ABSTRACTS 245

Cyclic Behaviour of Saturated Sands of Kalpakkam Region, Tamil Nadu

V. Jaya, G. R. Dodagoudar and A. Boominathan

(Indian Geotechnical Journal, Vol. 38, No. 2, April 2008, pp. 171-186)

Proper characterization of dynamic behaviour of soil deposits is a critical step in

the solution of earthquake geotechnical engineering problems. In this paper, the
strain dependent dynamic properties such as shear modulus and damping ratio of
the saturated sand are evaluated by laboratory experiments. The undrained strain-
controlled cyclic triaxial and bender element tests are conducted on reconstituted
sand samples at a wide range of shear strain levels. The sand is collected from
the site of a nuclear power plant located in Kalpakkam region, 65 km south of
Chennai city. Predictive relations are developed for estimating normalized shear
modulus and damping ratio of the sand based on hyperbolic model and cyclic
triaxial test results. Variables used in the equations are shear strain amplitude and
effective confining stress. The results presented in the paper can provide
guidance on the dynamic behaviour of sand deposits in the south-east coast of
India. The modulus reduction and damping ratio curves developed can be used
directly as the input parameters to represent the material nonlinearity of the sand
layers for free-field ground response analysis of the sites in the south-east coast
of India.

KEYWORDS: Shear modulus, damping ratio, cyclic triaxial test, bender element,
modulus reduction curve, hyperbolic model

Seismic Response of Concrete Gravity Dams

Considering Foundation Flexibility

B. V. Reddy, Avijit Burman and Damodar Maity

(Indian Geotechnical Journal, Vol. 38, No. 2, April 2008, pp. 187-203)

The dynamic response of dam-foundation coupled system in time domain is

presented in this paper. Both the dam and foundation are modeled by finite
elements and coupled them in a direct manner. The infinite soil domain has been
truncated at a certain distance away from the structure by imposing an artificial
radiation boundary for the finite element analysis. The radiation boundary
conditions in the foundation are incorporated using spring-dash pot models. The
foundation has been modeled considering a significant portion of the foundation
rock with nonlinear properties. The proposed algorithm has been validated by
comparing the results available in the literature. The effect of foundation flexibility
on the response of dam is highlighted. The parametric study shows the
importance of the effect of foundation flexibility on the seismic response of
concrete dams.

KEYWORDS: Concrete gravity dam, dam-foundation interaction, radiation

boundary conditions, seismic response, Wen elasto-plastic model,
Koyna dam
Indian Geotechnical Journal, 38(2), 2008, 187-203

Seismic Response of Concrete Gravity Dams

Considering Foundation Flexibility
B. V. Reddy * , Avijit Burman ** and Damodar Maity ***

Introduction

T
he seismic behavior of concrete dams has been the subject of extensive
research during the past decade because of concern for dam safety
during earthquakes. Concrete dams are distinguished from other
structures because of their large size and their interactions with the reservoir
and foundation. The response of a dam during an earthquake depends on the
characteristics of the ground motion, the surrounding soil and reservoir, and the
dam itself. For the structure founded on soft soil, the motion of the base of the
structure will be different from the free-field motion because of the coupling of
the soil-structure system. This is due to the following reasons: First, the inability
of the foundation to conform to the deformations of the free-field motion which
would cause the motion of the base of the structure to deviate from the free-field
motion. Second, the dynamic response of the supporting structure itself would
induce deformation of the supporting soil. i.e., the soil on which a structure is
constructed may interact dynamically with the structure during earthquakes,
especially when the soil is relatively flexible and the structure is stiff. This kind of
dynamic soil-structure interaction (SSI) can sometimes modify significantly the
stresses and displacements of the whole structural system from the values that
could have been developed if the structure were constructed on a rigid
foundation. If the foundation is rigid, the energy received by the structure from
the base during an earthquake, can be dissipated only through material
damping mechanisms, such as viscous damping. In the case of flexible soils,
some energy is fed back to the base and radiated away giving rise to the so-
called geometric damping or radiation damping.

Review of Previous Works

The general methodologies for soil-structure interaction are direct and
substructure approaches, depending on the modelling method for the soil
around the structure (Dutta and Roy 2002). In the substructure method, the soil-
structure system is divided into two substructures: a structure that may include a
portion of nonlinear soil adjacent to it and the unbounded soil. The unbounded
soil region is usually represented by an impedance matrix, which may be

*
Assistant Engineer, Structures, WS Atkins (India) P. Ltd., Bangalore, India - 560 009
**
Research Scholar, Department of Civil Engineering, IIT Guwahati, India - 781039
*** Associate Professor, Department of Civil Engineering, IIT Kharagpur, India - 721302
Email: [email protected]
188 INDIAN GEOTECHNICAL JOURNAL

attached to the dynamic stiffness matrix of the structure. Though simple, this
method is restricted to simple geometry and linear soil (Wolf, 1985). In the direct
method, the structure and the soil adjacent to it are modeled directly and
analyzed in a single step. A consistent free-field ground motion is being applied
to the boundaries of the discrete model and the response of the combined soil-
structure system is computed. The response of the soil and the structure
obtained was used as input in a second stage analysis to obtain the detailed
structural response. Direct SSI analyses were more commonly performed using
equivalent linear methods to approximate the effects of soil nonlinearity. This
technique limits the extent of soil domain to be considered (Wolf 1985, 1987).
Wolf and Song (1996) also developed consistent infinitesimal finite element cell
technique which proved to be very useful in modeling the unbounded media.
This method is exact in nature in radial direction and is able to simulate the
nonlinear behavior of soil. The detailed description of this method can be found
in this literature.

Much of the reported research on the dynamic analysis of dam

foundations assumes linear behaviour of the foundation media. Adnan and
Wilson (1990) developed an efficient computational technique for the dynamic
analysis of large linear structural systems with local non-linearities. They
developed a rational approach to the earthquake-resistant design of structure-
foundation systems with predetermined non-linearities occurring along the
structure-foundation interface. They had analyzed the dam-foundation model by
considering local nonlinearities of soil. Yazdchi et.al. (1999) presented a
computational method for the transient soil-structure interaction analysis using
the coupled finite element-boundary element method. In the analysis, the half
space soil was represented by the boundary element method (BEM) with linear
material properties. Estorff and Firuziann (2000) investigated the transient
inelastic response of structures coupled with a half space using a general
coupled boundary element and finite element formulation. The inhomogenieties
and elastoplastic behavior with hardening effects were accounted for in the near
field of the surrounding soil. The far field was modeled using boundary
elements. At the interface, the nodal forces resulting from the BE were treated
as additional loads in each iteration. Since non-linearities may occur only in the
FE sub domain, the geometrical linearity was checked along the interfaces by
observing the strains at the interface nodes. There is a common belief that
boundary element method is superior over finite element for the modelling of
infinite or semi-infinite domains. However, in the reported literature, (Yang et al.
1993) the efficiency of boundary element method in time domain analysis is not
ascertained. This is because of the presence of the convolution integral and
singularity of the kernels of the formulation which requires large storage space
and computational time for the evaluation of the effect of past time history and
numerical integration of the kernels. Moreover, use of boundary element
method requires the solution of an unsymmetrical and unbounded matrix. Hence
this method does not possess any significant advantage over the finite element
method. On the other hand, the FEM are well-established procedures
(Zienkiewicz and Taylor 1991, Bathe 1996). These methods routinely
accommodate complex geometries and non-linear phenomena. Kocak and
Mengi (2000) proposed a simple three-dimensional soil-structure interaction
model in which the layered soil medium was divided into thin layers and each
thin layer was represented by a separate parametric model. The parameters of
this model were determined, in terms of the thickness and elastic properties of
the sub layer.
 SEISMIC RESPONSE OF CONCRETE GRAVITY DAMS 189

Two important characteristics that distinguish the dynamic soil-structure

interaction system from other general dynamic structural systems are the
unbounded nature (Yun et al 2000) and the nonlinearity (Halabian et al 2002) of
the soil medium. The coupled response of the dam-foundation system not only
depends on their material properties, but also on the characteristics of the
ground motion (Maity and Reddy 2007). That is why it is important to model the
structure and the soil in such a way so that the actual behavior of the structure
and the soil is represented. While the structural models are well established in
the literature, soil models involve complicated analysis due to their unbounded
nature. Soil being heterogeneous, anisotropic and nonlinear in force-
displacement characteristics is very difficult to model physically (Kim and
Roesset 2004) and at the same time to represent mathematically.

The focus of the present paper is to analyze concrete gravity dam

considering soil-structure interaction for dynamic excitation considering
nonlinear material properties for the foundation. The interaction effect due to
foundation has been adopted by the direct method of soil-structure interaction.
The results show the need of consideration of foundation flexibility while
analyzing massive structures like concrete gravity dam.

Theoretical Formulation
Modeling of Dam
The structural system considered for the present investigation, has been
analyzed using two dimensional plane strain formulations. Since the problem
involved here is a long body, whose geometry and loading do not vary in the
longitudinal direction, can be analyzed by this idealization appropriately. The
equation of motion of the dam under seismic excitation in time-domain can be
expressed as:

Mu&& + Cu& + Ku = − M u&&g + F f (1)

Where, M and K are the mass and stiffness matrices of the dam
respectively. The parameter u, u& and u
&& are displacement, velocity and
acceleration vectors respectively, u&&g is the vector of ground acceleration, and
Ff is the force vector generated from the foundation-dam interaction. The
damping matrix C represents viscous damping in the structure. In this
formulation, a popular spectral damping scheme, called Rayleigh or proportional
damping is adopted. The damping matrix C is formed as a linear combination of
the stiffness and mass matrices as

C = αM + βK (2)

where, α and β are called proportional damping constants. The use

of this damping matrix is equivalent to damping effects that vary with frequency
(Bathe 1996).

Modeling of Foundation Domain

The foundation nonlinearity is incorporated through Wen (1976) elasto-
plastic model. The model is described in some details in the following section.
190 INDIAN GEOTECHNICAL JOURNAL

Wen Plasticity Model

The plasticity model is based on the hysteretic behavior proposed by
Wen (1976) as shown in Figure 1.

Fig. 1 Wen Plasticity Property Type for Uniaxial Deformation

All internal deformations are independent. Therefore, the yielding at one

degree of freedom does not affect the behavior of the other deformations. The
nonlinear force-deformation relationship is given by

f = r k d + (1 − r ) y z (3)

where k is the elastic spring constant, y is the yield force, r is the

specified ratio of post-yield stiffness to elastic stiffness (k), and z is an internal
hysteretic variable.

This variable has a range of z ≤ 1 , with the yield surface represented

by z = 1 . The initial value of z is zero, and it evolves according to the
differential equation
.
z=
k .
y
(
d 1− z
exp
) .
if d z > 0
(4)
k .
= d otherwise
y

where exp is an exponent greater than or equal to unity. Larger values of

this exponent increase the sharpness of yielding as shown in Figure 2. The
practical limit for exp is about 20 (Wen, 1976). The equation for z& is equivalent
to Wen’s model with A = 1 and α = β = 0.5. Nonlinear Model Time-History
Analysis method is used to perform the analysis.
 SEISMIC RESPONSE OF CONCRETE GRAVITY DAMS 191

Fig. 2 Definition of Parameters for the Wen Plasticity Property

Replacing Foundation Material with Nonlinear Wen Link Elements

Foundation soil is replaced with nonlinear Wen elasto-plastic link
elements (1976). The stiffness and damping values for Wen link elements are
calculated by using the following formulas shown in Table 1.

Table 1 Empirical Formulas for Replacing Foundation Soil

with Equivalent Springs

Direction Stiffness (K) Damping Mass

Vertical 4Gr 1 − v 1.79 K ρ r3 1.50 ρ r

3

18.2Gr (1 − v 2 ) (2 − v )2 K ρ r3 0.28 ρ r
3
Horizontal 1.08

K ρ r3 5
Rotation 2.7Gr
3 0.47 0.49 ρ r

K ρ r5 5
Torsion 5.3Gr
3 1.11 0.70 ρ r

Where,
ν - Poisson’s ratio, r - equivalent radius, G-shear modulus, ρ- mass
density of foundation

Absorption Boundary on Foundation

In the conventional approach of foundation-structure interaction analyses,
the effect of far field (unbounded media) is not considered. For wave
propagation analysis, the usual finite boundary of the finite element model will
cause the elastic waves to be reflected and superimpose with the progressing
waves. Besides modeling the foundation stiffness up to infinity, reflections of the
192 INDIAN GEOTECHNICAL JOURNAL

outgoing propagating waves on the artificial boundary at finite distance from the
structure must be avoided also. In this case, some numerical treatment is
needed to introduce artificial boundary so as to simulate the unbounded nature
of the soil foundation and yet maintain a finite computational domain. An artificial
boundary is required to be imposed at the truncated boundary using a non-
reflecting, absorbing, radiating or transmitting condition that can prevent
spurious reflections. Viscous damper components normal and tangent to a far
field boundary are used to simulate the radiation condition. The dashpot
coefficients are determined in terms of the material properties of the semi-infinite
domain as proposed by Wilson (1995).

Solution Scheme for Coupled Dam-Foundation System

In dam-foundation interaction problems, the foundation and the dam do
not vibrate as separate systems under external excitations, rather they act
together in a coupled way. Therefore, these problems have to be dealt in a
coupled way. To develop the fundamental SSI dynamic equilibrium equations,
the two-dimensional foundation structure system is considered as shown in
Figure 3.

Dam (d)

Common Nodes (c)

Foundation (f)

Fig. 3 Foundation Structure Interaction Model

The SSI model is divided into three sets of node points. The common
nodes at the interface of the dam and foundation are identified with “c”; the
nodes within the dam are “d” nodes; and the nodes within the foundation are “f”
nodes. From the direct stiffness approach in structural analysis, the dynamic
force equilibrium of the coupled system may be given in terms of the absolute
displacements, U, by the following sub-matrix equation:

⎡ M dd 0 0 ⎤ ⎧U&&d ⎫ ⎡ K dd K dc 0 ⎤ ⎧U d ⎫ ⎡0⎤
⎢ ⎥ ⎪ && ⎪ ⎢ ⎥ ⎪ ⎪ ⎢ ⎥ (5)
⎢ 0 M cc 0 ⎥ ⎨U c ⎬ + ⎢ K cd K cc K cf ⎥ ⎨U c ⎬ = ⎢0⎥
⎢ 0 M ff ⎥⎦ ⎪U&& ⎪ ⎢ 0 K ff ⎥⎦ ⎪U ⎪ ⎢0⎥
⎣ 0 ⎩ f⎭ ⎣ K fc ⎩ f⎭ ⎣ ⎦

where the mass and stiffness at the contact nodes are the sum of the
contributions from the dam (d) and foundation (f), and are given by:
 SEISMIC RESPONSE OF CONCRETE GRAVITY DAMS 193

(d ) (f )
M cc = M cc + M cc and K cc = K cc (d ) + K cc ( f ) (6)

In terms of absolute motion, there are no external forces acting on the

system. However, the displacements at the boundary of the foundation must be
known. To avoid solving this SSI problem directly, the dynamic response of the
foundation without the structure is calculated. In many cases, this free-field
solution can be obtained from a simple one-dimensional site model. The three-
dimensional free-field solution is designated by the absolute displacements v
&& .
and absolute accelerations v

By a simple change of variables, it is now possible to express the

absolute displacements U& and accelerations U&& in terms of displacements u
relative to the free-field displacements v. Or

⎡U d ⎤ ⎡u d ⎤ ⎡ v d ⎤ ⎡U&&d ⎤ ⎡u&&d ⎤ ⎡ v&&d ⎤

⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ && ⎥ ⎢ ⎥ ⎢ ⎥ (7)
⎢U c ⎥ = ⎢ u c ⎥ + ⎢ v c ⎥ and ⎢U c ⎥ = ⎢ u&&c ⎥ + ⎢ v&&c ⎥
⎢U f ⎥ ⎢u f ⎥ ⎢v f ⎥ ⎢U&& f ⎥ ⎢u&& f ⎥ ⎢v&&f ⎥
⎣ ⎦ ⎣ ⎦ ⎣ ⎦ ⎣ ⎦ ⎣ ⎦ ⎣ ⎦

The above equation can now be written as

⎡ M dd 0 0 ⎤ ⎧ u&&d ⎫ ⎡ K dd K dc 0 ⎤ ⎧u d ⎫
⎢ ⎥ ⎪ ⎪ ⎢ ⎥ ⎪ ⎪ (8)
⎢ 0 M cc 0 ⎥ ⎨ u&&c ⎬ + ⎢ K cd K cc K cf ⎥ ⎨ uc ⎬ = R
⎢ 0 0 M ff ⎥⎦ ⎪u&& ⎪ ⎢ 0 K fc K ff ⎥⎦ ⎪u ⎪
⎣ ⎩ f⎭ ⎣ ⎩ f⎭

where,

⎡ M ss 0 0 ⎤ ⎧v&&d ⎫ ⎡ K dd K dc 0 ⎤ ⎧v d ⎫
⎢ ⎥ ⎪ ⎪ ⎢ ⎥ ⎪ ⎪ (9)
R=−⎢ 0 M cc 0 ⎥ ⎨ v&&c ⎬ − ⎢ K cd K cc K cf ⎥ ⎨ vc ⎬
⎢ 0 0 M ff ⎥⎦ ⎪v&& ⎪ ⎢ 0 K fc K ff ⎥⎦ ⎪v ⎪
⎣ ⎩ f⎭ ⎣ ⎩ f⎭

But this approach is numerically inconvenient. Therefore, in order to

reduce the numerical difficulties, the following change of variables is introduced
as suggested by Wilson (1995).

⎡U d ⎤ ⎡u d ⎤ ⎡ 0 ⎤ ⎡U&&d ⎤ ⎡u&&d ⎤ ⎡ 0 ⎤
⎢ ⎥ ⎢ ⎥ ⎢ ⎥ ⎢ && ⎥ ⎢ ⎥ ⎢ ⎥ (10)
⎢U c ⎥ = ⎢ u c ⎥ + ⎢ v c ⎥ and ⎢U c ⎥ = ⎢ u&&c ⎥ + ⎢ v&&c ⎥
⎢U f ⎥ ⎢u f ⎥ ⎢v f ⎥ ⎢U&& f ⎥ ⎢u&& f ⎥ ⎢v&&f ⎥
⎣ ⎦ ⎣ ⎦ ⎣ ⎦ ⎣ ⎦ ⎣ ⎦ ⎣ ⎦

Substitution of this change of variables into equation (8) yields the

following dynamic equilibrium equation in terms of absolute displacements
{ud } of the dam body. Therefore, equation (8) can be expressed as
194 INDIAN GEOTECHNICAL JOURNAL

⎡ M dd 0 0 ⎤ ⎧u&&d ⎫ ⎡ K dd K dc 0 ⎤ ⎧u d ⎫
⎢ ⎥ ⎪ ⎪ ⎢ ⎥ ⎪ ⎪ (11)
⎢ 0 M cc 0 ⎥ ⎨ u&&c ⎬ + ⎢ K cd K cc K cf ⎥ ⎨ uc ⎬ = R
⎢ 0 0 M ff ⎥⎦ ⎪u&& ⎪ ⎢ 0 K fc K ff ⎥⎦ ⎪u ⎪
⎣ ⎩ f⎭ ⎣ ⎩ f⎭

The right hand side of equation (11) can be calculated as per the
suggestions of Wilson (1995). Thus, the vector R is expressed as follows:

⎡ M dd 0 0⎤ ⎧ 0 ⎫ ⎡ K dd K dc 0⎤ ⎧0⎫
⎪ ⎪ ⎢ ⎪ ⎪
R = − ⎢⎢ 0 0⎥⎥ 0⎥⎥
(d ) (d ) (12)
M cc ⎨v&&c ⎬ − ⎢ K cd K cc ⎨vc ⎬
⎪0⎪ ⎢ 0 ⎪0⎪
⎣⎢ 0 0 0⎦⎥ ⎩ ⎭ ⎣ 0 0⎦⎥ ⎩ ⎭

Validation of Proposed Algorithm

The results of the present model are compared with the results of dam
model analyzed by Yazdchi et. al. (1999) for Koyna ground motion. A dam of
height 15.0 m, crest-width 2.0 m and base width 10.0 m discretized with
isoparametric elements. The foundation size of 250 m × 100 m has been
considered in this study (Figure 4). The dam and the foundation are assumed to
be linear elastic with the following material properties:

Poisson’s ratio = 0.2

modulus of elasticity Ed = 3 × 107 kN/m2

mass density = 2600 kg/m3

The modulus of elasticity of the foundation was varied from 0.5 to 4.0
times the modulus of dam as considered in the literature. The Poisson’s ratio
and the mass density of the foundation were assumed to be the same as those
of the dam. The 1967 Koyna earthquake motion (Figure 5) has been used for
the analysis.

The maximum displacements at dam crest for different foundation

material has been calculated in time domain and presented in Table 2. The
comparison of the results obtained by the present model and Yazdchi et. al.
(1999) confirms the correctness of the proposed algorithm.

Table 2 Validation of Present Model with Yazdchi et. al. 1999 Model

Flexible base for different

Ef /Ed ratio
Dam with
0.5 1.0 2.0 4.0
Maximum displacement at crest (mm) 8.6 4.6 4.5 3.9
(Present model)
Maximum displacement at crest (mm) 7.53 4.41 3.9 3.70
Yazdchi et. al. (1999)
% of deviation 12 4 13 5
 SEISMIC RESPONSE OF CONCRETE GRAVITY DAMS 195

2.0m
(a)

6.0m
(120)

9.0m
Ed

(1)
10.0m

(b)

82.0m
100.00m
100.0m

250.0m

Fig. 4 Discretization of Coupled Dam-foundation System (Yazdchi et. al.1999)

196 INDIAN GEOTECHNICAL JOURNAL

0.6
0.4
0.2
Acceleration (g)

0
-0.2 0 2 4 6 8 10

-0.4
-0.6
-0.8
Time (Sec)

Fig. 5 Koyna - Longitudinal Earthquake

quake Motion (1967)

Analysis of Dam-Foundation Coupled System

Statement of the Problem
The Koyna dam has been chosen in the present study for the extensive
analysis using finite element technique. The tallest non-overflow monolith of
height 103 m, width at the top of the dam 14.8 m and at the base 70.0 m is
considered for the present study. The two-dimensional finite element idealization
for this monolith is shown in Figure 6.

36.5 m

66.5 m z

70 m

70 m

140 m

Fig. 6 FE Discretization of Dam-foundation System with Absorbing Boundary

 SEISMIC RESPONSE OF CONCRETE GRAVITY DAMS 197

The eight nodded, isoparametric finite elements have been used for the
discretization of the structure and foundation domain throughout the present
analysis. The mass concrete in the dam is assumed to be homogeneous,
isotropic, linear elastic solid with the following properties. Modulus of elasticity
7 2 3
(Ed) =3×10 kN/m ; Poisson’s ratio = 0.2 and mass density = 2400 kg/m . The
7
material properties of the foundation is: Modulus of elasticity (Ef) =2.5×10
kN/m2; Poisson’s ratio = 0.33 and mass density = 2400 kg/m3.

The dam is analyzed to get the response subjected to the seismic

accelerations of 1940 El-Centro (N-S component) earthquake. The entire
solution has been done in time domain with application of Direct Method of soil-
structure interaction.

Selection of Size of Foundation Domain

The dam and its foundation are assumed to be in a state of plane-strain
condition. A convergence study is carried out in order to arrive at a suitable
foundation domain of finite dimension by varying horizontal and vertical extent of
the foundation. The time periods and the maximum displacement at the crest
have been calculated and compared for the convergence study. The material
properties of dam and foundation are considered as stated in section 5.1.

In the first step, the foundation length in horizontal direction is taken

arbitrarily equal to 1.5 times the width of dam (b) at its base, and the foundation
length in vertical direction equal to half the width of dam at its base. By keeping
height of foundation in vertical direction as constant, the length in horizontal
direction is increased by 0.5 times the base width of dam (b), till the results are
converged. It is observed that in between the foundation widths of 1.5b and 2b
the variation in displacements of two models is negligible. Therefore, the
horizontal length of the foundation is fixed as 2b. In the second step, length of
foundation in vertical direction is increased by keeping horizontal length as
constant. It is observed that in between the foundation depths of 0.5b and b the
variation in displacements of two models is negligible. So the vertical length of
foundation is fixed as b. Thus, the size of the foundation domain may be
considered for further analysis as 2b× b (Figure 6) where the results are
converging sufficiently. Some of the results of the convergence study are shown
in Table 3. The above conclusions are made with an assumption that no
incoming waves propagating from infinity towards the structure exist at the
boundary i.e., there is no energy associated with the waves may radiate from
the infinity into this truncated area. The displacements at the edges of the
foundation are taken care by the absorbing boundary conditions that are
incorporated.

From the table it is observed that the time periods of dam with different
foundation models are increased compared to that of the dam with fixed base
condition. Similarly displacement at crest of dam is also increased in case of
coupled system when compared to fixed base system. These results clearly
show the importance of foundation flexibility to be taken into consideration
during the analysis.

Variation of Time Periods and Frequencies

Table 4 shows the comparison of time periods as well as corresponding
fundamental frequencies of rigid base and the flexible base system for different
modes. It is observed that the frequencies of the dam are significantly reduced
198 INDIAN GEOTECHNICAL JOURNAL

when the dam-foundation interaction effect is taken into account. The time
periods of the coupled system are elongated with the introduction of foundation
flexibility.

Table 3 Effect of Foundation Size on the Response of Koyna Dam

(Ef/Ed=0.833)

Time Period Max. Direct and Shear stress

Dam With (Sec) Disp. at values at heel (kN/m2)
Mode1 Mode2 Crest (m) (σxx) (σzz) (σxz)
Rigid foundation 0.358 0.132 0.032 315.38 1756.47 207.77
Foundation size
0.393 0.157 0.042 913.87 2596.79 451.05
1.5 b × 0.5 b
Foundation size
0.397 0.161 0.043 1033.78 2697.52 495.99
1.5 b × 1.0 b
Foundation size
0.391 0.157 0.042 819.38 2605.31 435.02
2.0 b × 0.5 b
Foundation size
0.40 0.167 0.045 1093.48 2829.61 520.91
2.0 b × 1.0 b
Foundation size
0.402 0.169 0.047 1186.39 2906.15 593.02
2.5 b × 1.0 b
Foundation size
0.402 0.169 0.045 1027.22 2719.67 511.38
2.0 b × 1.5 b
Foundation size
0.403 0.172 0.049 1060.09 2895.72 543.3
3.0 b × 1.0 b
Foundation size
0.405 0.177 0.049 1162.54 2867.84 592.13
3.0 b × 1.5 b
Foundation size
0.406 0.180 0.049 1161.12 2822.23 582.95
3.0 b × 2.0 b

Variation of Displacements and Stresses

By considering foundation material as elasto-plastic (Wen 1976) the
corresponding variation in displacements are shown in Table 5 for different
nonlinear parameters (i.e., yield force and stiffness ratio). From the tabular data
it is observed that for higher yield forces and stiffness ratios, the displacements
become almost constant (37 mm in this case). Selection of these nonlinear
parameters depends on the foundation material properties and one can get
these values from experimental data. For the present case, these parameters
have been considered on the basis of works done by Park et.al. (1986).

Table 6 shows the variation of crest displacement, direct and shear

stresses for rigid and nonlinear foundation material properties. The material
nonlinearity for foundation is accounted for by incorporating an advanced
plasticity-based soil model called Bouc-Wen elasto-plastic modal (Sections 3.2.1
and 3.2.2). The results show that the maximum crest displacement in nonlinear
case is more than that of rigid case.
 SEISMIC RESPONSE OF CONCRETE GRAVITY DAMS 199

Table 4 Variation of Frequencies and Time Periods

Rigid-Base System Flexible System

Mode T Ω T Ω
(Sec.) (rad/sec.) (Sec.) (rad/sec.)
1 0.358 17.55 0.40 15.688
2 0.132 47.373 0.167 37.665
3 0.087 72.304 0.127 49.396
4 0.065 96.811 0.092 68.092
5 0.043 147.27 0.071 88.215
6 0.041 153.26 0.065 96.573
7 0.031 206.12 0.054 115.79
8 0.027 233.22 0.048 130.97
9 0.025 250.72 0.047 131.79
10 0.024 254.44 0.044 141.37
11 0.022 289.81 0.040 155.43
12 0.021 290.81 0.038 164.70

Table 5 Effect of Yield Force and Stiffness Ratio on Displacements

Yield Displacement at crest (mm)

force for stiffness ratio of
( kN) 0.05 0.1 0.15 0.2
2
2×10 40 39 39 39
3
2×10 38 37 37 37
2×104 37 37 37 37
5
2×10 37 37 37 37

Table 6 Response of Dam with Nonlinear Foundation Material Properties

Direct and shear stress values at

Max. disp. at 2
heel (kN/m )
Dam with
crest (mm)
(σxx) (σzz) (σxz)
Rigid foundation 32 315.38 1756.47 207.77
Nonlinear foundation
40 491.04 2778.55 400.84
material properties

Figure 7 shows the crest displacement of dam considering the foundation

as rigid and flexible. Due to the flexibility effect of the foundation, the
displacement amplitudes are increased compared to that of the rigid base.
200 INDIAN GEOTECHNICAL JOURNAL

0.04
Rigid base
0.03 Flexible base (Nonlinear)
0.02
Displacement (m) 0.01
Horizontal

0
-0.01
-0.02
-0.03
-0.04
-0.05
0 2 4 6 8 10
Time (Sec)

Fig. 7 Variation of crest displacement under El-Centro excitation

Effect of Modulus of Elasticity of Foundation

The modulus of elasticity of the foundation is varied from 0.25 to 5.0
times the modulus of elasticity of dam to study the influence of the foundation
material properties on the response of the dam. The variation of horizontal
displacements at the crest, direct and shear stress values at heel for different
Ef/Ed ratio are shown in Table 7. It is observed from the tabular data that with
the increase of foundation stiffness, the crest displacements, stresses and the
vibration periods of the coupled system decrease significantly. This indicates
that with the increase in stiffness of the unbounded foundation media, the
system behaves like a structure on a rigid foundation. From Table 7, it is
observed that the maximum crest displacement for the case of Ef/Ed = 5.0 is 36
mm, while for cases Ef/Ed = 0.5 and Ef/Ed = 0.25 is 45 mm and 56 mm,
respectively. This indicates that there is about 25% and 56% increase in the
magnitude with decrease of the rigidity of the foundation material. This indicates
that assumption of a rigid base of the dam can underestimate the displacements
in the dam when compared to dam with flexible base. Similarly the maximum
2
direct stress value for the case of Ef/Ed = 5.0 is 2061.35 kN/m , while for cases
of Ef/Ed = 0.5 and Ef/Ed = 0.25 is 2684.92 kN/m and 3596.64 kN/m2 respectively.
2

This indicates that there is about 30% and 74% increase in the magnitude of the
stress value.

Table 7 Response of Dam for Different Ef/Ed Ratio

Time Period Max. Crest Direct and Shear stress

Ef/Ed ratio (Sec) Displ. values at heel (kN/m2)
Mode 1 Mode 2 (mm) (σxx) (σzz) (σxy)
0.25 0.468 0.215 56 993.71 3596.64 802.77
0.5 0.42 0.183 45 648.22 2684.92 539.54
1.0 0.395 0.162 39 464.95 2250.71 349.72
2.0 0.383 0.149 37 393.88 2123.4 254.99
3.0 0.378 0.144 36 372.09 2096.65 231.07
4.0 0.376 0.141 36 369.24 2076.42 219.4
5.0 0.375 0.140 36 366.84 2061.35 214.6
 SEISMIC RESPONSE OF CONCRETE GRAVITY DAMS 201

In general, it may be concluded that by including the effects of foundation

flexibility in the analysis, the displacement and the stress distribution in the body
of the dam will be significantly different. Therefore, the foundation stiffness
should be included in the response analysis of dams to increase the accuracy in
results. Figure 8 shows the variation of crest displacement with time for Ef/Ed
ratio of 0.25 and 5.0.

0.08
0.06 Ef/Ed 0.25 5

0.04
Displacement (m)

0.02
0
-0.02
-0.04
-0.06
-0.08
0 2 4 6 8 10
Time (Sec)

Fig. 8 Variation of Displacement for Different Ef/Ed Ratio (Nonlinear Analysis)

Base Shear and Base Moment

Table 8 shows the comparison of base shear (kN) and base moment (kN-
m) for first three fundamental modes with different Ef/Ed ratio along with the rigid
base. It is observed that the magnitudes of base shear and base moments
obtained from the rigid base model are much higher than that of the elastic
base. Consideration of flexibility of the supporting foundation reduces the
magnitude of the base shear and base moment of the dam.

It is observed that the magnitudes of base shear and base moments get
reduced as Ef/Ed ratio reduces. Incorporation of dam-foundation interaction
effects has the direct result of reducing the base shear applied to the structure,
and consequently the lateral forces and overturning moments.

Table 8 Comparison of Base Shear (kN) and Moment (kN-m) for Different
Ef/Ed Ratios

Ef/Ed = 0.25 Ef/Ed =1.0 Ef/Ed =5.0 Rigid base

Mode
base base base base base base base base
shear moment shear moment shear moment shear moment

1 11093 16641 12849 19274 12960 19440 13922 20883

2 47140 70710 86495 129743 99683 49524 99902 149853

3 68593 102890 140442 210664 191227 286839 179771 269656
202 INDIAN GEOTECHNICAL JOURNAL

Conclusions
The effect of foundation flexibility on the seismic response of concrete
gravity dam is investigated using the method described above. The proposed
algorithm has been validated with the results available in the literature. The
responses of the soil-structure system with absorbing boundary indicate that the
most of the incident energy is absorbed at the truncation boundary. By the use
of absorption boundary for the finite element analysis of unbounded foundation
domain, a base size of 2b × b will produce sufficiently accurate results compare
to the large foundation size. The parametric study shows that the consideration
of foundation flexibility may alter the response of the dam significantly. The
magnitude of base shear and base moment reduces with the increase of
foundation flexibility. The results show that the fundamental time period of the
coupled system is being elongated if the foundation becomes more flexible. The
magnitude of the displacements and stresses on dam under seismic excitation
becomes less if the foundation becomes rock type in nature. Consideration of
soil-structure interaction effect is necessary if the dam is founded on soft soil.

References
Adnan, I. and Wilson, E. L. (1990): ‘A Methodology for Dynamic Analysis of
Linear Structure-Foundation Systems with Local Nonlinearities’, Earthquake
Engineering and Structural Dynamics, Vol.19, pp. 1197-1208.

Bathe, K. J. (1996): ‘Finite Element Procedures in Engineering Analysis’,

Englewood Cliffs, NJ: Prentice-Hall.

Dutta, S. and Roy, R. (2002): ‘A Critical Review on Idealization and Modelling

for Interaction among Soil-Foundation-Structure System’, Computers and
Structures, Vol. 80, pp.1579-1594.

Halabian, Amir M. and El Naggar, M. Hesham (2002): ‘Effect of Non-Linear Soil-

Structure Interaction on Seismic Response of Tall Slender Structures’, Soil
Dynamics and Earthquake Engineering, Vol. 22, pp. 639-658.

Kim Yong-Seok and Roesset Jose M. (2004): ‘Effect of Nonlinear Soil Behavior
on Inelastic Seismic Response of a Structure’, International Journal of
Geomechanics, Vol. 4, pp. 104-114.

Kocak, S. and Mengi, Y. (2000): ‘A Simple Soil-Structure Interaction Model’,

Applied Mathematical Modeling Vol. 24, pp. 607-635.

Maity, D. and Reddy, B. V. (2007): ‘Influence of Nonlinear Foundation Flexibility

on the Seismic Response of a Concrete Gravity Dam’, International Journal of
Dam Engineering. Vol. 18, No. 2, pp. 75‐100.

Park, Y. J., Wen, Y. K., and Ang, A. H-S. (1986): ‘Random Vibration of
Hysteretic Systems under Bi-Directional Ground Motions’, Earthquake
Engineering and Structural Dynamics, Vol. 14, pp. 543-557.

Von, E. O. and Firuziann, M. (2000): ‘Coupled BEM/FEM Approach for

Nonlinear Soil Structure Interaction’ Engineering Analysis with Boundary
Elements, Vol. 24, pp. 715-725.
 SEISMIC RESPONSE OF CONCRETE GRAVITY DAMS 203

Wen, Y. K. (1976): ‘Method of Random Vibration of Hysteretic Systems’, Journal

of Engineering Mechanics Division, ASCE, Vol. 102, pp. 249-263.

Wilson, E. L. (1995): ‘Three -Dimensional Static and Dynamic Analysis of

Structures- a Physical Approach with Emphasis on Earthquake Engineering’,
Computers and Structures, Inc., University Avenue Berkeley, California, USA.

Wolf, J. P. (1985): ‘Dynamic Soil-Structure Interaction’, Prentice Hall:

Englewood Cliffs, NJ.

Wolf, J. P. (1987): ‘Soil-Structure Interaction Analysis in Time Domain’, Prentice

Hall: Englewood Cliffs, NJ.

Wolf, J. P. and Song, C. (1996): ‘Finite Element Modelling of Unbounded

Media’, Prentice Hall: Englewood Cliffs, NJ.

Yang, R., Tsai, C. S. and Lee, G. C. (1993). ‘Explicit Time-Domain Transmitting

Boundary for Dam-Reservoir Interaction Analysis’, Int. J. Numerical Methods. In
Engg, Vol. 36, pp. 1789-1804.

Yazdchi, M., Khalili, N., and Valliappan, S. (1999): ‘Dynamic Soil–Structure

Interaction Analysis via Coupled Finite-Element–Boundary-Element Method’,
Soil Dynamics and Earthquake Engineering, Vol. 18, pp. 499–517.

Yun, Chung-Bang., Kim, Doo-Kie., and Kim, Jae-Min (2000). ‘Analytical

frequency-dependent infinite elements for soil-structure interaction analysis in
two-dimensional medium’, Engineering Structures, Vol. 22, pp. 258–271.

Zienkiewicz, O.C. and Taylor, R.L. (1991). ‘The Finite Element Method’, Vol.2,
Solid and Fluid Mechanics, Dynamics and Nonlinearity, McGraw-Hill Book Co.
U.K.