A Dissertation Report on

Seismic Analysis of Koyna Dam with Dam-Soil Interaction

Submitted in fulfilment of the requirements for the degree of

MASTER OF TECHNOLOGY

IN

EARTHQUAKE ENGINEERING

By

ALABHYA SHARMA (2019PEQ5119)

Under the Supervision of

Prof. S.D. Bharti and Dr. T.K. Datta

DEPARTMENT OF NATIONAL CENTRE FOR DISASTER MITIGATION AND MANAGEMENT

MALAVIYA NATIONAL INSTITUTE OF TECHNOLOGY JAIPUR-302017

JULY 2021

1
 CERTIFICATE

This is to certify that the work presented in the thesis entitled Seismic Analysis Of Koyna Dam

With Dam-Soil Interaction submitted by Alabhya Sharma (2019PEQ5119) is a record of

original research carried out by him under my supervision and guidance in the fulfillment of the

requirements of the degree of Master of Technology in Earthquake Engineering in the

Department of National Centre for Disaster Mitigation and Management at Malviya

National Institute of Technology, Jaipur. The matter embodied in this thesis is original and has

not been submitted to any other University or Institute for the award of any other degree.

Prof. M. K. Shrimali Prof. S. D. Bharti

Dept. of N.C.D.M.M. Dept. of N.C.D.M.M.
MNIT Jaipur MNIT Jaipur

Date:
Place: MNIT, Jaipur

2
 DECLARATION

I, Alabhya Sharma this thesis titled, “Seismic Analysis of Koyna Dam with Dam-
Soil Interaction” and the work presented in it, are my own. I confirm that:

 This work was done wholly or mainly while in candidature for a research degree at this
university.
 Where any part of this thesis has previously been submitted for a degree or any other
qualification at this university or any other institution, this has been clearly stated.
 Where I have consulted the published work of others, this is always clearly attributed.
 Where I have quoted from the work of others, the source is always given. Except for such
quotations, this thesis is entirely my work.
 I have acknowledged all main sources of help.
 Where the thesis is based on work done by myself, jointly with others, I have made clear
exactly what was done by others and what I have contributed myself.

Date: Alabhya Sharma

(2019PEQ5119)

3
 ACKNOWLEDGEMENT

I would like to express my special thanks of gratitude to my teacher Dr. S.D. Bharti (Head of
Department, N.C.D.M.M, and MNIT Jaipur) for providing a friendly environment and support
which motivated us to carry out the project.
I would like to say thanks of gratitude to my teacher Dr. M.K. Shrimali (Professor, N.C.D.M.M,
MNIT Jaipur) for his continuous for my research work.

My special quote of thanks to Dr. T.K. Dutta (Former Professor, Indian Institute of Technology,
Delhi) for sharing his invaluable knowledge with us.

I am extremely thankful to Dr. Saurabh Vern for being there for me throughout day and night
whenever I required his help and without whom this project wouldn’t be possible.

At last, I would like to thank Faizur Rahman and Bhartesh Rawat for helping me throughout my
project work.

4
 ABSTRACT

Nowadays earthquake occurrence on huge infrastructure is becoming more common due to the

increase in construction projects and urbanization. One such issue is seismic occurrence on

concrete gravity dam. Even though concrete gravity dam is huge strong structures, they are still

vulnerable to earthquakes due to the presence of large reservoir bodies that can generate

enormous stress on the structure. Hence, dam-foundation interaction is one of the most important

aspects of the seismic analysis of concrete gravity dams. In this thesis, ABAQUS software is

used to analyze the effect of concrete gravity dams subjected to earthquake loads. Moreover, the

interaction between the dam–foundation systems is explicitly taken into account by modeling the

soil. To understand the seismic effect on the Koyna dam with the dam-foundation system,

seismic analysis is performed to investigate the Displacement and stress values of the Koyna

dam–reservoir–foundation system under the influence of the 1967 Koyna earthquake.

5
 CONTENTS

1 INTRODUCTION.........................................................................................................1
1.1 General.....................................................................................................................1
1.2 Koyna Dam...............................................................................................................3
2 LITERATURE REVIEW…........................................................................................7
3 METHODOLOGY…..................................................................................................14
3.1 Mathematical Model Of Coupled Dam–Foundation System..................................15
3.2 Effect Of Hydrodynamic Pressure..........................................................................16
3.3. Absorbing Boundary..............................................................................................17
3.4 Modeling Of Unbounded Soil................................................................................22
3.5 A Comparison Among SSI Modeling Strategies …................................................25

4 RESULTS AND ANALYSIS.....................................................................................32

4.1 Creating Infinite Mesh In Abaqus..........................................................................32
4.2 Changes We Get In The Model..............................................................................33
4.3 Short Summary.......................................................................................................35
4.4 Ground Motions......................................................................................................38
4.5 Displacements.........................................................................................................43
4.6 Validation...............................................................................................................45
5 CONCLUSION…........................................................................................................47
6 REFERENCES.............................................................................................................59

6
 CHAPTER 1

INTRODUCTION

1.1 GENERAL

A gravity dam is a solid structure, made of concrete or masonry, constructed across a

river to create a reservoir on its upstream. The section of the gravity dam is approximately

triangular, with its apex at its top and maximum width at the bottom. The section is so

proportioned to resist the various forces acting on it by its weight. Concrete gravity dams serve

electricity generation, water supply, flood control, irrigation, recreation, and other purposes.

They are an integral component of society’s infrastructure system. Concerns about their safety in

a seismic environment have been growing over the past few decades, partly, because earthquakes

may impair their proper functioning and trigger catastrophic failure causing property damage and

loss of life. Though gravity dams can survive moderate earthquake motion they present a

difficult problem if they are built in seismically active areas as little is known about the response

of the dam to severe levels of excitation. High compressive stresses resulting from impacts

during crack closure and from small contact zones during the maximum crack opening. It should

be mentioned that the occurrence of cracks does not imply complete failure which is proved by

the survival of the 103 m high Koyna dam during an earthquake (1967) with a magnitude of 6.5

Richter scale and with the peak ground acceleration in the stream direction of 0.49 g [1,31]. The

duration of the strong shaking lasted about 4s and the water level stood 11.278 m below the

crest.

After the earthquake, a major crack was noted at a level of 36.576 m below the crest,
1
which coincides with the level of slope change on the downstream face. The other concrete dams

known

2
to have suffered cracking as the results of an earthquake is Hsinfengking Dam (China,1962), a

104.851 m high buttress dam, and the Shih-Gang Dam which was also severely damaged by fault

movements and ground shaking during the 1999 Chi-Chi earthquake in Taiwan.

The issues of seismic safety of dams have been looked at with increased attention

in various parts of the world in recent years. It has become a major factor in the planning and

designing of new dams, proposed to be built in seismic regions and for safety evaluation of

existing dams in these regions. To prevent the failure of a dam, it is important to assess its

behavior at any age during its lifetime, so that remedial measures can be undertaken to strengthen

or perhaps decommission the dam at the right time. For the design of an earthquake-resistant

dam and the evaluation of the safety of an existing dam, it is important to use a rational and

reliable dynamic analysis procedure. The analysis procedure should be capable of evaluating the

deformations and stresses in a dam subjected to a given ground motion. Various numerical

techniques have been used by several authors to investigate the evaluation of damage in dams.

The present study

3
 Figure 1.1 image showing element type used in Abaqus

Concentrates on evaluation dynamic response of concrete gravity dam-reservoir system.

Numerical modeling of the dam with the hydrodynamic effect of reservoir water has been done

using ABAQUS. The stresses at various points and paths of propagation of tensile damages are

computed and validated with the existing resources.

Here the fig shows the dam with rigid foundation and koyna dam with its

dimensions the graphs of transverse ground acceleration and vertical ground acceleration record

are the results which we get from Abaqus while doing the analysis

Figure: 1.1 koyna dam with fixed base

4
1.2 KOYNA DAM

Koyna Dam, constructed during the years 1954 to 1963, is a straight gravity structure

made of rubble concrete. It is about 2800 ft. long, 280 ft. high above the river bed, and 338 ft.

high above the deepest foundation. The dam is constructed in 50-ft wide monoliths, and the

contraction joints between monoliths are provided with copper water seals. The spillway or

overflow part of the dam is about 300 ft. long. To provide a more impervious zone,

conventional concrete with maximum.

An aggregate size of 3 in. is used for 6 ft. thickness at the upstream face of all monoliths, and

also at the downstream face of the overflow monoliths. The no overflow and overflow sections

of the dam, presented in Figure 1, were designed to satisfy (1) no tension in the section, (2)

maximum compressive stresses to be less than allowable stresses for the concrete used, and (3)

the shear friction factor to be less than allowable values. A seismic coefficient of 0.05, uniform

over the height, was considered in the design loadings. The design criteria adopted for Koyna

Dam are similar to the practice in the United States for gravity dams. Departure from a typical

section was the result of changes in design that had to be introduced while construction was in

progress. At first sight, the Koyna section would appear to be more vulnerable to earthquake

damage than a typical section.

5
 Figure 1.2 Koyna and typical gravity dam sections

The Koyna earthquake of 11 December 1967 had a Richter magnitude of about 6-5, the

epicenter was reported to be within 8 miles of the dam, and the focal depth was estimated as 5-8

miles. This earthquake was recorded on a strong motion accelerograph located in a gallery within

a monolith near the right bank (Figure 3). The fundamental period of vibration of this monolith is

about 0.05 sec, and it, therefore, appears reasonable to consider the accelerogram as representing

the ground motion at the site. The peak accelerations for the ground motion recorded at Koyna

Dam are 0.63 g for the longitudinal component, 0.49 g for the transverse component, and 0.34 g

for the vertical component. Typical values for Magnitude 7 earthquakes recorded at moderate

6
Epicentral distances in California are smaller; for example, the maximum accelerations recorded

during the El Centro (1940) and Tehachapi (1952) earthquakes were 0.33 g and 0.18 g

respectively. The Koyna accelerogram has higher frequency components compared to typical

accelerograms recorded in California; for example, the number of zero crossings per second is

about twice as great as for California earthquakes. The Koyna earthquake had a strong phase of

about 6-sec duration in comparison to about 24 sec for the El Centro earthquake. The response

spectrum intensity of ground shaking at Koyna Dam was about two-thirds of that at El Centro

(1940),4.5 slightly smaller than that at Taft (1952),5 and probably less than one-half of what

would be expected in the vicinity of the causative fault in a great (Richter magnitude 8 or greater)

earthquake.

The most important structural damage to the dam was horizontal cracks on either the

upstream or the downstream face or both faces of several monoliths. The principal cracking was

in the taller no overflow monoliths on both sides of the spillway section around elevation KRL

(Koyna Reduced Level) 2060, which is the level at which the slope of the downstream face

changes abruptly (Figure 1). There was evidence of relative movement between adjacent

monoliths during the earthquake. Although the dam did not appear to be in danger of a major

failure, the damage was serious enough to result in the lowering of the reservoir for inspection

and repairs and to require permanent strengthening. The structural damage caused to Koyna Dam

by the earthquake of December 1967 is of unusual engineering significance. A modern dam,

designed according to criteria and procedures similar to those employed in many countries, had

been damaged by ground shaking of moderate intensity. This experience raises several important

questions. However, in this study, we focus on the seismic analysis of the koyna dam with dam-

reservoir interaction, and the analysis is done using ABAQUS.

7
1.3 SOIL-FOUNDATION-STRUCTURAL INTERACTION STATEMENT &

CONTRIBUTION TO KNOWLEDGE.

The unavailability of standards or validated analysis techniques for estimating the soil-foundation-

structure interaction (SFSI) leads to either ignoring or simplifying this interaction. The structural and

geotechnical analysis is usually conducted individually. The geotechnical engineer may simplify a

multi-degree of freedom to a single degree of freedom oscillator. Moreover, the structural engineers

may ignore the SFSI or use simple linear springs to represent the nonlinear SFSI and neglect the

nonlinear interaction between the superstructure and the substructure (Hokmabadi et al., 2013).

During earthquake excitations, the dam interacts with the surrounding soil. The dynamic behavior of

dam and soil should be studied at the same time when dynamic loads in a particular time act on the

superstructure and surrounding soil. It has been established that dams can be designed carefully and

constructed safely against several seismic performance criteria to prevent collapse during earthquakes.

The nature of foundation, structural system, and ground motion duration and characteristics are the

primary functions of structure response (Deepa and Nandakumar, 2008).

Some simple theoretical assumptions are considered in any operation assessing the reaction of several

types of foundations under seismic action. Consequently, a simplified method represents the subsoil by

proposing a series of linear springs, while the superstructure is simulated as a single degree of freedom.

The oscillator is adopted in the code regulation regardless of the foundation type. Furthermore, the

linear equivalent behaviour for the subsoil is selected without considering the nonlinear behaviour of

soil (such as soil damping, shear modulus). The soil responses are represented directly with a constant

value of stiffness and damping during the design procedure. Therefore, in the seismic design of the

buildings, more research on soil is required considering the influence of SFSI with a rigorous
8
 accounting of the higher modes of response and different foundation types (Yegian, Mullen, and

Mylonakis, 2001).

Considering the nonlinear response of the soil under earthquake motions, the foundation-structure

interaction under the seismic action can be determined in a process involving inertial interactions

between foundation and structure, and dynamic interactions between the foundation and the soil

underneath (Tabatabaiefar and Mansoury, 2016). However, in engineering practice linear springs are

used to model soil-pile interaction in simple methods such as the Winkler model.

Due to the limitations of Winkler methods, the researchers utilized advanced analytical tools to perform

fully nonlinear mathematical models to study the seismic effects on the pile foundations. However, the

adopted numerical models need to be verified against the experimental measurements before utilizing

them as a tool for nonlinear time-history of soil-foundation-structure interaction analysis. Therefore,

efforts are required to develop a verified numerical modeling procedure to be capable of considering the

significant aspects of SFSI analysis. Thus, this model can be used for further investigation of the

influence of SFSI on the seismic response of buildings.

1.4 NUMERICAL WORK

 Development of an enhanced nonlinear three-dimensional soil-foundation-structure model.

 Detailed study of the response of the regular multi-storey dual structural system supported by

different types of foundations to the seismic events.

 Examining the adequacy of conventional design procedures excluding the influence of foundation

type to guarantee structural safety.

 Acquiring a better understanding of the fundamental parameters that affect the soil- foundation–

structure interaction under seismic loads of superstructure regarding shear distribution, the

9
 rocking of the superstructure, lateral deformations, and foundation depths.

 Studying and comparing the effects of the foundation type on the superstructure’s seismic

response about shear distribution, the rocking of structure, lateral deformations, and foundation

depths.

1.5 THE THESIS LAYOUT

Chapter 1 outlines an introduction to the aims and objectives and the organization of the thesis.

Chapter 2 presents a literature survey on the soil foundation structure interaction under seismic effects.

The dynamic behavior of structures was investigated, the available modeling techniques for (SFSI)

simulation were discussed, and the available seismic building codes related to soil-foundation structure

interaction were summarized. Furthermore, previous numerical and experimental investigations of (SFSI)

are reviewed and discussed.

Chapter 3 illustrates the modeling procedure, the scaling methodology, and the scaling factors utilized in

the simulation of the soil container and superstructure. Furthermore, instrumentation setup and the soil

container test preparation, and experimental structural models are described. The proposed numerical

model for soil foundation structure was verified using the laboratory shaking table tests. Finally, the

influence of different foundations, structure height, and soil types on the response of the superstructure

was investigated.

Chapter 4 presents the three-dimensional numerical simulation of soil foundation structure by ABAQUS

software. The numerical model’s different components such as soil elements, structural elements, pile

elements, soil container, interface elements, and boundary conditions, and the dynamic loading were

10
described.

Chapter 5 is followed by references and appendices.

11
 CHAPTER 2

LITERATURE REVIEW

Some important studies are carried out in the past years about the seismic analysis of

gravity dams and some papers also consider dam-foundation interaction. The summary of these

studies about the topic of the research is given below.

JAMES J JOHNSONS (1993) [1] the response of the structure during the earthquake depends

on the characteristics of the ground motion, the surroundings, and the structure itself. For the

structure founded on the rock or very stiff soils, the foundation motion is essentially that which

would exist in the soil at the level of the foundation in the absence of the structure and any

excavation; this motion denoted the free-field ground motion. For the soft soil, foundation motion

differs from that in the free field due to the coupling of the soil and structure during the

earthquake. The interaction results from the scattering of waves from the foundation and the

radiation of energy from the structure due to the structural vibrations. Because of this, the state of

deformation in the supporting soil is different from that in the free field. The dynamic response of

a structure supported on a soft soil may differ substantially in amplitude and frequency content

from the response of an identical structure supported on a very stiff soil or rock. The coupled soil-

structure system exhibits a peak structural response at a lower frequency than would an identical

rigidly supported structure. Also, the amplitude of the structural response is affected by the

additional energy dissipation Introduced into the system through radiation damping and material

damping in the soil.

12
A.DE FALCO, M.MORI AND G.SEVIERI [2] during earthquake shaking, the dam-reservoir-

foundation system has to be considered a coupled system. In this paper Soil-Structure Interaction

(SSI) effect is investigated on a 2D plane model of a concrete gravity dam under earthquake

excitation. Firstly, different approaches to simulate the unboundedness of soil domain are explored:

the Perfectly Matched Layer (PML) technique, the Low Reflecting Boundary (LRB) condition, and

the Infinite Elements (IEs). Different options are compared in the time domain in the case of linear

elastic material. The importance of taking into account the SSI in the seismic assessment of concrete

dams is also highlighted by the energy balance during the time. Successively, the effects of SSI are

analyzed on a full interacting nonlinear plane model. The results which are obtained in terms of

material damage and dissipated energy through a parametric SSI simulation in the time domain show

the importance of the choice of the damage constitutive law of the material.

NICHOLAS STROMBLAD (2014) [3] the mechanical behaviour of structures that are in contact

with soil is affected by the interaction between the soil and the structure. This interaction is called

Soil-Structure Interaction, abbreviated SSI. Some structures where SSI is especially important are

buildings, bridges, and oil rigs. In this thesis, a laterally loaded component of an offshore oil rig has

been studied, and in particular how to model the soil to best capture the interaction with the soil. The

component is located in the sea bed, meaning that the interacting soil is completely submerged. A

simple and conventional approach is to model the soil using lateral non-linear springs, where the

behaviour of the springs is dependent on soil type, soil properties, loading, and depth below the sea

bed. It is a widely accepted method in geotechnical engineering, but the one-dimensional spring’s

from certain limitations, and an alternative modeling strategy is needed to truly capture the complex
13
three-dimensional behaviour of soils, such as modeling the soil as a continuum with solid elements.

In this thesis, a literature study was carried out to understand the mechanical behaviour of soils, and

especially CONTINUED… the method used to determine the behaviour of the non-linear springs.

Then, a finite element model of the specific component and the surrounding soil was created in

Abaqus/CAE, for three different soil compositions. The soil was modeled both with non-linear

springs and as a continuum, and the results from the two modeling strategies were compared, along

with the identification of limitations and benefits for the two approaches. For the static load case

studied it can be concluded that the results are similar, but there are also some notable differences

mainly dependent on soil composition and material properties. Continuum soil modeling is more

complex and computationally expensive than non-linear springs modeling but is advantageous in the

prediction of long-term erects. Future work should focus on the determination of a soil's properties,

as these highly influence the results for both modeling strategies. With an accurate characterization

of a soil's behaviour, the response of the structure will be predicted as correctly as possible.

SARKAR (2007) [4] used the profile of the Koyna dam for the study of the response of a dam

subjected to dynamic loading. For simulating the damage induced in the dam under an earthquake

motion, the nonlinear concrete properties have been taken into account through the concrete

damaged plasticity model. Compressible reservoir water and linear elastic homogenous foundation

material are considered in the model and analyzed by ABAQUS. The result shows that the behavior

of dam structure under an earthquake largely depends on the foundation and the reservoir. Also, the

tensile damage patterns of the systems considering only the dam structure, and the dam with the

foundation, are almost the same but it varies when the reservoir is considered along with the dam

and foundation.

14
15
 A. K. CHOPRA AND P. CHAKRABARTI (1995) [5] investigated several studies

about the impounded water and foundation rocks that significantly affect the response behavior

of the dam. The results obtained from the separate analysis of dam-water interaction and dam-

rock interaction will be invalid because the two problems are coupled. This paper provides a

developing technique for analyzing earthquake response of dam considering both impounded and

foundation rock effects together. This was achieved by using the substructure method, combined

with the transformation of displacements to generalized coordinates. This procedure is developed

mainly for two- dimensional analyses of gravity dam monoliths, where the dam was discretized

as a finite element system but the foundation and water bodies are treated as continua.

HANCHEN TAN AND ANIL K. CHOPRA (1995) [6] extended the available substructure

method and computer program for the earthquake response analysis of arch dams to include the

effects of dam-foundation rock interaction with inertia and damping of the foundation rock. The

objective of this study is to develop an effective procedure for analyzing the response of concrete

arch dams to earthquake ground motion, including the effects of dam-foundation rock interaction

with inertia and damping of the foundation rock considered, dam-water interaction, and reservoir

boundary absorption. In dam-foundation rock interaction, the computer method assumes

massless foundation rock ignoring foundation material damping and even significant interaction

effect. In this paper, the foundation impedance matrix is computed by the direct boundary

element approach. The resulting computational procedure represents the most efficient tool for

calculating the earthquake response of arch dams.

16
ALEMBAGHERI (2016) [7] proposed a new methodology to estimate earthquake damage of

gravity dams. In this methodology, the static pushover analysis is used to formulate a systematic

and rational procedure to estimate the level of damage from the linear seismic analysis results.

The tensile cracking of concrete is considered as the main potential damage. Examples of three

existing concrete gravity dams are provided to illustrate the methodology and discuss the

probable nonlinear response and failure mechanisms. The damage state of the dams under twelve

proper earthquake ground motions scaled to increasing intensity levels is estimated using the

newly proposed criteria and verified using the actual nonlinear time-history analysis. Finally, a

damage index is defined to quantitatively predict the earthquake damage of gravity dams.

WANG (2017) [8] carried out the three-dimensional nonlinear finite element analyses for the

Guandi dam-reservoir-foundation system considering the effects of contraction joints and cross-

stream excitation. The reservoir water is modeled with three-dimensional fluid finite elements in

the lagrangian formulation for accounting for the dynamic interaction between impounded water

and the dam-foundation system. The experimental results of the Koyna concrete dam were used

as a validation for the effectiveness of the numerical framework. From the results, it is observed

that the existence of contraction joints could reduce the integral rigidity of the monolithic model,

thus leading to larger nonlinear demands in the dam-reservoir-foundation system. Several case

studies are examined and the results reveal the significant influence of contraction joints.

17
 CHAPTER 3

METHODOLOGY

Dam–foundation and dam–reservoir interactions are two important aspects in the

dynamic analysis of dams. The first one can be considered by simplified or rigorous methods.

Solutions depending on the simplified methods are approximate. More realistic results may be

obtained by using rigorous methods, which require finite element or boundary element modeling

of foundation domain and result in increasing computation time and computer storage

requirement. Many DOFs are not desired in the non-linear analysis of dams because of the

necessity of an iterative solution and using very small-time integration steps.

However, dam–foundation rock interaction effects in the dynamic response of dams are

strongly dependent on the elastic modulus ratio of the foundation to the dam. The solutions

including flexible foundation approximate those of the infinitely rigid foundation case as this

ratio increases. Therefore, the foundation can be modeled as being rigid if the foundation rock is

much stiffer than the dam concrete, and also if the earthquake acceleration components are

recorded near the dam foundation.

The latter aspect, dam–reservoir interaction, may be considered in three ways: added

mass, Eulerian and Lagrangian approaches. The added mass approach is the simplest and first

technique and it employs the results of rigid dam–incompressible fluid interaction. More realistic

solutions for dam–reservoir interaction can be obtained by using Eulerian and Lagrangian

approaches. In the Eulerian approach, displacements and pressures are the variables in the

structure and reservoir domains, respectively, while displacements are variables in both domains

for the Lagrangian approach.

18
 In the finite element analyses of dam–reservoir systems using these approaches, a

radiating condition (for example, Sommerfeld radiation condition) may be applied to the

truncated boundary of the reservoir as one way for considering radiating damping. However, the

effect of the radiating condition on the solutions is generally negligible if the reservoir length

taken three or more times its height. Many researchers investigated the dynamic interaction of

dam–reservoir systems taking the reservoir length as three times its height.

The soil effects on the seismic behavior of buildings are seldom explicitly taken into account in

finite element models. More often, to overcome the difficulties involved in modeling soil-

structure interaction, code-provided response spectra depending on suitable soil categories are

used. The particular characteristics of the retaining structures do not allow us to use the simplified

methods which are commonly applied to ordinary buildings, so dam earth-quake safety

assessment requires the direct modeling of soil-structure interaction.

Soil-Structure Interaction is described to have two main components: kinematic and inertial

interaction [1]. In the literature, two contributions are relevant in approaching the problem of

soil-structure interaction, those by Wilson [2] and Wolf [1]. Wilson describes the “massless

foundation” method, basing on the consideration that the recorded ground motions are acquired at

the terrain surface where the response has already experienced the effects of the soil. This model

is governed by soil flexibility. The massless foundation model proposed by Clough in 1980 [3]

has been extensively used in seismic analysis of dam-foundation problems. In this context,

recorded displacements are imposed at the boundaries of the domain and the input motion reaches

instantaneously the base of the dam disregarding the inertial interaction. In the massless model,

the wave velocity in the foundation becomes infinite, so the input motion reaches instantaneously

the base of the dam and the structure takes all kinetic energy.

19
Inertial interaction is generated by elastic waves that develop under dynamic loads, promoting the

energy transport through the soil volume. Such a phenomenon that carries energy away from the

structure is often referred to as “radiation damping”. So, while in static SSI analysis the simple

truncation of the far-field with the setting of appropriate boundary conditions gives very often

good results, in dynamic cases it makes results to be erroneous because of reflection waves.

The present paper addresses Soil-Structure Interaction (SSI) for existing concrete gravity dams,

investigating its effects numerically on a 2D plane system under earthquake excitation. To

simulate the unboundedness of the soil domain, different modeling approaches are explored: the

Perfectly Matched Layer (PML) technique [5], the Low Reflecting Boundary (LRB) condition

[6], and the Infinite Elements (IEs) [7]. Moreover, the analysis in the time domain is performed to

compare the response of the model in terms of base shear, using different modeling options. The

energy balance during time shows the importance of modeling the SSI in the seismic assessment

of concrete dams. The results in terms of base shear values for several modeling approaches are

compared to highlight the differences among different modeling strategies.

Finally, the effects of SSI are analyzed on a full interacting nonlinear plane model with nonlinear

material constitutive behavior. The results in terms of material damage are obtained from a

parametric SSI simulation in the time domain, to take into account the presence of soil beneath

the dam.

The main contribution of this work is to compare different modeling strategies and to

highlight the importance of SSI modeling for the evaluation of the seismic behavior of concrete

gravity dams.

20
3.1 MATHEMATICAL MODEL OF COUPLED DAM–FOUNDATION

In dam–foundation interaction problems, the foundation and the structure do not vibrate as separate

systems under external excitations, rather they act together in a coupled way. Therefore, these

problems have to be dealt with in a coupled way. The most common SSI approach used is based on

the ‘‘added motion’’ formulation. This formulation is mathematically simple, theoretically correct,

and is easy to automate, and is used within a general linear structural analysis program. In addition,

the formulation is valid for free-field motions caused by earthquake waves generated from all

sources. The method requires that the free-field motions at the base of the structure be calculated

before the SSI analysis. To develop the fundamental SSI dynamic equilibrium equations, the soil–

structure system, as shown in Fig. 1, is considered. The absolute displacements of the structure are

considered to be the sum of two parts, viz. free field displacements and added part of the

displacements. Free field displacement is found out by analyzing the foundation

Figure 1.3 Coupled Dam Model Of Structure

21
Domain with no structure present on it against the earthquake forces. The added part of the

displacement is found out by carrying out coupled soil-structure interaction model. The SSI model

here is divided into three sets of node points, viz. the common nodes at the interface of the structure

and soil are identified with the subscript ‘‘c’’; the nodes within the structure are with ‘‘s’’ and the

nodes within the foundation area with ‘‘f’’. From the direct stiffness approach in structural analysis,

the dynamic force equilibrium of the system is given in terms of the absolute displacements, U, by the

following sub-matrix equation:

Where the mass and stiffness at the contact nodes are the sum of the contributions from the structure

(s) and foundation (f) and are given by

To solve the coupled soil-structure interaction problem, we would require to solve Eq. (1). Having

solved Eq. (1) using Newmark’s integration method, one would obtain the absolute displacements,

velocities, and accelerations of the coupled SSI problem. To avoid solving the SSI problem directly,

the dynamic response of the foundation without the structure is calculated. The free-field solution is

designated by the free-field displacements v, velocities _ v, and accelerations € v. Here, €U g is the

ground acceleration vector. By a simple change of variables, it becomes possible to express the
22
absolute displacements U, velocities _U and accelerations €U in terms of displacements u, relative to

the free-field displacements v. Or,

After replacing the values of €U, _U and U from Eq. (3), Eq. (1) is expressed as

23
This is a numerically cumbersome approach; hence, an alternative approach is necessary to formulate

the solution directly in terms of the absolute displacements of the structure. Since the analysis is now

for the foundation part only (free field analysis), hence the corresponding values of the displacement,

velocity, and acceleration for the structural part are taken as zero. This involves the introduction of the

following change of variables:

To calculate the free field displacements v, only the foundation domain is solved by considering no

structure is present on it. The foundation domain is subjected to earthquake motion and the free-field

displacement for the common and other foundation nodes is obtained.

After obtaining the free field response (i.e. v, _ v, and € v) the interaction force R is calculated using

Eq. (9) in the following simplified manner:

24
After obtaining the interaction forces R, the added responses of the dam and foundation domain are

calculated using Eq. (10). And then the added responses (i.e. u, _ u and € u) are added to the free field

responses to get the absolute responses of the coupled soil and structure domain, following Eq. (7):

The main assumptions used in this model are that the input motions at the level of the base rock are not

considered to be affected by the presence of the dam and that all interface nodes will be subjected to

the same free-field accelerogram. In theory, any desired spatial variation of the free-field components

could be considered at the interface, but there is seldom sufficient information to specify such

variation. In this case, the mass of the foundation is taken into account in the analysis such that it will

represent the dam–foundation interaction in a relatively more realistic manner.

3.2 EFFECT OF HYDRODYNAMIC PRESSURE

The effect of hydrodynamic pressure is considered according to added mass technique originally

proposed by Westergaard. Assuming the reservoir water to be inviscid and incompressible and its
25
motion to be of small amplitude, the governing equation for hydrodynamic pressure is expressed as

∆ 2 P=0. The solution of this equation is proposed by Westergaard and is used in the present work to

calculate the hydrodynamic pressure imposed on the upstream face of the dam body during any

earthquake

3.3 ABSORBING BOUNDARY

A way to eliminate waves propagating outward from the structure is to use Lysmer and Kuhlemeyer

boundaries. This method consists of simply connecting dashpots to all degrees of freedom of the

boundary nodes and fixing them on the other end (Fig. 2). Lyster and Kuhlemeyer boundaries are

derived for an elastic wave propagation problem in a one-dimensional semi-infinite bar. The damping

coefficient C of the dashpot equals

C=Aρc

Where A is the cross-section of the bar, ρ is the mass density and c is the wave velocity that has to be

selected according to the type

26
 Figure 1.4 Viscous Dashpot Connected To Each Degrees Of Freedom Of A Boundary Node.

Of wave that has to be absorbed (shear wave velocity C s or compressional wave velocityC p). In two

dimensions Eq. (12) takes the following form, which results in the damping coefficient C n and C❑ in

the normal and tangential direction, respectively.

C n= A1 ρC p ………………………………………………. (13)

C t= A 2 ρ C s … … … … … … … … … … … … … … … … … … ..(14)

The shear wave velocity C s and compression wave velocity C p is given b

G
C s=
√ P
……………………………………………………………….(15)

27
 E
G= 2(1+V )

Here E is Young’s modulus and n is Poisson’s ratio. However, in general, the directions of the incident

waves are not known in advance. In these cases, it’s advantageous to use a ‘diffused’ version as

suggested by White et al. [18]. Assuming that the wave energy arrives at the boundary with equal

probability from all directions, effective factors A1 and A2 are evaluated by minimizing the ratio

between the reflected energy and the incident energy over the range of incident angles. For an

isotropic medium this results.

3.4 MODELING OF UNBOUNDED SOIL

Dam models seldom take into account full interaction effects, because of the lack of adequate

numerical implementations or computational resources required by three-dimensional detailed models.

Recently, SSI for concrete retaining structures is addressed by many authors searching for a reliable

simulation of wave propagation in a semi-infinite medium, modeling the far field part of the

foundation. The unboundedness of the terrain was first considered by Lamb [8] in its classical

problem of a point load on a half-space, for which he provided an analytical solution. Wolf, on the

other hand, developed a formulation for appropriate spring-dashpot coefficients and boundary

conditions. Some worth noting methods are Lysmer boundary conditions, hyper elements, infinite

elements, rational boundary conditions, boundary element method, scaled boundary element method,

and high order non-reflecting boundary conditions.

To simulate the unboundedness of both solid and fluid domains, three different modeling options are

explored in this work, the Perfectly Matched Layer (PML) technique, the Low Reflecting Boundary

(LRB) condition, and the Infinite Elements (IEs). PML has been widely used for simulating wave

propagation in unbounded media to effectively avoid spurious wave reflections from the computational
28
domain boundaries. This technique can absorb incident waves under any angle and frequency,

preventing them from returning to the medium after incidence to the model boundaries. The procedure,

which was first introduced by Berenger in 1994, maybe applied to different physical problems. It

comes to a complex coordinate stretching of the domain to introduce a decay of the oscillation

avoiding any reflection in the source domain, thus simulating a perfectly absorbing material. The

rational scaling of PML is expressed by the following function of the dimensionless coordinate ξ

Where p is the curvature parameter and s the scaling parameter. Implementation of PML into most

commercial FE software packages is performed only in the frequency domain because the majority

of the formulations in the time domain highly affect the computational time and resources.

Which are used to incorporate unbounded domains into the finite element method, have a

formulation similar to those of FEM, except for the infinite extent of the element region and shape

function in one direction. I am based on a function that maps the global to the local coordinate.

Where ∆p is the pole distance,

And ∆s is the scaled thickness [7].

29
Finally, the LRB condition is obtained by imposing a mechanical impedance on the foundation

boundary of the model, following the equation

Where u is the displacement vector, [T] the stress tensor, n the unit vector of the boundary tangent

plane, [D] is the impedance matrix [17]

Where K is the bulk modulus of the soil, G the shear modulus, ρ the material’s density, and [I] the

identity matrix.

3.5 A COMPARISON AMONG SSI MODELING STRATEGIES

Different modeling strategies are analyzed and compared in terms of the resultant base shear force of

a 2D plane strain system under earthquake excitation. The analyses are carried out in the time

domain using COMSOL Multiphysics [18].

The dam and soil domains are modeled by applying the standard Solid Mechanics equations. Three

different techniques which have been discussed above are used to simulate the un-boundedness of

the soil: the Perfectly Matched Layer (PML) technique, the Low Reflecting Boundary (LRB)

condition, and the Infinite Elements (IEs). The study in the time domain is performed on an Italian
30
concrete gravity dam 65 meters tall, with a base of about 45 m, using four plane strain models, as

shown in figure 1. Reference model 1 simulates the dam on a rigid terrain. The solid mesh is

composed of 453 default second-order serendipity elements and the displacements of the base nodes

are restrained along with both directions.

In this case, the basin was simulated by the added mass model [19].

Figure 1.5 Different SSI Modeling Strategies for a Plain Strain Model of a Dam

The other three models account for massed foundation as an unbounded half-space. The second

model simulates the dam with the soil equipped with low reflection boundaries. The solid mesh

is composed of 3146 default second-order serendipity elements for both the soil and the dam.

The low reflecting boundary condition is defined on the boundary of the soil domain. Its

impedance parameter values are derived from the elastic properties of the soil domain. The third

model is similar to the second one, but in addition, 550 infinite elements are surrounding the

physical soil region. The solid mesh is thus composed of 3696 default second-order serendipity

elements. The infinite element pole distance is 400 m and the scaled thickness is 400 m.

31
 The fourth model of fig. 1 simulates the unboundedness of the soil through the PMLs applied at

the bottom and on the sides. Unfortunately, the PML technique is not always available in each

program code or, if present, may not be compatible with time-domain analyses. For this reason, a

suitable 2D simplified model with horizontal and vertical springs and dashpots at the base was

created in [20] to take into account the SSI by using tools and standard boundary conditions.

Two viscous damping coefficients and two spring stiffness coefficients were calibrated in [20],

starting from a model which is similar to that of case 2 of figure 1 but is equipped by PML

boundary elements around the soil domain, instead of low reflecting boundaries. Calibration has

been performed to provide the same resultant shear force at the base of the dam in the frequency

domain for both models: a simplified model and the interacting 2D model with PML boundary

condition. The result is that there is a set of coefficients that are capable to simulate SSI in terms

of resultant shear force at the base of the dam. Their values are independent of the frequency in

the range between 0 Hz and 25 Hz with good approximation and they are reported in table 1.

Figure 1.6 Dam Modeling Using Spring/Dashpot Strategies

STEP 1: PARTS - SKETCHING

32
 Figure 1.7 Sketch of Dam in Abaqus

The first and foremost part of Abaqus is sketching the model in the part section. Here we create a part

and by using the editing techniques and the proper dimension we can make a model. This model is fully

based on the finite element approach and used for the analysis by further action. In the part menu there

are several options for the sketching of the dam we can use a straight line, directly create a rectangle

with proper dimension, can create a circle, and many more shapes. There is an option of free editing

also here in this figure 1.7 I have taken a dimension of koyna dam which are standard values.

The sketching of this model is further use in doing modeling and analysis-

33
 Figure 1.8

34
 Figure 1.9

Fig 1.8 and Fig 1.9 are the dam with rigid base and dam with the infinite foundation during the

part menu

STEP 2 – PROPERTY

In the property menu, we put the property of the material used in the structure like

35
 Figure 1.10 Properties in Abaqus

STEP 3 – ASSEMBLY

In assembly, we connect each part of the model using the tools given in Abaqus.

Figure 1.11 assembly tools

36
 Figure 1.12 with a rigid foundation

Figure 1.13 with infinite foundation

STEP 4 – STEP

Step contains step manager, history output request manager, and field output requests manager

37
 Figure 1.14 step view

Here the type of analysis has to be described. In this case, nonlinear dynamic analysis is selected

for a period of t=10seconds. In the initial time step, displacements of the nodal points on the left

and right truncated boundary of the reservoir and foundation are assumed to be zero in the

normal direction. In addition, the bottom boundary of the foundation is fully constrained. After

static analysis, displacement constraints are released, and the horizontal and vertical components

of the Koyna earthquake accelerations are applied to the base of the foundation as input load.

38
which is scaled to have 0.34 g spectral acceleration in the stream direction, is used as seismic

excitation. The implicit Newmark-β method is used in the incremental dynamic analysis for time-

stepping, which is unconditionally stable. Any time step will lead to stable results. However, to

ensure accurate time integration, a sufficiently small-time increment is needed. On the other

hand, excessive time steps result in a high burden and poor computational efficiency. Initial

increment size and maximum increment size are 0.001 s and 0.01 s, respectively. Meanwhile,

automatic time incrementation is used for the dynamic analysis to further improve efficiency.

STEP 5 – INTERACTION MANAGER

Once the assembling is done the parts have to be able to interact with each other. Hence the

interaction step has to be done. Here the solid section dam and base are connected using a tie

Connection. Whereas water and dam are connected by surface to surface contact using the

penalty method. We don’t have to use interaction manager as we are not creating a surface to

surfaces interaction it is only used in the hydrodynamics effect.

Figure 1.15 interaction manager and constraints menu

STEP 6 – LOAD

39
 In this step, load and boundary condition has to be given. Gravity load is assigned to the dam

section and hydrostatic load is applied to the water. Also, the boundary condition inputs are given.

Applied loads include self-weight of the dam, hydrostatic, uplift, and earthquake forces. The static

solutions of the dam due to its gravity loads and hydrostatic loads are taken as initial conditions in

the dynamic analyses of the system.

Figure 1.16 with rigid base gravity load

Figure 1.17 with infinite soil gravity load and EQ

40
STEP 7 – MESH

In this section, the meshing of every part has to be done. The finite element model of the dam–

foundation interaction system. While modeling the mesh, the element type has to be defined.

Figure 1.18 mesh menu

Figure 1.19 dam with infinite foundation meshing

Figure 1.20 dam with rigid base meshing

41
STEP 8 – JOB

This is the former step where the data check can be done and input will be submitted for

processing. Once the processing is done the results of the model can be seen in the visualization

part. Also, there is an option ‘write input’ which can be used to generate an input for the current

model. This input file will be in doc format and can be imported in ABAQUS to convert it into a

model file.

Figure 1.21 job view

Here the job manager is used to contain information about the job activity which we have

created all job activities are store in the form of .cae .jnl and .odb formate in the computer

temp folder. We can see these results whenever we want and also we can monitor the

42
 analysis if some error arises then it also displays that error

CREATING INFINITE MESH IN THE ABAQUS (PART OF JOB)

For creating infinite mesh we make use of the input file of that model. The changes we want to do

are mandatory to do in the input file.

Input file:-

Figure 1.22 input file to be changed

43
CHANGES WE GET IN THE MODEL-

Figure 1.23 dam soil with infinite foundation

44
 Figure 1.24 view after removing the elements and nodes number

SHORT SUMMARY
PARTS USE FOR DOING SKETCHING OR CREATING THE MODEL ACCORDING
TO DIMENSIONS
DAM HEIGHT-103 M
TOE WIDTH – 14.2 M
HEEL WIDTH – 70.2 M
BOTTOM NECK TO TOP DISTANCE – 36.5 M
TOP SLOPE – 15.56
RIGID SOIL 90.2X10
INFINITE SOIL 976X244
PROPERTIES USE FOR PROVIDING MATERIAL PROPERTIES.

DAM FOR DAMPING-

RIGID BASE BETA - 0.00323
DENSITY
MASS DENSITY – 2643
ELASTIC-
YOUNGS MODULUS – 31027000000 POISSONS RATIO -0.2
CONCRETE DAMAGE PLASTICITY –
DILATION ANGLE – 36.61(PLASTICITY)
ECCENTRICITY – 0
K–0
VISCOSITY PARAMETER – 0
COMPRESSIVE BEHAVIOUR
YIELD STRESS INELASTIC STRAIN
13000000 0
24100000 0.001

DAM FOR DAMPING-

INFINITE SOIL BETA - 0.00323
DENSITY
MASS DENSITY – 2643
ELASTIC-
YOUNGS MODULUS – 31027000000 POISSONS RATIO -0.2
CONCRETE DAMAGE PLASTICITY –
DILATION ANGLE – 36.61(PLASTICITY)
ECCENTRICITY – 0
K–0
45
 VISCOSITY PARAMETER – 0
COMPRESSIVE BEHAVIOUR
YIELD STRESS INELASTIC STRAIN
13000000 0
24100000 0.001

RIGID SOIL SAME CONCRETE PROPERTIES

INFINITE SOIL DAMPING-

ALPHA – 1.64
BETA – 0.0012
COMPOSITE – 0
DENSITY-
MASS DENSITY – 2643
ELASTIC-
YOUNGS’S MODULUS POISSONS RATIO
27580000000 0.333

Table 4.1 short summary of the values used in the modeling

46
 CHAPTER 4
RESULTS

During earthquake excitations, the rocking of a structure may occur due to the inertial forces generated

within the soil-foundation-structure system. The rocking of the structure causes compression stress on

one side and tension stress on the other side of the foundation. As a result of these stresses, settlement

may occur on one side and possible uplift on the other side of the foundation. Each type of foundation

undergoes a different experience incurred by the structure rocking. The rocking component amplifies the

lateral structure displacement and may influence the total stability of the soil structure system. However,

a significant amount of ground motion energy dissipates due to the structure rocking which resulting in

lower shear forces indirectly applied to the structure. Comparing the behavior of different foundation

types concerning the soil-foundation-structure interaction helps the engineer to design the proper

foundation to resist the impact of soil-foundation-structure under dynamic effects (Poulos et al., 2015).

FOR THE RIGID SOIL-

47
 Figure 1.25 results for rigid foundation

FOR INFINITE SOIL-

Figure 1.26 dam with infinite soil

4.1 Ground Motions

The ground motion of both the longitudinal horizontal and vertical components of the 10 th December

1967 Koyna earthquake with Magnitude 6.5 and focal depth 12.0 km, recorded at the foundation
48
gallery of the dam at an epicentral distance of 12.6 km shown in Figure 2, has been used to compute

the numerical results. These components are characterized by corrected peak ground acceleration

(PGA) values 0.49g and 0.24g, respectively.

transverse ground acceleration record

0.6
0.4
Acceleration (g)

0.2
0
-0.2
-0.4
-0.6
0 1 2 3 4 5 6 7 8 9 10
Time(sec)

Figure 1.27

vertical ground acceleration record

0.5
0.4
0.3
0.2
Acceleration (g)

0.1
0
-0.1
-0.2
-0.3
-0.4
-0.5
0 1 2 3 4 5 6 7 8 9 10
Time (sec)

Figure 1.28 horizontal and vertical components of koyna dam accelerograms used in the present study

4.2 GRAPHS

The peak values of the horizontal displacements of the upstream nodes of the dam section are

plotted along with the height of the dam for different values of displacement on upstream nodes for

different values of the time.

49
 FOR RIGID FOUNDATION

Figure 1.29 top crest displacement u1

Figure 1.30 bottom crest displacement u1

50
Figure 1.31 resultant for rigid foundation

FOR RIGID FOUNDATION

51
Figure 1.32 displacement upper crest u2

Figure 1.33 displacement bottom crest u2

RESULTANT

52
Figure 1.34 resultant displacement for the upper crest and lower crest u2

FOR INFINITE SOIL

Figure 1.35 upper crest displacement for infinite soil u1

53
Figure 1.36 bottom crest displacement for infinite soil u1

RESULTANT

Figure 1.37 resultant displacement for infinite soil u1

54
 FOR INFINITE SOIL (U2)

Figure 1.38 upper crest displacement for infinite soil U2

Figure 1.39 bottom crest displacement for infinite soil U2

55
 RESULTANT

Figure 1.40 resultant displacement for infinite soil U2

For Dam-Foundation Interaction with Rigid Foundation –

56
figure 1.41 Time history graph of maximum principal stress occurred at heel

Figure 1.42 Time history graph of minimum principal stress occurred at heel

Figure 1.43 Time history graph of maximum principal stress occurred at toe

57
 Figure 1.44 Time history graph of minimum principal stress occurred at toe

Table 4.2 Peak value of Maximum and minimum principal stress at heel and toe of dam

58
 (RIGID FOUNDATION)

4.3 VALIDATION

In order to validate this ABAQUS koyna dam model we have taken maximum displacement of rigid

base and maximum displacement of infinite soil. By using origin we create these graph which truly

resembles the data.

Figure 1.41 this graph is showing the validation

Here red line showing peak displacement of infinite soil while the black line showing the peak
59
displacement of rigid base

CHAPTER 5
CONCLUSION

The paper presents a methodology for the analysis of concrete gravity dam subjected to seismic

excitations considering the soil– structure interaction effect. The proposed method is validated from

the literature which shows the accuracy of the developed algorithm. The dam like structure, having the

coupling effect due to the underlying foundation material during earthquake excitations is analysed.

The numerical results presented here prove the efficiency of the present algorithm to solve a soil–

structure coupled problem of massive structures such as concrete gravity dams. The advantage of using

the present direct method is that it requires less computational effort, in terms of both time and

memory. The responses of the soil–structure system considering an absorbing boundary indicate that

the incident energy is effectively absorbed at the truncation boundary. Another advantage of this

method is that it requires less computational effort since it avoids evaluation of convolution integrals

and Fourier transforms to calculate soil–structure interaction forces. The algorithm presented here is

simple so that it may be programmed easily. The results show that the displacements and stresses have

increased for the elastic as compared to the rigid base. Hence it is advisable to carry out the interaction

analysis for massive structures like dams under flexible base. It is also observed that the neck is the

most severely stressed zone; hence one may expect the appearance of cracks around the neck region of

the dam.

60
 CHAPTER 6

REFERENCES

1. Wang G, Wang Y, Lu W, Zhou C, Chen M, Yan P. XFEM based seismic potential failure mode analysis of concrete

gravity dam-water-foundation systems through incremental dynamic analysis. Eng Struct 2015;98:81–94.

https://doi.org/10.1016/j.engstruct.2015.04.023.

2. Zhang S, Wang G, Yu X. Seismic cracking analysis of concrete gravity dams with initial cracks using the extended

finite element method. Eng Struct 2013;56:528–43. https://doi.org/10.1016/j.engstruct.2013.05.037.

3. M.A.Lotfollahi Yaghin and M.A.Hesari. Dynamic Analysis of the Arch Concrete Dam under Earthquake Force with

ABAQUS 2008.

4. Bayraktar A, Altunişik AC, Sevim B, Kartal ME, Türker T, Bilici Y. Comparison of near- and far-fault ground

motion effect on the nonlinear response of dam-reservoir-foundation systems. Nonlinear Dyn 2009;58:655–73.

https://doi.org/10.1007/s11071-009-9508-x.

5. Calayir Y, Dumanoǧlu AA, Bayraktar A. Earthquake analysis of gravity dam-reservoir systems using the Eulerian

and Lagrangian approaches. Comput Struct 1994;59:877–90. https://doi.org/10.1016/0045-7949(95)00309-6.

6. Wang G, Wang Y, Lu W, Yu M, Wang C. Deterministic 3D seismic damage analysis of Guandi concrete gravity

dam: A case study. Eng Struct 2017;148:263–76. https://doi.org/10.1016/j.engstruct.2017.06.060.

61
7. Sarkar R, Paul DK, Stempniewski L. Influence of reservoir and foundation on the nonlinear dynamic response of

concrete gravity dams. ISET J Earthq Technol 2007;44:377–8.

8. Küçükarslan S, Coşkun SB. Transient dynamic analysis of dam-reservoir interaction by coupling DRBEM and

FEM. Eng Comput (Swansea, Wales) 2004;21:692–707. https://doi.org/10.1108/02644400410548341.

9. Hai-tao W, Jiayu S, Feng W, Zhiqiang A, Tianyun L. Experimental study on elastic-plastic seismic response

analysis of concrete gravity dam with strain rate effect. Soil Dyn Earthq Eng 2019;116:563–9.

https://doi.org/10.1016/j.soildyn.2018.09.020.

10. Anil K. Chopra and P. Chakrabarti. Earthquake analysis of concrete gravity dams including dam-water-foundation

rock interaction. Earthq Eng Struct Dyn 1995;24:673–86. https://doi.org/10.1002/eqe.4290240505.

11. Nguyen D Van, Kim D, Park C, Choi B. Seismic Soil–Structure Interaction Analysis of Concrete Gravity Dam

Using Perfectly Matched Discrete Layers with Analytical Wavelengths. J Earthq Eng 2019;00:1–22.

https://doi.org/10.1080/13632469.2019.1595222.

12. Mridha S, Maity D. Experimental investigation on nonlinear dynamic response of concrete gravity dam-reservoir

system. Eng Struct 2014;80:289–97. https://doi.org/10.1016/j.engstruct.2014.09.017.

13. Hanchen Tan and Anil K. Chopra. EARTHQUAKE ANALYSIS OF ARCH DAMS INCLUDING DAM-

WATER-FOUNDATION ROCK INTERACTION. Earthq Eng Struct Dyn 1995;24:1453–74.

14. Alembagheri M. Earthquake damage estimation of concrete gravity dams using linear analysis and empirical

failure criteria. Soil Dyn Earthq Eng 2016;90:327–39. https://doi.org/10.1016/j.soildyn.2016.09.005.

62
15. Melenk JM, Babuška I. The partition of unity finite element method: Basic theory and applications. Comput

Methods Appl Mech Eng 1996;139:289–314. https://doi.org/10.1016/S0045-7825(96)01087-0.

16. Moës N, Dolbow J, Belytschko T. A finite element method for crack growth without remeshing. Int J Numer

Methods Eng 1999;46:131–50. https://doi.org/10.1002/(sici)1097- 0207(19990910)46:1<131::aid-nme726>3.3.co.in

63