Impact Analysis of Overflow Spillway on U/S Flows & Hydraulic
Structure using CFD Technique – A Case Study of Marala HPP
Ali Nawaz Khan1, Muhammad Kaleem Sarwar2, Dr. Sajid Mehmood3, Azhar Bashir Magsi4
1.
2.
3.
4.

Research fellow and corresponding author, Centre of Excellence in Water Resources Engineering, University of
Engineering and Technology Lahore, Pakistan. E-mail: alinawaz.ce@gmail.com,
Assistant Professor, Centre of Excellence in Water Resources Engineering, University of Engineering and Technology
Lahore, Pakistan. E-mail: eng_Kaleem@yahoo.com
Assistant Professor, Centre of Excellence in Water Resources Engineering, University of Engineering and Technology
Lahore, Pakistan. E-mail: smahmoodpk@yahoo.com
Project Manager, Sinotec Co., Ltd., Lahore, Pakistan. E-mail: azhar_magsi@yahoo.com
Abstract
Barrages and canal falls are considered as a readily available option for hydropower generation as the pre-requisites of water and head are conveniently available on such sites. Most important aspect of such scheme is to set the levels of hydraulic structures so that there is absolutely no disturbance to the irrigation flows which is the basic purpose of the barrage and canal network. At the same time finding the optimum level for the proposed structures so that the maximum hydropower benefits are yielded through the scheme without compromising the safety. Present study intends to investigate the same for Marala Hydropower Project (MHP) proposed on Upper Chenab Canal (UCC) off-shooting from Marala Barrage on River Chenab. In order to define the optimum crest level of the spillway such that there is no negative impact on the discharge passing capacity of the UCC head regulator, two-pronged strategy has been applied
i.e. using computational flow dynamics (CFD) computer software and analytical approach. The results of modelling are first compared with the physical model of the said scheme for validation. This study will be helpful for any future hydropower development schemes on irrigation canals close to barrages.
Key Words: Low Head Hydropower Development, Canal Regulation, Numerical Modelling, FLOW-3D

1. INTRODUCTION
Water is diverted into a canal from a pool behind a barrage through a structure called the canal head regulator. This structure is also used as a regulation device for controlling the amount of water passing into the canal with the help of adjustable gates [1].
Spillway is one of the foremost important structures of a dam project. It enables the project to dispose off the excess water or negotiate floods in either controlled or uncontrolled manner in order to ensure the safety of the project. Spillway should be designed with utmost care and importance must be given to assign a design discharge capacity for spillway structure to avoid overtopping.
Conventional spillway, which is also called Ogee spillway, has four main parts; the upstream crest profile, downstream crest profile, the sloping face, and the energy dissipater at the toe. Upstream of the crest, the flow is subcritical (or gradually varying); the flow changes its state from subcritical to supercritical after the crest because the crest is followed by a steep sloping face. Spillway flows are essentially rapidly varying flows with pronounced curvature of the streamlines. Two processes simultaneously occur in the flow over the crest: formation and gradual

thickening of the turbulent boundary layer along the profile and gradual increase in the velocity and decrease in the depth of main flow [2].
To obtain most of reliable and acceptable hydraulic parameters involving various hydraulic structures including dams, river training works, spillway operation, flow conditions upstream and downstream of a hydraulic structure, physical modelling is considered to be the most widely used research technique around the globe [3].
In order to develop a hydropower scheme on the canal very near to the barrage/headwork, the most important aspect is to set the levels of the hydraulic structures in consideration such that there is no disturbance to the irrigational flow of the canal which is the prime function of an irrigation canal.
To conduct such a study, physical and numerical models are developed and flow behavior is observed and analyzed for the canal reach in question [4].
Physical modelling comprise of scaling down the project components in a controlled environment of hydraulic laboratories with the best possible projection of the proto type. Generally, physical modelling is applied at the detailed engineering design stage to confirm the viable operation of

1|Page

the proto type and to develop best possible operational guidelines for the project [5].
Physical modelling does not facilitate the study of different alternatives to be done simultaneously. In order to do so in physical modelling, such modifications are modelled and conducted one after another which is laborious and very expensive job.

2
Material and Methods
2.1
Data Collection
Data was collected including UCC canal cross and longitudinal sections, existing and proposed hydraulic structures and project layout 2D drawings, hydrological data and results of physical modelling which was conducted at Hydraulic Research Station, Nandipur, Gujranwala.

In the modern era, with the use of high performance and efficient computers and CFD techniques, the solution of the above problem lies in numerical modelling of the hydraulic structures. In which, various modifications and alternatives can be modelled, simulated and studies with a lot lesser effort and expense than physical modelling [6].
1.1. Marala Hydropower Project
Marala Hydropower Project is proposed on Upper Chenab
Canal which off takes from Marala Headworks. Marala hydropower project was proposed to be designed for the discharge of 420 cumecs and net rated head of 2.16 m resulting four (04) pit type Kaplan turbines to be installed with a total power potential of 7.64 MW and mean annual energy of 43.65 GWh. A power channel is proposed on the right bank side of UCC which will house the powerhouse for the said scheme. The spillway is proposed in the main
UCC channel to cater for the surplus flows [7].
Adjustment/Fixation of crest level of proposed spillway of
Marala HPP will involve its impact on upstream flow levels which will be ultimately effect the discharging capacity of head regulator of UCC and eventually the pond level of
Marala barrage due to close proximity of both hydraulic structures. So, an in-depth analysis is required to analyze the impact of various crest levels of the proposed spillway with reference to the discharge passing capacities of existing canal head regulator and proposed spillway.

Figure 2: Birdseye view of 3D model
2.2

Setting Up of CFD Model

Three dimensional (3D) drawings were developed from 2D drawings and then converted into stereo lithographic (stl) files in order to be used in Flow 3D software. Three different files were created in order to study and analyze three scenarios of different spillway crest levels. In next step, meshing of imported Stl. file images was carried out.
The extent of mesh domain on upstream and downstream of structure is defined in such a way that it could show the fluid movement and its impact properly. Also, to increase the accuracy, nested mesh option was utilized [8]. The boundary conditions applied for this model include, volume flow rate, specified pressure, symmetry, wall and outflow.
After specifying boundaries of the model, fluids were added on the upstream and downstream sides of the structure as initial condition [9]. After pre-processing explicit and fluid flow solver option were selected to solve Reynold Average
Navier Stokes equation (RANS) which is prime equation used by Flow-3D for simulation [10].
2.3
Sensitivity Analysis
Generally Flow 3D reacts very sensitive towards the
Turbulence Model and Boundary Conditions. There are various turbulence models available in Flow 3D which include RNG model, Large Eddy Simulation model, K-ƹ model and RSN model. Due to flexibility of RNG model and lesser time required for its simulation, it is preferred in this study [11]. Four scenarios of boundary conditions were used for sensitivity analysis as shown in Table 1.

Figure 1: Plan and of proposed Marala Hydropower Project

2.4
Model Validation
Validation of numerical model is most important aspect of
CFD modelling to evaluate the level of accuracy of the
2|Page

model. Validation process indicates the degree of accuracy of the model. A proper validation process will involve comparing the numerical model results with actual performance of the actual structure. For this study no actual structure has been built yet so physical model study results are used for the validation of CFD model.
Table 1: Different Scenario of Boundary Conditions

Set

X min X max Y min Y max Z min Z max 1

SP

O

S

S

S

SP

2

V

O

W

W

W

SP

3

SP

SP

S

S

S

SP

4

V

SP

W

W

W

SP

SP:
O:
S:
V:
W:

Specified Pressure
Outflow
Symmetry
Volume Flow Rate
Wall

2.5
Flow Modelling Procedure
Case-1 was developed for validation of the CFD model by using the same parameters of flow conditions as adopted in physical model study. In addition to that, three cases (Case2 to Case-4) were developed and simulated independently for analyzing the impact of various crest levels of proposed spillway i.e. for Case-2 crest level was fixed at 243.688 masl, for Case-3 it was set at 243.535 masl and Case-4 it was fixed at 243.840 masl. In order to simulate the undisturbed pond level of Marala barrage, the pond level was fixed at 247.25 masl. Size 3 mesh was selected for the whole reach and nested mesh of size 1 was used at head regulator and spillway areas for increased accuracy.

situation, X max boundary was set to Specified Pressure as well and level was given 244.5 for the first run. By doing this, the model was able to generate the required flow conditions and then in the restart simulations, the Outflow boundary condition was selected so that the model may attain the level on the downstream of spillway following its own pattern. Set 1 and 3 were found to be relatively more accurate and were used for further analysis.
3.2
Model Validation
Results of CFD model were compared with that of physical model results and also with the results of analytical computations. Summary of comparison is presented in
Table 2 and Figure 3 below which reveals that percentage difference of 0.157, 0.163 and 0.117 (lower than physical) is recorded for the water levels upstream of spillway, downstream of head regulator and upstream of head regulator. These results appear to be quite acceptable and reason for such low values of difference is there is less margin of level difference in case of canals so the difference in CFD and physical models is also on lesser side.
Table 2: Comparison of CFD and Physical Models

Parameter
CL of SW
Max Q
U/S of SW
D/S of HR
U/S of HR
Q:
U/S:
D/S:
SW:
HR:
CL:

Moreover, analytical computations were also performed using standard procedures in order to make the comparison more comprehensive.
3
Results and Discussion
3.1
Sensitivity Analysis
Total four sets of boundary conditions were used to first simulate the model and then analyze under defined parameters and environment. Fixation of Marala barrage pond level was done by using Specified Pressure option for
X min. However, for Case-1 555 m3/s was also entered as
Volume Flow Rate in order to generate the physical model conditions. In General, Outflow is selected for X max boundary but it is successful for a model where the difference of level is big. In case of canal falls, the difference of level is comparatively smaller and if the
Outflow option is selected, it may require many simulations to become stable and produce the results. To overcome this

Physical
FLOW
Model
3D
243.688
243.688
555
555
246.40
246.013
246.75
246.346
247.20
246.91
Discharge in m3/s
Upstream
Downstream
Spillway
Head Regulator
Crest Level

%
Difference
0.157
0.163
0.117

All levels have masl units.

248

Water Levels in UCC

246
Analytical Computations
Flow 3D
244

Physical Model

CANAL REACH

Figure 3: Comparison of Water Levels

3|Page

3.3
Results of Flow 3D Scenarios
After successful validation of CFD model, the three scenarios (Case-2, Case-3 and Case-4) each having different crest level of proposed spillway were simulated and results were obtained for analysis.

3.3.2
Case-3 (Crest at 243.535 masl)
For Case-3, the spillway crest was set to the elevation of
243.535 masl (799 feet). Following figures show the rendered images at the end of simulation.

3.3.1
Case-2 (Crest at 243.688 masl)
For each case, pond level of Marala barrage was fixed to
247.25 masl. For Case-2, the spillway crest was set to the elevation of 243.688 masl (799.5 feet). Following figures show the rendered images at the end of simulation.

Figure 7: Graphical Presentation of Water Levels

Figure 4: Preview of Head Regulator

Figure 8: Preview of Case-3 Spillway

Following results including water levels at various locations and discharge passed through the system are probed through the analysis of Case-3
Figure 5: Preview of Spillway

Following results including water levels at various locations and discharge passed through the system are probed through the analysis of Case-2

Figure 9: Results of Case-3

Figure 6: Results of Case-2

3.3.3
Case-4 (Crest at 243.840 masl)
For Case-4, the spillway crest was set to the elevation of
243.840 masl (800 feet). Following figures show the rendered images at the end of simulation.

4|Page



FLOW 3D results are lesser than analytically computed results
FLOW 3D results are higher than analytically computed results



Discharge in Cumecs

Spillway Crest Level vs Discharge through
H/R

Figure 10: Preview of Spillway

644.2
640
631
623
Analytical Computations

616
601.7

Flow 3D

243.5352

243.6876

243.8400

Figure 13: Comparison of CFD & Analytical Computations

3.5

Impact on Hydropower Potential of MHP

As three scenarios in terms of different spillway crest level have been developed for this study. Their impact on the hydropower generation for the proposed MHP are also looked into. As the crest level of proposed spillway is increased, the available head is increased as well which ultimately increases the power potential of the project. Crest level of proposed spillway and power potential are directly proportional. But in previous discussion, it was discovered that the crest level of proposed spillway and discharge passing capacity of UCC head regulator are inversely proportional. Figure 11: Preview of Head Regulator

Table 4: Summary of Q & P for all scenarios
Alternative
Case-3
Case-2
Case-4

Figure 12: Results of Case-4

Table 3: Comparison of CFD and Analytical Computations

Q (CFD)
640
631
616

Q (AC)
644.2
623
601.7

% Diff
-0.65
1.26
2.32

Spillway Crest Level vs Q & P
8.5

660
640
620
600
580

8.18
7.64

7.5

7.1

7

Analytical Computations
Flow 3D
Power Potential

243.5352

243.6876

6.5
243.8400

Figure 14: Spillway Crest vs Discharge & Power

5|Page

8

Power MW

As depicted above, analytical computations were also performed for each case to establish a comparison in order to verify the results of CFD model. The comparison in terms of discharge passing capacity of UCC head regulator is tabulated as under:

Case
3
2
4

Power Potential
(MW)
7.10
7.64
8.18

CFD Model vs Analytical Computations

Discharge in Cumecs

3.4

Max Q (FLOW 3D) m3/s 640
631
616

4

Conclusion and Recommendations
I.

II.

III.

IV.

V.

VI.

VII.

VIII.

5

Analysis of results obtained from the study reveal that the there is an inverse relationship between crest level of the proposed spillway and the discharge passing capacity of UCC head regulator.
On the contrary, it can be said that the relationship between the power potential of the proposed MHP and crest level of proposed spillway is directly proportional to each other.
Analysing different scenarios with respect to various crest levels of the proposed spillway, it is safe to conclude that the crest level at 243.688 efficiently serves the purpose of generating enough head for hydropower generation without disturbing the discharge passing capacity of existing UCC head regulator and hence irrigational requirements.
The existing capacity of head regulator of UCC is
792.75 m3/s whereas, the existing capacity of UCC system is 477 m3/s. By development of power potential of 7.64 MW using head across UCC head regulator, the capacity of UCC head regulator will be reduced from 792.75 m3/s to 622.87 m3/s.
Hence UCC head regulator will have about 23% higher capacity than the system.
By development of power potential of 7.10 MW, the capacity of UCC head regulator will be reduced from 792.75 m3/s to 645 m3/s which is about 26% higher capacity than the system.
By development of power potential of 8.18 MW, the capacity of UCC head regulator will be reduced from 792.75 m3/s to 604 m3/s which is about 21% higher capacity than the system.
For future development of hydel power near head regulators on barrages, the capacity of the system should be considered about 20% higher for future irrigation system enhancements.
In order to perform CFD modelling, it is recommended to use a computing machine with high-end specifications. This will reduce the simulation time considerably and also will enable to use smaller mesh size to increase the accuracy of the results.
In complex flow situations, physical modelling is considered to be the basic source of study.
However, once a numerical model of the same is developed and calibrated with the physical model, only then it can be used for various analysis with certain level of accuracy. Therefore, it is concluded that hybrid modelling technique is the answer to the modelling question.

[1] IIT Kharagpur, “Water Resources Engineering” Module
3 Irrigation Engineering Principles, Version 2, pp 505, 2008

References

[2] Mays, L. W., “Hydraulic Design Handbook”, McGrawHill book company, 1999
[3] Eric Lesleighter, Bronson McPherson, Karen Riddette,
Jon Williams, “Modelling Procedures used for the Spillway
Upgrade for Lake Manchester Dam” GHD Publication, pp:
08, 2008
[4] Andaroodi, M. R, “Standardization of civil engineering works of small high-head hydropower plants and development of an optimization tool”, Laboratory of
Hydraulic Constructions (LCH), Ecole Polytechnique
Fédérale de Lausanne (EPFL), Switzerland publication no.
ISSN 1661-1179, 2006
[5] M R Bhajantri, T I Eldho,P B Deolalikar,
“Hydrodynamic modelling of flow over a spillway using a two-dimensional finite volume-based numerical model”
Sadhana Vol. 31, Part 6, pp: 744-745, 2006
[6] Chanel, P.G and J.C, Doering, Assessment of spillway modelling using computational fluid dynamics, Canadian
Journal of Civil Engineering, pp: 1481-1485, 2008.
[7] Irrigation and Power (I & P) Department Punjab,
“Marala (UCC 0) HPP Feasibility Study”, pp. III-1, 2005
[8] D.K.H. Ho, K.M. Boyes and S.M. Donohoo,
“Investigation of Spillway Behaviour under Increased
Maximum Flood by Computational Fluid Dynamics
Technique” 14th Australasian Fluid Mechanics Conference
Adelaide University, Adelaide, Australia, pp: 577, 2001
[9] Kevin M. Sydor and Pamela J. Waterman, “Engineering and Design: The Value of CFD Modeling in Designing a
Hydro
Plant”
Hydro-Review,
http://www.hydroworld.com/articles/hr/print/volume29/issue-6/articles/, 2010
[10] USBR, Folsom Dam Joint Federal Project, Existing spillway modeling, Discharge Capacity Studies, California,
USA, 2009.
[11] Flow Science, Inc., Flow-3D user manuals, Version
9.2. Flow Science, Inc., Santa Fe, N.M, 2007.

6|Page