated by using different
The voltage level through long line transmission approaches utilizing the available flow data
is low. Either Chitral valley is to be connected record of station with the similar catchment
with high voltage transmission or medium size characteristics. For generation of the temporal
power plant is connected to Chitral city to have distribution of the daily, 10-daily, monthly and
reliable power supply. In Garam Chashma, hydel annual flows the discharge data of Shah-re-Sham
power station of 100 kW is being operated and Stream Gauging Station for one year, from June
maintained by SHYDO (WAPDA). Another 1987 to June 1988 in combination with Chitral
hydel power station of installed capacity 4.2 MW Gauging Station Data from 1964 to 1980 and
Reshun is in operated by SHYDO to provide 1982 to 2010 (46 years) have been used. The
electric power to Upper Chitral areas. In lower annual model used to generate daily flows at weir
Chitral, the electric power generation, site and power house sites, as given below and
transmission and distribution system is being the mean monthly flows is shown in Fig.: 2.
operated by SHYDO. A hydel power station of
300 kW installed capacity at Shishi Gol supplies 𝑸𝒘𝒆𝒊𝒓 = [𝟎. 𝟐𝟓𝟐𝟔(𝑸𝒄𝒉𝒊𝒕𝒓𝒂𝒍 )𝟎.𝟖𝟖𝟖 ]
power to Drosh area. Chitral hydel Power Station
1 MW has been in operation since 1975,
providing electric power to Chitral town.
Optimization of Chitral Hydropower Station

To exploit the available indigenous resources of
Lutkho River, the capacity enhancement of
Chitral power station is direly needed. The Fig. 2 Mean Monthly Flows of Lutkho River
existing structures would be remodeled for
enhanced capacity. Construction of new For the flow study of the weir and powerhouse
permanent weir instead of old stone weir and sites by regional analysis approach, flow data of
remodeling of power channel may enhance the seven stream gauging stations of the Chitral
power generation capacity to meet the growing watershed were selected having similar
demand of Chitral city. The topographic features Climatological and hydrological characteristics.
and availability of flows in river allow increasing These seven stream gauging stations are Chitral,
design discharge by remodeling the hydraulic Reshun Gol, Golen Gol, Shishi Gol, Ojhor Gol,
structures and more power can be generated Phander and Shah-Re-Sham (Lutkho River). The
rather than installed capacity i.e. 1 MW. The graphical relationship between average flows and
enhanced capacity would be exclusively used to watershed area is shown Fig. 3. Fig. shows that
meet the growing power demand of Chitral city. there is increasing trend in the average flows with
the increase in watershed area.
Methodology Adopted After transforming the daily flow data at the weir
Data Availability& Data Analysis site, daily, 10-daily, monthly and annual flows
In order to carry out the proposed study, the were computed at the weir site. Power and energy
relevant data viz. discharge data, climatological have been estimated for discharges of 5.60 m3/s
data (temperature, precipitation, and humidity), and 15.8 m3/s. Power and energy outputs have
topographic survey, and geological data were been calculated for a period of forty six years
collected from Water and Power Development using 10 daily flow records.
Authority (WAPDA) and processed by using
spread sheets where needed. However, the Estimation of Power and Energy
studies were purely theoretical and no Power and energy have been estimated for
experimental work was undertaken. On Lutkho discharges for the design discharge of 4.60 m3/s
river, discharge observations were performed and 15.8 m3/s. Power and energy outputs have
from 1987-88. So it is necessary to correlate it been calculated for a period of forty-six years
using 10 daily flow records. The design capacity sites at same locations, an option has been studied
is computed with following formula; P= to make use existing channel structure. With the
*g*Q*H/1000 Whereas P is Capacity (MW) construction of new permanent weir, the water
estimated from 10 days discharge and level at channel would be raised and discharge
corresponding head, Q is the 10 days discharge capacity increased 4.64 m3/s to 5.60 m3/s, which
with maximum value of design discharge Qd provide an installed capacity of 1.8 MW. The rise
(m3/s), H is Gross head (m). It is the difference of in water level is to be maintained within the limits
elevation between operating head and tail water of all existing hydraulic structures, accordingly
level, is combined efficiencies of turbine, the weir height is fixed to 3m high.
generator, transformer and that of the hydraulic
system upstream of powerhouse, g is gravity Salient Features: Alternate-I
acceleration = 9.81 m/s2. The mean annual energy Design capacity (P) =1.80 MW
is estimated on the basis of 10 daily flows using Design discharge(Q100) =5.60 m3/sec
the following formula, E =*g*Q*H*t/106 Gross head (H)=37.8 m
Whereas E is mean annual energy in Mean annual energy (E)=15.72 GWh
GWh/annum, Q is average design discharge Plant Factor = 100.0%
(m3/s), t is time in hours over 10 days. The plant Headrace channel length = 3675 m
factor which provides the basis for the installed Energy gradient in channel = 1:1000
capacity of the project is calculated as follows: pf Penstock length = 55m
= (E/(P*8.760))*100 Whereas pf = Plant factor Tailrace length = 30 m
(%), E is mean annual energy (GWh/annum), P No. and type of units = 2, Francis
is Installed Capacity (MW).
Alternate Ii: Project Layout with Extended
Layout Planning Channel
The various alternatives for the project were The channel section can be widened to certain
studied before the final selection. The final layout limits which the present topographic features of
was selected based on the technical aspects, the the area allow. The buildings, graveyards, road
maximum possible head is exploited, the alignment, river steep slope, weak and hard
waterways have the shortest possible length and formation restrict the channel section to certain
the most suitable structures are selected for dimensions. A rectangular concrete channel
prevailing geologic conditions. The topographic section 4.5 m wide and 2.4 m high can be
survey sheets were used to evaluate a suitable constructed and the capacity of canal can be
layout and to place the various components of the increased to 16.0 m3/s, which provide an installed
project. The existing power project was planned capacity of 5 MW.
and designed with limited capacity channel and
without permanent diversion structure about 40 Salient Features: Alternate-II
years ago. The topographic conditions from Design capacity (P) = 5.0 MW
existing intake to powerhouse area allow Design discharge (Q75) = 16.0 m3/sec
capacity of power channel and surface Gross head (H) = 37.8 m
powerhouse to be enhanced upto certain limits. Mean annual energy (E) = 42.70 GWh
The weir axis within 50 m of existing intake has Plant Factor = 97.50%
been considered where geological conditions are Headrace channel length = 3675 m
very good. With weir intake at same area and Energy gradient in channel = 1:1000
alternate location of powerhouse, the following Velocity in channel = 1.95 m/s
two alternate project layouts have been studied. Penstock length = 70 m
Alternate I: Project layout with existing Tailrace length = 22 m
channel No. and type of units = 2, Francis
Alternate II: Project Layout with extended
channel Advantages of Alternate Layouts
• Alternate-I with power channel can be
Alternate I: Project Layout with Existing enhanced to utilize the low flows with same
Channel gross head.
The existing channel and power station can be • Alternate-I would utilize the same project
upgraded with permanent weir and by diverting area as that of existing. Some area would be
increased flows towards power channel and required to be acquired where realignment of
powerhouse. With weir axis and powerhouse power channel may be required.
• Alternate II provides twice the energy (42.70 The flood study is essential for the design of
GWh) of Alternate I (15.72 GWh) various hydraulic structures at the weir and
• The weir height is 3.0 m. The geological powerhouse sites including coffer dams and
conditions for small size structure of weir, spillways etc, has been carried out. The flood has
headrace channel and powerhouse are very been estimated at weir and powerhouse sites by
good. These are more favorable for small and using flood frequency methods i.e. Gumbel, Log
surface structures. Pearson (Type-III). Maximum daily flow in a
year was converted into instantaneous peak to
Results and Discussion estimate the flood at weir and powerhouse site.
Design Discharge Log Pearson Type III has been adopted for Weir
A diversion concrete weir structure of 3 m high and Powerhouse sites due to best fit distribution
and 26 m long on the Lutkho River is to be of observed data. The flood frequency results for
designed that it will divert a flow of 16.0 m3/s to weir site are also shown in tabular form in Table
headrace channel. Two units of Frances Turbine 2 and graphical representation of floods in Fig. 5.
for the powerhouse have been proposed,
therefore, the design discharge for each unit will
be 8.0 m3/s.
Table 1 Summary of Comparison of

Alternatives
Alternat
Existin Alternat
e-II
Alternat g e-I
(Extende
e Scenari (Existing
d
o Channel)
Channel)
Discharg
4.64 5.60 16.0
e (m3/s)
Net Fig. 4 Flow duration curve at weir site of
Head 35.7 37.8 37.8 Lutkho River
(m)
Capacity
1.0 1.8 5.0
(MW)
Energy
8.74 15.72 42.1
(GWh)
Plant
factor 100 100 97.0
(%)
Weir
Stone
height 3.0 3.0
weir
(m)
Headrac
3675 3675 3675
e (m)
Flow Duration Curve Fig. 5 10,000 years return period floods in the
Flow duration curves of Lutkho River have been region
prepared using the rank-ordered technique for
the years 1964 to 2010 to assess the availability Table 2 Floods at proposed weir site by flood
of flows. Flow duration curve based on the frequency analysis
generated average daily flow data at the weir site
is shown in Fig.: 4. The maximum flow rate is Return Flood (m3/sec)
195 m3/s, minimum, which will remain available Period Gumbel Log
throughout the 100% 75% and 50% exceedance (years)
2 distribution
238 Pearson
239
time is 9.5 m3/s 16 m3/s and 23.39m3/s 2.33 245 Type-III
245
respectively. 5 272 270
10 294 290
20 315 307 off-peak hours, the volume of water available
25 322 313 for storage and inflow volume in 24 hours is
50 343 329 used to calculate the corresponding
discharge. The power during peak and off-
100 364 345
peak hour is used to calculate the
200 384 361 corresponding energy.
500 411 382
1000 432 397
10000 500 449
Basic Assumptions in the Calculations

The following basic assumptions have been made
in estimating power and energy of the identified
schemes.
• The residual water in the Lutkho River Fig. 6 Mean Monthly Power
during low flow period has been taken as 1.0
m3/s. It is based on sediment flushing The recommended design discharge is 16.0 m3/s,
requirement and minimum flows required in for the net head of 37.2 m the power and energy
the river for environmental purpose. have been estimated on the basis of 10-day flows
• The variations in the efficiencies of for all years from 1964 to 2010, the optimum
generating units due to change of net heads installed capacity would be 5.0 MW with average
have not been considered. For design mean annual energy of 42.70 GWh. The plant
discharge available to powerhouse, an factor has been taken as 97.50%. The firm
efficiency of 91.0%, 96.0% and 98% have capacity available 95% time of the year is 4.0
been used for the turbine, generator and MW. The calculated mean monthly power and
transformers, respectively. annual energy are presented in Fig. 6 and Fig. 7
• The head loss has been considered as 0.35 m respectively.
in penstock, 0.10 m at intake and 0.05 m in
tailrace. The losses at bend and in draft tube
are taken as 0.05 m. The total losses are taken
as 0.55 m to estimate net head for turbine.
The losses would be less for a discharge less
than the design.
• Energy is calculated by multiplying the
number of days in a month and 24 hours in a
day. For power and energy during peak and Fig. 7 Mean annually energy
Conclusions
Diversion Weir and Headrace Channel Due to limitation in extension of power channel,
As per topographic map, the bed level at weir the design discharge has been proposed as 16.0
axis is 1531.5 m and Lutkho River flows with m3/s, available 75% time of the year. The
mild gradient of about 1 %. To divert the design capacity of existing channel section can be
discharge of 16.0 m3/s into the power channel, a improved with lining in its entire length. With
concrete weir of 3.0 m height above the nullah construction of diversion weir, some remedial
bed has been proposed. A rectangular R.C.C works at hydraulic structure, the capacity can be
channel has been proposed on the right bank of enhanced to 1.8 MW. This option is not attractive
Lutkho River. The channel would be designed as as diversion weir draw additional flows for power
rectangular R.C.C section with width of 4.5 m channel. The alternate project layout has been
and height 2.4 m with free flow condition. The considered for optimum utilization of available
length of headrace channel has been estimated as potential. As power channel has limitation for the
3675 m. It has design capacity of 16.0 m3/s and design discharge to be conveyed through channel
maximum velocity upto 1.95 m/s. The alignment upto powerhouse, therefore capacity
longitudinal slope of channel has been taken as has been restricted to 5 MW. Two Horizontal
1:1000. Francis units each of 2.5 MW has been proposed
to generate 5.0 MW. As the design flow is
available for 75% time of the year and firm
capacity is 4.0 MW, therefore the proposed units
would be capable of generating the power round analysis of sediment carried by the Lutkho River
the year. As the project would be developed using should be carried out as is important for the
the land of existing structures, therefore social layout and design of the sand trap.
and environmental issues have been assessed
minimum. References
GTZ-WAPDA. 1999. Layout and Sizing; High
Recommendations Head Hydropower.
The project area has a history of generating GTZ-WAPDA. 2000. Hydropower in General,
power of 1 MW. The hydrological, topographical High Head and Low Head Hydropower
and geological condition favors the extension of Projects.
structure and capacity enhancement for 5 MW, GTZ-WAPDA. 2000. Power Plant Components;
with high plant factor power station to meet the High Head Hydropower.
power requirement of remote area. The existing HEPO-WAPDA. 2009. Reconscience Report of
powerhouse structure is old and too small for new Chitral Hydel Station.
units. It has to be dismantled and new building HEPO-WAPDA. 2011. Updated Feasibility
would be constructed at the same place. The Study Harpo Hydropower Project.
remodeling of existing hydraulic structures is HEPO-WAPDA. 2012. Updated Feasibility
essentials as these are about 40 years old. Study Basho Hydropower Project.
Detailed investigations are required along the HEPO-WAPDA. 2013. Plans for Development
power channel to adopt the slop stability of Hydropower in Pakistan.
measures for the safety of power house also
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Impact of Baglihar Hydroelectric Power Plant on Design Flood of Pakistan

Adnan Majeed1*, Rana Zain Nabi Khan2, Neelam Pari1, Ahsan Naseem1
1
Civil Engineering Department, University of South Asia, Lahore, Pakistan
2
*
[email protected]
Abstract: Water Resource is a live and sensitive issue in Pakistan, both internally between its provinces and
externally with India, Pakistan after separation from India the Indus Waters Treaty was signed in 1960 between
India and Pakistan but now India is trying to scrap this treaty by constructing Baglihar Hydroelectric Power
Plant on Chenab river, that is the allocated river of Pakistan. In this study Flood frequency analysis approach
(Gumbel’s Distribution) is used to calculate the return periods and decrease in water due to the construction of
Baglihar Hydroelectric Plant. Different return periods are calculated at different four gauging stations (Marala,
Kahanki, Qaidirabad, and Trimmu) at Chenab river. Gumbel’s distribution shows that at 1000- year return
period Chenab rivers facing huge decrease in flow. At Marala , Khanki, Qadirabad and at Trimmu the percentage
reduction are 11.33%, 28.91%, 30.6% and 40% respectively. The average reduction in flows occurs which is
almost 27.71 % at Chenab river, which will not only cause energy losses but also serious environmental damage
on Pakistan. The percentage reduction in flow at different return periods can be further analyzed by using other
flood frequency approaches.
Keywords: Baglihar Hydroelectric Power Plant, Indus water Treaty, Gumbel’s Distribution
Introduction exceeding 1.25 MAF and use the rivers for run of
The Indus river system, one of the major systems river hydro power generation with storage not
in the world, comprises the main Indus and its exceeding 1.6 MAF and nominal flood storage
five important tributaries: Jhelum, Chenab, Ravi, capacity of 0.75 MAF. (Raina, 2017). Pakistan
Beas and Satluj rivers. Indus river, along with has become one of the driest countries in the
Jhelum and Chenab, are referred to as the world because of the recent shortages of water.
'Western rivers', while the other three tributaries Pakistan criticizes Indian hydroelectric projects
(Ravi, Beas and Satluj) are termed as 'Eastern like Ratle,, Baglihar o n Chenab river and
rivers'. The common features of all of them is that Kishanganga hydroelectric plant on Neelam-
they originate in the Himalayas, pass through Jhelum river which have enabled India either to
Indian territory and, after Indus has received the reduce water flows to Pakistan or to release
waters of its five tributaries in Pakistan, it flows store waters and cause floods. According to
through Pakistan provinces of Punjab and Sind Sahai. M (2006) “The maximum pondage in the
and falls in the Arabian Sea, South of Karachi. operating pool shall not exceed twice the
The Indus Waters Treaty (IWT) was signed in pondage required for firm power as mentioned in
Karachi on September 19, 1960, by India’s Prime IWT. However, the design shall take due account
Minister Jawaharlal Nehru, Pakistan’s President of the requirements of Surcharge storage and of
Muhammad Ayub Khan, (Iftikhar, 2011). secondary power. Pakistan feels that the planned
According to this treaty, Ravi, Beas and Sutlej, pondage of 37.7 MCM in Baglihar project is
which constitute the eastern rivers, are allocated violation of the Treaty provisions. India denies
for exclusive use by India before they enter this based on its own calculations. Full Pondage
Pakistan. Similarly, Pakistan has exclusive use of Level is defined in the Treaty as the level
the western rivers Jhelum, Chenab, and Indus. corresponding to maximum pondage allowed
(Garg, 1999). under it. Pakistan asserts that in the Baglihar
From the rivers flowing in India, India got nearly design, the space between the maximum water-
33 Million Acre Feet (MAF) from eastern rivers level and top of the dam is unduly large and
whereas Pakistan got nearly 125 MAF from would enable India to have a greater storage than
western rivers. However, India can use the full pondage level.”. In this study, Gumbel’s
western river waters for irrigation up to 701,00 Distribution is used to calculate the peak flows at
acres with new water storage capacity not different return periods and determine the
percentage reduction in flows due to construction construction of dam i.e. Pre-discharges and Post
of this dam. discharges of river, respectively.
Site Description Gumbel’s Distribution Method
The Baglihar Dam that is a run-of-the- That extreme value distribution was introduced
river power project located on the Chenab by Gumbel (1941) and commonly known as
River in the Southern Doda district of the Indian Gumbel’s distribution. Gumbel distribution is a
occupied state of Jammu and Kashmir . The statistical method and It is one of the most widely
Chenab river which is of 960 km, a major river used probability. The Gumbel distribution is used
of India and Pakistan, forms in the upper to observe series of flood peaks, maximum
Himalayas in the Lahaul and Spiti rainfalls, maximum wind speed, etc at different
district of Himachal Pradesh, India, and flows return periods (Subramanya, 2013). According to
through the Jammu region of Jammu and Gumbel’s theory of extreme events, giving the
value of the Variate X (peak discharge value)
with a return period T is used as
XT = x̄+ K σ n-1 (1)

Where in Eq (1)
x̄ = Mean value of X
σ n-1 = Standard deviation of the sample size N.
K = Frequency factor and expressed as
𝑦𝑇 −Ῡ𝑛
K= (2)
𝑆𝑛
Where in Eq (2)
𝑇
y T = [ In. In 𝑇−1 ] (3)
or yT can be calculated as
𝑇
yT = [0.834 + 2.303 log log 𝑇−1] (4)
Kashmir into the plains of the Punjab, Pakistan . Ῡn = Reduced mean a function of sample size N
Fig. 1 Location of Kishanganga and Baglihar Sn = Reduced standard deviation a function of
Hydroelectric Plants sample size N
The values of Ῡn and Sn are to be find in literature
Data Collection given by Gumbel.
A 36-years (1980 to 2016) data from different
gauging stations is collected from Pakistan Results and Discussions
Meteorological Department, Lahore. The data In Pre-discharges considered those discharges
includes annual peak discharge of a river at which are on the gauging station before the
different gauging stations. The gauging stations construction of Baglihar (1980-2004)
at Chenab river are Khanki, Trimmu, Qadirabad,
hydroelectric project whereas in Post- discharges
and Marala. considered those discharges which occur after the
construction of Baglihar hydroelectric plant
Methodology (2005-2016). The Gumbel’s distribution is used
Flood frequency analysis is a probability model on both peak flows (Pre and Post discharges).
to the sample of annual flood peaks recorded over Peak flows determined at four gauging stations at
a period of observation, for a catchment of a 1000-year return period. The analysis of
given region and period. The model parameters Gumbel’s distribution is shown in below mention
established can then be used to predict the graphs Figs. 2 – Fig 5. At Marala gauging station,
extreme events of large recurrence interval the average percentage reduction at 1000- year
(Pegram and Parak, 2004). Flood frequency return period is 11.33% as shown in Fig 2. At
results are vital for floodplain management; to Khanki gauging station, the average percentage
protect the public, minimize flood related costs to reduction 28.91 % occurs as shown in fig 3.
government and private enterprises, for Similar reduction in peak flows can be seen in
designing and locating hydraulic structures other gauging stations that are at Qaidrabad and
(Tumbare, 2000). Trimmu. At Qaidirabad and trimmu gauging
In this study, Gumbel’s frequency analysis stations average percentage reduction are 30.6%
method used to estimate the return period at and 40%, respectively, as shown in fig 4 and fig
different gauging stations before and after the
5. The overall reduction in flow 27.71 % is
observed
Table 1 Summary result of Percentage

Reduction on Chenab River due to Baglihar
Hydropower Plant
Gaugin Mar Kha Qadira Trim Aver
g ala nki bad mu age
Station
Percen 11.3 28.9 30.6 40 27.71

tage 3 1 Fig. 5 Gumbel’s Distribution at Trimmu
reducti Gauging station
on
Conclusions
Table A, shows the percentage reduction in
flow at Chenab river at each gauging station
due to Baglihar hydropower plant. There is a
significant reduction in post discharges at all
gauging stations. The average reduction in flow
at Chenab river due to mentioned Indian dam is
27.71% by Gumbel’s distribution at 1000- year
return period. The results can be verified and
Fig. 2 Gumbel’s Distribution at Marala compared by using other flood frequency
Gauging Station approaches.
Acknowledgement
The authors would like to highly acknowledge
Pakistan Metrological Department Lahore, to
provide data used in this study.
References
Garg, S.K (1999). “International and interstate
river water disputes”.
Iftikhar, M. (2011). “Tackling the Kishanganga
Fig. 3 Gumbel’s Distribution at Khanki knot. The Nation. Islamabad”
Gauging Station Raina, A.N (2017) "Geography of Jammu and
Kashmir State"’
Sahai, I.M. (2006). "The Baglihar dispute".
International Water Power & Dam
Construction Pegram, G and Parak, M (2004).
“A review of the regional maximum flood and
rational formula using geomorphological
information”.
Subramanya, K (2013). A book on
“Engineering Hydrology” Fourth Edition
Tumbare, M. J. (2000). “Mitigating floods in
Fig. 4 Gumbel’s Distribution at Qaidrabad Southern Africa”. Paper presented at the 1st
Gauging Station WARSFA/WaterNet Symposium: Sustainable
Use of Water Resources, 1-2 November,
Maputo
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Risk of Indus Basin Water Transfer in Violation of Indus Basin Treaty

Ahmad Raza1*, Muhammad Waseem1, Navid Tahir2, Ijaz Ahmad1
1*
2
Department of Irrigation and Drainage, University of Agriculture, Faisalabad 38000, Pakistan
*
[email protected]
Abstract: Indus Water Treaty (IWT) was signed in 1960 between Pakistan and India to resolve the water conflict.
This study was carried to study threat of unlawful transfer of Western River’s water through tunnels. Catchment
area of the study area was determined by Arc GIS software. Locations for water diversion were selected
considering following criteria: large elevation difference available between source river and recipient river, River
have narrow cross section to allow tunnel on steep river section face, and location where river has enough water
depth so that the tunnel intake mostly remain under water even during low flow conditions. Cross sections of all
possible locations were also determined by Arc GIS software. Tunnel route was marked considering shortest
length and largest elevation difference between the source and recipient river and thus tunnel length was
determined. Tunnel design was done considering total energy available between source and recipient river, head
losses at intake and control gates, friction losses of tunnel surface, tunnel length and exit losses. Tunnel diameter
was selected such that tunnel intake remain mostly under water even during low flow conditions. Annual water
transfer wass estimated from minimum of 141 million cubic meter (MCM) to as much as 2131 MCM from Indus
River. Five sites Tso Moriri, Gaik, Nimmo Bazgo, Dah and Batalik on Indus and two sites Jangalwar and Kund
Bara on Chenab River are identified as high-risk sites. The Kund Bara tunnel from Chenab River seems to be first
adopted due to be shorter length (32.3 Km) and large flow diversion (3532 MCM). Technical capabilities,
financial gains and most important socio-political-diplomatic hegemony of India reflect high risk of Indus basin
water transfer by India in violation of Indus Water Treaty. There are more than twenty possible locations where
unlawful river flow diversion could be made.
Keywords: IWT Indus Water Treaty, Conflict Resolution, Digital Elevation Model (DEM)
Introduction seasons and water shortage in the dry seasons.

Pakistan is amongst those countries having The storage capacities of Indian dams affect the
shortage of water (Akther, 2010). The major pattern of river flows into Pakistan. India has
resources of water in Pakistan are rivers, ground announced 135 dams on the western rivers which
water, rainfall and glaciers. The annual average include 24 dams on Indus River, 77 dams on
rainfall of Pakistan is less than the 240 mm. Indus Jhelum River and 34 dams Chenab River. Social
River has five main tributaries Sutlej river, Bias media in Pakistan have viewed that India is
River, Ravi river, Jhelum river and Chenab river. violating IWT, but India always claimed that all
Indus river total length is 3,200 kilometres and projects are run of river hydropower projects
calculated flow volume is 180 billion cubic such as Wullar Barrage and other projects
meters. The catchment area of the Indus River constructed within the regularity framework of
from the originating point (Himalaya) to the IWT.
terminating point (Arabian Sea) is 965,120 India has started construction big hydropower
square kilometers. Indus basin catchment area dams on the Indus river. India point out nine
found in China (Tibet) is 92,160 square dams on Indus river which will be constructed in
kilometers and in Afghanistan 74,240 square future. Three of them are Nimmoo Bazgo, Chtak
kilometers. Catchment area of the basin found in and Dumkhar big dams explained below. (Haq,
northern areas of Pakistan, Azad Kashmir and 2010). India has completed many dams and many
other area are 24,3200 square kilometers. Major other dams are under construction on Chenab
portion 58% (55,5520 square kilometers) of River. Three big dams Salal, Baglihar and Dul
Indus basin lie in Pakistan. Hasti are completed. These three dams are now
Indian all run of hydropower project on Indus, in operation. India has planned to construct 3
Chenab and Jhelum Rivers have serious effect on large and 8 small dams and other 24 projects are
river flows. Pakistan worries that Indian project conduct (Tunnel) type on Chenab River. (Haq,
can cause flood in the western rivers during wet 2010)
India has completed 13 dams on the Jhelum River Assessment of possible routes and location of the
in which the three major dams are lower Jhelum diversion tunnels were located where the
with 105 MW capacity of power house, Uri dam diversion could be possible. Twenty locations
with 480 MW capacities of power house and Sind were selected of which eighteen were on Indus
phase II with 105 MW capacity of power house. River and two were on Chenab River. Locations
India also point out 74 other projects on Jhelum were selected by considering following
river which include Gangbal dam with 100 MW parameters:
capacity of power house, Ujh dam 280 MW, • Significant elevation difference between
Sonamarg dam 165 MW and Kishanganga 330 Source River (Indus River or Chenab River)
MW. (Haq, 2010) and Recipient River.
India was spending 200 billion dollars for the • Narrow cross section at diversion site to
construction of water tunnels on Indus River, allow tunnel on steep section face.
which ultimately affect the fertility of land • Location where enough water depth so that
(Khalid Mustafa, 2008). India is starting the the tunnel intake (mouth) remain under
world 3rd largest dam (Kargal) on Indus river, water and is liable to be not detected due to
which will decrease flow of Indus river up to 45 being covert and unlawful activity.
% (News & Views Mohammad Jamil) and other
NGO’s as Indus Basin Commission Water River Water Depth by Hydraulic Model HEC-
Pakistan (IBCWP), World Water Council RAS
(WWC) pointing out threat of illegal transfer of HEC-RAS model was used to determine the
Indus basin water in violation of Indus basin depth of water in the cross section of selected
treaty. However, no specifics are provided by the locations. Three discharges were estimated on
media and NGO, s about water transfer. each location by the relationship (Discharge per
This research would aim to determine the unit Catchment area). The flow depth of water in
alternative location and route that could be used each cross section was determined for maximum,
for un-authorized diversion of river Indus water. minimum, and average discharges.
So, by determining the routes we can check the Under covert operation conditions tunnels can be
tunnel configuration and length of tunnel. Hence, best way for diverting water from one river
there is need to be vigilant unlawful water system to other. The tunnel is considered to
transfer of Indus river water which is the major adhere following hydraulic requirements.
source of water supply of Pakistan. This study is • Bottom of the tunnel should be at some slope
based on the data which was easily available • At small discharge the tunnel will perform as
from office files, internet resources and Google open channel flow and at large discharges the
Earth images. Flows at different locations were tunnel behave like pipe flow
worked on proportionate basis. The low-level • Minimum Diameter of the tunnel should not
accuracy of such data could have affected the less than 5 ft from construction point of view
study output.
• Location and design of the tunnel is affected
by the Geology (Geology is not considered in
Methodology
this study due to unavailability of geologic
India could transfer the water of western rivers details, but a favorable geology is
(Indus, Jhelum and Chenab rivers) to the eastern considered).
rivers (Ravi, Beas and Sutlej Rivers). These un- Typical working sheet showing detail of tunnel
authorized diversions are possible within design is given in Table 2.1.
territorial boundaries of Jammu and Kashmir.
Study area of research is Indus River, but a small
Annual Flow Diversion
portion of Chenab river is also included.
Following procedure was adopted for the
Indus River bed profile was determined by using calculation of annual flow diversion from
Google Earth software. Base point was selected tunnels.
at Skardu (35.27° N, 75.25º E). A working map i. Catchment area of all possible locations
of the whole Indus River and its tributaries was were determined from DEM processing
prepared from G.T sheets (25 in number) with the
through GIS
help of CAD. Terrain processing was done by ii. Average monthly discharge of 21years
using Arc GIS Tools for catchment extraction. of Indus and 14 years of Chenab River
Terrain preprocessing was carried out in a was calculated at Skardu and Dhamkund
definite sequence for catchment extraction respectively
purpose.
iii. Available monthly discharge (Jan to iv. Water depth at all possible locations
Dec) of all possible locations were during whole year (Jan to Dec)
calculated on the basis of catchment area determined with the help of HEC-RAS
model.
Table 1 Typical working sheet of tunnel design

Input
Discharge Q (m³/s) 25 n 0.013
Length of Tunnel (m) 56900 u/s Invert level (m) 4497
Water elevation/depth (m) 4527 d/s Invert Level (m) 3833
Assumed dia of tunnel (m) 2.50
Computation
Area of tunnel (m²) 4.91 Gate and valve losses (Kg) 0.19
Velocity of water (m/s) 5.09 Exit losses (Kv) 1
Velocity head (V²/2g) 1.32 Total head loss (HL) 692.39
Friction losses (fL/D) 523.72 Exit energy head (m) 3835
Entrance losses (Ke) 0.05 U/S energy head (m) 4527
Results
Water elevation/depth = U/S energy head 4527
Results and Discussions to December for each tunnel calculated. Annual

The aim of study was to identify the possible volume diverted from tunnels also calculated
locations where the Indus River and Chenab river from the available discharges.
water could be diverted to Ravi, Beas or Sutlej Twenty locations of flow diversion were
rivers under the covert operation in violation of identified in which eighteen were on Indus River
Indus basin treaty. and two were on Chenab River. The diversion
Indus River Bed profile was determined from the sites are given in Table 3.1. The leading tunnels
Skardu (35.27° N, 75.25º E) to the boundary line are shown in Fig. 3. Catchment area and
of Jammu and Kashmir and China as shown in discharges (maximum, minimum and average)
Fig. 1. Marked river are shown in Fig. 2. From available at all possible locations were
graph it is clear that at Skardu the catchment area calculated.
is maximum, but the elevation of river is Following arrangements were considered to
minimum. Catchment area decrease with length divert flow and minimize the sediment entry to
at 112,665 km² at Skardu to 10,108 km² near line tunnels.
of control between J&K and China. River bed i. Diversion carried on a steep sloping river
level increases from 2200 m at Skardu to 4150 m face
near line of control between J&K and China. ii. No external structures
Catchment area of Indus and Chenab River was iii. Intake constructed as underwater activity
determined at Skardu and Dhamkund gauging iv. Flow control gate/valves housed in an
stations by using Arc GIS. The coordinate of underground chamber. Presumed diversion
Kachura and Dhamkund gauging stations are tunnel intake and flow control arrangement
35.27 N, 75.25 E and 33.24 N, 75.14 E. are required.
Catchment area obtained from Arc GIS are v. To reduce the bed, load the invert level of
shown in Figs. tunnels was designed up to 0.5 m from
Indus River discharge was collected from Kachra river bed.
gauging station and Chenab River discharge was vi. Submerged trash rack will use to eliminate
collected from Dhamkund gauging station. Indus trash load.
River and Chenab River maximum, average and vii. Under water dyke at the upstream of tunnel
minimum discharges were based on the historic to ward off sediment load.
data. Monthly discharge available from January
viii. During flood in river the tunnel will be
closed
Fig. 1 Indus River Bed Profile
Fig. 3 Twenty possible diversion tunnel sites
Fig. 2 Indus River catchment area

Table 2 Detail design of Tunnels (T1-T20)
Q per
u/s d/s Annual
Length Slope Dia Discharg Length
Location Elevation Elevation Volume
(km) (m/m) (m) e (m³/s) (MCM/
(m) (m) (MCM)
km)
T1 Tso
4527 3833 56.9 0.012 2.5 25 765 13.44
Moriri
T2 Indus 4127 3633 120 0.004 2.5 15 425 3.54
T3 Indus 4118 3873 106 0.002 3.0 17 522 4.92
T4 Chenab 916.5 629 58.8 0.005 4.5 74 2245 38.18
T5 Indus 4016 3873 109 0.001 3.1 14 417 3.83
T6 Indus 3908 3306 150 0.004 2.7 18 531 3.54
T7 Indus 3775 2250 162 0.009 2.9 32 881 5.44
T8 Indus 3602 2443 170 0.007 2.7 23 657 3.87
T9 Indus 3503 3138 169 0.002 3.0 17 489 2.90
T10 Indus 3389 2513 183 0.005 3.2 30 829 4.53
T11 Indus 3316 2519 188 0.004 3.0 24 639 3.40
T12 Chenab 2450 1357 32.3 0.034 4.2 165 3532 109.3
T13 Indus 3172 2857 170 0.002 2.7 12 375 2.20
T14 Indus 3040 2866 164 0.001 4.5 36 1055 6.44
T15 Indus 2907 2750 165 0.001 3.6 18 540 3.27
T16 Indus 2784 2035 155 0.005 3.0 25 666 4.30
Q per
u/s d/s Annual
Length Slope Dia Discharg Length
Location Elevation Elevation Volume
(km) (m/m) (m) e (m³/s) (MCM/
(m) (m) (MCM)
km)
T17 Indus 2730 1927 158 0.005 4.5 76 2116 13.39

T18 Indus 2590 1867 160 0.005 4.5 73 2131 13.32
T19 Suru 2618 2436 181 0.001 3.1 13 361 1.99
T20 Suru 2786 2585 166 0.001 2.3 6 141 0.88
Risk ranking of tunnels

The other important risk factor is cost, which is
reflected by tunnel lengths and diameter. Smaller
length tunnels are more viable. Tunnel length
varies from 30 to 180 Km. It is clear that the
tunnels which have more volume per tunnel
length are more viable/attractive route for river
water diversion. Considering that minimum 1m Fig. 6 Excavation volume of tunnels at River
extra tunnel radius cutting will be required to Indus
accomplish tunnel
lining/stabilizing/smoothing, tunnel total
excavation volume is shown in Fig. 6 varies from
0.91 to 5.47 MCM. Considering extraction
volumes, tunnel length, and diversion volumes,
the most likely activities includes:
Indus River T1, Tso Moriri, T7 Gaik, T14
Nimmo Bazgo, T17 Dah and T18 Batalik
Chenab River T4 Jangalwar and T12 Kund
Bara T12
Kund Bara tunnel from Chenab River seems to
adopt due to shorter length (32.3 Km) and large
flow diversion (3532 MCM). Fig. 7 Volume of water/volume of excavation of
tunnels
Conclusions
Technical capabilities, financial gains and most
important socio-political-diplomatic hegemony
of India reflect high risk of Indus basin water
transfer by India in violation of Indus Water
Treaty.
Fig. 4 Flow volume of tunnels
India can divert Indus River water at many
locations in J&K occupied region and the annual
water transfer is estimated from possible
locations minimum of 141 MCM to as much as
2131 MCM.
Unusual long tunnel lengths make covert and
unlawful water theft/transfer less attractive.
Water transfer by India through tunnels from
Indus River is very difficult but cannot be ruled
Fig. 5 Possible tunnels lengths at River Indus out.
References
Akther,S., (2010). Emerging challenges to Indus
waters treaty issues of compliance &
transboundary impacts of Indian
hydroprojects on the western rivers, Journal Ghanbarpour, Shokoufeh Salimi, Mohsen
of the Institute of Regional Studies, Mohseni Saravi and Mehdi Zarei (2011)
Islamabad, Pakistan, V. 28., No.4, Autumn Calibration of River Hydraulic Model
2010, pp. 3-66 Combined with GIS Analysis using Ground-
Abbasi, A., (2012). Indus water treaty between Based Observation Data, Research Journal of
India and Pakistan, Paper of Pakistan Applied Sciences, Engineering and
Institute of legislative development and Technology 3(5): 456-463, 2011
transparency, Haq, N.U., (2010). Pakistan's water concerns,
[http://www.pildat.org/publications/publicat [http://ipripak.org/factfiles/ff127.pdf]
ion/FP/IndusWaterTreatybetweenPakistanA Hayat, (2012). Indus waters treaty and resolution
ndIndia_PakIndiaDialogueIII.pdf] of water conflicts between two nuclear
Abbasi, A.H., (2009). 'www.defence.defence.pk' nations Pakistan and India,
The News, Saturday,14 February, 2009 [http://www.feem-
Autobee, R., (1996). Colorado-Big Thompson web.it/ess/ess12/files/papers/hayat.pdf]
Project, Hicks, F.E., and Peacock, T., (2004). Suitability
[http://www.usbr.gov/projects//ImageServer of HEC-RAS for flood forecasting, Canadian
?imgName=Doc_1303159857902.pdf] Water Resources Journal, V. 30., pp. 159–
Ahmad, F., (2007), Probabilistic flood risk 174 (2005)
analysis in Chenab Riverine area in Kiss, R., (1994). Determination of drainage
Muzaffargarh district. M.Sc Thesis, Center network in digital elevation models, utilities
of Excellence in Water Resources and limitations, Journal of Hungarian
Engineering, University of Engineering and Geomathematics Water Resources Research,
Technology, Lahore. v.30., p. 1681-1692,
Ahmad B, Muhammad Shumail Kaleem, Mohsin Latif, Q., (2007), Effect of channel improvement
Jamil Butt, Zakir Hussain Dahri (2010) on water levels for Kabul River under flood
Hydrological modelling and flood hazard conditions. M.Sc Thesis, Center of
mapping of Nullah Lai, Pakistan Agricultural Excellence in Water Resources Engineering,
Research Council (PARC), Islamabad, University of Engineering and Technology
Pakistan Lahore.
Bozdag, A., and Gocmez, G., (2010). Digital Miller, S.N., (2002). Estimating channel
elavetion model at determining of some morphologic properties from a high
basin characteristics: a case study on the resolution DEM,
cihanbeyli (Konya) basin (balwois 2010 - [http://www.tucson.ars.ag.gov/icrw/proceedi
ohrid, Republic of Macedonia 25, ngs/miller.pdf]
[http://balwois.com/balwois/administration/f Shah, S.J.A., (2011). Indus waters treaty under
ull_paper/ffp-1492.pdf ] stress: imperatives of climatic change or
Bartak, V., (2008). How to extract river networks political manipulation, Margalla Papers 2011
and catchment boundaries from DEM: a Special Edition, Vol. 15., Issue I., 2011
review of digital terrain analysis techniques. Taylor, T.C., (1975). Tunnel construction and
Journal of Landscape Studies 2 NAWAPA,
[http://fzp.czu.cz/~bartakv/thesis] [http://www.larouchepac.com/node/17641]
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Subsurface Flow Analysis of Hydraulic Structures using Bligh’s Theory,

Lane’s Weighted Creep Theory and Modern Khosla’s Theory
Hasnain Jillani*1, Muhammad Haris Asghar Khan1, Hammad ur Rehman1
1*
*
[email protected]
Abstract: Hydraulic structures on permeable foundation requires special attention towards subsurface flow
analysis compared to surface flow as the safety of structure is concerned. The seeping water underneath the
foundation endangers the safety of hydraulic structures in two ways. Firstly, it sets up pressure gradient at the
toe of structure, if excessive, may cause movement of the bed material with consequent failure of the work. This
phenomenon is known as piping. Secondly, uplift pressures are created under the work itself which, if not
adequately contained, will burst through it, providing a short path for subsoil flow and leading to total failure of
the work. Present study includes the subsurface flow analysis of selected hydraulic structures using Bligh’s,
Lane’s and Khosla’s theories. The effect of location and depth of sheet piles on uplift pressures will also be
investigated in this study. The preventive measures adopted to control the exit gradient or uplifts will be evaluated
and discussed. The results of exit gradient and uplift pressures will be calculated and compared with safe values
recommended by the above three researchers.
Keywords: Sub-surface analysis, hydraulic structures, seepage theories
Introduction undermining in case of u/s flexible apron failure.

Hydraulic structures on permeable foundations The basic purpose of intermediate sheet pile is to
are susceptible to seepage beneath the structures. protect the main components of barrage i.e. u/s
This seepage water undermines the foundation glacis, crest, d/s glacis, piers, bridge, etc. in case
and ultimately results in failure of structure. of failure of u/s or d/s sheet pile. Singh (2011)
Hence the hydraulic structures on permeable stated in his study that the effectiveness of
foundations must be checked against possible intermediate sheet pile stands nil against
pipping and undermining. Sufficient floor length reduction in uplift pressures and it only adds to
and depth of sheet piles ensures the safety of the cost of barrage. Recently Khosla’s theory of
structure against seepage water. Presently there independent variables is widely used for the safe
are three approaches, namely Bligh’s, Lane’s and exit gradient of structures. According to Khosla’s
Khosla’s Theory, to analyze hydraulic structure exit gradient formula given as 𝐺𝐸 = 𝑑
𝐻 1
, it’s
for subsurface flow. In recent practice, Lane’s 𝜋√𝜆
the head and depth of d/s sheet pile that influence
and Khosla’s approaches are frequently adopted
the exit gradient value. Although depth of d/s
because of their reliable results. Subsoil flow
sheet pile is decided by the d/s scour depth, but
threatens the stability of the work in two ways:
the value of safe exit gradient is the real
1. Sets up pressure gradient at the toe of work,
governing factor for fixing the depth of d/s sheet
if excessive, may cause movement of the bed
pile (Haigh and Fraser, 1941). Shri (1995)
with consequent failure of the work. (Piping)
engraves that the excessive length of d/s sheet
2. Pressures (uplifts) are created under the work
pile may cause the uplifts to increase under the
itself which, if not adequately contained, will
d/s floor while a shorter length reduces the exit
burst through it, providing a short path for
gradient, hence a suitable length of d/s sheet pile
subsoil flow and leading to total failure of the
is provided in order to keep the above problems
work.
minimized. During 19th century the value of exit
There are two major functions for the sheet pile
gradient between 1/4 to 1/5 was considered safe
to execute, one being protection of barrage floor
against piping and undermining (Kapur, 1941).
against any danger of undermining and second is
A lot of investigation were made for local scour
to increase the creep length in order to make the
in cohesionless soils. However, river banks and
exit gradient safe. Haigh and Fraser (1941)
beds are often composed of mixture of gravel,
highlighted the importance of u/s sheet pile as u/s
sand, silt and clay. The location, extent and
sheet pile may protect the structure from possible
geometry of scour hole around bridge piers may
vary extensively in cohesive soils from those in structure is very small compared to its effect on
cohesionless soils (Raju et al., 2002). exit gradient.
Scour depth may also vary depending upon the Failure of hydraulic structure by piping is
shape of structure to be encountered. Oliveto and enhance by the enlargement of tunnel, formed
Hager (2002) investigate the characteristics of underneath the structure, as time passes. In order
scour holes around cylindrical and square piers to estimate the time to downstream flooding i.e.
and vertical-wall abutments with uniform bed failure of structure one must know the coefficient
material. Their findings at the end of of erosion. Pipe diameter prior to breaching can
experimental program were that the maximum also serve as an indicator for the peak flow and
scour depths occurred around the centreline of consequently time to failure (Bonelli and
the upstream face of the piers and around the Benahmed, 2011). Markus et al. (2007) assessed
upstream corner of the vertical-wall abutments. It the stability of dams, dykes and levees.
was also concluded that the maximum accretions Embankment dams and dykes can fail due to
took place around the rear face of these elements. three possible failure threats namely external
The location of sheet pile underneath a hydraulic erosion (overtopping), internal erosion and
structure effects the seepage through foundation sliding. Internal erosion is more hazardous for
soil. Alsenousi and Mohamed (2008) studied the hydraulic structures being progressive in nature.
effect of location and inclination of sheet pile on To minimize the risk of failure risk management
seepage underneath a structure. According to is recommended to understand the internal
their findings the exit gradient d/s of hydraulic erosion process.
structure decreases with the increase in sheet pile Hydraulic structures formed on permeable
inclination. If the exit gradient and the seepage foundation can undergo seepage underneath
behind the hydraulic structure are considered as which will endangers its stability. Subsurface
the major factors in the design of the structures, analysis for existing and proposed sections of
the optimum location of the sheet pile is at the toe Panjnad Barrage was carried out using Bligh’s
of structure with inclination angle equals to 120°. approach, Lane’s approach and modern Khosla’s
While if the uplift head is considered as the major approach. For the scour calculations, well known
factor, the optimum location of the sheet pile is Lacey’s scour formula was used. The key plan
at the heel of the structure with inclination angle of Panjnad Barrage is shown in Fig. 1.
equals to 120°. Similarly, Alnealy and Alghazali
(2015) use of inclined cut-off towards the
upstream side with 45° inclination is beneficial in
decreasing the uplift pressure to 40.3%, seepage
quantity to 28.5%, as compared with the general
case of no cut-off. Also, the use of inclined cut-
off towards the downstream side with 120°
inclination is beneficial in decreasing value of
exist gradient to 5.0% and increasing the safety
factor against piping phenomenon to 3.18. It was
also concluded that the reduction in uplift
pressure, exist gradient and seepage when using
double cut-off at U/S with 45° inclination and
D/S with 120° inclination is more than that of
using double cut-off at the up and downstream at
right angle that given decreasing in value of exist
gradient to 8.03%, uplift pressure to 42.3%, Fig. 1 Key plan of Panjnad Barrage
seepage to 31.15% and increasing the safety
factor against piping phenomenon to 2.43. Methodology
Al-Saadi et al., (2011) reported that use of Hydraulic structures formed on permeable
downstream cut-off inclined towards the foundation can undergo seepage underneath
downstream side with angle less than 120° which will endangers its stability. Subsurface
increases the safety factor against the piping analysis was carried out using Bligh’s approach,
phenomenon. It was also concluded that the Lane’s approach and modern Khosla’s approach.
effect of downstream cut-off inclination angle in For the scour calculation Lacey’s scour formula
reducing uplift pressure head under the hydraulic was used. Different methods employed are
discussed as under:
2 Fine sand 7 1/7
Bligh’s Creep Theory 3 Coarse sand 5 1/5
As per Bligh’s approach the seeping water 3 to
follows the bottom contour of structure. The total 4 Gravel and sand 1/3.5 to 1/3
3.5
creep length is calculated as L = 2(Vertical Boulders,
Length) + Horizontal Length. Head loss per unit 2.5 to
5 gravels and 1/3 to 1/2.5
length is given as HL / L, where HL is the head 3
sand
loss between upstream and downstream, also 1.6 to
known as hydraulic gradient. To make structure 6 Clayey soils 1/3 to 1/1.6
3
safe against piping and undermining sufficient
creep length should be provided i.e. L = C.HL, Khosla’s Theory
where C is Bligh’s creep coefficient. Different Khosla gave the concept of independent
values of C against different soil types are listed variables. According to Khosla’s theory the
in following Table 1. seepage water doesn’t creep along the bottom
contour of structure instead it moves in a set of
Lane’s Creep Theory streamlines. Laplacian equation giving the steady
Extending Bligh’s Theory Lane suggested that seepage in vertical plane is expressed as follow:
the vertical creep is more effective in reducing
the uplifts as compared with horizontal creep. 𝑑2 ∅ 𝑑2 ∅
Hence a weightage factor of 1/3 was suggested + =0
𝑑𝑥 2 𝑑𝑧 2
for horizontal creep as against 1.0 for vertical
creep. Now the total creep length is given as L = Where ϕ is flow potential and is equal to Kh.
2(Vertical Length) + 1/3(Horizontal Length). In K = Coefficient of permeability and h = residual
this case the safe creep length is given as Ll = head at any point.
Cl.HL, where Cl is the Lane’s creep coefficient. Exit Gradient
Different values suggested by Lane for Cl against Khosla defines Exit Gradient as follow:
different soil types are tabulated in following
Table 2 𝐻 1
𝐺𝐸 =
𝑑 𝜋√𝜆
Table 1 Bligh’s Safe Hydraulic Gradient for
Different Types of Soil
Where,
Safe
H = Head loss between upstream and
Hydraulic
Sr. Value downstream
Type of Soil Gradient
No. of C d = Depth of downstream cut-off
should be
less than
Fine micaceous 1 + √1 + 𝛼 2
1 15 1/15 𝜆=
sand 2
Coarse grained
2
sand
12 1/12 α = b/d, b being total floor length
The Exit Gradient computed from above relation
Sand mixed
must lie within safe limit as given in following
with boulder
3 5 to 9 1/5 to 1/9 Table 3.
and gravel, and
for loam soil
Table 3 Khosla’s Safe Exit Gradient for
Light sand and Different Types of Soil
4 8 1/8
mud
Type of Soil Safe Exit Gradient
Shingle 1/4 to 1/5
Table 2 Lane’s Safe Hydraulic Gradient for
Coarse sand 1/5 to 1/6
Different Types of Soil
Fine sand 1/6 to 1/7
Safe
Hydraulic
Sr. Value To estimate the percentage pressures at various
Type of Soil Gradient
No. of Cl key points of sheet pile one can either make use
should be
of developed equations or developed curves.
less than
Percentage pressure at key points of downstream
Very fine sand
1 8.5 1/8.5 sheet pile are estimated as follow:
or silt
1 𝜆−2 Correction for the mutual interference of
∅𝐸 = cos −1 ( ) piles
𝜋 𝜆
The percentage pressure correction due to mutual
1 𝜆−1 interference of pile is given by
∅𝐷 = cos −1 ( )
𝜋 𝜆
𝐷 𝑑+𝐷
Where, 𝐶 = 19√ [ ]
1 + √1 + 𝛼 2 𝑏′ 𝑏
𝜆=
2
Where, b’ = Distance between two pile lines
𝑏 D = Depth of pile line, the influence of which has
𝛼= (respective) to be determined on the neighbouring pile of
𝑑
Percentage pressure at key points of upstream depth d. D is to be measured below the level at
sheet pile are calculated as which interference is desired.
∅𝐶1 = 100 − ∅𝐸 d = Depth of pile on which the effect is
∅𝐷1 = 100 − ∅𝐷 considered
For an intermediate sheet pile, the percentage b = Total floor length
pressures at key points are estimated by the The correction is positive for the points in the rear
following relations: or back water, and negative for the points forward
1 𝜆1 − 1 in the direction of flow.
∅𝐸 = cos −1 ( )
𝜋 𝜆
Correction for thickness of floor
1 𝜆1 𝑡𝑓
∅𝐷 = cos −1 ( ) 𝐶𝑓 = ( ) (∅1 − ∅2 )
𝜋 𝜆 𝑑
1 𝜆1 + 1 Where, tf = Floor thickness

∅𝐶 = cos −1 ( ) d = Depth of sheet pile
𝜋 𝜆
Where, ∅1 − ∅2 = Difference of uplift pressure
√1 + 𝛼1 2 + √1 + 𝛼2 2 between two consecutive key points
𝜆= The correction is negative for the points in the
2 rear or back water, and positive for the points
forward in the direction of flow.
√1 + 𝛼1 2 − √1 + 𝛼2 2
𝜆1 =
2 Correction for the slope of floor
Correction for the slope of floor is applicable
𝑏1 only to the key points of piles located at the start
𝛼1 =
𝑑 or end of slope. The correction is estimated by the
following relation:
𝑏2
𝛼2 =
𝑑 𝑏𝑠
𝐶𝑠 = 𝐹𝑠 ( )
𝑏1
b1 = Length of pucca floor upstream of sheet
pile
Where, bs = Horizontal length of slope
b2 = Length of pucca floor downstream of sheet
b1 = Distance between two piles
pile
Fs = Correction factor which depends on
The above calculations are for the line diagram
the ratio of slope and is given in Table 4
of structure and need corrections for the
following three cases: Table 4 Slope Correction Factors
a. Correction for the mutual interference of
piles Slope (H : V) Correction Factor
b. Correction for thickness of floor 1:1 11.2
c. Correction for slope of floor 2:1 6.5
3:1 4.5
4:1 3.3
5:1 2.8
6:1 2.5
7:1 2.3 Results and Discussion
8:1 2.0 There were mainly four different sections of
Panjnad barrage which were analysed, namely
The correction is positive for the downward slope Left Pocket, Main Weir, Annexe Weir and
and negative for the upward slope following the Additional Depressed Bays.
direction of flow. Creep results for different barrage sections, as per
Bligh’s and Lane’s Weighted Creep Theory, for
Scour Calculations existing and remodelled barrage are compared
Scour calculations were performed as per with safe values in Table 5
Lacey’s Scour Formula which is given as below:
Exit gradient for different sections of Panjnad
1⁄ Barrage, calculated using Khosla’s Exit Gradient
𝑞2 3 approach, are tabulated in following Table 6
𝑅 = 0.9 ( )
𝑓
The remaining seepage pressure at end of pucca
Where, q = Discharge per unit width (cusecs/ft) floor, as calculated by equations proposed by
F = Lacey’s silt factor = 1.75√𝐷50 Khosla, and required and provided thicknesses
D50 = Mean particle size (mm) for existing and remodelled barrage are tabulated
in Table 7 and Table 8 respectively.
Table 5 Creep Analysis Results as per Bligh and Lane Creep Theories
Existing Barrage Remodelled Barrage
Sr. Creep Ratio Creep Ratio Safe Bligh’s Safe Lane’s
Description
No. Bligh’s Lane’s Bligh’s Lane’s Creep Ratio Creep Ratio
Theory Theory Theory Theory
1 Left Pocket 16.4 8.5 22.0 13.8 15.0 7.0
2 Main Weir 18.4 10.5 21.8 13.6 15.0 7.0
Annexe 19.9 12.9
3 14.3 9.1 15.0 7.0
Weir
Additional 11.9
4 18.9 11.9 18.9 15.0 7.0
Bays
Table 6 Exit Gradient Results as per Khosla Theory

Remodelled Safe Exit Gradient as per Soil
Existing Barrage
Sr. Barrage Condition
Description
No. Exit Gradient Exit Gradient
(1/GE) (1/GE)
1 Left Pocket 8.03 9.34 6.0
2 Main Weir 8.08 9.34 6.0
3 Annexe Weir 5.41 8.33 6.0
Additional 7.92
4 7.92 6.0
Bays
Table 7 Uplift Analysis Results as per Khosla Theory

Existing Barrage
Sr. No. Description Remaining Seepage Pressure Required Thickness Provided Thickness
1 Left Pocket 24.90 4.21 2.50
2 Main Weir 23.69 4.01 2.50
3 Annexe Weir 24.36 4.30 4.00
Table 8 Uplift Analysis Results as per Khosla Theory

Remodelled Barrage
Sr. Remaining Seepage Required Provided
Description
No. Pressure Thickness Thickness
1 Left Pocket 26.77 1.42 3.50
2 Main Weir 26.77 1.42 3.50
3 Annexe Weir 29.08 1.27 3.50
Additional Depressed
4 27.06 5.59 6.50
Bays
Note: Required thickness with 4-5 ft. water cushion on downstream floor.
Table 9 Lacey’s Scour Results

Scour Below Existing Cut-off Proposed Cut-off
Sr.
Description Bed Depth Depth
No.
U/S D/S U/S U/S
1 Left Pocket (q = 250.82 cusecs/ft) 25.09 32.76 5.00 25.50 22.50 33.50
2 Main Weir (q = 250.82 cusecs/ft) 25.09 32.76 22.00 25.50 22.50 33.50
Annexe Weir (q = 250.82
3 25.09 32.76 22.00 20.00 22.50 32.50
cusecs/ft)
Additional Bays (q = 310.58
4 33.12 35.97 - - 31.75 35.00
cusecs/ft)
The adequacy of sheet piles was checked on basis compared with provided or existing
of scour calculations. The results of scour as per thickness. The reason for less provided
lacey’s scour formula for existing and thickness being sealing of grooves with
rehabilitated conditions are tabulated in Table 9. pressure relief holes in downstream block
apron instead of rigid concrete floor. The
Conclusion existing impervious floor thicknesses are
The following conclusions are drawn from the satisfied by taking water cushion over stilling
present study based on the results discussed basin, however the new concrete floor is
above: provided with thicknesses for severe case of
• The creep coefficient, as per Bligh’s Theory, no water cushion. The floor thickness for the
for annexe weir section falls short from additional bays is recommended according to
recommended safe value. However, these gravity design considerations.
coefficients are within limit for the existing • He provided sheet piles in existing sections
left pocket and main weir sections. of barrage confirms the scour calculations
• The creep coefficients, as per Lane’s according to Lacey’s scour formula with
Weighted Creep Theory, for all existing factor of safety being 1.25 and 1.50 for the
barrage sections are well above the upstream and downstream respectively. The
recommended limiting value of 7.0. lacey’s silt factor for the existing barrage
• The required creep length was achieved by sections was adopted as 1.0. However, for the
sealing the downstream block apron scour calculations of remodelled barrage
provided with pressure relief holes at regular with additional bays lacey’s silt factor being
spacing. adopted as 0.85 (as per recent site soil
• The creep coefficients are further improved conditions). On the other hand, the factors of
by the remodelling of existing barrage in safety are same as 1.25 and 1.50 for the
which the sloping sealed block apron is upstream and downstream respectively. As
replaced by the levelled concrete floor with per the results of scour calculations a new
end sill. The creep length for additional sheet pile is proposed upstream of left
depressed bays is well above the pocket. The depths of proposed sheets piles
requirement. downstream of left pocket, main weir and
• Exit gradient calculations for different annexe weir are fixed in order to satisfy both
existing barrage sections are safe as per the the scour and exit gradient criteria.
site soil condition except for annexe weir
where this value is below the limiting safe References
value of exit gradient. New sheet piles are Alnealy, H.K.T., N.O.S. Alghazali (2015),
proposed at end of proposed downstream Analysis of Seepage Under Hydraulic
impervious floor in order to satisfy the exit Structures Using Slide Program, American
gradient criteria. Journal of Civil Engineering. Vol. 3, No. 4,
• The required floor thicknesses for the pp. 116-124.
existing sections of barrage (left pocket, Alsenousi, K.F., G.M. Hasan (2008), Effects of
main weir and annexe weir) are more Inclined Cut-offs and Soil Foundation
Characteristics on Seepage Beneath Dams, Dykes and Levees, Intermediate
Hydraulic Structures, Department of Civil Report of the European Working Group of
Engineering, University of Garyounis, ICOLD Contributions to the Symposium,
Libya. Freising, Germany.
Ansari, S.A., U.C. Kothyari, K.G.R. Raju (2002), Oliveto G, W.H. Hager (2002), Temporal
Influence of Cohesion on Scour Around Variation of Clear-Water Pier and Abutment
Bridge Piers, Dept. Of Civil Engineering, Scour, Journal of Hydraulics Engineering.
Aligarh Muslim University, Aligarh, India. 128: 811–820
Bonelli, S., N. Benahmed (2011), Piping Flow Saleh I., K. Al-Saadi, H.T.N. Al-Damarchi,
Erosion in Water Retaining Structures: H.C.D. Al-Zrejawi (2011), Optimum
Inferring Erosion Rates From Hole Erosion Location and Angle of Inclination of Cut-off
Tests and Quantifying the Failure Time. 8th to Control Exit Gradient and Uplift Pressure
ICOLD European Club Symposium Dam Head under Hydraulic Structures, Jordan
Safety - Sustainability in a Changing Journal of Civil Engineering, Volume 5, No.
Environment, Sep 2010, Innsbruck, Austria. 3
Fraser, F., Haige (1941), The Emerson Barrage, Shri, N. (1995) Manual on Barrages and Weirs
Journal of the Institution of Civil Engineers, on Permeable Foundation, Central Board of
Paper No. 5227. Irrigation and Power, New Delhi, India.
Kapur, S.N. (1941), River Diversion at Trimmu, Volume 1, Publication No. 179.
I.S.E. Paper No. 229. Singh, R.J. (2011), Optimal Hydraulic Structures
Markus, A., J.J. Fry, G. Matthias, P. Sebastian Profiles Under Uncertain Seepage Head,
(2007) Assessment of the Risk of Internal Motilal Nehru National Institute of
Erosion of Water Retaining Structures: Technology, Allahabad-211004, India.
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Assessing Hydropower Resilience under Changing Climate: Murum-Bakun

Cascade in East Malaysia
Khamaruzaman B. Wan Yusof1*, Mubasher Hussain1,2, Muhammad Raza-ul-Mustafa1
1
Department of Civil and Environmental Engineering, Universiti Teknologi Petronas, Malaysia
2
Hydro Department, Sarawak Energy Berhad, Malaysia
*
[email protected]
Abstract: This study assessed the climate change impact on the Murum Hydropower Plant (HEP) and Bakun
HEP in Sarawak state of Malaysia. Multi-model ensemble of three General Circulation Models (GCMs) from
Coupled Model Intercomparison Project Phase 5 (CMIP5) was used to simulate the daily precipitation over the
catchment for the period of 2011–2100 under two future scenarios (RCP4.5 and RCP8.5). A hydrological model
was developed using Australian Water Balance Model (AWBM) approach; and historical and future river inflows
were simulated at Murum Dam and Bakun Dam. Reservoir operation model was developed for the hydropower
system to simulate the operation under the historical inflows as well as for the future projected inflows. It was
noted that the mean river inflow will improve in the future at both Dams especially under RCP8.5. The similar
results were noted while simulating the reservoir operation and it was noted that the annual energy from Murum
HEP and Bakun HEP would increase up to 12 % and 7% respectively under RCP8.5 of 2080's.
Keywords: climate resilience, hydrologic modeling, hydropower, reservoir operation, CMIP5
Introduction study, we assessed the potential impacts of

When we talk about the renewable energy, climate change on the operation of Murum
hydropower is the major source of renewable reservoir and Bakun reservoir using General
energy with relatively lower cost of energy. It Circulation Models (GCMs) from the Coupled
integrates well with the other intermittent sources Model Intercomparison Project Phase 5
such as solar and wind power to ensure the grid (CMIP5).
system stability. Currently, hydropower The objectives of this paper were; 1) to project
contributes about 16% of the global energy and future changes in precipitation over the Rajang
from 1973 to 2011, the global hydropower River Basin in Sarawak utilizing CMIP5
production increased by 175 percent [1]. projections, 2) projection of river inflows to the
Continuous increase of greenhouse gas emission Murum Reservoir and Bakun Reservoir using the
in the atmosphere, positive radiative forcing and projected future precipitation over the catchment,
global warming are evidence of anthropogenic 3) cascaded reservoir operation modeling to
influence on the global climate system [2]. assess the climate change impact on hydropower
During 1951 to 2010, emission of greenhouse energy yield from these hydroelectric plants.
gases caused a global surface warming ranging
from 0.5 °C to 1.3 °C. It will cause further Study Area and Data Description
warming and change the future global climate Study Area
system [3]. Climate scientists noted that, it would Murum HEP
alter the hydrological systems of major river Murum HEP is the recently commissioned
basins as exhibited by [4-10]. In Malaysia, very hydroelectric plant in the Sarawak state of
few studies explored the climate change impact Malaysia, it is located in upper Rajang River
on water resources such as [11-15] and they Basin (RRB) and regulates about twenty percent
revealed that there will be notable changes in of the Bakun catchment. The project location
river inflows during the 21st century. map is as shown in Fig. 1.The plant generates a
Stream flow into a reservoir play a vital role for maximum output of 944 MW with 4 vertical
hydroelectric projects as it is the fuel to run the Francis turbines (each of 236 MW). The project
hydroelectric plants. Climate change results in has 141 m high Roller Compacted Concrete
temperature and precipitation change (i.e. (RCC) gravity Dam as shown in Fig. 2. The dam
intensity and pattern) and it significantly impact has world tallest stepped spillway structure (54 m
on regional hydrological processes [16]. In this wide) with a total discharge capacity of 2160
m3/s. The project has a surface powerhouse with four radial gates to discharge flood water
connected through 5.5 km long twin tunnels. A during extreme flood events (as shown in Fig. 3).
catchment of 2750 km2 drains into the reservoir The Bakun HEP has a surface powerhouse
to provide water resources to hydroelectric plant. consists of eight Francis turbine each of 300 MW.
The catchment is covered with the tropical The Bakun reservoir is the largest fresh water
rainforest where precipitation in the form of storage reservoir in Malaysia with surface area of
rainfall drives the water resources system. about 640 km2 masl and storage capacity of
37,521 million m3 at full supply level of 228
masl. A catchment of 14,850 km2 drains into the
Bakun reservoir to supply inflow to the Bakun
HEP.
Fig. 3 Bakun Dam in Sarawak state of Malaysia

state of Malaysia
Data Description
Daily rainfall record of eleven rainfall stations (as
Fig. 1 Location of Bakun HEP and Murum HE shown in Fig. 1) for the period of 1976–2005 was
Catchments in Sarawak state of Malaysia acquired from Department of Irrigation and
Drainage (DID), Sarawak for this study. There
were some missing periods in rainfall records at
all stations and these were in filled using
multilevel regression method. For climate change
impact assessment, the precipitation ensembles
for three GCMs from the CMIP5 have been
selected in this study as listed in Table 1. The
precipitation time series for the controlled run
(1976-2005) and future run (2011-2100) of each
GCM were downloaded from Intergovernmental
Panel on Climate Change (IPCC) World Data
Center for Climate website; http://www.ipcc-
data.org/sim/gcm_monthly/AR5/Reference-
Archive.html. Only one ensemble member
r1i1p1 was used in this study. These three models
were declared satisfactory among the others
Fig. 2 Murum Dam in Sarawak state of during evaluation over the Rajang River Basin
Malaysia and therefore selected in this study. The mean
ensemble of these three selected models
Bakun HEP (B3MMM) was also developed for the control
Bakun HEP is the largest hydroelectric plant in and future period to assess the climate change
Malaysia with total installed capacity of 2,400 impact in this study.
MW. The plant commissioned in 2012 and
supply power to the Sarawak state grid. The Table 1 Selected GCMS from CMIP5 used in
Bakun Dam is 205 m high concrete face rockfill the study
dam (CFRD) and a side channel gated spillway
CMIP5 Institute Modeling Countr Hydrological modeling using Aquarius
Model Group y Forecast
ACCESS CSIRO- Commonwe Austral A rainfall-runoff model was developed using the
1.0 BOM alth ia Australian Water Balance Model (AWBM)
Scientific approach. Aquarius Forecast tool was used to
and
develop the hydrologic model for the Bakun
Industrial
Research catchment. As the Murum HEP lies in the Bakun
Organization Dam catchment and control about twenty percent
(CSIRO) and of the Bakun catchment. The river inflows at
Bureau of Bakun Dam are available for the period of 2003
Meteorology to 2007, therefore a hydrologic model was
(BOM), calibrated at the Bakun Dam and after
Australia calibration; the model was run for Murum
ACCESS CSIRO- Commonwe Austral catchment to simulate the daily inflow at Murum
1.3 BOM alth ia Dam. The meteorological model was established
Scientific
using the Thiessen polygon gauge weight method
and
Industrial
for precipitation calculation. During the model
Research calibration, the simulated daily inflows were
Organization compared (as shown in Fig.. 4) with the observed
(CSIRO) and daily inflow using the coefficient of
Bureau of determination (R2), percent deviation (D), and
Meteorology Nash-Sutcliffe efficiency (E). The model’s
(BOM), performance parameters—R2, D, and E were
Australia calculated using the following equations:
GFDL- NOAAGF Geophysical USA
ESM2M DL Fluid ∑((𝑄𝑜𝑏𝑠 − 𝑄̅𝑜𝑏𝑠 )−(𝑄𝑠𝑖𝑚 − 𝑄̅𝑠𝑖𝑚 ))
Dynamics 𝑅2 = (1)
√∑((𝑄𝑜𝑏𝑠 − 𝑄̅𝑜𝑏𝑠 )2 −(𝑄𝑠𝑖𝑚 − 𝑄̅𝑠𝑖𝑚 )2 )
Laboratory,
USA ∑(𝑄𝑠𝑖𝑚 )−(𝑄𝑜𝑏𝑠 )
𝐷 = 100 × ∑ 𝑄𝑜𝑏𝑠
(2)
Methodology
Downscaling future precipitation ∑(𝑄𝑠𝑖𝑚 −𝑄𝑜𝑏𝑠 )2
A weather generator approach was used to 𝐸 =1− ∑(𝑄𝑜𝑏𝑠 −𝑄̅𝑜𝑏𝑠 )2
(3)
simulate the future precipitation over each
selected rainfall station. A weather generator The value of R2 close to 1, value of D close to 0%
called WeaGETS developed by [17] was used in and the value of E close to 1 implies good
this study. Third order Markov chain was used calibration. During this study, the coefficient of
for the precipitation occurrence and gamma determination (R2) was noted as 0.68, percent
distribution to generate the precipitation amount deviation (D) as 0.03 %, and Nash-Sutcliffe
for the wet days. The historical daily efficiency (E) as 0.68. These results were
precipitation for the period of 1976-2005 was satisfactory and complement well with some
used as input to the weather generator to generate previous studies such as [4, 19-22]. All these
the future time series of daily precipitation. The studies simulated river inflow for the climate
detailed methodology of this method is provided change assessment with E ranging from 0.48 to
in [17]. The stochastically generated future 0.83 and R2 from 0.63 to 0.84. The performance
precipitation at each rainfall station in the of hydrological model during calibration is as
catchment was perturbed using the quantile shown in Fig. 4.
perturbation technique (a change factor After calibration of the hydrological model, the
technique) employing the controlled (historical) long term daily inflow were simulated at the
and future runs of mean ensemble of three GCMs Murum Dam and also for the eighty percent of
under RCP4.5 and RCP8.5. During this study, the unregulated catchment of Bakun Dam for the
Climate Perturbation Tool developed by [18] was historical period pf 1976–2005 as well as for the
used for perturbation of daily precipitation at three future period of 2011–2040 (2020's), 2041–
each station. 2070 (2050's) and 2071–2100 (2080's). And then
the inflow at Bakun Dam were derived from the
following equation;
𝑄𝐵𝑎𝑘𝑢𝑛 = 𝑄𝑢𝑛𝑟𝑒𝑔𝑢𝑙𝑎𝑡𝑒𝑑 + 𝑄𝑟𝑒𝑔𝑢𝑙𝑎𝑡𝑒𝑑 𝑀𝑢𝑟𝑢𝑚
(4)
where QBakun is total inflow at Bakun, Qunregulated is Results and discussion
inflow from the unregulated Bakun catchment Projected river inflows under future
and Qregulated Murum is total outflow from the Murum precipitation scenarios
HEP in the result of reservoir operation. The future river inflows were simulated under the
projected precipitation for both climate scenarios
5000 i.e. RCP4.5 and RCP8.5; and were compared
4500 observed flow with the historical inflows as shown in Table 2. It
was noted that the mean inflow at Murum Dam
4000
would remain same and increase by 4% during
daily inflow (m3/s)
3500 2020's; decrease by 2% and increase by 3%

3000 during 2050's; and increase up to 4% and 11%
2500 during 2080's under RCP4.5 and RCP8.5
2000
respectively. It was noted that the mean inflow at
Bakun Dam would remain same and increase by
1500
4% during 2020's; decrease by 1% and increase
1000 by 3% during 2050's; and increase up to 4% and
500 11% during 2080's under RCP4.5 and RCP8.5
0 respectively.
The mean seasonal inflow was also calculated at
1/1/2003
1/1/2004
1/1/2005
1/1/2006
2/1/2007
2/1/2008
each hydroelectric plant and their variability in

the context of future climate changes were
assessed under the both RCPs. It was noted that
Fig. 4 Observed and simulated inflows at Bakun
under RCP4.5, there would be reduction in DJF
Dam during the model calibration inflow at Murum HEP and Bakun HEP during
2050's and 2080's. Under RCP8.5 scenario,
Reservoir operation considerably increase in MAM inflow is noted at
A reservoir operation model was developed for Murum HEP and Bakun HEP in future. For the
the Murum reservoir and Bakun reservoir using JJA, both HEP would receive increase in JJA
the HEC-ResSim model. All project features inflow in future. The JJA is the period when these
including reservoir storage, surface area, river catchments yield lowest inflow in result of lowest
inflows, evaporation, dam height, spillway precipitation over the catchment. It is noted that
discharge rating, station’s installed capacity, the lowest flow during JJA would improve in
turbine rated discharge, turbine efficiency, future. MAM is the late North-West monsoon
environmental flow requirement etc. were period and receive the average rainfall over the
provided in the reservoir operation model. The upper RRB, but it is projected that the inflow
station dispatch rules were defined for the daily during this period would also improve in future.
power requirement and additional output to avoid But reduction in river inflow in noted during the
spillway discharge. The reservoir operation was DJF in the RRB, which indicates that there would
simulated using the historical inflows (1976– be a seasonal shift over the upper RRB in future
2005) and also for the future generated inflows due to potential climate changes which would
for all of three future periods under the both result in decrease in inflow during DJF and
RCPs. increase during MAM.
Table 2 Changes in seasonal availability of Table 3: Projected changes in future

water resources at each Dam hydropower energy
RCP4.5 RCP8.5
RCP4.5 RCP8.5 202 205 208 202 205 208
0's 0's 0's 0's 0's 0's
202 205 208 202 205 208
0's 0's 0's 0's 0's 0's Murum HEP
Murum Ann
HEP 1 -1 5 5 4 12 ual 0 -2 4 4 3 11
Bakun DJF 3 -9 -9 -1 -3 5
HEP 1 -1 4 4 4 7 MA
M -4 0 4 10 2 12
JJA 2 4 15 7 15 17 projected to have 7% increases in future energy
SON 1 -4 6 2 -3 11 production under RCP8.5.
The hydrological models are commonly being
Bakun HEP
developed using the current hydrological
Annu
al 0 -1 4 4 3 11
behavior and catchment characteristics. The
model parameters are calibrated with the
DJF 2 -7 -8 -1 -1 5
observed historical inflows. When these models
MA
are employed to simulate the future inflows, it is
M -5 1 2 7 3 10
assumed that the catchment characteristics and
JJA 2 2 14 7 13 17 hydrological behavior of the river will be
SON -1 -3 5 1 -1 10 stationary in future. Even though the Sarawak
state of Malaysia is less populated and
Projected changes in hydropower energy anthropogenic impacts are comparatively lesser,
It is noted that the annual energy from Murum but the deforestation is being witness in the upper
HEP would increase by 1% and 5% during Rajang river basin. These activities would alter
2020's, decrease by 1% and increase by 4% the hydrological behavior of these catchments in
during 2050's and increase by 5% and 12% future and there is a need to investigate the future
during 2080's under RCP4.5 and RCP8.5 changes in the catchment characteristics/ land use
respectively. Significant increase in annual changes and their impact on river water yield.
energy is noted during 2080's under RCP8.5 (as
shown in Table 3) due to projected increase in Acknowledgment
water resources availability under RCP8.5 of This study was funded by Sarawak Energy
2080's. Berhad, Malaysia under a climate change
For the Bakun HEP, it is also projected that the research project. The authors would like to
annual energy would increase by 1% and 4% express thanks to Department of Irrigation and
during 2020's, decrease by 1% and increase by Drainage, Sarawak (DID) for providing the
4% during 2050's and increase by 4% and 7% relevant observed precipitation data.
during 2080's under RCP4.5 and RCP8.5
respectively. Significant increase in annual References
energy is noted during 2080's under RCP8.5 (as [1] M. A. Cole, R. J. R. Elliott, and E. Strobl,
shown in Table 3) due to projected increase in "Climate Change, Hydro-Dependency, and
water resources availability under RCP8.5 of the African Dam Boom," World
2080's. Development, vol. 60, pp. 84-98, 2014.
[2] M. Hassan, P. Du, S. Jia, W. Iqbal, R.
Conclusions and Recommendations Mahmood, and W. Ba, "An Assessment of
While assessing the climate change impact on the South Asian Summer Monsoon
future precipitation over study area, it was Variability for Present and Future
concluded that the annual Precipitation is Climatologies Using a High Resolution
expected to increase significantly over the river Regional Climate Model (RegCM4.3)
basin under RCP4.5 and RCP8.5. Similar trend is under the AR5 Scenarios," Atmosphere, vol.
projected for the seasonal precipitation, 6, pp. 1833-1857, 2015.
especially during the March to May, when the [3] IPCC, "Summary for Policymakers. In:
increase in precipitation is more robust for both Climate Change 2013: The Physical
scenarios. As these catchments are projected to Science Basis. Contribution of Working
receive increase in future precipitation especially Group I to the Fifth Assessment Report of
for the period of 2080s of RCP4.5 and 2050s and the Intergovernmental Panel on Climate
2080s of RCP8.5, therefore the water resources Change [Stocker, T.F., D. Qin, G.-K.
availability is also increasing during these Plattner, M. Tignor, S.K. Allen, J.
periods. The projected increase in precipitation Boschung, A. Nauels, Y. Xia, V. Bex and
under future climate provides water security to P.M. Midgley (eds.)]," Cambridge
the hydroelectric plants. The Murum HEP and University Press, Cambridge, United
Bakun HEP would expect higher increase in Kingdom and New York, NY, USA., 2013.
hydropower production under RCP8.5 compared [4] R. Mahmood, S. Jia, and M. Babel,
to RCP4.5 due to increase in projected inflows "Potential Impacts of Climate Change on
under the future climate. The Murum HEP is Water Resources in the Kunhar River Basin,
projected to have 12% and Bakun HEP is Pakistan," Water, vol. 8, p. 23, 2016.
[5] H. G. Hidalgo, J. A. Amador, E. J. Alfaro, [14] M. Hussain, S. Nadya, K. W. Yusof, and M.
and B. Quesada, "Hydrological climate R. Mustafa, "Potential impact of climate
change projections for Central America," change on inflows to the Batang Ai
Journal of Hydrology, vol. 495, pp. 94-112, reservoir, Malaysia," International Journal
2013. on Hydropower & Dams, vol. 24, pp. 44–
[6] C. Zhang, B. Zhang, W. Li, and M. Liu, 48, 2017.
"Response of streamflow to climate change [15] M. Hussain, K. W. Yusof, M. R. Mustafa,
and human activity in Xitiaoxi river basin in R. Mahmood, and J. Shaofeng, "Projected
China," Hydrological Processes, vol. 28, changes in temperature and precipitation in
pp. 43-50, 2014. Sarawak state of Malaysia for selected
[7] S. W. Kienzle, M. W. Nemeth, J. M. Byrne, CMIP5 climate scenarios," International
and R. J. MacDonald, "Simulating the Journal of Sustainable Development and
hydrological impacts of climate change in Planning, vol. 12, pp. 1299 - 1311, 2017.
the upper North Saskatchewan River basin, [16] D. Labat, Y. Goddéris, J. L. Probst, and J.
Alberta, Canada," Journal of Hydrology, L. Guyot, "Evidence for global runoff
vol. 412-413, pp. 76-89, 2012. increase related to climate warming,"
[8] H. Chang and I.-W. Jung, "Spatial and Advances in Water Resources, vol. 27, pp.
temporal changes in runoff caused by 631-642, 2004.
climate change in a complex large river [17] J. Chen, F. P. Brissette, and R. Leconte, "A
basin in Oregon," Journal of Hydrology, daily stochastic weather generator for
vol. 388, pp. 186-207, 2010. preserving low-frequency of climate
[9] M. G. Grillakis, A. G. Koutroulis, and I. K. variability," Journal of Hydrology, vol. 388,
Tsanis, "Climate change impact on the pp. 480-490, 2010.
hydrology of Spencer Creek watershed in [18] V. Ntegeka, P. Baguis, E. Roulin, and P.
Southern Ontario, Canada," Journal of Willems, "Developing tailored climate
Hydrology, vol. 409, pp. 1-19, 2011. change scenarios for hydrological impact
[10] D. Khadka, M. S. Babel, S. Shrestha, and N. assessments," Journal of Hydrology, vol.
K. Tripathi, "Climate change impact on 508, pp. 307-321, 2014.
glacier and snow melt and runoff in [19] R. Meenu, S. Rehana, and P. P. Mujumdar,
Tamakoshi basin in the Hindu Kush "Assessment of hydrologic impacts of
Himalayan (HKH) region," Journal of climate change in Tunga-Bhadra river
Hydrology, vol. 511, pp. 49-60, 2014. basin, India with HEC-HMS and SDSM,"
[11] M. Z. M. Amin, A. J. Shaaban, N. Ohara, M. Hydrological Processes, vol. 27, pp. 1572-
L. Kavvas, Z. Q. Chen, S. Kure, et al., 1589, 2013.
"Climate change assessment of water [20] A. K. Verma, M. K. Jha, and R. K. Mahana,
resources in Sabah and Sarawak, Malaysia, "Evaluation of HEC-HMS and WEPP for
based on dynamically-downscaled GCM simulating watershed runoff using remote
projections using a regional hydroclimate sensing and geographical information
model," Journal of Hydrologic system," Paddy and Water Environment,
Engineering, vol. 21, 2016. vol. 8, pp. 131-144, 2010.
[12] M. L. Tan, D. L. Ficklin, A. L. Ibrahim, and [21] G. Yimer, A. Jonoski, and A. V. Griensven,
Z. Yusop, "Impacts and uncertainties of "Hydrological Response of a Catchment to
climate change on streamflow of the Johor Climate Change in the Upper Beles River
River Basin, Malaysia using a CMIP5 Basin, Upper Blue Nile, Ethiopia," Nile
General Circulation Model ensemble," Basin Water Eng. Sci. Mag., vol. 2, pp. 49-
Journal of Water and Climate Change, vol. 59, 2009.
5, p. 676, 2014. [22] A. García, A. Sainz, J. A. Revilla, C.
[13] M. R. Mustafa, M. Hussain, and K. W. Álvarez, J. A. Juanes, and A. Puente,
Yusof, "Climate change impact on water "Surface water resources assessment in
resources of Bakun hydroelectric plant in scarcely gauged basins in the north of
Sarawak, Malaysia," Proceedings of the 3rd Spain," Journal of Hydrology, vol. 356, pp.
International Conference on Civil, Offshore 312-326, 2008.
and Environmental Engineering, vol.
(ICCOEE 2016, Malaysia, 15-17 Aug
2016), pp. 283-287, 2016.
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Rainfall-Runoff Relationship for Small Watershed in Potohar Area

Moien Ahsan1*, Sonia Zafar2, Ghulam Nabi1
1
*
[email protected]
Abstract: Rainfall-Runoff relationship is essential for planning and design of soil and water conservation
structures, flood estimation and control. Runoff is a complex interaction between precipitation and landscape.
Water is one of the most important resource available and it is controlled by a natural phenomenon of
“Hydrological Cycle”. Hydrological modeling is a complex phenomenon in which a relationship is created
between the input and output parameters. The objective of study is to develop the rainfall runoff relationship for
the small watersheds in Potohar regions of Pakistan. In this study the detailed analysis of runoff mechanism was
done using Semi distributed approach (TOPMODEL) for three small watersheds (Dharabi, Domeli and Lehri
Dam watershed) in Potohar region of Pakistan. The model was simulated for Low, medium and peak flow events
to check the applicability of model. TOPMODEL showed satisfactory results for the existing three small dam’s
watersheds in Potohar area. Comparison of low, medium and peak flows were made between simulated and
observed. Model efficiency for Dharabi, Domeli and Lehri dam was 74%, 78 and 79 % respectively. Runoff
fraction was determined was 0.73. Rainfall runoff relationship was developed for the Potohar region y = 12359x
+ 81183, R² = 0.844 in which X=Rainfall (mm), Y=Volume (mm) The outcome of this study is quite helpful for
future work in water resources development techniques such as construction of small dams, ponds, soil and water
conservation structures for agriculture purposes.
Keywords: Rainfall Runoff Relationship, Runoff Coefficient, Semi-distributed approach
Introduction of catchment modeling technique causes a lot of

Rainfall-Runoff relationship is very important variations in the result, which may cause design
role in Planning, Development and management failure. In order to harvest rainwater, several
of water resources. Attempts have been made in water resources development schemes such as
the past to determine rainfall-runoff relationship construction of small dams, mini dams, ponds
in the for the Potohar area of Punjab, Pakistan and soil & water conservation structures should
catchment. Runoff occurs when parts of the be constructed to store the water for hydropower,
landscape are saturated or impervious. Runoff is agriculture, etc. During this study the detailed
a complex interaction between precipitation and analysis of runoff mechanism was done, and the
landscape factors. There are many factors which outcome is helpful for water resources
effects the peak and volume of flow generated. development techniques such as construction of
These may be meteorological and hydrological small dams, pounds, soil & water conservation
factors, land use and cover, topography, soil structures, and irrigation and agriculture
characteristics, etc have been defined for urban, purposes. The purpose of the study is the
rangeland, and agricultural drainages. There are estimation of runoff by Semi-distributed
different techniques which may be used for approach and development of Rainfall-Runoff
estimation of peak and volume of flow. There coefficient for Potohar area in the Punjab,
may be Empirical, Rational, Lumped, Distributed Pakistan. The study was also conducted to
and Semi-distributed approaches. establish the guidelines / Rainfall-Runoff
Rainfall-Runoff relationship is essential for relationship for runoff estimation for the soil and
design and planning of soil and water water conservation structures.
conservation structures, flood estimation and Haye and Youngs (2005) compared the design
control. Mostly empirical approach is used for characteristics of observed hydrological data of
such purpose, no comprehensive work has been eight small catchments in central Virginia whose
done for estimation of Rainfall-Runoff for area ranges from 2.5 to 52.7 acres. Rational
Potohar area of Indus basin. The land resources Method was used for estimation of peak
of Potohar area are characterized with discharge. Time of concentration and runoff
fragmented land holdings. An improper selection coefficients were also used in this study.
The temporal changes in spatial distribution of extensive flat areas. The height of Potohar
land cover, runoff coefficients, and runoff Plateau mainly varies from 300 to 600 meters
volume was estimated. A digital elevation model above sea level. Residual hills such as Kala
and runoff hydrograph was made for study area Chitta and Khairi Murat rise to just over 1000
which based upon above mentioned data meters and are the result of resistant rock being
information. From the land cover pattern of the left behind the erosion of resistant rock. A large
study area it was concluded that with the passage part of Plateau has been dissected and eroded by
of time urban area increases and vegetation area the action of running water. It presents a varied
decreases. Due to this change the runoff volume landscape. This dissected land termed as a
and peak flow increased. It was concluded that by badland topography and landscape of ridges,
increasing vegetation in catchment area floods ravines and troughs. The Soan river is the most
can be reduced Zhan (2005). dominant feature of the region. It forms gullies
Norbiato(2009) has studied 14 catchments whose and alluvial plains, mainly used for agriculture.
area ranges from 7.3 to 608.4 km2. All the The Potohar Plateau is rich in minerals like rock
catchments were mountainous and analysis the salt, gypsum, limestone, coal and oil.
effect of different factors e.g. flood type, Three small dams (Domeli, Dharabi and Lehri
geology, temperature, initial soil moisture Dam) were selected in the Potohar area to
conditions and land use on the distribution develop the rainfall runoff relationship and for
functions of the event runoff coefficients. Runoff the the estimation of runoff coefficient. TLehri
coefficients were determined from hourly and Domeli are located in the District Jhelum and
rainfall, and runoff data. Five hundred and thirty- Dharabi Dam is located in district Chakwal.
five (535) events were from 1989-2004 and these Catchment area and location of dams is shown
events were analyzed. The results of this study given in Table .The Total area of Potohar area is
indicate that the correlation between spatial 22,254 km². Elevation of the project area varies
distributions of runoff coefficients with mean between 2365 to 165 m, out of which elevation in
annual precipitation is approximately same. approximately 80 % area varies between 1000 to
Mean annual rainfall well explained the spatial 165 meters as shown in Fig. 4.
variability in average runoff coefficient value
which ranges from 0.40 to 0.48). Runoff Table 1 Catchment Area of Selected Small
coefficients for the micro catchments calculated Dams
by using the rainfall runoff relationship Sr Name Catchment District
(Sepaskhan and Fard; 2010). No area (Sq.km)
Rainfall runoff relationship developed for 1 Domli Jhelum
Prequeno River catchment, located in Brazil was 170
Dam
by using two semi-distributed approaches 2 Lehri Jhelum
(TOPMODEL and HYCYMODEL). The models 52
Dam
were applied series 3360 hourly observed data. 3 Dharabi Chakwal
The ratio between total runoff and total 196
dam
precipitation was 0.79, 0.81 and 0.74(Chaffe et
al; 2008).
Silva et al (2010) concluded that semi-distributed
(TOPMODEL) approach seems to more reliable
Hydrological approach to estimate the stream
flow discharge in large river basins. Monte Carlo
simulations were used to find the best set of
parameters in terms of Nash-Sutcliffe coefficient.
For the two decades runoff is likely to decrease,
so a good water management strategy will be
necessary as a possible adaptation measure stated
by Parinda; 2004.
Description of Study Area

Potohar Plateau Fig. 4 Study Area
The Potohar plateau is located in the south of Methods and Materials
Islamabad between the River Indus and River The required input data for the selected semi
Jehlum. It is an open undulating land with distributed technique (TOPMODEL) has been
used in this research is collected from different
organizations. the data required to the rainfall
runoff relationship was developed using
TOPMODEL. The data required for Semi
distributed approaches (TOPMODEL) is
Precipitation data of catchment, Stream flow data
of catchment, Climatic data, Topographic data,
Land cover/ land use data.
TOPMODEL has four basic input files that must
be completed to run the model. All these input
files are made in notepad program of windows.
(a) Catchment data file (b) Hydrological Input
data file (c) Topographic Index Map file (d) Text Fig. 2 Main screen of TOPMODEL
description file.
CROPWAT software is also used to estimate the 3. Give the path of the project files that were
evapotranspiration of the study area. The previously copied in the WINDOW’S drive
evapotranspiration is used in the hydrological as “C:\TmodVB123\projects”. The
input file as an input data. software will automatically detect all the
input files and their path will be shown in
TOPMODEL Parameters their respective places.
The basic five parameters for the TOPMODEL
are given below.
M; Parameter of the exponential transmissivity
function or recession curve. (Units of depth, m)
ln(To); Natural logarithm of the effective
transmissivity of the soil when just
saturated. A homogeneous soil throughout the
catchment is assumed. (Units of m2 /h)
Srmax; Soil profile storage available for
transpiration, i.e. an available water capacity.
(Units of depth, m)
Srinit; Initial storage deficit in the root zone (an Fig. 3 TOPMODEL Project Loading
initialisation parameter. (Units of depth, m)
ChVel; Effective surface routing parameter. 4. Press “Continue” button from Fig. and
(Units of m/h) next window will appears as shown in Fig.
Model Run and Flow Simulation
TOPMODEL is purely a hydrological model and
simulates hydrological affluxes of watershed.
The model is run for the three selected dams
(Domeli, Lehri and Dharabi) The importance of
the current version of is its runoff mechanism.
TOPMODEL works on the variable concept of
contributing area. Main screen of the
TOPMODEL is given in Fig. . The procedure that
was adapted to run the model is given below
1. First copy all the input files to the
WINDOW’S drive with the path Fig. 4 Control panel screen of TOPMODEL
“C:\TmodVB123\projects”
2. Open the TMOD9701 Program then main 5. Select “Hydrograph Simulation” option.
screen of TOPMODEL will open as shown A new window will open. Select “RUN” from
in Fig. 2 and load the file by selecting the this window. It will give the flow simulation
Load Project. hydrograph for the selected time interval.
6. Model calibration and validation was
done for the selected dams watershed.
Punjab, Pakistan and rainfall runoff relationship
Evaluation Criteria of TOPMODEL was developed.
The evaluation criteria of TOPMODEL are
model efficiency and model errors. Model Rainfall runoff relationship for the Dharabi
efficiency should closer to 100% and model Dam
errors closer to zero. The minimum acceptable 1) The TOPMODEL calibration was a time-
limit for modeling was that minimum modeling consuming process. Hundreds of trails
efficiency should be greater than 50%. were done for different values of model
Followings are the TOPMODEL evaluation parameters; finally, that set of parameter
criteria errors i.e SSE: Sum of squared residuals values was selected that gave the
over all time steps, SLE: Sum of squared log maximum efficiency with less error.
residuals over all time steps, SAE: Sum of TOPMODEL calibration is shown in Fig.
absolute errors over all time steps. 5 TOPMODEL Caliberation, Simulated Vs
Results and Discusions Observed flow and Rainfall-Dharabi dam
Semi-distributed (TOPMODEL) has been 2)
applied on the three small dams (Domeli, 3) and efficiency was obtained 74 % which is
Dharabi and Lahri dam) in the Potohar area of satisfactory.
Fig. 5 TOPMODEL Caliberation, Simulated Vs Observed flow and Rainfall-Dharabi dam
Fig. 6 TOPMMODEL validation-Dharabi Dam
TOPMODEL validation was also carried for estimating the future flows as well. The
out. The set of parameters showed the graph given below shows the observed and
maximum matching of the historical and simulated flows for the validated period.
simulated flows. Then the model for these Model validation is shown in Fig. .
set of parameter values was validated by 4) Fig. 5 TOPMODEL Caliberation,
extending the simulation time for next six Simulated Vs Observed flow and Rainfall-
months to ensure that the model is capable Dharabi dam
5) 21stAugust ,2010 and rainfall was 21.47
6) shows that first selected low event having mm. The volume calculated by
volume 26348 m3 was observed on 5th TOPMODEL was only 498089 m3with
June 2010 with rainfall 0.40 mm. The percentage difference of 6%. Similarly, for
volume calculated by TOPMODEL was the selected high flow events percentage
23348 m3with percentage difference of difference obtained was 12, 3, 8 and 13.
11%. Similarly, for the selected low flow Thus, it was concluded that TOPMODEL
events percentage difference was obtained also gave good simulation for high flow
to be 2, 6, 16 and 5. Hence, it was events.
concluded that the TOPMODEL gave 11) It was concluded from the TOPMODEL
good simulation for low flow events. simulation that this model estimates the
7) Fig. 5 TOPMODEL Caliberation, low, medium and high flows accurately.
Simulated Vs Observed flow and Rainfall- For any storage structure peak flow is
Dharabi dam important. A comparison of observed and
8) simulated flows was made for
9) shows that first selected medium event TOPMODEL. The observed flood peaks
having volume 130808 m3 was observed for the same year 2010 were compared
on 2nd July 2010 and rainfall was 4.33 with the simulated ones. The comparison
mm. The volume calculated by of different events simulated by
TOPMODEL was only 111260 m3with TOPMODEL with observed flows and the
percentage difference of 15%. Similarly, percentage difference are shown in Table
for the selected medium flow events 2.
percentage difference was obtained 10, -4 12) Similarly, TOPMODEL was calibrated
and 4. Hence, it was concluded and validated for the Domeli and Lehri
TOPMODEL also gave good simulation dam.
for medium flow events. .
10) The first selected high flow event having
volume 530778 m3 was observed on
Table 2 Comparison of Observed flows, TOPMODEL simulated flows and Percentage Errors
Simulated by
Date Observed(m3) %age difference Events
TOPMODEL
6/5/2010 26348 23348 11
5/7/2010 50981 49869 2
7/20/2010 53999 50999 6 Low
6/8/2010 61624 51624 16
12/30/2010 95649 90742 5
7/2/2010 130808 111260 15
7/28/2010 230586 201586 10
Medium
6/29/2010 285027 296035 -4
2/8/2010 335786 320948 4
8/21/2010 530778 498089 6
8/10/2010 565900 498634 12
8/24/2010 621888 601888 3 High
7/20/2010 722166 662166 8
7/29/2010 2054988 1789669 13
Fig. 5 TOPMODEL simulated, Observed flow
and Percentage errors
Determination of rainfall runoff fraction /

coefficient
Rainfall runoff fractions were determined for
Dharabi, Domeli and Lehri dam watersheds.
Runoff fraction was estimated by comparing the
simulated and observed runoff of selected events.
In order to obtain a single value of runoff fraction
for Dharabi, Domeli and Lehri watershed
average of runoff fractions of all selected events
was taken as 0.72, 0.73 and 0.74, respectively.
Runoff fraction for the Potohar area was obtained
to be 0.73. Fig. 6 Rainfall runoff relationship
Development of Rainfall runoff relationship Conclusions

Rainfall-Runoff relationship for runoff • According to the formulation of model, it is
estimation for soil and water conservation not suitable for watersheds with long dry
structures. Rainfall runoff relationship developed season resulting in very low flows.
for the three dams is shown in TOPMODEL simulates efficiently for low,
medium and peaks flows with lesser dry
span.
• TOPMODEL is more suitable for small
watersheds (Dharabi, Domeli Dam and Lehri
Dam) as it has shown very good simulation
results with less errors.
• Flow simulations by using TOPMODEL
give efficiency upto 74%.
• Rainfall runoff fraction was obtained to be
0.73. (Appendix C)
• Rainfall runoff relationship was obtained
y = 0.196x + 1.288
R² = 0.844, y=Volume (mm), x=Rainfall
(mm)
Fig. 6. • Only four (4) years data was available so
y=0.196x+1.288, R² = 0.844 rainfall runoff Fraction and rainfall runoff
x= Rainfall(mm), y=Runoff(mm) relationship can be verified using long time
step data.
• As the topography of Dharabi catchment is
mountainous so distributed approach by
using GIS gridded based data is highly
recommended.
Reference
1. Hayes D. C and Young. R. L. 2005.
“Comparison of Peak Discharge and Runoff
Characteristic Estimates from the Rational
Method to Field.
2. Zhan. Y. 2005. “Remote Sensing and GIS in
Runoff Coefficient Estimation in Binjiang
Basin” Geo science and Remote Sensing
Symposium, IGRSS, 05. Proceeding, 2005
IEEE International, volume 6, Page
No:4403-4406.
3. Norbiato, D., Borga M., Merz R., Blöschl G., hydrological models” An Interdisciplinary
and Carton A. 2009. “Controls on event Journal of Applied Science, Vol. 3.
runoff coefficients in the eastern Italian 6. Silva R V D; Yamashiki Y; Tatsumi K and
Alps” Journal of Hydrology 375 (2009) 312– Takara K. 2010.” Distributed TOPMODEL
325. approach for rainfall runoff
4. Sepaskhan A.R and fard S R M. 2010. routingmodeling’s in large scale basins”
“Determination of Rainfall-runoff Department of environmental Engineering,
relationship based on soil physical properties Kyoto University.
for use in microcatchment for water 7. ParidaB.P(2004) “Forecasting impact of
harvesting system design” Journel of Science climate change on runoff coefficients in
and Technology,2010, vol 34, issue 4, Limpopo basin using Artificial Neural
August 010. Shrazuniversity Iran. Network”. University of Botswana,
5. Chaffe, P. L. B., Silva. R. V., and kobiyama. Gaborone. Welcome to the Second
M. 2008. “Rainfall-runoff process analysis of AiaccWrkshop, Dakar. Af_42 Research
theePequeno River catchment, Curitiba Team Botswana March 24 -27, 2004.
metropolitan region, Brazil, with two
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Selection of Best Diversion Facility for the Construction of Kohala

Hydropower Dam
Muhammad Yaseen1*, Muhammad Naveed1, Muhammad Kaleem Sarwar2, Muhammad Hassan
Rehman2, Muhammad Imran Azam3
1*
Centre for Integrated Mountain Research, University of the Punjab, Lahore, Pakistan; Email:
2
Lahore 54890
3
College of Hydraulic and Environmental Engineering, China, Three Gorges University, Yichang
443002, China
*
[email protected]
Abstract: Diversion arrangement plays vital role in construction schedule planning and overall safety of the
hydropower project in rivers where frequent flash floods and perennial flows of sufficient magnitude can occur.
Failure of a diversion scheme results in major damages to the project in terms of manpower, machinery, loss of
time and increase in project cost. The main objective of this research work was to study the possible diversion
facilities for the construction of Kohala Hydropower Dam and to recommend the best solution for the diversion
facility. The project area lies in Muzaffarabad district of Azad Jammu Kashmir (AJK). Different diversion
alternatives which applied to this project were diversion through temporary flume and pipe line, diversion through
conduits through or under the body of dam, diversion through multiple stages, diversion through impounding of
flow behind the dam, diversion through temporary channel through the dam and diversion through tunnel through
the abutments. The inflow 4765 m3/s in 100 years return period and 575 m3/s in 30 years return period were
selected as the river diversion criteria during the first dry season. By considering the hydrological, topographical
and geological conditions of site, provision of diversion tunnel through the abutment with u/s and d/s coffer dams
is an optimum solution for the Kohala Hydropower Dam Project. The concrete lined diversion tunnel with 7.5 m
diameter on right bank is a better choice. The invert level of the tunnel is at 857m with bell mouth transition at
entrance for smooth entry of flow. The crest levels of U/S and two D/S cofferdams were 869m, 859.5m and 856m
respectively determined.
Keywords: Hydropower; coffer dam; diversion tunnel; design flood; return period
Introduction scheme can minimize potential flood damage to

Shortage of energy is becoming an obstacle in the work in progress and overall project
economic growth and development of Pakistan. completion period. The 10 or 25-year frequency
The country has sufficient surface water but flood is generally selected as the design flood for
unfortunately, Pakistan’s energy market diversion scheme based on the previous analysis
investment in hydel-power generation has been or past experience for sizing the diversion works
caught up in confusion and no significant and risk factor. Thus, diversion scheme need to
progress has been achieved so far. The be capable of diverting the peak flow of selected
Government is trying to facilitate private frequency flood. Any higher flood is liable to
investors to promote hydel power generation in overtop the cofferdam leading to its complete
the country to meet the ever-increasing demand failure / destruction along with large damage to
of electricity in a cost-effective manner. the under construction works and delays in
Construction of the hydro-power plant is project completion. The choice of diversion
normally done in the river stream. For creating through cofferdam, diversion channel or
dry working conditions, the river flow is diversion tunnel also depend upon the geological
temporarily diverted by construction of a conditions of the area, if the geological
diversion weir or diversion dam across the main conditions are not favorable for off channel
river course. Size and type of the diversion diversion, part of the main stream/river can also
structure depends upon the flood characteristics be used as a diversion channel for phased
of the river and topographical conditions at the construction. The main objectives of this
diversion site. It can be a diversion tunnel if the research work are to study the possible diversion
topography is favorable. A planed diversion facilities for the construction of Kohala
Hydropower Dam and to recommend the best about 140 Km from Rawalpindi / Islamabad.
solution for the diversion facility. Muzaffarabad town, capital of AJK (Azad
Jammu Kashmir), lies further about 35 km from
Materials and Methods Kohala Bridge. Chakothi and Kohala villages are
Study Area situated on upper and lower limbs of the river,
Selection of best diversion works for Kohala respectively. The dam site will be located on the
Hydropower project will be carried out through upper limb and power house on the lower limb of
this study. The project area lies in Muzaffarabad Jehlum River. The selected layout of Kohala
district of AJK. The area of interest for the study Hydropower project comprises a concrete dam at
lies between Chakothi and about three kilometers Siran, about 18 km long transfer tunnel passing
downstream of Kohala bridge on the jehlum river under Mir Fateh and an underground power
between Longitude 73°25’ to 73°50’ East and house near Barsala.
Longitude 34°00’ to 34°25’ North. The bridge is
Fig. 1 Project Location Map
The Siran Dam will be designed for a full supply monsoon, the flow in the tunnel is pressure flow
of 900masl creating a reservoir of approximately in 100 years return period.
14.9 million m³ storage. The minimum draw
down level has been taken as 890masl. The live i. Discharge capacity in river cut and first
operating storage between full supply level and dry season
minimum draw down level is 6million m³. The The discharge capacity in river cut and first dry
location of Kohala Hydropower project is shown season can be determined by:
in Fig. 1. Q=Cb(2g)0.5He1.5
Where:
a. Discharge capacity of Diversion Tunnel Q, discharge, m3/s; C, comprehensive discharge
The discharge capacity of diversion tunnel is coefficient, 0.34; b, width of tunnel, m; g, gravity
different in the first dry season and the second acceleration, m/s2; He, actual head being
monsoon. In river cut and the first dry season, the considered on the invert of tunnel, including
flow in the tunnel is free discharge. In the second velocity of approach head, m
ii. Discharge capacity in second monsoon and the topographical limitation does not permit
The discharge capacity in second monsoon is to avail this option.
determined by:
Q=C(2gH)0.5 ii. Diversion through Conduits or Under the
Where: Body of Main Dam
Q, discharge, m3/s; C, comprehensive discharge The outlet works for an embankment dam often
coefficient; g, gravity acceleration, m/s2; H── entails the construction of conduit that may be
head, m used for diversion during construction of the
dam. This method for handling the diversion
b. Head Losses in the Tunnel flows is an economical one, especially if the
After selection of appropriate diameter head conduit for the outlet works is large enough to
losses through the tunnel are calculated. carry the diversion flows. As the topographic
Following are the different head losses in the data shows that the slope of terrain steep as
tunnel. shown in Fig. 3 and the water carrying channel is
• Entry Losses (hen) narrow where the dam is to be constructed, hence
• Friction Losses (hf) the conduit through or under the body of dam is
• Bending Losses (Hb) not recommended.
• Exit Losses (hex)
• Total head losses (Hl) iii. Diversion through Multiple Stages
hen = (1 / c2 - 1) * ( v2 / 2 g ) The multiple-stage method of diversion over the
top of alternate low construction blocks or
2 2/3 2
hf = 29.1 n ( L / r ) v / 2 g through diversion conduits in a concrete dam
requires shifting of the cofferdam from one side
2
Hb = Kb * ( v / 2 g) of the river to the other during construction.
During the first stage, the flow is restricted to one
hex =1.0 * V2 / 2 g portion of the stream channel while the dam is
constructed to a safe elevation in the remainder
Hl = hen + hf + Hb + hex of the channel. In the second stage, the cofferdam
is shifted, and the stream is carried over low
RESULTS AND DISCUSSIONS blocks or through diversion conduits in the
a. Hydrological Data Analysis constructed section of the dam while work
i. Inflow in First Dry Season proceeds on the unconstructed portion. The dam
The maximum monthly inflows and their return is then carried to its ultimate height, with
periods in dry season in 35 recorded years are diversion finally being made through the
shown in Fig. 2. The dry season is from October spillway, penstock, or permanent outlets.
to February, a total of 5 months. According to the topographic data river section is
narrow and slope is steep and also flow is very
high which is 575 m3/sec, therefore construction
of dam cannot be made in multiple stages. So,
this option is unsuitable to divert the peak flow
during the construction of dam keeping in view
the safety.
iv. Diversion through Impounding of Flow

behind the Dam
Fig. 2 Flood Flow and Return Period in The flow may be impounded behind the dam
Dry Season during its construction period. Pumps are used, if
b. Possible Diversion Facilities for Kohala necessary, to control the water surface. This
Hydropower Dam Construction option is suitable for those dams’ areas where
i. Diversion through Temporary Flume and flow is small and wider river cross section is
Pipe Line available for storage of water during of
The flow may be bypassed by the installation of construction of main dam. As the flow is 575
a temporary flume or pipeline. As the flow is 575 m3/sec for 30 years return period. So, need a high
m3/sec with 30 years return period which is so storage reservoir behind the main dam. The
high for flume. For this discharge large cross topographic of dam site clearly portray that space
section and more cost for structure is required is not available to store this large amount of
incoming flow and also cost will be very high to off by a gate or a bulkhead, and a concrete plug
construct this large impounded water body. If we is placed in the intake structure for permanent
make arrangement for pumping of water, it will closure. In addition, gates and second-stage
further increase the cost. So, this option also not concrete are added in the gate chamber and
suitable due to topographic, hydrological and control structure, where applicable, to complete
economic point of view. the outlet works for permanent operation.
Temporary diversion tunnels that are not part of
a spillway or outlet works can be lined or unlined.
The advisability of lining the diversion tunnel is
influenced by
• The cost of a lined tunnel compared with that
of a larger unlined tunnel of equal carrying
capacity;
• The nature of the rock in the tunnel,
specifically, whether it can stand
unsupported and unprotected during the
Fig. 3 River view for proposed site of Kohala passage of the diversion flows.
Hydropower Project. • The permeability of the material through
which the tunnel is carried, because it affects
v. Diversion through Temporary Channel the amount of leakage through or around the
through the Dam abutment.
At sites where it may not be economical to Some means of shutting off the diversion flows
provide a tunnel or conduit large enough to pass must be provided. This can be accomplished
the diversion design flood, a temporary channel through the use of closure devices such as bulk
involving a gap through the embankment dam heads, slide gates, or stop logs. Permanent
may be used to divert stream flows while the closure of the diversion tunnel is made by placing
remainder of the embankment is being a concrete plug in the tunnel. Keyways may be
constructed. This method is adaptable where excavated into the rock to ensure adequate shear
wider river cross section is available; obviously resistance between the plug and the rock or
it cannot be used in narrow canyons. This lining. After the plug has been placed and
diversion facility is not possible because of sufficient time has elapsed for concrete shrinkage
narrow river cross section and high inflow. grout is forced in through previously installed
grout connections to the contact between the plug
vi. Diversion through Tunnel through the and the surrounding rock to ensure a watertight
Abutments joint. Keeping in all views of hydrological,
It is usually not feasible to do a significant topographical, geological and economical
amount of foundation work in a narrow canyon concern, this option of providing diversion tunnel
until the stream is diverted. In this situation the through the body of abutment is most suitable for
tunnel may prove the most feasible means for the Kohala Hydropower Dam Project.
diversion, either for a concrete dam or for an
embankment dam. The stream flow is bypassed c. River Diversion Arrangement for Dam
around the construction area through tunnels in Construction
one or both abutments. If tunnel spillways or i. Dam Construction Activity
tunnel outlet works are designed, it Usually The headrace tunnel will be 16 km long and its
proves economical to use them in the diversion construction will occupy key route, i.e., the total
plan. If the upstream portion of the permanent construction schedule will be controlled by
tunnel is above the river bed elevation, a construction of headrace tunnel. Thus, the dam
temporary upstream diversion adit can be construction can be stopped during the first
provided to affect a stream-level bypass. When monsoon and resume construction in the second
there is a river outlet works tunnel, particularly dry season. It will not increase the total time of
for embankment dams, it is generally used for the project construction schedule. In the end of
diversion. Normally the diversion works tunnel is the second dry season, the dam height can be
placed at an elevation near river level. When a placed to the elevation higher than 882m based
tower or drop inlet intake is used, a temporary on the rate of concrete placement. The elevation
adit upstream of the base of the intake structure of 882m is the elevation of spillway weir. In the
is necessary. After diversion, this adit is closed second monsoon, the dam should retain the flood
in the reservoir, prevent the flood from shortened. In consideration of dam concrete
overtopping over its crest and the concrete placement, cost of river closure and the ability of
placement can be continued. The flood will be contractor for river cut construction, the inflow
discharged through diversion tunnel two sluices for river cut can be selected between 100m3/s and
and spillway. 200m3/s. In October, the maximum inflow is
581.50m3/s in 35 years (from 1970 to 2004). In
Diversion Criteria 35 years recorded inflow, there are 71 days of
The diversion criterion is different for the first inflow more than 150m3/s from 11th to 20th in
dry season and the second monsoon. It depends October. The probability of daily inflow less than
on the dry season, monsoon and concrete 150m3/s is 93.5%; there are 7 years of inflow
placement for dam. more than 150m3/s from 11th to 20th in October.
The probability of inflow less than150m3/s in
First dry season October yearly is 80%. Based on the above
If concrete placement can be stopped and the dam analysis, the inflow of 150m3/s can be selected
is allowed overtopping during the first monsoon, for river cut. The time of river cut can be arranged
the diverted flood will be small during the first from 11th to 20th in October.
dry season. But, during the first dry season, the
flood should not overtop the crest of upstream d. Hydraulic Calculations for Diversion
cofferdam. If there is a flood warning system, Tunnel and Crest Elevation of Coffer
even in first monsoon, when flood is smaller than Dams
the discharge capacity of diversion tunnel, the First Dry Season
concrete placement can be continued. The diverting inflow is 575 m3/s in first dry
season. One upstream cofferdam and two
Second monsoon downstream cofferdams will be built. The
In second monsoon, the dam will be placed to calculations are to determine the size of diversion
certain level, for example, 882m. The dam height tunnel and the crest elevation of upstream and
is 36m, and overtopping is not allowed. But the downstream cofferdams. The calculation also
flood can pass through the diversion tunnel, considers that the discharge capacity of diversion
sluices and spillway. The flip buckets of all tunnel is enough for the first monsoon in order to
sluices and the downstream surfaces of spillway avoid flood overtopping the main dam.
should be formed in the second dry season, which
is not difficult. In this case, the return period of Optimization of tunnel size
flood control should be raised to 100 years in all The diverting inflow for the first dry season is
year. If the inflow is smaller than the flood of 100 575m3/s. The calculations are carried into
years return period, the concrete placement can comparison among the elevation of inlet invert,
continue. In Later monsoon, the flood control dimensions of cross section, velocity in the
will be still the flood of 100 years return period. tunnel, height of upstream cofferdam. The results
are given in Table 1.
Selection of Inflow in First Dry Season for
River Diversion From Table 1, the higher elevation of inlet invert
Two main factors will affect the selection of is, the smaller the section area of the tunnel
inflow in first dry season: relative lower cost and becomes. But the average flow velocity in the
higher resistance to flood risk. If a smaller return tunnel will increase. Because the diversion tunnel
period is selected, the cost of river diversion will is fully concrete lined, the velocity in the tunnel
be low, but the risk of overtopping upstream can be between 15m/s and 18m/s. The flow
cofferdam is higher. The upstream cofferdam velocity in the tunnel is not more than 15 m/s.
built of rockfill is not allowed overtopping during The maximum flow velocity is determined by
first dry season. From Fig. 2, the maximum 4765m3/s in 100 years return period. In this case,
inflow in 10, 20 and 30 years return period is 4765m3/s will be discharged through diversion
separately about 450m3/s, 540m3/s and 575m3/s. tunnel, sluices and spillway, and the flow
velocity in tunnel will not more than 18 m/s.
River Cut Thus, the option of the elevation of inlet invert at
The river cut will be performed after monsoon 857m and with 7.5m width is an optimum
and in early dry season. If it is too early, it will selection from the point views of economy,
cost more. If it is too late, the duration of concrete height to width and velocity in tunnel. The
placement in the first dry season will be proportion of height to width should not be too
high, or cost will be higher. In light of the axis of section 4.3.4, the flow is 150 m3/s. The levels of
part diversion tunnel parallel to the rock strike or cofferdam are given in Table 2.
with small angle to the rock strike, the smaller
tunnel width will be a better selection. RCC coffer dam
The tunnel length from the entrance to the exit The height of downstream RCC cofferdam is
portal is 320m. The slope of the tunnel is 1.614%. determined by diverting inflow of 575m3/s. The
In some area of developed joints and broken rock, corresponding downstream water level is
shotcrete, rockbolts, meshes and consolidation 858.70m. By adding 0.80m safety allowance, the
grouting will be applied. Excavation slopes at crest elevation is 859.50m.
tunnel entrance and tailrace channels will be In first monsoon, the upstream cofferdam will be
applied with shotcrete, rockbolts and meshes. flushed away. The space between Main dam and
cofferdam will be no use for dam concrete
Crest elevation of upstream cofferdam placement after the first dry season. But Dam
The upstream cofferdam is only for dam construction needs space between dam and
construction during the first dry season. It will be cofferdam for delivery of steel bars, forms and
flushed away in first monsoon. The crest other facilities and construction materials. The
elevation is 869.00m. During river cut, the two space between dam and downstream cofferdam
small rockfill cofferdams will be built first. After can be used for this purpose. Thus, the
the river is closed, and the inflow is diverting by downstream cofferdam should be an overflow
diversion tunnel, the clay core can be placed. cofferdam constructed with RCC. In order to
Based on the inflow of 150 m3/s during river cut, building RCC cofferdam, a rock fill dam will be
the reservoir level is 858.06m. built downstream of RCC cofferdam. The last
downstream cofferdam is built of rock fill. The
Crest elevation of downstream cofferdams crest elevation is 856.00 m. The cofferdam and
There are two downstream cofferdams. The its gravel foundation will be grouted with spin
height of the last downstream cofferdam is injection for waterproof. The elevation of river
determined by the inflow when river is cut. In floor is 852m.
Table 1 Hydraulic Calculation for Dimension of Diversion Tunnel
Elevation of Tunnel Cofferdam Cost

Flow Velocity Height: Width
Inlet Invert Height Height
M m M m/s US$
Tunnel Width=8.0m
854 10.4 16.6 9.65 1.30 3487171
855 9.1 16.6 11.42 1.14 3104971
856 8.3 16.6 12.38 1.04 2869771
857 7.7 16.6 14.14 0.96 2693371
Tunnel Width=7.5m
854 10.95 15.9 9.8 1.46 3663659
855 9.55 15.9 11.54 1.27 3266759
856 8.65 15.9 12.94 1.15 3011609
857 8.05 15.9 14.11 1.08 2841509
Tunnel Width=7m
854 11.7 15.3 9.65 1.67 3872697
855 10.1 15.3 11.42 1.44 3435897
856 9.1 15.3 12.83 1.53 3162897
857 8.5 15.3 14.02 1.21 3162897
Table 2 crest levels of cofferdam. Downstream

Upstream Downstream Crest Elevation Cofferdam
Level Level of the Last m m m
859.20 854.80 856.00
lithotypes. Empirical and numerical methods
Second Monsoon were used for tunnel design. Single diversion
In the second monsoon, the diverting inflow is tunnel on the right abutment of dam is
4765m3/s. Before the monsoon comes, the recommended for diversion system as it consists
spillway chute should be formed for flood of good rock. Detail design of upstream and
discharge. The weir crest elevation of spillway is downstream cofferdams is required for cost
882.00m. The invert elevation of sluices is estimates of the diversion arrangements.
862.00m. Added with diversion tunnel, three
facilities will jointly discharge flood of 4765m3/s. References
By hydraulic calculation, the reservoir level is Benmebarek, N. (2005), “Numerical Studies of
885.93m. Based on the analysis of the rate of Seepage Failure of Sand within a Cofferdam”,
concrete placement, the spillway chute can be Biskra University, Biskra, Algeria, p 264-273.
formed before the second monsoon. It is
unnecessary to design extra temporary bottom Casagrande. A. (1973), “Embankment Dam
outlet in dam or increase the size of diversion Engineering” Jhon Wiley and Sons, Inc., USA, p
tunnel for diverting flood. 97-103.
Conclusions Davis, C.V and Sorensen, K. E. (1993), “Hand

The inflow is 4765 m3/s in 100 years return Book of Applied Hydraulics”, Third Edition,
period. It is suggesting that 575 m3/s in 30 years McGraw Hill, New York, p, 14-7
return period be selected as the river diversion
criteria during the first dry season. Keeping in Fell, et al. (1992), “Geotechnical Engineering of
views the hydrological, topographical and Embankment Dams”, A.A. Bakema, Rottedam,
geological conditions, provision of diversion Netherlands, p 450-455.
tunnel through the abutment is most suitable for
the Kohala Hydropower Dam Project. It is Gerodetti, M., (1981), “Model Studies of an
observed that at 575m3/sec design flood one Overtopped Rock fill Dam,” International Water
tunnels of 7.5 m diameter is required to the river Power &. Dam Construction, p. 25-31
flows during Kohala dam construction. The
option of the elevation of inlet invert at 857m Moiseev, S. N., (1975), “Passage of Flows
through rock-earth dams during construction”,
and with 7.5m width is a proper selection from Journal of Engineering, Springer, New York,
the point views of economy, height to width and Vol. 9, No. 1, PP 15-21.
velocity in tunnel. The top shape of the entrance
is an elliptical curve. Neghabat, F and Robert M. Stark, (1972), “A
To reduce the losses in diversion tunnel, the Cofferdam design optimization”, Journal of the
concrete lining of appropriate thickness is Mathematical Programming, Springer Berlin,
recommended. As Desilting Chambers are Heidelberg, Vol 3, No. 1, PP. 263-275.
proposed on left bank of river according to
Kohala Hydropower Project, Feasibility Study Novak, P, A.I.B. Moffat and C. Nulluri, (1990),
Report. Detail study is recommended after “Hydraulic structures” University of Newcastle
proposing and relocating of desilting chambers upon Tyne, Chapman and hall, London, p. 301-
for provision of Diversion tunnel on left bank. 302.
Detailed engineering geological, hydrological,
topographical studies should be carried out for USBR. (1987), “Design of Small Dams”, Water
the better design of Diversion Facility for the Resources Technical Publication, Third Edition,
Kohala hydro Power Project. Detailed Studies SSOP, Washington, D.C , p 491-496
should be carried out both in field and laboratory
like discontinuity surveying, Drilling and Yanmaz, A. M., (2000), “Overtopping Risk
sampling for laboratory testing. Laboratory Assessment in River Diversion Facility Design”,
studies should be performed on core specimens Canadian Journal of Civil Engineering, Vol. 27,
taken from the boreholes, in order to determine No: 2, PP. 319-326.
physical and mechanical properties of the main
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Hydropower Optimization Station using PSO and GA Techniques

Ijaz Ahmad1, Muhammad Zaman2,3*, Shouqi Yuan2, Liu Junping2, Muhammad Saifullah4
1
Center of Excellence in Water Resources Engineering, University of Engineering and Technology,
Lahore 54890, Pakistan.
2*
Research canter of Fluid Machinery Engineering & Technology, Jiangsu University, P.R. China
3
Department of Irrigation and Drainage, University of Agriculture, Faisalabad, Pakistan; Email:
4
Department of Structures & Environmental Engineering, University of Agriculture, Faisalabad,
Pakistan
*
[email protected]
Abstract: Optimization of Xin’anjaing hydropower is a complicated problem. This paper tried to get the optimal
hydropower generation using two commonly used methods GA and PSO. The model for the optimization is
developed and applied to the Xin’anjiang hydropower station. It is observed from the results that for the
optimization of the hydropower and for getting maximum benefits, when simple genetic algorithm techniques have
applied, it showed some drawbacks such as slow and pre-matures convergence during the process. The results
revealed that the particle swarm optimization gave more efficient and optimal results. Findings revealed that PSO
technique not only retain the advantages over GA but also showed the high-speed convergence, and search
accuracy. We can get more profitability by optimization of the Xin’anjiang hydropower station by the application
of this model.
Keywords: PSO, GA, PSO-GA, Optimization, Xin’anjiang, Hydropower generation
Introduction: methods GA and PSO. This paper discusses the

Water resources allocation and its appropriate and feasibility and effectiveness of GA and PSO
optimal use is becoming important and techniques and their comparisons in optimization
challenging for humans due to its shortage and of these hydropower stations.
climate change factors. Optimal allocation of
water resources is complicated and important for Mathematical model for xin’anjiang
water resources management and this work has hydropower scheduling
become mathematical model based [1]. After the The mathematical model consists of two-part,
increase of computer performance, some objective function and constraints. This model
algorithms were developed in 1970 for getting the takes water levels of the reservoirs as decision
optimal solution of daily life problems. Different variables represented by 𝐻𝑗𝑡 in equation 1 and
kind of programming approaches (linear, maximizations of the electricity is objective
nonlinear and dynamical) were developed and function over the 12 months periods.
applied for the solution of reservoir operation in The objective function is given below
the past. [2-7]. Many researchers utilized these
methods for the solution of different problems [8- E=Max=∑𝑇𝑗=1 ∑𝑀
𝑗=1 𝐴𝑗 𝑞𝑗𝑡 𝐻𝑗𝑡 𝛥𝑡 (1)
10]. Among of these programing approaches,
genetic algorithm is important one, which was The constraints in the mathematical model are
developed by Goldberg [11]. Genetic algorithm given as below
(GA) is also a global optimization technique used Water balance equation
in the optimization of water resources. [12-15].
𝑉𝑗,𝑡+1 = 𝑉𝑗𝑡 + (𝑄𝜆𝑗𝑡 −𝑄𝑗𝑡 )𝛥𝑡 (2)
Another important optimization technique which
Reservoirs discharge limits
is called Particle swarm optimization (PSO) were
𝑄𝑗𝑡.𝑚𝑖𝑛 ≤𝑄𝑗𝑡 ≤ 𝑄𝑗𝑡.𝑚𝑎𝑥 (3)
developed by Kennedy and Eberhart (1995) as an
optimization technique and it’s becoming one of Reservoirs storage volume limits
the most important algorithm in optimization 𝑉𝑗𝑡.𝑚𝑖𝑛 ≤𝑉𝑗𝑡 ≤ 𝑉𝑗𝑡.𝑚𝑎𝑥 (4)
problems.[16]. Optimization of Xin’anjaing and Hydropower station power generation limits
Fuchunjiang hydropower stations is a complicated 𝑁𝑗𝑡.𝑚𝑖𝑛 ≤ 𝐴𝑗 𝑞𝑗𝑡 𝐻𝑗𝑡 ≤ 𝑁𝑗𝑡.𝑚𝑎𝑥 (5)
problem. This paper tried to get the optimal T = total period count within a year, T=12
hydropower generation using two commonly used M = total number of reservoirs
𝐴𝑗 = Power generation coefficient 𝑣𝑖𝑡+1 = 𝑤 ∗ 𝑣𝑖𝑡 + 𝑐1 𝑟𝑎𝑛𝑑1 (𝑝𝑏𝑒𝑠𝑡𝑖 − 𝑘𝑖𝑡 ) +
E = maximum power generation output from 𝑐2 𝑟𝑎𝑛𝑑2 (𝑔𝑏𝑒𝑠𝑡𝑖 − 𝑘𝑖𝑡 ) (6)
hydropower 𝑘𝑖𝑡+1 =𝑘𝑖𝑡 + 𝑣𝑖𝑡+1 , (7)
𝑄𝜆𝑗𝑡 = Inflow of reservoir j at time period t, m3/s Where 𝑘 𝑚𝑖𝑛 𝑡+1
< 𝑘𝑖 <𝑘 𝑚𝑎𝑥
𝐻𝑗𝑡 = Average head of reservoirs j at time period
t, m .
𝑉𝑗,𝑡+1 = Volume of reservoir j at the end of time
period t
𝑄𝑗𝑡.𝑚𝑖𝑛 , 𝑄𝑗𝑡.𝑚𝑎𝑥 = minimum and maximum water
discharge of reservoir j at time period t, m3/s
𝑉𝑗𝑡.𝑚𝑖𝑛 ,𝑉𝑗𝑡.𝑚𝑎𝑥 = minimum and maximum volume
of reservoir j at time t
𝑁𝑗𝑡.𝑚𝑖𝑛 = Minimum hydropower generation
constraint of reservoir j at time period t
𝑁𝑗𝑡.𝑚𝑎𝑥 = Installed plant capacity kW
Suppose that inflow has obtained 𝑄𝜆𝑗𝑡 and the

relation of the water discharges follow equation 2
and if the storage capacity of the reservoir does not
satisfy the constraints (3) then there is need to
dispose the excessive flow. This will lead towards
a complicated and dynamical programming
problem.
Optimization techniques
a. Genetic algorithm
Genetic algorithm was developed by Goldberg
(1989) Basic difference between GA and other
methods are search of optimal results. GA can Fig. 1 flowchart of the GA
solve discrete, discontinuous and non-convex
problems without differentions due to search from Where i= (1,2…….population
whole population [11]. In GA search, during each size/swarm),t=number of reproduction steps,
iteration individuals are rated for their effective 𝑣𝑖𝑡 ::the velocity vector of particle i in the ith
evaluation and selection, crossover and mutation reproduction step; w=inertial weight,
operator are used to generate new population. The 𝑐1 , 𝑐2 :learning rates, , 𝑟𝑎𝑛𝑑1 , 𝑟𝑎𝑛𝑑2 :
evaluation loop is repeated until desire independent random variables from (0,1)
termination results reached as shown in Fig. 1. uniformly distributed, 𝑝𝑏𝑒𝑠𝑡𝑖 is best solution
produced by particle i, 𝑔𝑏𝑒𝑠𝑡𝑖 is the best solution
b. Particle swarm optimization produced by the swarm. Particle swarm
Kennedy and Eberhart (1995) developed particle optimization is easy and simple for
swarm optimization as a continuous application implementation, computationally efficient as
optimization technique. compared to other programming techniques
It has two parts in first phase particles are The PSO algorithm is summarized in the
randomly distributed within the search space that following steps
is called initialization while the second phase is i. The current position of the ki and velocity
evolutionary phase in which particles change and is initialized with [0 1] random number
adjust their position in search of optima by range.
following the most successful particles until the ii. Generate initial swarm k(0)
termination. Suppose particles are moving in a D iii. Generate Initial velocity v(0)
dimensional space with a velocity v. Particles iv. Set t=0 where t is generation number.
position is represented k j = k j1 , k j2 , k j3 ,k jD ) v. Repeat the procedure
while the velocity is denoted as vi = (vi1 , vi2 vi. Compute fitness value of each particle in
,…… viD ). The velocity and position after t+1 the swarm
time is given as vii. End the procedure again
viii. Compute PBest(i) and GBest
ix. Perform PSO operations
x. Compute v(t+1) from equation 6 electricity using optimal techniques than
xi. Compute k(t+1) from equation 7 conventional methods.
xii. Perform Elitist mutation
xiii. Set t=t+1 Table 1 Results with different generations
xiv. Repeat the whole procedure until the Gene
GA PSO
termination criteria reached Energy Execut Energy Execut
ratio
output (10 ion output (10 ion
n 8 8
Case study kWh) time(s) kWh) time(s)
The Xin’anjinag and Fuchunjiang hydropower 41.209 7.086 37.187 1.844
100
stations are located at the Xin’an River, Zhejiang
Province in east China. One hydropower station 200 41.291 13.504 37.187 2.499
has installed at Xin’anjiang reservoir which is 400 41.712 25.175 37.751 3.641
called Xin’anjinag hydropower station. The
Xin’anjiang Hydropower Plant has 9 power Table 2 Results of different algorithms
generating units with a total installed capacity of Convent
845,000 kW, generating annual output of 1.86 ional GA PSO
billion kWh. The Xin’anjiang Dam is first 105 method
meters high reinforced concrete dam in Chin with Ene Ene
466.5 billion m3 reservoirs capacity. Dam has a rgy rgy
Energy Execu Execu
maximum flood discharging capability of Year outp outp
output tion tion
14,000m3/s. While Fuchunjiang hydropower ut ut
(10 time(s time(s
station is a run off river plant and its electricity 8 (10 (10
kWh) 8 ) 8 )
generation is highly variable and dependent at the kW kW
h) h)
flow released from Xina’anjinag reservoir and
Rain
Lan River. 41.7 37.7
y 26.413 25.17 3.64
Based on the monthly inflow data of past 48 years 12 51
year
from Xin’anjiang and Fuchunjiang reservoirs, Aver
model is run and calculates the energy of rainy 33.3 30.1
age 18.614 25.85 3.69
year, average year and normal years by taking the 54 25
year
dead level as the initial water level of scheduling. Dry 25.0 23.5
13.075 25.64 3.69
Genetic algorithm run with different generation year 45 12
value=100,200 and 400 in order to know with avera 33.3 30.4
19.367 25.57 3.67
which generation value it will give better results. ge 69 62
Different parameters used in GA are described
here as: population size=200, Cross rate =0.8;
mutation rate =0.1; variable dimension =24. 45
Particle swarm optimization performed with
randomly initialization and with different 40
generation values=100,200 and 400. Different

35
best parameters found are: c1=1.4962; c2=1.4962;
objective( 108 kWh
process of maximum value of objective

process of average value of objective
w=0.7298; variable dimension D=24; 30
Genetic algorithm and particle swarm
optimization techniques have been applied at the 25
mathematical model and the results obtained are 20

given in the table I and II and shown in Fig. 2, 3
and 4.Results revealed that (table II) we can 15
0 50 100 150 200 250 300 350 400
produce more electricity using optimization generation
techniques as compared with conventional Fig. 2 Electricity generation (a using GA and

method. We can get approximately 33% more using PSO) for rainy year of Xin’anjiang and
Fuchunjinag
Rainy Years using GA
Fig. 3 Monthly total Power generation of Hydropower station with different generation values (a=100, b=200,
c=400) using GA and (d=100, e=200, f=400) using PSO for rainy year.
Fig. 4 Monthly water release ((a) using GA and (d) using PSO) and water level ((b, c), using GA and (e, f) using
PSO) for rainy year of Xin’anjiang and Fuchunjinag
It has been observed that different generation convergence of both methods but a better value of
values gave different value of objective function electricity production has achieved by using GA.
as shown in Fig. 3. Small generation number give The findings reveal that generation value also
a little less value of objective function (electricity effect the convergence speed and objective
generation) but the time of convergence is little. it function. For getting the better objective function
can be seen from Fig. 3 and table I that electricity value we should use high value of generation.
production using GA with different generation Results of the model reveals that by managing our
rates is more than PSO but the time of water level and water release equal to optimal
convergence of PSO is lesser as compared to GA. water level and water release, we can get
Electricity generation (objective function) is also maximum electricity. With the application of
affected due to different discharge rate and water Genetic algorithm, we can get more benefits in the
levels as given in table II. It was observed that for form of power generation. Results showed that
different years (rainy, average, dry) electricity genetic algorithm is a better algorithm for
production is different. It can be seen that genetic hydropower optimization. However further study
algorithm gives more value of objective function is needed with the application of different other
as compared to particle swarm optimization techniques and by the combination of different
technique but consuming a little more time than optimization techniques.
PSO.
Results revealed that there is not a big difference References
in convergence time using two methods, but a Hejazi MI, Cai X (2011) Building more realistic
significance difference can be seen in electricity reservoir optimization models using data
production as GA produces more electricity as mining – a case study of Shelbyville
compared to PSO. Results of the analysis showed reservoir. Advance Water Resources 34:701
that the water level and discharge have an 717.doi:10.1016/j.advwatres.2011.03.001)
important relationship for the energy output as Loucks, D. P. , Stedinger, J. R. , and Haith, D. A.
shown in Fig. 4. Results of the analysis revealed (1981). Water resource systems planning and
that in order to get the optimal electricity analysis, Prentice-Hall, Englewood Cliffs,
generation, we should manage the discharge and N.J.
water level of Xin’anjiang and Fuchunjiang Yakowitz, S. (1982). “Dynamic programming
reservoirs as shown in Fig. 4. If we supply applications in water resources.” Water
optimum discharge to the power station by Resources Research. volume, 18 (4 ), pages
controlling water level of the reservoir as 673–696
presented in Fig. 4, we can produce optimum Yeh, W. W.-G. (1985). “Reservoir management
electricity; consequently, we can get maximum and operations models: A state-of-the-art
benefits from this hydropower station. review.” Water Resources. Researc. , volume
21 (12 ), pages 1797–1818.
Conclusion Wurbs, R. A. (1993). “Reservoir-system
Optimization of Xin’anjaing and Fuchunjiang simulation and optimization models.” Journal
hydropower is a complicated problem. This paper of Water Resources Planning and
tried to get the optimal hydropower generation by Management. volume119 (4 ) pages 445–472
building a mathematical model and using Genetic Mousavi SJ, Karamouz M, Menhaj MB (2004)
algorithm and particle swarm optimization Fuzzy-state stochastic dynamic programming
techniques. Mathematical model for the for reservoir operation. Journal of Water
optimization is developed and applied to the Resources Planning and Management,
hydropower stations. Results showed that we can volume 130(6):pages 460–70.
get more electricity using optimal methods than Huang WC, Yuan LC (2004) ) A drought early
conventional methods. It is observed from the warning system on real-time multireservoir
results that for the optimization of the hydropower operations. Water Resources Research
and for getting maximum benefits, when simple 40:W06401. doi:10.1029/2003WR002910
genetic algorithm techniques have applied, it gave Ganji A, Khalili D, Karamouz M (2007b)
better objective function value than PSO. Results Development of stochastic Nash Game model
reveled that when PSO technique have applied at for reservoir operation I. The symmetric
the current mathematical model, it searches stochastic model with perfect information.
objective function with high convergence speed, Advances in Water Resources volume 30(1)
but value of objective function is less than GA. pages 528-542.
There is not a significance difference in time of
Kucukmehmetogl M (2009) A game theoretic Water Resources Management volume 24,
approach to assess the impacts of major pages 321–331
investments on Trans boundary water Hejazi MI, Cai X, Borah D (2008) Calibrating a
resources: the case of the Euphrates and watershed simulation model involving
Tigris. Water Resources Management human interference –an application of multi-
volume 23(15):pages 3069–3099 objective genetic algorithms. Journal of
Labadie JW (2004) Optimal operation of multi- Hydro-information volume 10(1):pages 97–
reservoir systems: state-of the-art review. 111
Journal of Water Resources planning and Tung C, Hsu S, Liu C, Li J (2003) Application of
Management, volume130(2):pages 93–111 the genetic algorithm for optimizing
Goldberg DE (1989) Genetic algorithms in operation rules of the LiYuTan reservoir in
search, optimization and machine learning. Taiwan. Journal of the American Water
Addison-Wesley, Reading Resources Association volume 39(3) pages
Fi-John C, Li C, Li-Chiu C (2005) Optimizing the 649–57
reservoir operating rule curves by genetic Kennedy J, Eberhart R (1995) Particle swarm
algorithms. Hydrological Process volume optimization. Proc. IEEE Int. Conf. Neural
19:pages 2277–2289 Networks, Perth, Australia, Dec: 1942–1948
Jian-Xia C, Qiang H, Yi-min W (2005) Genetic
algorithms for optimal reservoir dispatching.
ISBN: 978-969- 8670-06- 01 19-20 December 2017
One Dimensional Numerical Simulation of Scour and Deposition in a

Channel using Finite Difference Method
Muhammad Zain Bin Riaz1*, Muhammad Masood1, Rana Zain Nabi Khan1
1*
Centre of Excellence in Water Resources Engineering, University of Engineering and Technology
Lahore, Pakistan
*
[email protected]
Abstract: A one-dimensional Saint Venant model is developed for simulating hydraulics and bed changes in
irrigation channels. The governing equations are discretized using Finite Difference Method. The Saint-Venant
equations describing unsteady flow in open channels and the continuity equation for the conservation of sediment
mass are numerically solved. These equations are highly nonlinear and therefore do not have analytical solutions.
For this purpose, the MacCormack explicit finite difference scheme is used. The scheme is second order accurate;
it is a coupled solution as it is a two-step predictor corrector method. Model gives results in terms of bed level
changes, flow depth and discharge provided physical boundaries of the system are valid for simulation time.
Model execution and accuracy is very sensitive to time step and stability. Simulations show good accuracy when
applied within the limitations of the model. Qualitative trends of results are in good agreement with previous
studies.
Keywords: Finite Difference Method, Bed level change, Saint-Venant equations, MacCormack Scheme.
Introduction once the hydraulic conditions became compatible

Solution of the equations describing gradually with the increased sediment load. Yen et al.
varied unsteady flow in open channels is of vital (1989) performed a series of overloading
importance in the design of hydraulic structures experiments with uniform coarse sediment and
and water resources system, in the analysis of found that both the aggradation wave speed and
channel mechanics problems and in the the mean sediment transport velocity increase
development of channel control works. Many of with the initial equilibrium bed slope and with
these problems in rivers and canals involve the decreasing loading ratio. Series of experiments
transport of sediment by flow and its scour or were conducted by Yen et al. (1992) to study the
deposition. Sediment transport phenomena are reversibility of an alternating aggradation-
often time variant, even when the flows are degradation process. The results show that the
steady. Proper quantification of aggradation recovery rate decreases as standard deviation of
and/or degradation and changes in channel form sediment gradation increases. Alam (1998)
of such channel still has been a subject of applied MacCormack scheme to the study of
considerable research. A systematic study on the aggradation-degradation in alluvial channels.
aggradation-degradation phenomenon of these Kassem and Chaudhry (2002) developed a two-
channels of Pakistan is, therefore, of utmost dimensional numerical model to predict the time
importance from the viewpoint of improved variation of bed deformation in alluvial channel
planning and design of various water bends. In this model, the depth-averaged
development projects. unsteady water flow equations along with the
A number of experimental studies had been sediment continuity equation are solved by using
conducted to study the effect of the long term and the Beam and Warming alternating direction
short-term bed level changes in alluvial channels implicit scheme. Deng and Li (2003) studied the
with different flow conditions. Begin et al. (1981) river channel equilibrium profile using one
experimentally studied degradation of alluvial dimensional unsteady flow and sediment
channels in response to lowering of the bed level. transport model. Some of the widely used one
Soni et al. (1980) conducted an experiment that dimensional models are MIKE11 (DHI, 2003)
covered a wide range of flow and sediment and HEC-6 (USACE, 1993) for sediment
loading conditions. They observed that after a transport, erosion and deposition in straight
long time, the aggradation in downstream of the channels and rivers.
section of increased sediment supply was stopped
Current study determines the rate and extent of 2) Momentum equation for water
bed level-changes due to transient conditions of ∂q ∂ q2 1 ∂z
+ ( + gy 2 ) + gy + ghsf = 0 (2)
flow and sediment (routing of sediment laden ∂t ∂x y 2 ∂x
water) by using explicit finite difference method. 3) Continuity equation for sediment
∂ q y ∂q
The continuity and momentum equations along
∂t
+ qs ] + ∂xs = 0
[(1 − p)z (3)
with sediment mass balance equation are solved
numerically by finite difference explicit method
of MacCormack.
The present work is aimed at investigating more
closely with coupled solution through
MacCornaack, where relatively rapid changes in
both fluid and sediment discharge are imposed at
the upstream boundary.
Computational Procedures
The simulation method consists of two parts in
each time step, i.e., flow computations, and Fig. 2. Definition Sketch of variable
sediment routing. Flow computation is executed Terms used in these equations are defined as,
first. Then, sediment routing is executed to q = Water discharge per unit width
compute the quantity of channel bed variations. y = Depth of flow
The computational flow chart is shown in Fig.1. z = Bed elevation
qs = Unit sediment discharge
Numerical Simulations g = Acceleration due to gravity
The effective application of a model basically Sf = Friction slope
depends on the accurate path to develop the x = Longitudinal distance along
model for a certain problem. The assumptions the channel
used in the numerical model development to t = Time
simplify a phenomenon are sometimes critical to p = Porosity of bed layer
the extent of their validity. Considering all these
limitations, a model is developed from the basic The friction slope, Sf is calculated using the
equations describing the problem. Manning equation (SI unit):
q2 n2
Sf = y3.333 (4)
Where, n = Manning roughness
coefficient.
The sediment discharge is predicted by an
empirical function of the flow velocity:
𝑞 b
qs = a ( ) (5)
𝑦
Where, a, b are empirical constants to be
adjusted. Values of a, b depend upon sediment
properties and channel geometry.
Numerical scheme
The Eq. (1) through Eq. (3) characterizes as
nonlinear hyperbolic partial differential
equations and closed form solutions of these
equations are not conceivable with the exception
Fig. 1 Computational Flow Chart of rare simplified cases. Therefore, they are
solved by numerical schemes. In this model,
Governing equations MacCormack (1969) is used and solved by a
The one-dimensional partial differential finite difference scheme. The MacCormack
equations describing unsteady flow in a channel scheme is a two level predictor-corrector scheme.
with movable bed (Fig. 2) are given by: For one- dimensional flow two possibility of this
1) Continuity equation for water scheme are conceivable. In one alternate, forward
∂y ∂q
+ ∂x = 0 (1) finite differences are used in the predictor part
∂t
and backward finite differences are used in the computed and that the variation in flow depth
corrector part. In the second alternative, would be negligible. The discharge and the bed
backward finite differences are used in the elevation at the downstream end were determined
predictor part and forward finite differences are by extrapolation from the values at the interior
used in the corrector part. The values of the nodes.
variables so determined during the predictor part
are used during the corrector part. The finite Model Performance Assessment
difference approximations for the derivatives of The Coding of the model has been checked by
a variable ‘f’ are as follows: comparing it results with the already computed
Predictor step: by the implicit model (see Ghumman, 1995)
∂f f∗i −fki namely, Chen (1974), Chang and Richard (1971),
= (6)
∂t ∆t and LUM, (developed by Ghumman, 1994). The
∂f fk k
i+1 −fi results are shown in Fig.. 3. It indicates that the
∂x
= ∆x
(7)
algorithm performs well and the overall
agreement is satisfactory.
Where, i= the node in spatial grid, k = the node
Channel section is wide rectangular; the length of
in time grid, Δx = the distance step and Δt = the
the reach, L=14 km; bed slope, S0 =0.0003; and
time step. An asterisk refers to the predicted
Manning's n=0.01.
values of the variable.
The value of porosity p was assumed as 0.528
Corrector step:
∂f f∗∗ ∗ and the sediment load was calculated by using eq.
i −fi
∂t
= ∆t
(8) (5) as given below:
∂f f∗i −f∗i+1 Q 3
( )
= (9) A
∂x ∆x Q s=0.00385 y
Q (10)
Where, the two asterisks denote the value of the
variables calculated in the corrector step. The inflow hydrograph was defined as a steady
flow of 6.2m3 per second per unit width. The
Boundary conditions following equation was described by Chen
The above equations determine the values of h, q, (1974) for second upstream boundary condition.
∆t
and z at the new time level k+1 at each interior ∆Ad (1) =
2p∆x
[Qn+1 n+1 n n
s2 − Q s1 + Q s2 − Q s1 ]
grid points (i = 2,3...N-1). The values at the (11)
boundary grid points (i = and N) cannot be Qs1 is the sediment input discharge at the
calculated by using these equations. Hence, they upstream node, Qs2 is sediment discharge at node
are calculated by using boundary conditions. For 2 calculated by eq. (10), n represent the time step.
one-dimensional flows, characteristic method The initial flow along the reach was assumed to
gives acceptable results and is employed herein. be 6.2 m3 per second per unit width. The initial
Two boundary conditions at upstream boundary depths used are given in the table 1 with upstream
and one boundary condition at downstream end and downstream depths 2.15 and 4.0 meters
are required for the proper functioning of the respectively.
model. In addition to these, initial conditions are
needed at every computational grid point. Table 1 The initial depths at different nodes
Uniform unit discharge (q0), uniform flow depth (Ghumman, 1994)
(h0) and bed levels as calculated from initial bed Nod Dept Nod Dept Nod Dept
slope (S0) were provided as initial conditions, i.e., e No. h (m) e No. h (m) e No. h (m)
q(x,0) = q0 1 2.15 6 2.227 11 2.957
h(x,0) = h0
z(x,0) = z0 for x=0 2 2.152 7 2.293 12 3.199
The flow depth, h at node 1 was determined by 3 2.158 8 2.399 13 3.456

extrapolation from the already calculated values
4 2.168 9 2.549 14 3.724
at the interior nodes using the MacCormack
scheme. The downstream boundary condition 5 2.184 10 2.738 15 4.00
was the constant depth, which was specified by
h(xn,t) = h0 for t ≥ 0. This was based on the
assumption that the channel was long, and the An assumed constant water stage was defined.
bed transients would not reach the downstream The equation used is;
end within the period for which conditions were yn = 4.0 − zn (12)
𝑦𝑛 and 𝑧𝑛 are flow depth and total bed elevation
change at downstream end of reach since start of
simulation, respectively. The sum of 𝑦𝑛 and 𝑧𝑛
remains constant. Initially 𝑧𝑛 at downstream is
zero.
Fig. 4 General Plan and Location Map of

Fig. 3 Bed level changes compared with CRB Canal
published solution These data, including water stages and channel
bed variations, can be used to calibrate and verify
Application to CRB Canal the model. Field data from 2000, including
Chashma Right Bank Irrigation Project (CRBIP) is geometric cross-sectional data and bed material
a large irrigation system spread over two provinces data, are used as the initial condition. Data from
(Punjab and North West Frontier Province) of 2000 to 2001 were used to calibrate the model
Pakistan. It is the 1st irrigation system in Pakistan, and data from 2002 to 2003 were used for
initially conceived as a crop-based supply system, verification.
and then followed by the design and construction
of its physical infrastructure. Parameter Examination
Chashma Right Bank Canal (CRBC) off-takes The total simulation time was divided into low
from Chashma Barrage on right flank of Indus sediment flow period and high sediment flow
River and runs almost parallel to the river along period. May to Sept was high sediment flow
the highest contour for about 273.5 Km (170 period, whereas rest of the period was low
miles). The total command area of the project is sediment flow period. The inflow sediment load
230,675 hectares (570,000 acres) of land on the was calculated from the measured hydraulic and
right flank of River Indus. The canal had been sediment data for the year 2001. The sediment
planned to deliver design discharge of 138.3 rating was prepared for the low sediment flow
cumecs (4879 cusecs). The distribution system is and high sediment flow separately. The outflow
about 1,100 Km (700 miles) long and 83 sediment rating was used as inflow sediment
secondary channels irrigate a narrow strip of land rating for the next reach and so on.
between the Chashma Right Bank Canal and To avoid an enormous load of calculations, a
Indus River. single representative cross section was chosen.
The first 38 Km (24 miles) of the canal are The geometric properties of all the individual
unlined with a slope of 1 in 8,000. A flat slope of sections over the entire reach were averaged for
1 in 14,000 has been provided in the next 201 Km this representative cross-section. The reach
(125 miles). Last 17.7 Km (11 miles) in the tail length, L=15 mile (24 km); bed slope,
reach have a steep slope of 1 in 8,000. The S0=0.000125; Manning’s n=0.016
general plan and location map shown in Fig. 4. The value of porosity, p=0.3; sediment discharge
was calculated using eq. (5) in which values of
empirical constants were taken as;
α=0.00000006; β=3.248; frictional slope was
estimated by Manning's formula.
Table 2 The initial values for discharge

No Discha No Discha No Discha
de rge de rge de rge
No. (cfs) No. (cfs) No. (cfs)
1 3443 7 3383 13 3323 A = (Y − 7.3)/0.000832 (13)
Pw=2*y + T; Pw is wetted perimeter and T
2 3433 8 3373 14 3313
Substituting these terms in the Manning's eq.;
3 3423 9 3363 15 3303 2
A A
Q = . ( ) . S 0.5 (14)
n Pw
4 3413 10 3353 16 3293
The downstream boundary was defined by the
5 3403 11 3343 17 3283 Manning's equation using a subroutine NDEPTH
in the MacCk. Couple Model.
6 3393 12 3333 Using table 3 of flow depth satisfies the
Manning's equation at initial conditions and thus
model runs successfully without facing any
The initial depths were assumed as normal depth complication of instability due to Manning's
is calculated by Manning's formula using the equation during computations. However, model
assumed cross-section as given below. execution is terminated immature if values of
initial flow depths other than table 3 were used.
Table 3 The initial depths at different node

Node No. Depth (ft) Node No. Depth (ft) Node No. Depth (ft)
1 6.997 7 7.488 13 8.089
2 7.0343 8 7.5877 14 8.0909

3 7.069 9 7.7615 15 8.1300
4 7.193 10 7.849 16 8.1764

5 7.292 11 7.894 17 8.2125
6 7.390 12 7.982
For dx =1000 ft.; dt=c*dt1; Any increase and Verification
decrease of dx and dt have tremendous effect on Using the parameters determined in the previous
stability and execution of model. If c=0.75, it analysis, the model is applied to simulate the
stops only after 48-time steps and for c=0.65 its channel bed evolutions of CRB canal from 2002
simulation is completed upto 90 days. Stability to 2003. The total simulation period is 330 days.
problem was overcome by using c=0.55 that The comparisons of the observed and computed
simulated for more than 198 days and 10841-time bed profile and bed change are shown in Fig. 7
steps when ram quota exceeded with the message and 8 respectively. The overall accuracy is good.
"Error during write, Disk quota failure" as output The computed and measured bed profiles are well
file is too large and big. Hence direct display of matching with each other as shown in Fig.7.
results was necessary to avoid excess of memory
error and for model debugging. However,
unnecessary output data could be avoided
through managed output file.
The difference between calculated and observed
bed level changes is ±0.82 ft. It is due to the
limitations of the model.
The bed profile computed with the model after
simulation was compared with the measured bed
profile. The computed and measured bed profiles
are well matching with each other as shown in Fig. 7 Observed and Computed bed profile
Fig. 5. after 330 days simulation
The deposited sediment resulted in raising the

bed of the canal at different locations. The graph
was plotted for bed changes and it is shown Fig.
8.
Fig. 5 Observed and Computed bed profile

after 325 days simulation
The sediment deposition pattern depends upon Fig. 8 Sediment deposited and bed changes
the sediment inflow, hydraulic discharge, flow after 330 days
velocity, roughness, bed slope and bed material
gradation. The sediment deposition pattern was Running MacCk. Couple Model with this data is
estimated by plotting the change in bed elevation humble effort of this study. Generally, set trends
at different time intervals. The Changes in bed of results are fairly responsive but need further
elevation at different time interval are plotted in minute model calibration and refinement which
Fig. 6. is expected in future research to make the model
a generalized one and viable for sediment as well
as for hydraulic routing.
CONCLUSIONS
In this study the change in bed level of channel
is studied with application of a one-dimensional
mathematical model based on MacCormack
scheme. The following conclusions can be made
after summarizing the present study:
Fig. 6 Bed change after simulation 1- Model developed in this study has fairly
good qualitative and quantitative accuracy
when applied within the limitations of the
model. However, quantitative estimates may Gill, M.A. (1987) “Nonlinear solution of
be demoralized sometimes. aggradation and degradation in channels.”
2- Results and model execution was more stable Journal of Engineering Mechanics, IAHR,
and prosper for relatively smaller Vol. 25, No. 5, pp. 537-547.
computational time intervals. Jain, S.C. (1981) “River bed aggradation due to
3- Employing different equations for similar over loading.” Journal of Hydraulic
conditions will yield transport rates that Engineering, ASCE, Vol. 107 No. HY1,
might differ by several orders of magnitude. pp. 120-124.
Therefore, it is difficult to reach at a Kassem, A.A. and Chaudhry, M.H. (2002)
universal transport equation. “Numerical Modeling of Bed Evolution in
4- The MacCormack scheme can be applied to Channel Bends”. Journal of Hydraulic
develop a two-dimensional model to study Engineering, ASCE, Volume 128, Issue 5,
the two-dimensional problems and also a pp. 507-514. MacCormack, R. (1969).
three-dimensional model to study three- “The effect of viscosity in hypervelocity
dimensional problem. impact cratering.” Paper 69 354, American
Institute of Aeronautics and Astronautics,
References Cincinnati, Ohio.
Alam, M. (1998) “Application of MacCormack Soni, J.P.; Garde, R.J.; and Raju, K.G.R. (1980)
scheme to the study of aggradation “Aggradation in streams due to
degradation in alluvial channels” M.Sc. overloading.” Journal of the Hydraulics
Engineering thesis work, Department of Division, ASCE, Vol. 106, No.HY1, pp.
Water Resources Engineering, BUET, 117-132.
Dhaka. Torres. W.F.J and Jain, S.C. (1984).
Begin, Z.B. (1981) “Development of longitudinal “Aggradation and Degradation of Alluvial
profiles of alluvial channels in response to Channel Beds”, IIHR Report No. 274,
base level lowering.” Earth Surface Iowa Institute of Hydraulic Research,
Processes and Land Forms, Vol. 6, No. 1, University of Iowa.
pp. .46-68. Yadav, S.M. and Samtani, B.K. (2008) “Bed
Deng, J. and Li, Y. (2003) “A study on the Load Equation Evaluation based on
equilibrium profile for the Luoshan- Alluvial River Data, India”. KSCE Journal
Hankou reach in the middle Yangtze of Civil Engineering, Volume 12, Number
river”. International Journal of Sediment 6, pp. 427-433.
Research, Vol. 18, No. 2, 2003, pp. 107- Yen, C.L; Chang, S.Y.; and Lee, H.Y. (1989).
114. “Recovery of channel bed in aggradation-
DHI Inc. (2003), 301 South State Street, degradation process” Proceedings, 23rd
Newtown, PA, 18940, USA. congress of IAHR, Ottawa, Canada, pp.
http://www.dhisoftware.com/ B323-B329
Ghumman, A. R 1994 The development and Yen, C.L; Chang, S.Y.; and Lee, H.Y. (1992).
optimization of numerical model’s foe the “Aggradation-degradation process in
sediment routing in rivers. Ph.D thesis, alluvial channels” Journal of Hydraulic
Queen Marry and Westfield College, Engineering, ASCE, Vol. 118, No HY12,
University of Landon pp. 1651-1669.
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Risks Involved on Hydropower Projects in Pakistan

Raheel Ahmed Rana1*
1*
*
[email protected]
Abstract: Hydroelectric power plant has complex structures and involves large amounts of capital with a long-
running construction period. This situation enacts uncertainty factors with considerably high risks. Failure to
manage project risks leads to completion time delays and cost overruns. Risk Management (RM) is a logical and
systematic process of identifying, analyzing and responding to risks associated with any activity, function or
process. The objectives of RM are to increase the probability and impact of positive events and decrease the
probability and impact of adverse event. The alternative to risk management is risky management! In the financial
world, risk management is the process of identification, analysis and acceptance or mitigation of uncertainty in
investment decisions. However, in developing countries like Pakistan, the observance of RM in the construction
industry is rare. In Hydropower Sector, such studies are very almost ZILCH as the Hydropower Market is
immature in term of indigenous sufficiency. This research is conducted to highlight the main risk factors causing
delays of hydropower construction projects subsequently causing the cost over runs, conflicts, Arbitrations and
loss to Economy. The research employs risk analysis to obtain feedback from experts participating having
experience in Hydropower in Pakistan. The research outcomes are as follows: identifying the risks that can cause
delays in the Hydropower construction projects in Pakistan, calculating and classifying the degree of impact of
each risk and ranking them along with the comparison of views of different stake holders. The practical
significance of this study is to ensure the timely completion of projects, ensuring early risk mitigations techniques
that benefits all the stake holder including Investor, Client and Contractor.
Keywords: Risk Management, Delays in Hydropower Projects.
Introduction countries. Investors will commit funds toward

Pakistan’s potential from hydro is at least the development of an energy project in general
60,000MW, (according to the Private Power and based upon the project's risks in which the
Infrastructure Board- PPIB) most of this from Pakistan is lacking. International Finance
Khyber-Pakhtunkhwa and Gilgit-Baltistan. But Cooperation (IFC of World bank) in its GUIDE
this is where the picture stops looking bright and TO INTVESTOR AND DEVELOPER OF
becomes depressing. The total installed capacity HYDROELECTRIC suggest that Risks can
of hydropower in Pakistan is just 6,800 MW as reduce project attractiveness for some lenders;
after the construction of the Tarbela Dam, not a other lenders are willing to accept more risks but
single big hydropower project has been carried for higher yields. But if the project risk profile is
out during the last 40 years, except Ghazi too high, project financing may not be feasible.
Barotha. According to the Water and Power Thus, a well-prepared financial analysis should
Development Authority of Pakistan (WAPDA), include a DETAILED RISK ANALYSIS.
two-thirds of the country’s electricity is today But the dilemma does not stop here as the
produced from fossil fuels, and just one-third projects that somehow comes from Paper to
from hydro. become reality phase, are also delayed not only
The biggest hurdle as “assumed generally” may in construction phase (construction risks) but also
be lack of political will, stability and consensus during bidding process for years and years.
along with hurdle is finances, considering Private sector analysts say that many of these
Pakistan’s dwindling ability to attract foreign projects will never see even their ground-
investors because of the security situation and breaking.
water dispute with India which constantly crops Very Recent examples are 69 MW Lawi
up. Added predicaments is non-observance and Hydropower (feasibility in 1990, bid for 3 times
non-provision to investors, the clear pathways and still not awarded), 84 MW Matiltan
from conception to return on investment (ROI) Hydropower (Feasibility studies in 1990, four
especially in Pakistan being developing time re-bid in recent 3 years), Jagaran II ( delayed
almost 3 years in bidding process), 34.5 MW is an important and integral element of project
harpo ( approved from ECNEC, foreign funded, management’. He stated that the success of a
still no sign of bidding process), Neelum Jhelum project manager was heavily influenced by
Hydropwoer ( long delays in construction), Dasu efficient and effective management of the risk
Hydropower Project (land acquisition and involved. Therefore, systematic risk management
Chinese contractor issue) Diamer Basha Dam needs to be applied in order to manage a project
(funding issue) and horrible list continues. effectively throughout its life cycle.
Risk management steps were also introduced by
Project Life Cycle A Guide to the Project Management Body of
Risk management Knowledge (PMBOK® Guide-PMI 2004) 2
PMBOK® Guide (PMI 2004) 2 suggested which stated there were four fundamental steps
dividing a project life cycle into four major stages of risk management: risk identification, risk
and defined the ‘typical’ project life cycle as: analysis (qualitative and quantitative), risk
feasibility, planning and design, construction, response planning, and risk monitoring and
and turnover and startup. In addition, it is well control.
known that construction activities within the
project life cycle are executed under ‘unique’ Systematic approach to Risk Management
circumstances and continually face a variety of At the first step, risks should be classified into
uncertainties which can result from ‘known’ (the different groups with certain criteria in order to
risk events which occur frequently and an clarify the relationship between them. This step
inevitable feature of all construction projects), is called risk identification (Zhi 1995) 6. He
‘known-unknown’ (the risk events whose explained that the background of the
occurrence is foreseeable, and their probability of identification of risk and the creation of a risk list
happening is known), and ‘unknown-unknown’ is dependent upon many factors, such as past
(the risk events whose probabilities of occurrence experience, personal tendency, and the
and effect are not foreseeable, considered possession of information. Further, the
moreover, as force-majeure event conditions). In researchers, Cohen and Palmer (2004) 7, stated
other words, the construction phase deals with there are two common techniques to employ in
many unknown, unexpected, and unpredictable risk identification: experienced-based risk and
factors (Akintoye, A. S & MacLeod, M. J brainstorming- based risk assessment.
(1997)), which sometimes put the project at risk. The second step entails the risk analysis and
As defined in ISO 31000 4, risk is the effect of evaluation of the risks pertaining to risk
uncertainty on objectives ; risk management is management. Risk analysis response
the identification, assessment and prioritization management may only be performed on
of risks followed by the coordinated and identified potential risks.
economical application of resources to minimize,
monitor and control the probability and / or
impact of unfortunate events 2 (negative events),
or to maximize the realization of opportunities
(positive events).
Paul Newton, in his book named “ Managing
Project risk” 5 suggested that everything that is
done in Business contains some Risk. No matter
what activity, there is element of risk that must
be analyzed and weighed against potential
reward. Most of the researchers have claimed that
risks in the construction projects can be defined
as anything that influences the construction
project in the planning phase as well as the Fig. 1 Risk Management Process
execution phase (Akintoye & MacLeod 1997)3.
However, few people view risk only as negative The third step is to put in place an appropriate
indicator. Despite the arguments surrounding method in order to treat the risks. At this step, the
risks in the construction project, Zhi, H project management team should decide and
(1995,’Risk management for overseas formulate risk treatment strategies or mitigation
construction projects’6, International Journal of measures. In short i.e. avoidance, transference,
Project) Management, stated ‘Risk management mitigation, and acceptance (PMI 2004).
Proposed Research Orientation
The Author, in this research study, identified the
Forty-Four (44) project delay attributes through
detailed literature review, brains storming and
discussion with Hydropower experts, that were
grouped into nine (9) categories keeping in view
risk to consultant related, contractor related,
design-related, equipment-related, externality -
Fig. 2 Risk treatment related, labour-related, material-related, owner-
related, project-related, engineer-related,
Risk Management Studies on Hydropower contract related and human-behavior related.
Project
The Author has referred few studies that are Delays Factors in Hydropower Construction
related to risk management in Hydropower Based on literature reviewed, and interview /
Sector. In order to prevent time delays and cost discussion with few Hydropower experts
overruns in hydropower construction in following Nine (9) broad Risk categories was
Indonesia (state owned power company of clustered in the study that are summarized in the
Indonesia like WAPDA here in Pakistan) below Fig. 3 and table 1;
conducted the study on project risk management
in the construction stage of hydropower plant
projects. (named: Project Risk Management in
Hydropower Plant Projects: a Case Study from
the State-owned Electricity Company of
Indonesia Author: Weddy, B.S &
Hardjomuljad,S,) 8. Similarly, a study in
Vietnam was conducted by MAI S.H, WANG
J.Q & HONG V.U 9 on Risk evaluation and
control of EPC hydropower construction project
in Vietnam). This study highlighted the need and
importance of RISK Management. Another
study on Risk assessment of river-type
hydropower plants by using fuzzy logic approach
(Kucukali, S, Cankaya University, Department of
Civil Engineering, Balgat, 06530 Ankara,
Turkey) 10 suggest that that the most concerned
risks are site geology and environmental issues. Fig. 3 Risk Groups
Table 1 Risk in Hydropower Project

Risk
Risk
Category
A Risk from contracts
A-1 Unfair contract terms
A-2 uncertain and unclear contract terms
EPC mode of contract in developing country like Pakistan where expertise level is
A-3 low
Second Language contracts with misleading clause (e.g. Chinese companies signing
A-4 contract in English
B Risk from politics and laws
B-1 Relationship among investor, contractor with authority and relevant department
B-2 Political influence / political Change
B-3 Law and regulations from management agencies / Government policies
C Risk from technology
C-1 Technical Design
C-2 Negative survey data
C-3 Lower expertise level in Developing countries in contractor / consultant
C-4 Electrometrical Equipment’s sourcing and performance
D Risk from Natural Conditions
D-1 Geology of Area (Subsurface condition and subsurface ground water)
Hydrological risk (lower or excess than expected water causing unusual seasonal
D-2 variation/ floods)
D-3 safety and security on remote areas
E Construction risks
E-1 Delayed Site Access
E-2 Availability of plant and equipment at site
E-3 Changes in the work or design at later stage
E-4 Availability of material
E-5 Late drawings and instructions
E-6 Availability of labor
E-7 Lack of local industry experience in Hydropower
F Performance risks category (construction related)
F-1 Poor Site managements and supervision
F-2 Delays in decision Making
F-3 Accidents / safety
F-4 Rain and flood
F-5 Unsuitable management structure and style of contractor
F-6 Lack of communication / trust / maturity between client, consultant, and contractor
F-7 Defective work
F-8 Lack of Expertise in Hydropower Sector
G Contractual and legal risks during execution of project
G-1 Insolvency of contractor or owner
G-2 Delayed dispute resolution
G-3 Delayed payment on contract and claims
G-4 Procedural delay from the client specially when LENDER is involved
G-5 EPC mode of contract in Pakistan and lack of EPC expertise in hydro sector
G-6 Chinese EPC contractors ( Project taken on low rates then fail to deliver)
H Financial and economic risks category
H-1 Wrong Costing
H-2 Inflation
H-3 Interest rate
H-4 National and international impact of monetary policies
H-5 Financial capacity of contractors
I Social and Environmental Risk
I-1 Resettlement and social unrest
I-2 International objection on social environmental or cultural grounds
I-3 Destruction during war
I-4 Environmental risk
Methodology for survey owner of contractors worked on hydropower,

Questionnaire survey were conducted across owner of consultants worked on hydropower,
hydropower specialist that included consultants, financial advisors, directors of big construction
contractors, clients and financial advisor on houses, foreigners from china, UK, Iran worked
hydropower projects to gather their views on in Pakistan, field specialist, planning engineers,
causes of delay in delivery of hydropower economist, Mechanical specialists, contract
projects. specialists etc.
In this research, only experts from Hydropower In order to get the realistic and representative
industry were selected worked in Pakistan on results, survey was conducted only through the
different hydropower related business in experts having exposure in Hydropower
different capacities that include but not limited to business. Every risk has to be rated by the
project directors, project managers, designers,
respondent from the scale 1-5 (5 being high risk Data Analysis’
and 1 as low risk). Researched Weddy, B.S & Hardjomuljad,S, 8
used the relative importance index to determine
Respondents Data: the risk tanks and same process has been adopted
1. Total respondents = 52 here. The contribution of each of the factors to
overall delays was examined and the ranking of
the attributes in terms of their criticality as
perceived by the respondents was done by use of
Relative Importance Index (RII) which was
computed using equation
Relative Important Index (RII) = ∑ W /

A∗N (0 ≤ RII ≤ 1)
Where: W – is the weight given to each factor by
the respondents and ranges from 1 to 5, (where
“1” is “Low Risk” and “5” is “High Risk”);
A – is the highest weight (i.e. 5 in this case)
and; N – is the total number of respondents =52
in this case.
Survey Results and Analysis

Ranking of Risk
The Author based on the survey result and
calculated Relative Important Index value
categorized the risk into extremely Important,
Very Important, Important, Less Important and
Not important risk. Table 2 shows the Raking of
risk based on the RII. The survey results show
that there are four risk factors (9.09%)
categorized as ‘extremely important’. These four
risks namely “Chinese EPC contractors (Project
taken on low rates then fail to deliver)”, Wrong
Costing, Insolvency of contractor or owner,
delayed dispute resolution” are the most
“critical” risk. Twenty-Eight risks (63.63%) risks
are considered as Very Important (RII = 60 to
<80). The average RII score for these Important
risks 0.698 so in the “Very Important” Categories
group the risks with RII more than the average
RII are considered to have significant impact.
These risks include form rank 5 to rank 21 in
Table 2.
Fig. 4
Table 2 Ranking of Risk based on RII

Rankin Code Risk RII Sco Relativ
gRank 1 G-6 Chinese EPC contractors (Project taken on low rates re
218 e2.876
%
Rank 2 H-1 then failCosting
Wrong to deliver) 0.838 213 %
2.810
Rank 3 G-1 Insolvency of contractor or owner 0.819 209 %
2.758
Rank 4 G-2 Delayed dispute resolution 0.804 208 %
2.744
0.800 %
Rank 5 F-2 Delays in decision Making 204 2.692
Rank 6 G-5 EPC mode of contract in Pakistan and lack of EPC 0.785 203 %
2.678
Rank 7 expertise in hydro sector
F-8 Lack of Expertise in Hydropower Sector 0.781 199 %
2.626
Rank 8 D-1 Geology of Area (Subsurface condition and subsurface 0.765 197 %
2.599
Rank 9 ground contract
A-1 Unfair water) terms 0.758 196 %
2.586
Rank G-3 Delayed payment on contract and claims 0.754 195 %
2.573
10
Rank A-2 uncertain and unclear contract terms 0.750 190 %
2.507
11
Rank C-3 Lower expertise level in Developing countries in 0.731 189 %
2.494
12
Rank A-3 EPC contractor
mode/ consultant
of contract in developing country like 0.727 188 %
2.481
13
Rank F-1 PoorPakistan
Sitewhere expertiseand
managements level is low
supervision 0.723 186 %
2.454
14
Rank E-1 Delayed Site Access 0.715 185 %
2.441
15
Rank G-4 Procedural delay from the client specially when 0.712 184 %
2.428
16
Rank LENDER is involved
H-5 Financial capacity of contractors 0.708 184 %
2.428
17
Rank C-2 Negative survey data 0.708 183 %
2.415
18
Rank F-6 Lack of communication / trust / maturity between client, 0.704 183 %
2.415
19
Rank consultant, and contractor
E-7 Lack of local industry experience in Hydropower 0.704 182 %
2.401
20
Rank I-1 Resettlement and social unrest 0.700 182 %
2.401
21 0.700 %
The above results depict the pictorial overview Frequency Analysis on SPSS (Statistical
regarding the major risk involved in the in- Package for the Social Sciences
Hydropower project in developing Country like The Author also conducted the frequency
Pakistan. Thus, it may be safely concluded here analysis of each risk based on count of survey
that Project Management Team must consider results on SPSS. Frequency table of top ranked
these potential risks in Hydropower Projects. risk is shown in table 3 and frequency of rank 2
and ranked 3 risks are depicted in chart 4 to 5 for
the 52 respondents of the survey.
Table 3 Frequency Analysis

Rank 1: Chinese EPC contractors (Project taken on low rates then Frequency Percent
fail to deliver)
very low risk 1 1.9

low risk 2 3.8
Neutral 10 19.2
high risk 12 23.1
very high risk 27 51.9
Total 52 100.0
Ranking of Group categorized in the group G were the contractual

The table 3 show the ranking of group based on risk during the execution of projects and most of
Combined RII score of survey results and Group these risks can be foreseen and mitigated with
G - Contractual and legal risks during execution good project management, involving the
of project was ranked number No 1 by the contractual experts, using the best in practice
respondents. The author found an interesting fact contract document like FIDIC, timely involving
from the results that in top Ten risks, five risk the dispute boards, taking ownership of issues
belongs to Group G that subsequently resulted and resolving in good faith.
Group G to be most influenced group to cause
issue in Hydropower Projects. The risk
Table 3 Risk Groups Overall Ranking
Ranki Group Combined Relative Importance
Group Name
ng Code Index (RII)
Contractual and legal risks during execution
Rank 1 G of project 0.780
Rank 2 D Risk from Natural Conditions 0.709
Rank 3 A Risk from contracts 0.674
Performance risks category (construction
Rank 4 F related) 0.657
Rank 5 C Risk from technology 0.651
Rank 6 H Financial and economic risks category 0.644
Rank 7 I Social and Environmental Risk 0.621
Rank 8 B Risk from politics and laws 0.621
Rank 9 E Construction risks 0.603
Another stimulating fact is that Ranked 3 group given to every person as of their experience
is related to Contracts risks that involve unfair difference, education differences, exposure
and unclear tern and conditions. One sided difference scope / role differences. So, based on
contracts terms and unclear/improper contract experience, exposure of Hydropower projects,
documents are always the concern for the educations and last but not the least experience of
Contractors and subsequently result in Potential Author working with the respondents, weightage
risk for the project. to their score is introduced in order to refine the
result to ensure authentic and making true
Weightage Assigned to the Respondents representation of issues.
Though the results were collected from the In comparison for actual and refined results,
experts from various trades and back ground in Relative Percentage was introduced to compare
Hydropower Industry, however the Author the degree and ranking. However, the results
considered that there may be some kind of show that there is not any substantial upset as first
biasedness, inconsistency and experience lack in top 10 ranked risk remained the same with up to
survey result from the respondents. The Author rank 6 no change in ranking.
also realized that the same weighted can’t be
Agreement and disagreement between coefficient, rs, can take values from +1 to -1.
Contractor, Client and TPS A rs of +1 indicates a perfect association of ranks,
The Author also analyzed the agreement and a rs of zero indicates no association between
disagreements in survey results between client, ranks and a rs of -1 indicates a perfect negative
contractor and third-party experts in order to association of ranks. The closer rs is to zero, the
highlight the perspective of different weaker the association between the ranks.
stakeholders. Third party respondents include the
experts from consultants, dual experience
contractor – client experiences, consultant –client
experience, contractor- consultant experience,
hydropower financial advisors.
Rank Correlation for risk categories for the three
groups of respondents was calculated to establish
the level of agreement and disagreement. The
author used the rank correlation using spearman
coefficient. The Spearman correlation
Table 4 Spearman Correlation (over all result)

Client vs Contractor 0.60
Contractor Vs Third Party 0.84
Client Vs Third Party 0.67
Broadly speaking the Client perspective is

different as compared to Contractor and third
party, whereas the Contractor and third party
experts have more harmony in the views.
Table 5 Ranking Comparison of Client, Contractor and TP
Third
Client Contractor
Code Risk Party
Ranking Ranking
ranking
A-1 Unfair contract terms 26 6 13
A-2 uncertain and unclear contract terms 9 17 8
EPC mode of contract in developing country
A-3 12 18 9
like Pakistan where expertise level is low
Second Language contracts with misleading
A-4 clause ( e.g. Chinese companies signing 38 43 37
contract in English
Relationship among investor, contractor with
B-1 28 26 28
authority and relevant department
B-2 Political influence / political Change 13 34 24
Law and regulations from management
B-3 17 35 33
agencies / Government policies
C-1 Technical Design 35 31 22
C-2 Negative survey data 33 19 4
Lower expertise level in Developing countries
C-3 14 11 15
in contractor / consultant
Electrometrical Equipment sourcing and
C-4 18 40 36
performance
Geology of Area (Subsurface condition and
D-1 15 10 2
subsurface ground water)
Hydrological risk (lower or excess than
D-2 expected water causing unusual seasonal 16 25 16
variation/ floods)
D-3 safety and security on remote areas 3 27 23
E-1 Delayed Site Access 5 22 18
E-2 Availability of plant and equipment at site 27 38 38
E-3 Changes in the work or design at later stage 31 14 25
E-4 Availability of material 42 42 39
E-5 Late drawings and instructions 29 21 30
E-6 Availability of labor 43 44 43
Lack of local industry experience in
E-7 34 15 17
Hydropower
F-1 Poor Site managements and supervision 11 20 14
F-2 Delays in decision Making 6 7 7
F-3 Accidents / safety 41 41 44
Rain and flood
F-4 36 39 42
Unsuitable management structure and style of

F-5 22 23 26
contractor
Lack of communication / trust / maturity
F-6 7 16 29
between client, consultant, and contractor
F-7 Defective work 19 28 34
F-8 Lack of Expertise in Hydropower Sector 23 3 19
G-1 Insolvency of contractor or owner 8 4 3
G-2 Delayed dispute resolution 10 5 1
G-3 Delayed payment on contract and claims 20 8 11
Procedural delay from the client specially
G-4 30 12 20
when LENDER is involved
EPC mode of contract in Pakistan and lack of
G-5 1 9 12
EPC expertise in hydro sector
Chinese EPC contractors (Project taken on
G-6 2 1 5
low rates then fail to deliver)
H-1 Wrong Costing 4 2 6
H-2 Inflation 32 32 31
H-3 Interest rate 44 36 40
National and international impact of monetary
H-4 39 37 32
policies
H-5 Financial capacity of contractors 24 13 21
I-1 Resettlement and social unrest 25 24 10
International objection on social
I-2 21 33 27
environmental or cultural grounds
I-3 Destruction during war 40 29 41
I-4 Environmental risk 37 30 35
The detail review of table 4 and 5 epitomizes the comprehensive analysis of the risk, and instigate
concern of contractor, client and consultants etc. a reasoned allocation and mitigation strategy, and
and the intensity of their concern, difference and this is employer or project sponsor. Many
prospective towards various risk. For instance, contractors, particularly on large projects,
Client ranked Risk named “EPC mode of contract undertake a risk assessment as part of tender
in Pakistan and lack of EPC expertise in hydro process, but this will be limited to consideration
sector” as No 1 but contractor and third-party of the risk within their own scope of work and
experts does not agree with the perspective of contractual arraignments and results of the
client. Risk A1 (Unfair contract terms) being survey second it. Often this will be limited to an
ranked 6 by the contractor as one-sided contract analysis of ‘what can go wrong?’ It will rarely
is always the concern for contractor and consider any of the employer’s risk. The prime
Contractor has to cater these risk at the bidding mover in this process has to be employer and
stage, however, as expectedly the client considers encompassing all aspects including finance,
it a less potential risk. The risk like Political and environmental, construction, operational etc.
law order (B2 and B3) are the always the concern In developing countries, the people’s culture
of client/ developer and ranked high by the client standard is low, causing various difficulties; the
in comparison to contractor and third-party immigration and resettlement for land
experts. The Risk like technical design and withdrawal and handover for the construction
negative survey data is normally responsibility of contractors are complex, Hydropower equipment
designer / consultants and subsequently they for the projects must be imported from abroad
have ranked them higher than the contractor and with complex procedures, difficult shipment, and
client. Delayed site access is one the risk slow assembly, dependency on foreign
attributed to employer / client risk and their Contractors, lenders involvement, high inflation
concern is depicted in the survey results. All rates, which affect the purchase of required
stakeholder appears to agree on delay Decision materials, machines, and equipment, Natural
Making risk (F-2) risk. Contractor has ranked conditions such as climate, hydrology,
Lack of Hydropower expertise in Pakistan quite topography, and geological conditions lead to
high in compared to the Client and third party. further complications, lack of expertise like
Delayed dispute resolution has been ranked 1 by tunnel / surge shaft construction, the sub-
the Consultants being the neural participants and contractors construction capacity being poor, the
client side has ranked it No 10. Disputes domestic construction technology has low
originates in every project but the lack of owner productivity, and is not up to standard, the
ship form stakeholders and improper infrastructure and traffic facilities for transport
contemporary records and delaying tactics results are poor; machine and equipment transportation
in delayed in dispute resolution and project suffer encounter many difficulties leading to delays and
costs and delays. Delayed payment is one of the cost overrun, However, timely risk management
main concern of the contractors and ranked on Hydro Power project can minimize / eliminate
higher as compared to client as expectedly. the effect of aforementioned summarized
Similar is the case with procedural delay where situation(s) and at the same time can attract the
contractor normally suffer and ranked higher in investor and financing agencies and make the
the results. The risk like Chinese contractor project SELLABLE.
factor and wrong costing brings all stake holder
on one page. Wrong costing itself include the References
lack of expertise, lack of technical staff, not 1 International Finance Cooperation (2014) (IFC
understanding of specification and project Of World Bank) “Guide To Investor And
requirements / scope from all stake holders etc. Developer Of Hydroelectric” (Page 105).
Resettlement issue is ranked higher by the Third 2 Project Management Institute, A Guide to the
party that includes consultants but surprisingly Project Management Body of Knowledge
the Client has not shown much concern on the (PMBOK® Guide), 2014 5th edition, Project
issue. Management Institute, Newton Square.
Conclusion 3 Akintoye, A. S & MacLeod, M. J (1997), “Risk

Risk analysis can be undertaken by any party to analysis and management in construction”,
a construction project, for that part of the project International Journal of Project Management,
with which they are concerned. There is, vol. 15, no. 1, pp.31-38.
however, only one party that can undertake a
4 ISO31000:2009. (Latest) Risk management - 8 Weddy, B.S & Hardjomuljad,S, 2011 “Project
principles and guidelines on implementation. Risk Management in Hydropower Plant Projects:
International Organization for Standardization. a Case Study from the State-owned Electricity
5 Paul N., 2015 in book named “Managing Company of Indonesia.
Project risk”
9 MAI S.H, WANG J.Q & HONG V.U Published
6 Zhi, H (1995), “Risk management for overseas Online April 2016 “ Study on Risk evaluation
construction projects”, International Journal of and control of EPC hydropower construction
Project Management, vol. 13, no. 4, pp. 231-237. project in Vietnam”
7 Cohen, M W & Palmer, G R 2004,’Project risk 10 Kucukali,.S , 2011 “Risk assessment of river-
identification and management’, AACE type hydropower plants by using fuzzy logic
International Transaction, pp. 13-15. Retrieved approach” Cankaya University, Department of
September 10, 2006, from ABI/INFORM Global. Civil Engineering, Balgat, 06530 Ankara, Turkey
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Analysis of regime behavior of lower Gugera branch canal after

rehabilitation/remodeling of the system.
Sajid Mehmood*1, Ghulam Nabi1
1
Project Management Unit, Punjab Irrigation Department, Lahore, Pakistan
*
[email protected]
Abstract: Pakistan has one of the world most effective irrigation systems which are comprised of canal network.
These canals are mostly unlined and are designed mainly either on Lacey or Kennedy design methods. Irrigation
canals essentially off-takes from a river and draw a fair share of silt moving in the river. The sediment is carried
either in suspension or along the bed of the canal. The basic criteria of design of irrigation channels are to aim
at a velocity which will keep the sediment in suspension without causing silting or scouring, such a velocity is
known as non-silting and non-scouring velocity. This is known as minimum stream power concept. Basically,
there are two concepts for alluvial channel design. In first concept, the channel design is based on the concept
of non-silting and non-scouring approach. Such a channel is known as regime channel. The most common design
methods based on this concept are of Kennedy method, Lacey regime and Simon & Albertson regime theory.In
the second concept, the design of irrigation channel is based on the properties of boundary material in which the
channel is flowing. The most important design method based on this concept is Tractive Force Method.
Keywords: Irrigation, canals, Lacey theory
Introduction significantly different from each other. The

The Lower Gugera Branch Canal Irrigation Lower Gugera Branch Canal had been generally
system lies between the rivers Ravi & Chenab i.e. receiving less sediment from the canal head
Rachna Doab. It is bounded by Qadirabad- compared to its sediment transporting capacity
Balloki (Q.B.) Link Canal on the eastern side and is showing scouring trends. However, in
while it terminates beyond Trimmun-Sidhnai contrast, the Burala Branch Canal had been
(T.S.) Link Canal on the western side. The Lower receiving more sediment compared with its
Gugera Branch Canal off-takes from the tail of transporting capacity and silted up even after five
Upper Gugera Branch canal at Head Buchiana years of its operation.Due to uneven distribution
near Jaranwala City as shown in fig 1. Lower of sediment load in the system the bed of Lower
Gugera Branch Canal off takes from the Right Gugera Branch has scoured up to 2ft in different
side and Burala Branch Canal from the front. An reaches. Water surface line has been depressed
Escape channel is also off taking from the left and off taking channel as well as the outlets are
side. The system has been maintained over the suffering badly. The bed of Burala Branch Canal
years with very limited resources, resulting in has been deposited and the capacity of the
backlog of various repair / remodeling works. A channel has been reduced.
large part of the system has already lived over a The objectives of study were re-evaluation of the
century of its life and has undergone considerable design of the canal system, analysis of the
deterioration. Punjab Irrigation and Drainage sediment budget of the system and evaluation of
Authority (PIDA) decided to rehabilitate the sediment carrying capacity of designd channel by
system and remodel/reconstruct various worn out using different sediment transport functions.
structures. The main object for rehabilitation of
the system was to improve conveyance
efficiency, reliability and durability of the system
as per designed discharges by
rehabilitating/remodeling the entire irrigation
system, so as to provide irrigation water to the
beneficiaries in a more sustainable and equitable
manner. The problem of sediment management
in off-taking channels became serious after the
rehabilitation of the system. The behavior of both
the canals in terms of sediment entry is
delivery schedules. The flow control structures
are designed at bifurcations for water level
regulation, discharge regulation and discharge
measurement.Habib (1998) used the SIC model
for the hydraulic simulation of CRBC for to
predict its design and operation.
Methodology
Three reaches of Lower Gugera Branch Canal
have been taken for research (1RD = 1000ft =
304.8m).
1. Reach RD 0+000 to 16+500
2. Reach RD 16+500 to 27+000
3. Reach RD 27+000 to 65+000
The design parameters and actual longitudnal
section were collected from respective irrigation
divisions. The existing canal cross-Sections were
Fig. 1 Line Diagram of Study Area also observed at site and the cross-Section were
also obtained from the consultants of
Understanding of the sediment transport rehabilitation project of Lower Gugera Branch
phenomena is inadequate and this makes it very Canal for the data prior to rehabilitation to
difficult indeed to predict (Sanmuganathan, ascertain the changes in the regime. For
1990). Impact of structural intervention on suspended sediment/ bed sediment observation
sediment management of large canals is the sediment sampler DH-48 sediment sampler
explicated by Shakir et al. (2007) in a case study and DH-24 sediment sampler has been used for
of Marala Barrage. He concluded that some of the determination of existing sediment load in the
past interventions have a positive effect on silt flowingwater at the specified sites. Sediment
control entry in the off-taking canals. The samples were analyzed by International
analysis of data regarding sediment entry at the Sediment Research Institute of Pakistan (ISRIP)
head of the canals indicate that the raising of crest Water and Power Development Authority
of MR Link in the year 2000–2001 has improved (WAPDA) laboratory. The results of samples are
the sediment intake in this canal to some extent, used in application of different design procedures
but its effect on the UCC is almost negligible.The and sediment analysis adopted for comparison of
data regarding sediment entry at the head of the different design methods used for the alluvial
UCC which had a scouring tendency does not canal design. The discharge/sediment data of
show any significant effect of this intervention as LCC (parent channel of Lower Gugera Brach
was perceived at the implementation stage. canal system) at head Khanki has been collected
However, it is worth mentioning that analysis of from the irrigation department.Four types of data
data related to cross-sections and L-sections of were collected for this study is geometric data,
the head reaches of these canals show a different hydraulic data, sediment data and structural
trend, which makes it difficult to assess the long drawing/data.
term and sustainable effect of this intervention on Daily recorded water discharges released from
both of the canals. A critical review of data Khanki Barrage to feed the LCC were used from
collection process and effective monitoring of the 2000 to 2013 for hydraulic study. Sediment data
system is compulsory for meaningful analysis was analyzed to construct a sediment rating curve
and future course of action if the present from the suspended sediment inflowing
tendency seems to be changing, the possibility of concentration at the head of the LCC canal. A
which does exist. sediment rating curve meant a relationship
The analysis was focused on transient bed between sediment load (including both
profiles which are determined, under steady flow suspended and bed loads) and water
conditions, by a decrease of the upstream discharges.Four design methods i.e. Kennedy,
sediment transport rate. Ankum (1995) defined Lacey, Simon & Albertson and Tractive Force
flow structures as the structures required to keep was applied to all these selected reaches and
the irrigation system in the desired state. The calculated canal parameters were plotted against
purpose of flow control structures is to distribute the existing observed cross- sections parameters.
and to divert the flows to the end users w.r.t the
Brief description of these methods is given channels, R  D , where ‘D’ is the normal flow
below, depth. Hence,  =  .D.S ,
0 0
The Manning Formula: The most frequently W tan  2
used formula for an open channel flow is the  = S

cos  tan  1 − ,
S a tan 2 
1
 tan 2 
Manning formula.The formula is; V = n R2/3 K = S
= cos 1 − which may be
1  tan 2 
S1/2 or; Q = A R2/3 S1/2 l
n sin 2 
Kennedy theory: simplified as, K = 1− ,
Kennedy Theory was utilized for a protracted sin 2 
period for the design of canal systems. In Indo- (  )max= 1.5γRS
0
Pak subcontinent some of the initial chief canal
systems were based on this theory. V° =D0.64 ,
=CDb.(ft/sec), V=m * 0.84 D0.64,where D = depth There are many universal accepted techniques to
V determine the sediment load entering at the head
of channel (ft),Where m = = critical velocity regulator of the canal. Many techniques have
V0 been proposed to access the sediment load entry
ratio (CVR) under a large range of flow conditions and
Lacey’s theory: sediment characteristics.In this study sediment
According to Lacey the silt factor ‘f’in the transport capacity was calculated by using
regime formula depends upon:Average size of (Ackers& White (1973), Engelund& Hansen
boundary material in the channel and its density; (1967), Van Rijn (1984), Yang (1973), and
Brownlie (1981)) formulae. The calculated
f = 1.76 D sediment carrying capacity was compared with
50 sediment load entering in the canal.
where ‘D50’ is the mean particle diameter of silt
in mm. i. Ackers and White Method (1973):
The Ackers and White method described the
Velocity of flowing water and the hydraulic sediment transport in terms of three
2.50V 3 dimensionless parameters; Ggr (transport
f = parameter), D* (grain size sediment
mean radius of the channel: R parameters), Fgr (Mobility parameter).
Where V=mean velocity in m/s.
𝟏⁄
(𝐒−𝟏)𝐠 𝟑
Simons and albertson regime method: D* = [ ] 𝐝35,
𝛝𝟐
The following equations were fitted to the 𝟏−𝐧
various straight line graphs plotted in between 𝐔∗𝐧 𝐕
various parameters: 𝐅𝐠𝐫 = √𝐠𝐝𝟑𝟓 (𝐒−𝟏) √𝟑𝟐 𝐥𝐨𝐠(𝟏𝟎𝐡)
[ ]
𝐝 𝟑𝟓
0.37
C2 V2 V 
= = K  b 
𝐆𝐠𝐫 = 𝐂 [ 𝐀𝐠𝐫 − 𝟏]
𝐅 𝐦
g 2 DS 4 V 
 
The Ackers and White function to determine the
Tractive force method total sediment transport read as:
In 1950’s investigators established the tractive 𝐕 𝐧
force technique of stable canal design and that is 𝐪𝐬 = 𝐆𝐠𝐫 S𝐝𝟑𝟓 (𝐔 )
∗
the solely firmly rational method in practice
today. According to tractive force technique Engelund and Hansen Method (1967):
when all forces acting on the particle are in This method is based on an energy approach;
equilibrium, there will be no scouring and no they established a relationship between transport
silting and the canal will remain stable. Unit and mobility parameters. The Engelund and
 . A.L.S Hansen function for the total sediment transport
tractive force or shear stress =  = 0
= is calculated by: Dimensionless transport
0 P.L
parameter
 .R.S Where, R= hydraulic radius. In very wide
0
𝑞𝑠
∅= 𝑤 2
3 𝛽 = 1 + 2 ( 𝑠) (4.34)
√(𝑆−1)𝑔𝑑50 𝑈∗
𝑈∗2 𝑊 0.8 𝐶 0.4
Dimensionless mobility, 𝜃 = ψ = 2.5 + ( 𝑈𝑠) (𝐶𝑎)
(𝑆−1)𝑔𝑑50 ∗ 𝑜
The relationship between those parameters is Brownlie Method (1981):

expressed by: Brownlie (1981) defined a method to compute
the sediment transport rate, the transport rate (in
0.1𝜃2.5 𝐶 2 ppm by weight) is calculated by:
∅= 2𝑔 R −0.3301
q s = 727.6 c f ( Fg − Fgcr )1.978 S0.6601( )
d 50
The total sediment transport is expressed by:
, Cf: Coefficient for the transport rate (Cf = 1 for
𝑞 ) laboratory conditions and Cf = 1.268 for field
0.05 𝑉2
𝑠= 0.5 𝑑
50 𝐶3
conditions.
(𝑆−1)2 𝑔
Yang method (1973):

This method is based on hypothesis that the
sediment transport in a flow should be related to
the rate of energy dissipation of the flow. The
total sediment transport is also expressed in ppm
by mass as a function of the unit stream power
by:
𝑙𝑜𝑔𝑐𝑡=𝐼+Jlog(VS−Vcr S)
w
With the Yang’s coefficients represented by:

𝑊𝑠 𝑑50
𝐼 = 5.435 − 0.286𝑙𝑜𝑔 ( 𝜗
)−
𝑈
0.457𝑙𝑜𝑔 (𝑊∗ )
𝑠
And
𝑤𝑠 𝑑50
𝐽 = 1.799 − 0.409𝑙𝑜𝑔 ( )−
𝜗
𝑈
0.314𝑙𝑜𝑔 (𝑤∗ )
𝑠
The critical velocity for initiation of motion:
Vcr = 2.05 ws,
The total load transport is calculated by:
qs = 0.001 Ct V h
ii. Van Rijn Method (1984a and 1984b):
The total sediment transport by the Van Rijn
method is computed by the summation of the bed
and suspended load transport
qs = qb + qsus
The bed load transport rate is calculated as:
qb = ubδbcb Particle, Ub = 1.5T0.6 [(S-1)
gd50]0.5
Saltation height, δb = 0.3 D0.7, T0.5d50 Fig. 1 Site visit
Bed load concentration, Cb=0.18C0T/D*,
Maximum Volumetric concentration C0 = 0.65 Results and Discussions
With, D* = 𝐝35 Lower Gugera Branch was designed on Lacey
𝑞𝑏 = 0.053(𝑠 − 1)0.5 𝑔0.5 𝑑501.5 𝐷∗−0.3 𝑇 2.1 method having a discharge capacity of 2,244 /
ds= [1+0.011) (σs-1)(T-25)]d50 2,643 cusecs for Rabi and Kharif respectively
and length of the channel is 387+566 ft off-taking
from tail of Upper Gugera Branch. First three 8,500 1,400
reaches RD 0+000 to RD 66+000 were selected 8,000 1,200
for observing cross sections. The existing cross 7,500
1,000
sections were observed at successive 5+000 RDs 7,000
and the comparison of design versus existing bed 6,500 800
profiles as shown in fig 2. 6,000 600
645
5,500
Designed Bed Level Existing Bed Level FSL
400
640 5,000
4,500 200
635
630
4,000 0
Elevation (ft)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
625
Month
620 Avg. Discharge (cfs) Avg. Total Sediment (ppm)
615
610 Fig. 4 Long-term monthly variations in

0
5,000
10,000
15,000
20,000
25,000
30,000
35,000
40,000
45,000
50,000
55,000
60,000
65,000
sediment concentrations with discharge
R.D.
During the analysis of collected
Fig. 2 Comparison of designed and existing bed discharge/sediment data series from 2000-2013,
profile of Lower Gugera Branch Canal it was found that variation in the sediment load
entering LCC system is mainly due to the inflow
LCC system is the oldest system of Punjab and at head Khanki rather than discharge off-taking
off-taking from left side of head Khanki having a in LCC system. The graphical representations of
design discharge of 7,000 cusecs. The daily aforesaid phenomena are shown in fig 5 and 6.
discharge/sediment data series from the
100,000
laboratory of head Khanki were collected for the
last fourteen years (2000 – 2013).The same data
LCC System Sediment Load (short ton)
series were analyzed. Yearly and monthly

10,000
variations in the sediment concentrations
entering LCC system with discharge are
presented in fig 3 and 4.
1,000
7,500 1,200
100
1,000
7,000 y = 0.0576x1.1181
800 R² = 0.4429
10
1,000 10,000 100,000 1,000,000
6,500 600
Khanki u/s - total inflow (ft3/sec)
400
6,000 Fig. 5 Suspended sediment rating curve of LCC
200
System versus head Khanki inflows
5,500 0
The existing design parameters of Lower Gugera
2006
2000
2001
2002
2003
2004
2005
2007
2008
2009
2010
2011
2012
2013
Branch canal for the selected three reaches are

Avg. Discharge (cfs) Avg. Total Sediment (ppm) presented below
Fig. 3 Annual variations in sediment

concentrations versus discharge
100,000
LCC System Sediment Load (short ton)

10,000
1,000
100
y = 0.0000x2.4382
R² = 0.2389
10
1,000 10,000 100,000
Canal Inflow (ft3/sec)
Fig. 6 Suspended sediment rating curve of LCC System versus LCC inflow
Table 1 Existing design data of selected reaches of Lower Gugera Branch Canal
Table 2 Re-evaluation of existing design by Tractive Force method

Reach 1 Reach 2 Reach 3
Existi Existi Existi
Descripti Unit
ng Differen ng Differen ng Differen
on s Lacey Lacey Lacey
Desig ce Desig ce Desig ce
n n n
Discharge ft3/se
2,643 2,643 0% 2,430 2,430 0% 2,347 2,347 0%
(Q) c
Bed 1 in 0.0001 0.000 0.0001 0.000 0.0001 0.000
0% 0% 0%
Slope (So) 1 4 14 4 14 4 14
Manning'
- 0.021 0.021 0% 0.021 0.021 0% 0.021 0.021 0%
sn
Horizonta
- 0.5 1.5 200% 0.5 1.5 200% 0.5 1.5 200%
l Slope
Vertical
- 1.0 1.0 0% 1.0 1.0 0% 1.0 1.0 0%
Slope
Flow
ft 7.5 6.6 -12% 7.3 6.6 -10% 7.2 6.6 -9%
Depth (D)
Bed
ft 115.0 135.9 18% 110.0 124.8 13% 108.0 120.5 12%
Width (B)
Table 3 Re-evaluation of existing design by Lacey theory

Existi Existi Existi
Descripti Unit
ng Differen ng Differen ng Differen
on s Lacey Lacey Lacey
Desig ce Desig ce Desig ce
n n n
Discharge ft3/se 2,643 2,643 0% 2,430 2,430 0% 2,347 2,347 0%
(Q) c
Bed 1 in 0.0001 0.000 -33% 0.0001 0.000 -32% 0.0001 0.000 -32%
Slope (So) 1 4 09 4 09 4 10
Manning' 0.021 0.021 0% 0.021 0.021 0% 0.021 0.021 0%
-
sn
Horizonta 0.5 0.5 0% 0.5 0.5 0% 0.5 0.5 0%
-
l Slope
Vertical 1.0 1.0 0% 1.0 1.0 0% 1.0 1.0 0%
-
Slope
Flow 7.5 7.9 6% 7.3 7.7 6% 7.2 7.6 6%
ft
Depth (D)
Bed 115.0 119.5 4% 110.0 114.3 4% 108.0 112.3 4%
ft
Width (B)
Table 4 Re-evaluation of existing design by Simon & Albertson method

Existi Simon Existi Simon Existi Simon
Descripti Unit
ng & Differe ng & Differe ng & Existing
on s
Desig Alberts nce Desig Alberts nce Desig Alberts Design
n on n on n on
Discharg ft3/s 2,643 2,643 0% 2,430 2,430 0% 2,347 2,347 Dischar
e (Q) ec ge (Q)
Bed 0.000 0.0000 -56% 0.000 0.0000 -56% 0.000 0.0000 Bed
1 in
Slope 14 6 14 6 14 6 Slope
1
(So) (So)
Manning' 0.021 0.021 0% 0.021 0.021 0% 0.021 0.021 Mannin
-
sn g's n
0.5 1.5 200% 0.5 1.5 200% 0.5 1.5 Horizon
Horizont
- tal
al Slope
Slope
Vertical 1.0 1.0 0% 1.0 1.0 0% 1.0 1.0 Vertical
-
Slope Slope
Flow 7.5 9.0 20% 7.3 8.8 20% 7.2 8.7 Flow
Depth ft Depth
(D) (D)
Bed 115.0 110.6 -4% 110.0 106.0 -4% 108.0 104.1 Bed
Width ft Width
(B) (B)
Table 5 Re-evaluation of existing design by Kennedy theory

Existi Simon Existi Simon Existi Simon Existi
Descripti Unit
ng & Differen ng & Differen ng & ng
on s
Desig Alberts ce Desig Alberts ce Desig Alberts Desig
n on n on n on n
Discharg ft3/s 2,643 2,643 0% 2,430 2,430 0% 2,347 2,347 0%
e (Q) ec
Bed 0.000 0.0001 0% 0.000 0.0001 0% 0.000 0.0001 0%
1 in
Slope 14 4 14 4 14 4
1
(So)
Manning' 0.021 0.021 0% 0.021 0.021 0% 0.021 0.021 0%
-
sn
Horizont 0.5 0.5 0% 0.5 0.5 0% 0.5 0.5 0%
-
al Slope
Vertical 1.0 1.0 0% 1.0 1.0 0% 1.0 1.0 0%
-
Slope
Flow 7.5 7.5 0% 7.3 7.3 0% 7.2 7.2 0%
Depth ft
(D)
Bed 115.0 112.6 -2% 110.0 107.6 -2% 108.0 102.3 -5%
Width ft
(B)
Re-evaluation of existing design parameters of average suspended sediment concentrations in

Lower Gugera Branch canal using Kennedy, Upper Gugera Branch (the parent channel) is 477
Lacey, Simon& Albertson and Tractive force ppm, whereas in the off-taking channels of Upper
methods is presented as follows Gugera i.e. Burala Branch and Lower Gugera
To evaluate the sediment load, the sediment data Branch, are 456 ppm and 400 ppm respectively.
was observed on the following channels at head The bed material samples were also collected by
Buchiana with the coordination of ISRIP using BM-54 bed sampler in all three aforesaid
(WAPDA), Upper Gugera Branch canal (tail) , channels. Allowable sediment transport capacity
Burala Branch canal (RD 1+200) and Lower of Lower Gugera Branch canal was estimated
Gugera Branch canal (RD 0+900). Suspended using different sediment transport functions and
sediment sampler D-49 was used to observe the the results are presented as follows in table 6.
suspended sediment load. It was found that
Table 6 Sediment transport capacity of Lower guger Gugera Branch Canal estimated by different
sediment transport function
Lower Design Measured Allowable Sediment transport capacity (ppm)
Gugera Discharge Sediment
Reach (ft3/sec) (ppm) Ackers& (ft3/sec) (ppm)
White
1 2,643 393 1,065 1 2,643 393
2 2,430 406 1,060 2 2,430 406
3 2,347 401 1,054 3 2,347 401
logical because the variation in the approaching

The analysis of existing cross sections of the sediment load is mainly due to the Chenab river
canal under consideration reaches shows that bed inflow rather than off-taking discharge of LCC
is scoured by more than 1 to2 ft in selected system.
reaches and the scouring trend increases from The existing canal design has been re-evaluated
head to tail reach. The bed has eroded up to 2 ft using different design methods of alluvial
at different sites and severe bank erosion is also channel and the results differ with the existing
in progress at the downstream of different fall design in terms of side slope, bed slope, water
structures.The analysis of sediment data (2000- depth and bed width etc. However, it is important
2013) at head Khanki shows that maximum to mention that difference in water depth and bed
sediment concentrations are in the months of July width is not much significant. The results of
and August which are about 1,350 ppm and 750 suspended/bed material sampling and sediment
ppm on average respectively i.e. in Monsoon. transport theories show that suspended silt load
Whereas, the same data shows that there are less in LCC System is less than the allowable limits.
sediment concentrations in December which are Furthermore, the distribution of suspended silt at
about 75 ppm. It has been found that relationship head Buchiana is not equitable which is due to
between the LCC system’s suspended sediment the difference of crest level of the head
load does not show a good trend with the regulators.The results of different sediment
variation in its discharge. However, when it is transport functions predict that the sediment
plotted against head Khanki’s inflow, the concentration carrying by the system is also less
relationship shows some good trend which is also
than the permissible limits calculated by these silt load and silt charge water is continuously
function. eroding the canal bed and sides. The results of
different sediment predictors also predict that the
Conclusions sediment concentration in the system is not up to
As most of the canal systems in the Punjab are the design limits, hence the bed is eroding.
either designed by Lacey or Kennedy, are the The sediment distribution in the off taking
modified forms of these method. The existing channels is not equitable due to difference in the
canal reaches of the selected channel show that crest levels, so more silt vanes should be
bed of channel is scoured up to 2 ft and the sides constructed at head of Lower Gugera Branch to
are also eroded at the downstream of fall balance the load in the off taking channels. The
structures. The scouring trend is significantly sediment data of Head Khanki reveals that
increasing towards the tail and water line has sediment entry into the Lower Chenab Canal is
been depressed. Resultantly the off-taking also not up to permissible limits so the
channels and the outlets are suffering badly.The functioning of silt excluder and crest of the said
results of suspended sediment samples collected channel should be analyzed. The irrigation
from the selected channels shows that sediment channel should be designed by considering the
concentration is less than the permissible limits sediment concentration in the system and silt
in the LCC System so the System is sediment carrying capacity should be checked by using the
hungry. Furthermore, the sediment diverting into different sediment transport formulae. Due to
Lower Gugera Brach canal at head Buchiana is shortage of water channels runs on rotational
also less which is due to the difference in the crest regulation pattern which affect the regime
levels of the off-taking channels. The analysis of conditions and morphology of these canals.
bed material samples are mostly composed of bed Therefore, at any cost channel could not run on
load depiction in the canal section shows that bed less supply. Periodic sediment measurement
material is composed of coarser silt and fine silt should be done throughout the year until the
particles in all the reaches.The sediment data at canal is in regime. Canal design/operational
head Khanki clearly reveals that major part of the software should be introduced to facilitate the
sediment load entering into the canal consists of decision-making authorities.
finer particles and total sediment concentrations
are less than the allowable limits. To balance the
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Impact Analysis of Floods on Fixation of Road Profile Grade Level for

Roads in Hilly Terrain
Zafar Iqbal1, Kaleem Sarwar2, Sajid Mahmood3, Muhammad Afzal4
1
Research Associate, Centre of Excellence in Water Resources Engineering, University of
Engineering and Technology, Lahore, Pakistan
2
Assistant Professor, Centre of Excellence in Water Resources Engineering, University of
Engineering and Technology, Lahore, Pakistan
3
Engineering and Technology, Lahore, Pakistan.
4
Senior Engineer, Water Resources Division, National Engineering Services (Pvt.) Limited, Lahore,
Pakistan.
*[email protected]
Abstract: Flood hazards are the significant threat to the human life as well as their property. Recent example
of such destruction is 2010 flash flood. In this study hydrodynamic modeling was performed on Swat River from
Kalam to Madyan. Floods for different return periods at Kalam were computed by flood frequency analysis of
flow data gauged at Kalam. Floods of lateral tributaries of Swat River downstream of Kalam were computed
using HEC-HMS rainfall runoff model with the help of available rainfall data at Kalam and catchment
characteristics of the study area. Steady flow simulations were carried out in HEC-RAS and water surface profiles
were computed for floods of different return periods. Model geometry was extracted from stereo pairs based
Digital Elevation Model (DEM). The computed water surface profiles were processed in HEC-GeoRAS and
inundation extents were marked. Finally, the profile of road section was fixed on the basis of water levels
computed from hydrodynamic model and inundation extents for 100-year return period flood. The research will
be useful for planning and design of bridges, culverts, causeways, river training and flood protection works,
channelization of streams and hill torrents to protect the cities and town against floods and road infrastructure
especially in hilly areas.
Keywords: Kalam, Madyan, Swat River, Water Surface Profile, Inundation Extents, Profile Grade
Introduction section were completely destroyed by the flood

National Highway N-95 road network starts from [1].
Chakdara intersects Mingora, Manglour, The road profile grade level (PGL) of previously
Khwazakhela, Madyan, Bahrain and ends at built National Highway N-95 was badly affected
Kalam. This is an important road and connects by the hydrological conditions which needs
Islamabad, the capital city to Kalam in Swat attention of the authorities to safeguard the
Valley. Swat River starts at Kalam from the important part of the communication.
confluence of Ushu and Uttror Rivers. Being
adjacent to Swat River this road faced severe
flooding in 2010 especially beyond Madyan
towards Kalam. Fig. 1 shows the layout plan of
the National Highway N-95 [1].
The road section from Madyan to Kalam is
approximately 44 Km where Swat River flows
through V shape valley lying on the right side of
the road whereas; on the left side are the steep
hills. This road section was completely washed
away by 2010 flood. Every possible way of
communication, either transportation or
telecommunication, with Kalam was terminated.
Bridges and culverts encountered on this road
Fig. 1 Layout Plan of National Highway N-95
with GIS. Water surface levels were simulated
For the safe passage of flood flows it is pertinent using HEC-RAS. HEC-GeoRAS was used for
to prepare hydrodynamic model for the study generation of floodplain maps. It was concluded
reach which will assist in determining the that HEC-RAS model along with GIS play an
inundation extents against different return period important role in determining the flood extent
floods. Thereby i.e. profile grade level (PGL) and flood zone mapping [7].
based on the hydrodynamic model with the
provision of certain free board will be designed. Materials and methods
Monsoon of the year 2010 brought the worst Hydrological Modeling
flooding in past 80 years in the region. Heavy Flood frequency analysis at Kalam was carried
monsoon rainfalls in Khyber Pakhtunkhwa, out using Gumbel Extreme Value Type-I (EVI)
lower Punjab, Sindh, and Baluchistan regions and Log Person Type-III. Gumbel Extreme
made the rivers to overflow which devastated the Value Type-I (EVI) is the best fit for the study
areas from Gilgit - Baltistan to Kotri, the last and area as per analytical applications.
biggest barrage on River Indus in Sindh. During Hourly rainfall data was analyzed and intensity
flood season of 2010, flows in western rivers duration frequency curves were developed for
were very high and some of the recorded different durations (i.e. 1hour, 2hour, 3hour,
discharge levels were very high than those 6hour, 12hour and 24hour) at Kalam rain gauge
observed during the past high floods of 1956, station using the point rainfall data for each year
1973, 1976, 1988 and 1992. Therefore, it is for representing the rainfall data for decision
unavoidable to redesign / re-section the flood support of hydrodynamic modeling as shown in
protective bunds according to new water levels Fig. 2.
obtained during 2010 Flood [2].
HEC-GeoHMS and HEC-HMS rainfall runoff
model was applied for Kaha Hill Torrent
Watershed Management of D.G.Khan District.
HEC-HMS model was used for estimating
realistic volume of runoff as a result of rainfall
for optimum utilization of water resources and
better watershed management [3].
HEC-RAS model was used to identify the flood
prone areas in Chiniot District along Chenab
River and concluded that HEC-RAS can be
efficiently used to simulate the floods. It was
suggested that Remote Sensing techniques along
with GIS should be involved to get more precise
results [4]. Fig. 2 IDF Curves for Kalam Derived from
Flooding in Nullah Lai in Rawalpindi was Recorded Data
studied using GIS and HEC-RAS for estimation
of flood extents. GIS model was used to find out Rainfall Runoff Model for Tributaries of Swat
the inundation extents and to delineate the River
variation in topography of the area. It was The Hydrologic Modeling System (HEC-HMS)
concluded that HEC-RAS model results are very software was used to simulate the precipitation-
close to the JICA surveyed results of the runoff processes of Swat River catchment. For
inundated area [5]. this purpose Swat River was divided into thirty
HEC-RAS model along with GIS and Remote sub-catchments. The time of concentration and
Sensing environment was used for the study of lag time for each sub catchment was computed
flood in Ping River basin in Thailand. Water separately. The data for computation of time of
levels were generated using HEC-RAS for concentration and delineation of sub catchment
determination of the inundation extents. HEC- areas were extracted by using Shuttle Radar
RAS results were verified by remote sensing Topographic Mission (SRTM) 90 m Digital
images. Finally, the model results were used for Elevation Model (DEM).
preparation of hazards maps [6]. Floods for 25, 50 and 100 years for Ushu and
A reach of Neka River of Iran was studied for Uttror Rivers were computed from flood
floodplain analysis and delineation of flood frequency analysis at Kalam using Gumbel
extents and water depths using HEC-RAS along Extreme Value Type-I. Floods for lateral
tributaries of Swat for the study reach were Identification of Study Area and Collection of
computed by HEC-HMS by using respective Stereo Pairs Image
frequency storms. Only magnitude of rainfall is Swat River flows in gorge throughout the length
required, when flood is estimated with SCS of study area. Area of interest (study reach of
distributions in HEC-HMS. Swat River) was marked on Google Earth and
Satellite image (0.5m resolution and 5 Km wide
Calibration and Validation of Rainfall Runoff strip) was collected from NESPAK. The image
Model was processed using Arc GIS for preparation of
The contribution of rainfall (of about 20%) is DEM. The prepared DEM was calibrated with
very low as compared to snow melt (of about Ground Control Points (GCPs) collected from the
80%) at Kalam stream gauge station. The results study area. Calibrated DEM on 2 m resolution
of flood frequency analysis at Kalam were used was used for the extraction of model geometry
in HEC-HMS as a constant discharge. HEC- with the help of HEC-GeoRAS.
HMS model was actually used to calculate the
flow of lateral tributaries of Swat River Identification of River Reach and Its
downstream of Kalam to Madyan (study reach). Tributaries for the Study Reach
So, HEC-HMS model was calibrated with As calibrated DEM on 2 m resolution was a
Chakdara stream gauging station. The peak limited strip of 5 Km width normal to the Swat
discharge on 29 June 2007 was 30,894 cusecs at River. So, for the identification of river reach and
Chakdara. The curve no 51 was taken at AMC-I its tributaries falling in area of interest a polygon
condition and adopted for calibration. The was marked on Google Earth and imported in Arc
specified hyetograph method of metrological GIS along with STRM data as shown in Fig. 3.
model was used. Hourly rainfall data of 28 and SRTM data was extracted within the limits of
29 June 2007 in all rain gauges selected for polygon using mask command in Arc GIS.
calibration was incorporated in the model. The Coordinates of study area was assigned to SRTM
rainfall was assigned to each sub basin by making data.
criteria, which rainfall station is nearest to the sub
basin. The peak discharge of 11,270 cusecs at
Kalam was taken as constant base flow for the
month of June.
The rainfall runoff model was validated with
Chakdara stream gauge station also. The peak
discharge on 10 October 1987 was 35,600 cusecs
at Chakdara. The hourly rainfall data of 9 and 10
October 1987 was incorporated in the model in
the same way as discussed above for model
calibration. The peak flow of 2150 cusecs was
adopted as a constant base flow for the month of
October.
Hydrodynamic Modeling Fig. 3 Identified River Reach and its Tributaries

The hydraulic modeling of Swat River was for the Study Area
carried out by making use of HEC-RAS which
requires various inputs to calculate water surface Extraction of Model Geometry
profiles. These inputs are layout of the river and HEC-GeoRAS was used for the extraction of
its tributaries, river cross sections, Manning’s model geometry in Arc GIS. Different layers
roughness for main stream and tributaries and were created, exported the cross-sectional data in
boundary conditions for the selected river reach. GIS format and finally imported in HEC-RAS for
Model geometry was generated by making use of further processing.
HEC-GeoRAS, an extension of Arc GIS. Steady
flow data was provided to HEC-RAS and Manning's Roughness
simulations were carried out. The Manning’s roughness for River Swat, it’s
over bank / flood plain was assessed from
satellite image and field visit of the study area
and adopted as 0.045 for main stream and 0.050
for left and right over banks respectively.
Table 1 Results of Flood Frequency Analysis
Boundary Conditions for Hydrodynamic for Selected Distributions
Model Calculated Flood Magnitude
Return
U/s boundary conditions were provided as (Cusecs)
Period
constant discharges (600, 650 & 700 cumecs) for (Years) Gumbel Extreme Log-Pearson
25, 50 & 100 years return periods respectively at Value Type-I Type-III
Kalam. Flood peak of 2010 was not recordable as 2.33 14,541 15,016
steam gauge at Kalam was washed away. So, 5 16,879 17,113
flood peak at Kalam was computed from rainfall
using rainfall runoff model. This flood peak was 10 18,782 18,495
used as upstream boundary in HEC-RAS for 25 21,188 19,942
computation of 2010 flood. The computation of
50 22,972 20,853
backwater requires known hydraulic conditions
at most downstream section of main stream. The 100 24,744 21,653
representative longitudinal slope in main stream 200 26,508 22,365
was considered as downstream boundary
500 28,837 23,201
condition for computation of normal depth for
steady flow simulations. 1,000 30,596 23,767
10,000 36,439 25,338
Model Simulation
Steady flow Simulations were carried out using Table 1 shows that computed flows for different
25, 50 and 100 years peak floods and 2010 flood return periods from Log Pearson distribution are
peak respectively and corresponding water slightly on lower side as compared to Gumbel’s
surface profiles were plotted. Flow change distribution. Chi square test was applied on both
locations in model were added for lateral the distributions. The values of chi square test
contribution from tributaries. Water levels, depth were 5.36 and 3.26 for Gumbel Extreme Value
of flow, average velocity and top width for 25, 50 Type-I and Log Pearson Type-III respectively.
and 100 year return period flood and 2010 flood As computed values for chi square for both the
at each cross section of Swat River for the study distributions are less than the standard chi square
reach were calculated. value which is 12.02. So, both the distributions
can be used for frequency analysis of the study
Flood Inundation Mapping area. Gumbel Extreme Value Type-I distribution
Based on simulation results, the estimated 100 value is close to the standard chi square value.
year flood levels were used to create inundation That’s why, this distribution is best fit for the
map. The layout and top elevations of existing study area. Furthermore, Gumblel Extreme Value
road was compared with simulated flood levels Type-I distributions are widely used in hydrologic
and inundation extents to check its safety against analysis. Storm rainfalls are most commonly
100 year flood event. Finally the road profile modeled by the EVI distribution [8, 9].
grade level (PGL) was suggested with the
provision of certain free board. Rainfall Frequency Analysis
The rainfall data of day of maximum daily peak
Results and discussions for every year was selected. Hourly data of
The aim of the analysis is to fix the safe profile selected storm was analyzed. 1 hour, 2 hour, 3
of road for the study reach depending upon hour, 6 hour, 12 hour and 24 hour maximum data
extents of inundations and water levels. was calculated from the selected storms and
frequency analysis was done using Gumbel
Flood Frequency Analysis Extreme Value Type-I. The results of rainfall
Flood frequency analysis was carried out at frequency analysis for different return periods are
Kalam using Gumbel Extreme Value Type-I shown in Table 2.
(EVI) and Log Pearson Type-III with
instantaneous maximum discharge data of each Table 2 Results of Rainfall Frequency Analysis
year. The computed discharge values for both the for Different Return Periods
distributions for different return periods are Retur
shown in Table-I. 12h 24h
n 1hr 2hr 3hr 6hr
r r
Period
(Year) Rainfall Depth (Inches) HMS at Madyan are listed in Table-III.
0.4 0.6 0.8 1.2 Hydrograph of 100-year return period at Madyan
2.33 1.83 2.62 is shown in Fig. 4.
4 5 1 1
0.6 0.9 1.1 1.6 Table-III: Peak flows at Madyan from HEC-
5 2.33 3.36 HMS for 25, 50 and 100 year return periods
8 4 3 1
0.8 1.1 1.3 1.9 Return Period Peak Discharge
10 2.74 3.97
7 8 9 3
(Year) (Cusecs)
1.1 1.4 1.7 2.3
25 3.25 4.74
2 8 2 4 25 138,327
50
1.3 1.7 1.9 2.6
3.63 5.32
50 162,923
0 1 6 4 100 187,649
1.4 1.9 2.2 2.9
100 4.01 5.88
9 3 0 4 Table 3 shows that the difference in computed
1.6 2.1 2.4 3.2
200 4.39 6.45 peak discharges (at most downstream location of
7 5 4 4
the study reach i.e. Madyan) is small (of about
1.9 2.4 2.7 3.6
500 4.89 7.20 12%) which represents the trend of rainfall and
1 4 6 3
2.0 2.6 3.0 3.9 stream flow data at Kalam. Fig. 4 shoes a peak
1,000 5.26 7.76 hydrograph at Madyan which clearly indicates
9 6 0 3
2.6 3.4 3.8 4.9 that discharge contribution from the lateral
10,000 6.51 9.63 tributaries is appreciable in the peak of resulting
9 0 0 2
hydrograph
Table 2 shows the magnitude of rainfall for
Sink "Swat River at Madyan" Results for Run "100 year"
different return periods. It is clear from the results 200,000
that the run-off generating rainfall at Kalam is up 180,000
to 6 hours Beyond 6 hour there is an appreciable 160,000
recession in rainfall. 140,000
120,000
Calibration and Validation of Rainfall Runoff

Flow (cfs)
100,000
Model 80,000
The rainfall runoff model was calibrated with

60,000
Chakdara stream gauge station. The peak
40,000
discharge on 29 June 2007 was 30,894 cusecs at
Chakdara. The hourly rainfall data of 28 and 29 20,000
June 2007 was incorporated in model. The 0

00:00 12:00
19Jan2014
00:00 12:00
20Jan2014
00:00
computed peak hydrograph at Chakdara is Run:100 YEAR Element:SWAT RIVER AT MADYAN Result:Outflow
Run:100 year Element:R13 Result:Outflow
32,015 cusecs which is very close to the observed Run:100 year Element:Olal Khwar Result:Outflow
Run:100 year Element:Shahgram Khwar Result:Outflow
discharge at Chakdara for the same day.

The rainfall runoff model was validated with Fig. 4 Hydrograph of 100 Year Return Period
Chakdara stream gauge station also. The peak Flood at Madyan
discharge on 10 October 1987 was 36,600 cusecs
at Chakdara. After incorporating hourly rainfall Flood Estimation for 2010 Extreme Flood
data of 9 and 10 October 1987, the model gives Event
peak hydrograph at Chakdara as 38,548 cusecs Recorded rainfall of 29 July is 84 mm (3.3 1inch)
which is very close to the observed flow at at Kalam, which is 24-hour maximum rainfall.
Chakdara. This rainfall was distributed over the entire
Same procedure for model calibration and catchment area of study reach of Swat River.
validation has been followed for development of Peak discharge at most d/s location of the study
design hyetograph of Swat basin [10] and study reach i.e. Madyan was estimated as 75,879
of Subarnarekha River basin [11]. cusecs. Hydrograph at Madyan is also presented
in Fig. 5.
Flood Computation using Rainfall Runoff
Model
Using the results of Table-I and Table-II floods
of various return periods was computed by
Hydrologic Modeling System software (HEC-
HMS). Peak discharges calculated from HEC-
Sink "Swat River at Madyan" Results for Run "2010 Actual Storm"
80,000
Table 4 Comparison of annual and seasonal

70,000
maximum frequency of 2010 rainstorm
60,000
Return Period Rainfall Depth (Inches)

50,000
Seasonal
Annual
Flow (cfs)
40,000
(Year) Max. (June-

Max.
30,000
Sept.)
20,000 2.33 2.52 0.80
10,000 5 3.31 1.04
0
00:00 12:00 00:00 12:00 00:00
10 3.95 1.23
28Jul2010 29Jul2010
Run:2010 ACTUAL STORM Element:SWAT RIVER AT MADYAN Result:Outflow
Run:2010 Actual Storm Element:R13 Result:Outflow
25 4.75 1.47
Run:2010 Actual Storm Element:Olal Khwar Result:Outflow
Run:2010 Actual Storm Element:Shahgram Khwar Result:Outflow 50 5.35 1.66
Fig. 5 Hydrograph of Estimated Storm of 2010 100 5.95 1.84
at Madyan 200 6.54 2.02
500 7.32 2.25
1,000 7.91 2.43
10,000 9.88 3.31
Hydrodynamic Modeling
Stereo Pairs based DEM calibrated with ground
control points (GCPs) on resolution of 2 m was
used for extraction of model geometry. Steady
state simulations were carried out with peak
discharges of 25, 50 and100 years return period
floods and estimated flood peak of 2010 extreme
flood event respectively. Water surface profile
and XYZ perspective plot for 100-year return
period for part of the study reach are shown in
Fig. 7 and Fig. 8.
Fig. 6 Mean Monthly Rainfall of Kalam Rain
Gauge Station Kalam to Madyan Plan: Plan 04 7/9/2014
Swat Kalam-Madyan
Legend
1440
Monsoon rainfalls at Swat basin are almost nil. In EG 100
Swat basin rainfall season usually starts in 1430

WS 100
Crit 100
December and prolongs to April/May as shown
Elevation (m)
1420 Ground
in Fig. 6. The storm of 2010 which prolonged
from 27-30th July 2010 is considered as the 1410
highest ever observed rainstorm not only in Swat 1400

basin but also in adjoining catchments. Stream
flow data was not available beyond 27 July 2010 1390
because stream gauge was washed away due to 1380

5000 6000 7000 8000 9000
extremely heavy flows. The rainfall record was Main Channel Distance (m)
also not available beyond 29 July 2010 due to Fig. 7 Water Surface Profile of 100 Year Return
heavy rains in Swat Basin Recorded rainfall of 29 Period Flood for Part of the Study Reach
July is 84 mm (3.31 inch) at Kalam, Annual
maximum frequency of this rainfall is 5 years One of the limitations of stereo pairs-based image
return period while the seasonal frequency (June is that it takes the free water surface as plain
to September) is 10,000-year return period. 2010 ground and cannot take the elevations below the
severe flood event occurred in Swat basin due to water surface. It gives better results for rivers in
the Monsoon breaches from Indus to Swat basin. gorges as compared to the rivers in plain areas
Comparison of annual and seasonal maximum because rivers are very wide in plain areas. Fig. 7
frequency of 2010 rainstorm is shown in Table 4. and Fig. 8 clearly shows that the longitudinal
slope of the Swat River is very steep of about 1 Fixation of Road Profile Grade Level
in 500. Depth of flow and width of every cross Profile of road section was fixed on the basis of
section is very small as compared to the actual water levels (water surface profile) computed
depth of cross section i.e. Swat River flows in a from hydrodynamic model and inundation
gorge throughout the study reach. So, the stereo extents for 100-year return period flood and
pairs-based image can safely be applied for the shown in Fig 10.
study reach.
Kalam to Madyan Plan: Plan 04 7/9/2014

Legend
WS 100
Ground
41627.16 40282.59
Bank Sta
40990.01
39722.79
38946.84
Fig. 8 X-Y-Z Perspective Plot for Part of the

Study Reach for 100 Year Return Period Flood
Fig. 10 Safe Profile of the Road for the Study
Inundation Mapping
Reach
Flood inundation extents were marked by
processing HEC-RAS results in HEC-Geo RAS.
Fig. 10 shows thalwage of the Swat River, water
The inundation extents for 100 years return
surface profile and profile grade level of road
period flood and existing road imported on
along the Swat River for the study reach. River
Google Earth image is shown in Fig. 9.
thalwage indicates that longitudinal slope of the
Fig. 9 shows the spread of flood waters for the
river is very steep, which is also an indication of
part of modeled stream reach. The difference in
very high velocities throughout the study reach.
computed water levels and inundation extents for
The computed profile of road contains a free
different years return period floods (25, 50, 100
board of 1.5 m for further safety of road and its
and 2010 flood) is of about 5% because Swat
infrastructure.
River flows in a gorge throughout the study
reach. Fig. 9 also shows that the existing road is
Conclusions
inundated at most of the locations throughout the
Recorded rainfall of 29 July 2010 is 84 mm
study reach. So, there is a dire need to raise the
(3.31inch) at Kalam (the rainfall record is not
profile of road.
available beyond 29 July due to heavy rains).
Annual maximum frequency of this rainfall is 5
years while the seasonal frequency (June to
September) is 10,000 years. The severe flood
event occurred in Swat basin due to the Monsoon
breaches from Indus to Swat basin. The soil was
fully saturated prior to the rainfall event due to
snow melts and the rainfall was completely
converted in to runoff. These were the major
causes of such heavy destruction of 2010 flood.
The difference in computed water levels for
different return periods i.e. 25, 50, 100 and 2010
floods is very small (of about 5%) because Swat
River flows in a gorge and the longitudinal slope
of the river is very steep (of about 1 in 500)
throughout the study reach. The water levels
from 100-year return period flood (being the
Fig. 9 Inundation Extents and Existing Road for
highest at the most downstream location i.e.
Part of Study Reach
Madyan) has been selected for design of road
profile grade level owing to its importance and
severity of floods. Existing shingle road is not [6] Duan Minya, Jixian Zhang, Zhengjun and
safe due to its low elevation and get inundated at Aekkapol. 2006. Use of remote sensing and
most of the locations. So, there is a dire need to GIS for flood hazard mapping in Chiang
raise the profile of road. Mai province, Northern Thailand: Chinese
Academy of Surveying and Mapping,
References Beijing, China.
[1] NESPAK. 2013. Study and Detailed Field [7] Ghanbarpour, Shokoufeh Salimi, Mohsen
Survey for Assessment and Mitigation of Mohseni Saravi and Mehdi Zarei. 2011.
Potential Flooding Locations on NHA Calibration of River Hydraulic Model
Network. Final Completion Report, May Combined with GIS Analysis using Ground
2013. Based Observation Data: Research Journal
[2] Annual Flood Report. 2010. Office of the of Applied Sciences, Engineering and
Chief Engineer Advisor and Chairman: Technology 3(5), 2011.
Federal Flood Commission, Islamabad. [8] Chow V.T. 1953. Frequency Analysis of
[3] Abdul Khaliq Hashmi. 2005. Rainfall Hydrologic Data with Special Application
runoff modeling for Kaha Hill Torrent to Rainfall Intensities, Urbana Ill:
Watershed Management in D.G.Khan University of Illinois Bulletin, 50(81).
District: M.Sc. Thesis, Center of Excellence [9] Tomlinson. A.I. 1980. The Frequency of
in Water Resources Engineering, UET, High Intensity Rainfall in New Zealand,
Lahore. Part I: MWD Technical Publication No.19,
[4] Tariq M.A. 2005. Determination and Willington.
Implementation of Flood Risk Zoning: [10] Sulman Maqsood Randhawa. 2012.
Employing Physiographic and Hydraulic Development of Design Hyetograph for
Parameters for Flood Prone Areas of Flood Estimation in Swat Region: M.Sc.
Chiniot District, M.Sc. Thesis, Center of Thesis, Center of Excellence in Water
Excellence in Water Resources Resources Engineering, UET, Lahore.
Engineering, UET, Lahore. [11] D.Roy, S. Begam, S. Ghosh and S. Jana.
[5] Ahmad B, Muhammad Shumail Kaleem, 2013. Calibration and Validation of HEC-
Mohsin Jamil Butt, Zakir Hussain Dahri. HMS model for a river basin in Eastern
2010. Hydrological modeling and flood India: ARPN Journal of Engineering and
hazard mapping of Nullah Lai: Pakistan Applied Sciences 8(1), January 2013.
Agricultural Research Council (PARC),
Islamabad, Pakistan.
ISBN: 978-969- 8670-06- 01 19-20 December 2017
Hydraulic Performance Assessment of an Orifice Spillways using CFD

Modeling
Zohaib Nisar1*, Muhammad Kaleem Sarwar2, Ghulam Nabi3
1*
Research fellow and corresponding author, Centre of Excellence in Water Resources Engineering,
University of Engineering and Technology Lahore, Pakistan
2
Engineering and Technology Lahore, Pakistan
3
Engineering and Technology Lahore, Pakistan
*
[email protected]
Abstract: An orifice spillway is normally gated and is used when substantial discharge capacity is needed at
low reservoir levels. Orifice spillways are designed for dual purpose of flood disposal as well as flushing of
sediments. Flow passing through the spillway shows the complex turbulent behaviour. To model the effect of
turbulence, Reynolds’s-averaged Navier-Stokes equations are commonly used which is an expanded form of
Navier-Stokes equation. Flow physics becomes more complex in case of orifice spillways due to short lengths of
spillway, large variation in reservoir levels, high flow depths and wide range of Froude numbers varying from 3
to 9. In view of this background, present study intends to numerically investigate the hydraulic behaviour of orifice
spillway. The objectives of this study include the numerical modeling of complex flows over the orifice spillways,
pseudo validation of numerical model results and to assess the flow parameters at different operating conditions.
Results showed that model is capable of simulating the orifice spillway flows. Model can measure flow parameters
at different operating conditions with an acceptable error of 0.23 to 1.5 %.
Keywords: Orifice Spillway, Flow Physics, Froude Number, FLOW-3D, Flow Parameters.
Introduction requirement of skilled labour and cannot readily

The large capacity outlets placed below the dam capture behavious like cavitation and air
crest and controlled by gates are called orifice or entrainment effects, which occur in reality [3]. In
submerged spillway. These outlets are openings in recent years, numerical modelling is extensively
the dam used to drawdown the reservoir level and being used to investigate hydraulic performance
can also serve to pass density currents. Large of spillways. Computational fluid dynamics
bottom openings serve as submerged spillways (CFD) is a numerical method used to solve fluid
and their capacity can be used for flushing flow problems. The computational fluid dynamics
sediment from the reservoir and disposal of flood (CFD) analysis can solve the Navier-Stokes
[1]. The flow is free surface flow for low reservoir equations in three dimensions and free surface
levels and orifice flow for reservoir levels above computation in a significantly improved manner.
the orifice opening. The crest profile of most of Computational fluid dynamics presents a cost-
the orifice spillway is flat (α < 30°). Breast walls effective solution that can be employed
are provided to hold 60 to 90 m of water heads throughout the entire design process. The use of
behind them. Because of narrow river gorges, CFD modelling for spillway application is quite
limited available waterways, and 6 to 8 m thick recent [4]. In last decade, many physical model
piers supporting huge breast walls, the discharge studies have been validated using CFD models.
intensities range between 100 and 300 m2/sec. Most of the numerical studies used the CFD code
The depths of flow are 10 to 12 m leading to low Flow-3D to analyze the spillway flows. Savage
Froude numbers between 3 and 6 [2]. Hydraulic and Johson [5] compared numerically generated
performance of spillways can be checked by discharge rating curves with physical model data
physical modelling and numerical modelling. and United States Bureau of Reclamation (USBR)
Traditionally, reduced scale physical spillway calculations. The study found that Flow-3D
models are used to study spillway hydraulic slightly overestimated the discharges. Gessler [6]
performance. But, it has some disadvantages such model the spillway flows and found 5% difference
as scaling effects, time consuming, expensive, between CFD and physical model results. Savage
and Johnson [5] did not confirmed application of
CFD model for all spillway configurations. Ho Mangla Dam Spillway
and Riddete [3] applied CFD model to evaluate To study the complex flows behaviour of orifice
the hydraulic performance of different spillways spillways, CFD modelling of Mangla dam main
for increased flood discharges and suggested spillway was carried out by operating the model
future work for cavitation, air entrainment, scour at different reservoir levels and gate openings.
modelling, air demand and dynamic interaction. The main spillway of dam consists of two-stage
Bhosekar et al. [2] studied the performance of the stilling basin and sloping side walls. The head
aerator of orifice spillway by varying the works of the main spillway are 444 feet long. It
discharge, gate openings and cavity sub-pressures consists of three monoliths separated by 24 feet
and presented the results with respect to jet length, wide piers. The head works is followed by
cavity pressure and air entrainment coefficients in parabolic chute and intermediate weir divides the
the form of non-dimensional plots and developed chute into two and creates a stilling basin and
equations for jet length and air entrainment water pool at an elevation of 1000 feet [9]. The
coefficient for the orifice spillway aerator. CFD spillway plan and the longitudinal section is
code, Flow-3D was selected for this research shown in (Fig. 2).
study due to its ability to model the free surface
flow by using true volume of fluid (true-VOF)
method developed by Hirt and Nichols [7] and
track the sharp interface between water and air.
This code also models the complex geometric
region by using fractional area/ volume obstacle
representation (FAVOR) technique [8]. This code
overlay the mesh on imported non-flow geometry
while FAVOR technique is used to determine the
void or flow region within each cell as shown in
Fig. 2 Plan and Longitudinal section of Mangla
(Fig. 1). With finer grid spacing, high resolution
Dam Spillway
of the non-flow region is achieved. The use of
multi block grids enable larger domains to be Material and Method
modelled and use of nested mesh technique enable Data Collection
more flow details to be captured in regions of Data collection for model set up sensitivity
interest [3]. Flow physics becomes more complex analysis and model validation include detail
in case of orifice spillway due to short lengths of engineering drawings, physical model study
spillway, large variation in reservoir levels, high results and discharge rating curve. Fig. 3 shows
flow depths and wide range of Froude numbers single bay three-dimensional model of Mangla
varying from 3 to 9. In view of this background, dam spillway up to intermediate weir. This model
present study intends to numerically investigate was imported into CFD model.
the hydraulic behaviour of orifice spillway. The
objectives of this study include the numerical
modeling of complex flows over the orifice
spillways, pseudo validation of numerical model
results and to assess the flow parameters at
different operating conditions.
Fig. 3 One Bay of Spillway 3D model
Setting Up of CFD Model

To model Mangla dam spillway, collected
drawings were first converted into three
VF = Open Volume/Volume of Cell dimensional drawings then imported in stereo
AF = Open Area/ Cell Edge Area lithographic (Stl) file extension image. Total four
Fig. 1 Conceptual diagram of FAVOR method Stl. file images were prepared for different gate
[5] openings. These Stl. file images were then
imported into CFD code (Flow-3D) for numerical
set-up. In next step, meshing of imported Stl. file
images were carried out. The extent of mesh
domain on upstream and downstream of structure Gated Flow Modelling
is defined in such a way that it could show the For sensitivity analysis, gated flow simulations
fluid movement and its impact properly. The were performed at vertical gate openings of 3.05
boundary conditions applied for this model m, 6.10 m, 9.14 m and 10.67 m to find out water
include, volume flow rate, specified pressure, levels by maintaining reservoir level at 378.54 m
symmetry, wall and out flow. After specifying and 384 m. Then flow rate was measured for
boundaries of the model, fluids were added on the model validation at same operating conditions.
upstream and downstream sides of the structure as Further vertical gated flow simulations were
initial condition. After pre-processing explicit and performed to find out the water surface profiles,
fluid flow solver option was selected to solve pressures and velocity at gate opening of 6.10 m
Reynold Average Navier Stokes equation and 9.14 m and at reservoir level of 380.07m,
(RANS) which is prime equation used by Flow- 381.59 m and 383.11 m respectively. Initially
3D for simulation. simulation was performed with 3 m × 3 m uniform
mesh but in order to get more refined results size
Sensitivity Analysis of mesh was reduced to 1 m. Fig. 4 shows gated
CFD code Flow-3D does not require calibration flow modelling.
[10] but it is sensitive to the turbulent models and
boundary conditions. Commonly used turbulence
models include large eddy simulation, RNG, K-ƹ,
RSM and DSM. Code user manual declares that
RNG model is robust and most accurate model
[11]. So, it was selected for further simulations.
Four scenarios of boundary conditions were used
for sensitivity analysis as shown in Table 1.
Model Validation
Validation of numerical model is extremely
important. Validation process indicates the degree Fig. 4 Flow Simulation at 10.67 m Gate Opening
of accuracy of the model. A true validation of
model involves the comparing of the model Results and Discussion
results with those from actual structure [5]. But for Sensitivity Analysis
this study model validation was carried out by Model was run at different operating conditions
using physical model study result due to the lack by applying all set of boundary condition for the
of actual performance measurements at prototype. sensitivity analysis of the model. The results of
sensitivity analysis are discussed below.
Table 1 Different Scenario of Boundary
Conditions Sensitivity Analysis of Model for Boundary
Scenario 1 Scenario 2 Conditions
X min. Specified X min. Volume Flow Model was operated at reservoir level of 378.54 m
Pressure Rate
and 384 m with different gate openings to
X max. Outflow X max. Outflow
Y min. Wall Y min. Wall compare the water levels with physically observed
Y max. Wall Y max. Wall one for all scenarios of boundary conditions.
Z min. Wall Z min. Wall Comparison of results shown in tables 2, 3, 4 and
Z max. Specified Z max. Specified 5 indicate that water levels computed with
Pressure Pressure
boundary condition scenario 3 are more close to
Scenario 3 Scenario 4 the physical model results as compared to other
X min. Volume Flow X min. Specified scenarios. Hence boundary conditions mentioned
Rate Pressure under scenario 3 were selected for further
X max. Specified X max. Specified simulation.
Pressure Pressure
Y min. Symmetry Y min. Wall
Y max. Symmetry Y max. Wall Model Validation
Z min. Symmetry Z min. Wall After sensitivity analysis, model validation was
Z max. Specified Z max. Specified carried out for gated and free flow condition
Pressure Pressure
which is discussed as under.
Flow Rate Comparison shows that percentage difference is in
Discharge passing through spillway was the range of 0.43% to 0.63% which is quite
calculated by operating the model at different acceptable. Similarly, percentage error at free
operating conditions for validation of model. A flow condition is within 0.40 % as shown in Table
comparison between CFD model discharge and 7 Which indicates that model has successfully
physically observed one at various gate openings validated for Mangla dam spillway.
and reservoir level is shown in Table 6.
Table 2 Numerical and Physical Model Results for Scenario 1 of Boundary Condition
Sr. Reservoir Avg. Physically Observed Water Avg. Numerically Calculated Water Avg. %
Level (m) Level (m) Level (m)
NO G/O G/O G/O G/O G/O G/O G/O G/O Difference
3.05 m 6.10 m 9.14 m 10.67 3.05 m 6.10 m 9.14 m 10.67
m m
1 378.54 324.15 324.59 325.37 327.13 321.97 322.49 323.25 324.77 0.67
2 384 - - - 327.17 - - - 324.46 0.83
Table 3: Numerical and Physical Model Results for Scenario 2 of Boundary Condition
3.05 m 6.10 m 9.14 m 10.67 3.05 m 6.10 m 9.14 m 10.67
m m
1 378.54 324.30 324.59 325.37 327.13 320.42 321.24 322.90 324.10 0.98
2 384 - - - 327.17 - - - 324.26 0.89
3.05 m 6.10 m 9.14 m 10.67 3.05 m 6.10 m 9.14 m 10.67
m m
1 378.54 324.15 324.59 325.37 327.13 322.57 322.93 323.67 324.74 0.56
2 384 - - - 327.17 - - - 325.36 0.55
3.05 m 6.10 m 9.14 m 10.67 3.05 m 6.10 m 9.14 m 10.67
m m
1 378.54 324.30 324.59 325.37 327.13 319.03 320.93 322.75 323.91 1.13
2 384 - - - 327.17 - - - 324.45 0.83
Table 6: Comparison of results between CFD model and Physical Model (Gated Flow)
Sr. Reservoir Physical Model Discharge (Cumec) CFD Model Discharge (Cumec) Avg.%
Level (m) G/O 3.05 G/O 6.10 G/O 9.14 G/O 3.05 G/O 6.10 G/O 9.14 Difference
NO m m m m m m
1 378.54 523.45 1044.09 1657.88 523.88 1058.58 1628.96 0.43 %
2 384 555.92 1117.22 1790.38 556.49 1117.93 1810.91 0.63%
Table 7: Comparison of results between Physical and CFD Model (Free Flow)
Sr. No Reservoir Level (m) Discharge (Cumec) Avg. % Difference
Observed CFD
1 378.54 2077.91 2069.81 0.39
2 384 2222.43 2227.48 0.23
Water Surface Profiles gated operation of the model at reservoir level of
On successful validation of model, water surface 380.07 m, 381.59 m and 383.11 m are shown in
profile at different vertical gate openings were Fig.s 8, 9 and Fig. 10 respectively. Significant,
computed at reservoir level of 380.07 m, 381. 59 pressure variation is observed due to variation in
m and 383.11 m as shown in (Fig.s 5, 6 and 7). flow depth and velocity between 50 m to 125 m
Flow surface profiles are important to ensure that length of the chute from dam axis at large gate
flood water is not interfering with other structures opening and at all reservoir levels. Further,
such as bridges at crest or raised gates or pressure remained high up to 175 m distance at
overtopping the chute walls. The fluctuation in large gate opening (9.14 m) in all cases. Beyond
water surface is negligible along the chute but 175 m, little higher-pressure values than large
show slight fluctuation due to variation of gate opening are noted at 6.10 m against all
discharge beyond 150 m at all operating reservoir levels. At all operating conditions,
conditions. Further, continuous drop in water gradual increasing trend in pressure value is
surface level is noted in all cases. Fluctuations in observed due to high flow depth from 175 m to
water surface levels are negligible and it is not the end of the chute. Generally, pressure
interfering the bridge at crest or raised gates or remained positive throughout the length of
overtopping the chute walls throughout the spillway chute which shows that Mangla dam
length of spillway chute. spillway chute is safe against the cavitation
damages.
Velocity
Velocity indicates the potential for erosion
damage downstream of the spillway. The erosion
assessment depends on the accurate prediction of
flow velocity. CFD modelling can reliably
predict the velocities provided the model is
properly prepared. Velocity profile for gated
operation of the model at different operating
condition is shown in Fig.s 11, 12 and 13
respectively. Fig.s shows that velocity values are
increasing with distance but near the entry point
of stilling basin, velocity values drops abruptly
Fig. 5 Water Level for Gated Flow at 380.07 m due to increase of water depth. Near the crest of
of Reservoir Level spillway, flow velocity is in the range of 10 to 15
m/sec but when water moves further downward
it attains maximum value of 30 m/sec on all
operating conditions. So, velocity varies from 10
m/sec to 30 m/sec along the chute of Mangla dam
spillway. Water depth increases due to the
presence of small height weir at end of spillway
chute which reduces the velocity 10 m/sec. The
existence of two stage stilling basin at spillway
site will further reduced this velocity value which
will ultimately reduce the potential for the
downstream bed erosion.
Fig. 6 Water Level for Gated Flow at 381.59 m

of Reservoir Level
Pressure Distribution
The computations of pressure along spillway
chute are used to examine the potential for
cavitation damage due to excessive sub
atmospheric pressure. Pressure distribution for
Fig. 7 Water Level for Gated Flow at 383.11 m Fig. 10 Pressure Distribution at Reservoir Level
of Reservoir Level of 383.11 m
Fig. 11 Velocity Profile at 380.07 m of

Fig. 8 Pressure Distribution at Reservoir Level
Reservoir Level
of 380.07 m
Fig. 12 Velocity Profile at 381.59 m of

Fig. 9 Pressure Distribution at Reservoir Level
Reservoir Level
of 381.59 m
Acknowledgements
The author (Zohaib Nisar) would extremely like
to thank Engr. Kaleem Sarwar (Assistant
Professor) for their guidance, help in this
research and writing this research paper. Author
also would like to thank Dr. Ghulam Nabi
(Assistant Professor) and Center of Excellence in
Water Resources Engineering, University of
Engineering & Technology, Lahore, Pakistan for
doing research and completion of my
specialization in Water Resources Engineering.
References
Fig. 13 Velocity Profile at 383.11 m of [1] Novak, P, A.I.B, Moffat, C. Nalluri and R.
Reservoir Level Narayanan, Hydraulic Structures, Taylor &
Francis Co. 4th Edition, ISBN 0-203-96463-
Conclusion and Recommendations 2, 2007.
Computational Fluid Dynamics (CFD) model of [2] Bhosekar, V.V, V. Jothiprakash and P.B,
Mangla dam spillway was operated by varying Deolalikar, Orifice Spillway Aerator:
the reservoir levels and gate openings. Model Hydraulic Design, Journal of Hydraulic
computed the pressures at different gate openings Engineering, Vol. 138, No. 6, pp: 563-572,
and reservoir levels. The computed pressures 2012.
were found in range of 10 KPa to 350 KPa at all [3] Ho, D. K.H. and K.M, Riddette, Application
opening conditions. Pressures over the chute of computational fluid dynamics to evaluate
remain positive which indicates no danger of hydraulic performance of spillway in
cavitation. Variation in velocity over the profile Australia, Australia Journal of Civil
was noted between 10 m/sec to 30 m/sec which Engineering, Vol. 6, No.1, pp: 81-104, 2010.
reduce to 10 m/sec near the entry point of stilling [4] Chanel, P.G and J.C, Doering, Assessment of
basin. Flow depth computed near gate slot was spillway modelling using computational
6.02 m which raise to 48.81 m at the end of chute. fluid dynamics, Canadian Journal of Civil
The maximum mean level of water surface Engineering, pp: 1481-1485, 2008.
computed at a distance of 51.72 m from dam axis [5] Savage, B.M. and M.C, Johnson, Flow over
was 343.24 m which is well below top level of ogee spillway: Physical and Numerical
the chute wall 376.10 m indicating the no danger Model Case Study, Journal of Hydraulic
of overtopping. Engineering, Vol. 127, No. 8, pp: 640-649,
Model computed the discharge both for vertical 2001.
gated and free flow (vertical full gate opening) [6] Gessler, D, CFD Modelling of Spillway
conditions. For gated flow conditions, maximum performance, Proceedings of world Water
flow passed through single bay of spillway was and Environmental Resources Congress
1810.91 m3/sec while at full vertical gate 2005, American Society of Civil Engineer,
opening, it was 2227.48 m3/sec which confirms Anchorage, Alaska, 2005.
the design discharge capacity 3173.04 m3/sec at [7] Hirt, C.W and Nichols, B.D, Volume of fluid
maximum flood level of 384 m. Computed flow method for dynamic of free boundaries,
parameters with low percentage error difference Journal of Computational Physics, Vol. 39,
in comparison with observed one confirms that pp. 201-225, 1981.
model is able to assess the flow behaviour [8] Hirt, C.W. and Sicilian, J.M, A Porosity
successfully. Simulations of flows with nested technique for the definition of obstacles in
mesh technique, other turbulence model, and rectangular cell mesh, Proceedings of 4th
numerical options are required especially for the international conference of ship dynamic,
smaller gate openings. Moreover, air entrainment National Academy of Science, Washington
effect on flow parameters is also required to be D.C, 1985.
model for refinement of results. Modelling of [9] WAPDA Model Study Cell, IRI, Hydraulic
flows in stilling basin of spillway is Model Studies for Mangla dam Raising
recommended for analysis of flow behaviour and Project, Final Technical Report, Hydraulic
to check the potential for erosion damage. Research Station, Nandipur, 2004.
[10] USBR, Folsom Dam Joint Federal Project, CFD = Computational Fluid Dynamics
Existing spillway modeling, Discharge RNG = Re-normalized Group Model
Capacity Studies, California, USA, 2009. RANG = Reynold Average Navier Stokes
[11] Flow Science, Inc., Flow-3D user manuals, Equation
Version 9.2. Flow Science, Inc., Santa Fe, Stl. = Stereo lithographic
N.M, 2007. G/O = Gate Opening
Cumec = Cubic meter per second (m3/sec)
Appendix – notation
Following symbols used in this paper:
ISBN: 978-969-8670-06-01
Centre of Excellence on Water Resources

Engineering
UET, G.T. Rd, Lahore- Pakistan. 54890
Tel: +92 42 99250257
Fax: +92-42-99250259
Email: [email protected]
www.cewre.edu.pk
For orders, please contact:

© 2017 Centre of Excellence on Water Resources Engineering, Conference Organizing Committee
Tel: +92 42 99250257
UET Lahore-Pakistan. Email: [email protected]
All rights reserved. Published 2018.

Printed on recycled paper.
Printed in the Pakistan.

A Vital Source of Sustainable Energy For Pakistan

Uploaded by

Copyright:

Available Formats

A Vital Source of Sustainable Energy For Pakistan

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

A Vital Source of Sustainable Energy For Pakistan

Uploaded by

Copyright:

Available Formats

International Conference On

A VITAL SOURCE OF SUSTAINABLE ENERGY FOR PAKISTAN

CENTRE OF EXCELLENCE IN WATER RESOURCES ENGINEERING

© 2017 Centre of Excellence in Water Resources Engineering, UET Lahore-Pakistan.

All rights reserved. Published 2018.

Centre of Excellence on Water Resources

Prof. Dr. Fazal Ahmad Khalid S.I

Prof. Dr. Niaz Ahmad Akhtar

It is my great pleasure to congratulate the team of CEWRE which effectively organized

Prof. Dr. Habib Ur Rehman

Prof. Dr. Noor Muhammad Khan

Engr. M. Kaleem Sarwar

Engr. M. Kaleem Sarwar

Dr. Ghulam Nabi

Engr. Mohammad Masood

Dr. Ijaz Ahmad

Dr. Muhammad Waseem

Engr. Rana Zain Nabi Khan

December 19, 2017 (Tuesday)

Raheel Ahmad Rana

Evaluation of Small Hydropower Scheme Operating in Chitral, Pakistan; a

Detection and estimation of sediment transport trends in the upper Indus

Introduction 675 Milliman et al. (1984)

Acknowledgements Gibbons, R. D., Bhaumik, D. K. and Aryal, S.,

Hydraulic Transient Analysis for Proposed Penstock of Khari Power

Javed Munir1*, Syed Abbas Ali2, Irfan Mahmood3

Keywords: Transient analysis; Penstock; Khari powerhouse

Fig. 1 Location Map

Fig. 2 Longitudinal Section of Jari Tunnel

Existing Steel Penstock

Fig. 6 Layout Plan of Proposed Powerhouse

Fig. 7 Longitudinal Section of Proposed Penstock for 3.5 MW Powerhouse

Application of Computational Flow Dynamics (CFD) Analysis for Surge

Keywords: Surge, FLOW-3D, Numerical Model, Taunsa, RANSE

Introduction of immense importance especially during the

A detail on VOF method related issues can be

Case Study of 135MW Taunsa Hydropower

Case Study Scenarios

Normal Flow Conditions

Statistical Analysis of inflows of River Indus at Tarbela in the Wake of

Dr. Muhammad Khurram Shahzad1*

Keywords: Statistical analysis, River inflows, Tarbela dam

Introduction some of the contributing glaciers will completely

∑(𝑞𝑖 −𝑞̅𝑖 )(𝑞𝑖+1 −𝑞̅𝑖+1 )

Results and Discussion

Fig-4: Characteristics of Seasonal flows of Indus

Fig 4. volumetric discharges and its fluctuations

Fig. 4 Normally, the volumetric discharges of

Table 1 Results of Mann Kendall (MK) Test

APR I 2.5 * OCT I 0.4

F. Seemab1, W. Sherani1, M. Abbas1, M.N. Alam1, R. Farooq1, Z. Ashfaq1, M.A.U.R. Tariq1*

Keywords: Water footprint methods, hydropower generation

Evaluation of Suitable Design Flood Frequency Approaches for

Keywords: Flood frequency analysis, Indus River basin, distributions

Introduction The flood discharge adopted for the design of

Table 2 Chi-Square Test Results of Upper Indus Basin(UIB)

Conclusions Kosi river breach: an evaluation”

Impacts of hydro climatic variables trends on water resources of Yihe

Introduction numerous scientists investigated the impact of

Jan 0.07 -0.65 0.9 -2.39 0.70 -1.68*