INTERNATIONAL JOURNAL OF OPTIMIZATION IN CIVIL ENGINEERING

Int. J. Optim. Civil Eng., 2015; 5(3):255-266

OPTIMAL DESIGN OF GRAVITY DAM USING DIFFERENTIAL

EVOLUTION ALGORITHM
R. Deepika and C.R. Suribabu*, †
School of Civil Engineering, SASTRA University, Thanjavur – 613 401, India

ABSTRACT

The shape optimization of gravity dam is posed as an optimization problem with goals of
minimum value of concrete, stresses and maximum safety against overturning and sliding
need to be achieved. Optimally designed structure generally saves large investments
especially for a large structure. The size of hydraulic structures is usually huge and thus
requires a huge investment. If the optimization techniques are employed in the design stage,
the project investment can be effectively minimized. There are many optimization
techniques were used to optimize the gravity dam. In the present work, optimization of
gravity dam is carried out using the differential evolution technique. Differential evolution is
an evolutionary algorithm which process iteratively to locate best solution in the large search
space. Searching of optimal solution to a problem is carried out by the process of mutation,
cross over and reproduction from the initial developed candidate solutions. After undergoing
a number of iterations, it is possible to get the minimum cross sectional area of dam which
can satisfy various constraints and thus the reduction in volume of concrete can be achieved.
From the results obtained, it is found that differential evolution is one of the efficient
techniques for solving such a problem over continuous space. The success of differential
evolution in solving a specific problem critically depends on appropriately choosing trial
vector generation strategies and their associated control parameter value. The optimum
solution obtained is compared with analytical method and it is found that there is 20.44 % of
reduction in the requirement of concrete is envisaged.

Received: 10 February 2015; Accepted: 20 May 2015

KEY WORDS: optimization; gravity dam; differential evolution algorithm.

*
Corresponding author: School of Civil Engineering, SASTRA University, Thanjavur – 613 401, India
†
E-mail address: [email protected] (C.R. Suribabu)
256 R. Deepika and C.R. Suribabu

1. INTRODUCTION

Optimization techniques plays an effective role in structural design in which the structure
can be designed in such a way that the area and volume can be optimized and hence the
requirement of the materials can be effectively reduced and thus it minimises the cost of
construction. It is an important means to save investment at the same time factor of safety
and stability of the structure need to be ensured without compromise. They are several
optimization techniques available to optimize a structure. In the present work, differential
evolution algorithm has been adopted to optimize the gravity dam. Gravity dam is a
hydraulic structure that is constructed using rigid materials. Construction of concrete gravity
dam is quite common and optimizing its size can result significant reduction in the
construction materials. The external forces acting on the gravity dam are entirely resisted by
the weight of the dam itself. The main advantages of gravity dam are, it is simple to design,
it can be constructed upto any height, failure of gravity dam gives sufficient warning and
moreover it requires least maintenance. Although it has several advantages it has a
disadvantage that the requirements of material to construct gravity dam is very high which
ultimately lead to high cost of construction. The main objective of this paper is to optimize
the shape of the gravity dam using differential evolution such that to minimize the cross-
sectional area of the dam and reduce the material requirements so that construction cost is
economized.

2. LITERATURE REVIEW

Concrete gravity dam is a well known hydraulic structure constructed across the river
valley to impound the water to use for various purposes. The economy and safety of the
structure are the key elements in shape design. To achieve the economical design
without safety violations, several attempts were made to optimize the shape of concrete
gravity dam that ultimately can bring the concrete reduction. Khatibinia and Khosravi
[1] attempted shape optimization of gravity dam using gravitational search algorithm
and orthogonal cross over. Kshirsagar [2] studied the effect of variation of earth quake
intensity on stability of Tilari gravity dam located in Maharastra, India. Nannan li et al.
[3], used particle swarm algorithm coupled with ANSYS software to carried out
optimum design of gravity dam. Ali et al [4] were developed genetic algorithm based
model for optimal design of a gravity dam under seismic excitation with reservoir-dam-
foundation interactions. Xinmiao Wu et al. [5] illustrate the application of improved
particle swarm algorithm to optimize the gravity dam and sluice gate. Salmasi [6],
designed the gravity dam by means of genetic algorithm. The optimal geometry of the
dam section is obtained changing the various values of top width, the values of bottom
width of upstream batter and bottom width of downstream batter are determined such
that the dam section is safe for reservoir empty as well as for reservoir full conditions.
The top width corresponding to the minimum cross section of profile will be the optimal
top width section. GA optimization tool in excel software is used for the design of dam
corresponding to the optimal top width through an iterative process involving a
 OPTIMAL DESIGN OF GRAVITY DAM USING DIFFERENTIAL EVOLUTION … 257

preliminary layout of the structure followed by stability and stress analysis. This process
here is repeated until an acceptable cross section is attained. Kaveh and Mahdavi [7]
presented an optimization model for minimizing the weight of arch dam subjected to
natural frequencies, stability and geometrical constraints using charged system search
and particle swarm optimization. Kaveh and Mahdavi [8] used colliding bodies
optimization algorithm for the design of arch dams.The objective of formulated
problems includes minimizing the total weight of arch dam subjected to natural
frequencies, stability and geometrical limitations. This paper proposes differential
evolution model for optimal shaping of Tilari gravity dam located in Maharastra, India.
The data available in Kshirsagar [2] is used to optimize the gravity dam and the obtained
shape is compared with its existing size.

3. OPTIMIZATION MODEL FOR GRAVITY DAM

The basic dam profile is presented in Fig. 1. It is divided into three portions as upstream
triangular section, central rectangular section and downstream side triangular portion. Here,
is the height of grade change point of U/S dam slope, is the bottom width of u/s
slope, is the bottom width of d/s slope, is the height of grade change point of d/s dam
slope, n is the slope of u/s face, m is the slope of d/s face, B is the top width of the dam,
is the bottom width of the dam and H is the total height of the dam. The objective is to
minimize the cross sectional area of the dam which is expressed as sum of the individual
area of three portions of the dam.
Objective function

Minimize A(x) = 0.5 + BH + 0.5 (1) Constraints

Constraint conditions including geometrical constraint, stability constraint and stress

constraint.
Geometric constraints

(2)
(3)
(4)
(5)
(6)

Here, are the seismic co-efficient.

Stress constraints
Stress at toe and heel should be less than allowable concrete compressive stress ( ) and
it is expressed as follows
Principal stress at toe (  pD )
258 R. Deepika and C.R. Suribabu

(7)

The principal stress is calculated as per equation given below

 pD   yD sec2 D   pH'  peH

'
 tan 2 D (8)

Figure 1. Cross- section of gravity dam

Normal vertical stress at toe,

 yD = [1+ ] (9)
= (10)

Hydrodynamic pressure intensity at the base of the dam for tail water side,

= (11)

Slope of downstream face of dam

Where, - Total vertical force acting on dam under seismic condition
 OPTIMAL DESIGN OF GRAVITY DAM USING DIFFERENTIAL EVOLUTION … 259

̅ (12)

- Specific weight of water

- Height of tail water

(13)

Slope of downstream face of dam

Principal stress at heel )

(14)
Here, (15)
Normal vertical stress at heel,  yU = [1 - ] (16)
(17)

Hydrodynamic pressure intensity at the base of the dam,

(18)
tan = (19)

Where, H- Full reservoir level height

(20)

Shear stress at toe ( )

(21)

Here,

[ ( )] tan (22)

Shear stress at heel ( )

(23)
Here, [ ] tan (24)

Stability constraints
The safety factor of sliding stability of dam foundation should be less than that required
as per IS 6512-1984.
260 R. Deepika and C.R. Suribabu

F.O.S against overturning,

∑ ⁄∑ (25)
Here,
- Net positive moment acting about toe
- Net negative moment acting about toe
Shear friction factor,

∑ ∑ (26)

Here,
- Co-efficient of friction
q - Permissible shear stress at foundation
- Total horizontal force acting on dam under seismic condition

4. DAM OPTIMIZATION PROBLEM

Kshirsagar [2] conducted theoretical study on the effect of variation of earthquake intensity
on stability of gravity dam of Tilari project located at village Dhamane, taluka Chandgad,
District Kolhapur of Maharashtra State. The same data is utilized to optimize its cross-
section. This dam was constructed in 1986 on Tilari river and the location of this dam falls
under seismic zone III as per I.S. 1893-1984. Length of the gravity dam structure is 485m
and maximum discharge is designed as 1028 m 3/s. Total height of the dam and maximum
reservoir level are 38.55m and 36.2 m respectively. Top width of dam is 4.9m, Upstream
face slope of 0.10:1, Downstream face slope of 0.85:1, 13 m height is reserved for silt
deposition. Kshirsagar [2] assumed unknown data and same is retained in the present study.
The tail water level is 3 m from foundation level and centre of drainage gallery from axis of
the dam is taken as 1m. The following data are used (Krhrisagar, 2014).
Specific gravity of dam material -2.40
Coefficient of friction (µ) 0.75
Design wind velocity - 80km/h
Permissible compressive strength of concrete - 3000 kPa,
Permissible shear stress at foundation - 1200 kPa.

5. DIFFERENTIAL EVOLUTION ALGORITHM

Differential Evolution (DE) algorithm is a population based stochastic optimization

algorithm proposed by Storn and Price [9]for solving complex continuous design variable
optimization problems as a major development over genetic algorithm. The DE algorithm
searches best solution through population exploration with the help of mutation, crossover
and selection operators. DE discovers the best candidate solutions for a formulated
optimization problem iteratively until the stopping criterion is reached. The initial
 OPTIMAL DESIGN OF GRAVITY DAM USING DIFFERENTIAL EVOLUTION … 261

population can be generated randomly with in design variables bounds. The fitness value of
each candidate solution obtained from the initial population is calculated according to the
chosen objective function. The weighted differential vector is obtained after selecting two
candidate solutions from the population and multiplying its vectororial difference between
them by multiplying factor called mutation (0 to 1) and resulting weighted vector is added
with the third randomly selected candidate solution which needs to be selected from the
population other than earlier selected two candidate solutions. Thus, the new candidate
solution so obtained is designated as noisy vector and it is now subjected to crossover
process with a target vector selected randomly from the population. The candidate solution
obtained at the end of crossover process is called trial vector. The vector having better
fitness among trial and target vectors is considered as a potential candidate solution to move
the next generation. The number of candidate solutions for next generation is filled up for
chosen population size by repeating the above mentioned procedure number of times equal
to the population size. The entire process is repeated either predefined number of
generations or until specified termination criteria is achieved.
262 R. Deepika and C.R. Suribabu

Figure 2. Flow chart on the process of differential evolution

The flow chart presented in Fig. 2 illustrate the various steps in differential optimization
process in addition to the various calculation involved in the convention design steps of
Gravity Dam.

6. RESULT AND DISCUSSION

The actual size of the dam is presented in Fig. 3. The dam is having total bottom width 36.43
m and the height of upstream side sloping portion (X1) and downstream side inclined (X2)
are 30.95 and 33.35 m respectively with the value of n and m are 0.1 and 0.85 respectively.
This geometry of dam cross section makes area of 709.493 . The developed optimization
model for minimizing the cross-section of the gravity dam as single objected subjected to
 OPTIMAL DESIGN OF GRAVITY DAM USING DIFFERENTIAL EVOLUTION … 263

various constraints listed in this study provides an optimal area of 564.496 m 2 at the expense
of 500 generation while size of population is being kept as 20. The geometrical values of the
dam that obtained in the optimization model (Fig. 4) n=0.1, m=0.6, and =28.96. It can be
seen from the study that there is a reduction in cross section size from 709.493 to
564.496 . Thus, 20.44 % reduction in volume of concrete can be achieved. Hence, the
optimized size obtained using differential evolution could be economical geometry at the
same time the obtained size does not violate any constraints. Fig 5 shows the how the
optimization progress in each generation. The result presented in graph denotes the
minimum objective value obtained in each generation. There is a gradual decrease of
function value and reaches minimum at 482th generation Same solution is obtained even
after selecting different combination of mutation and cross over. Fig. 6 illustrates the
variation of slope parameters n and m over the generation. There is a clear indication that
both n and m values are decreased which resulted the reduction in the cross sectional area of
the dam. The seismic co-efficient values corresponding to optimal design is obtained as α H =
0.064 and αV = 0.053. Both these values are closer to the lower limit of the seismic co-
efficients. The factor safety against overturning and sliding and shear friction factor will be
higher if the seismic co-efficients are closer to lower limit. Hence, the obtained geometry of
the gravity dam satisfies the both economic aspect as well as safety constraints.

4.9
m

0.85
1

33.35m
30.95m

3.095
m 28.348
mm
36.43
m before optimization
Figure 3. Cross Section of dam
264 R. Deepika and C.R. Suribabu

4.9

0
.60
1

33.35
m
28.96m

2.896
m 20.01
mm
27.806m
Figure 4. Cross Section of dam after optimization

0.8

0.7

0.6
Slope parameter

0.5

0.4

0.3

0.2 n

0.1 m

0
0 100 200 300 400 500 600
Generation Number
Figure 5. Plot showing best solution obtained in each generation
 OPTIMAL DESIGN OF GRAVITY DAM USING DIFFERENTIAL EVOLUTION … 265

800
Objective Function (Area in sq.m)

700

600

500

400
0 100 200 300 400 500 600

Generation Number

Figure 6. Variation of slope parameters (n and m) over the generation

7. CONCLUSION

This paper presents the optimization model for economical design of gravity dam section.
The optimization model is solved by differential evolution. Minimization of cross-sectional
area is kept as objective function and constraints used in the model are height of grade
change point of U/S dam slope, the slope of U/S face, the slope of D/S face and seismic co-
efficient. On comparing with manual calculation based design area, there is significant
reduction of size is achieved through optimization. For this dam, volume of reduction in
concrete of 20.44% has been achieved while comparing with the existing dimensions. As the
material requirements certainly get reduced due to reduction in the cross-sectional geometry
of the dam, which ultimately leads to reduction in the construction cost of the dam.

REFERENCES

1. Khatibinia M, Khosravi S. A hybrid approach based on an improved gravitational

search algorithm and orthogonal crossover for optimal shape design of concrete
gravity dams, Appl Soft Comput 2014; 16: 223-33.
2. Kshirsagar DY. Effect of variation of earthquake intensity on stability of gravity
dam, J Indian Water Resour Soc 2014; 34(3): 1-6.
3. Nannan Li, Zhihong Qie, Xinmiao Wu, Panpan Gao. A method based on PSO RBF
to the optimization of dam structure, IEEE, 2011, 1846-50.
4. Ali AAM, Al-Suhaili RHS, Behaya SAK. A genetic algorithm optimization model
for the gravity dam section under seismic excitation with reservoir- dam- foundation
interactions, Amer J Eng Res 2014; 3(6): 143-53.
266 R. Deepika and C.R. Suribabu

5. Wu X, Qie Zh, Zhou Zh, Zhang H. Application of improved PSO to optimization of

gravity dam and sluice gate, Proceedings of the 7th World Congress on Intelligent
Control and Automation, IEEE 2008; 8(5): 6178-82.
6. Salmasi F. Design of gravity dam by genetic algorithm, World Acad Sci Eng Technol
2011; 5(8): 755-60.
7. Kaveh A, Mahdavi VR.Optimal design of arch dams for frequency limitations using
charged system search and particle swarm optimization, Int J Optim Civil Eng 2011;
1(4): 543-55.
8. Kaveh A, Mahdavi VR. Colliding bodies optimization for design of arch dams with
frequency limitations, Int J Optim Civil Eng 2014; 4(4): 473-90.
9. Storn R, Price K. Differential evolution - A simple and efficient adaptive scheme for
global optimization over continuous spaces. Technical Report, Int Comput Sci Instit
1995; 8: 22-5.