RESEARCH ARTICLES

A comparative assessment of the seismic

response of an earthen dam using analytical
simulation and empirical methods
Srijit Bandyopadhyay1, Raj Banerjee1, Aniruddha Sengupta2,*, Y. M. Parulekar1
and G. R. Reddy3
1
Bhabha Atomic Research Centre, Trombay, Mumbai 400 085, India
2
Civil Engineering Department, Indian Institute of Technology Kharagpur, Kharagpur 721 302, India
3
Bhabha Atomic Research Centre, Trombay, Mumbai 400 085, India

that region from the safety and serviceability point of

This article presents the permanent deformation of an
earthen dam located in the vicinity of a safety related view.
structure for Mw = 6.5 design basis earthquake. A non- The determination of permanent deformation of a dam
linear 2D dynamic analysis using a real earthquake during a probable ground motion is one of the seismic
motion compatible with the design spectrum was per- safety criteria. However, estimation of permanent defor-
formed to check the earthquake-induced deformations mation of a dam due to seismic loadings is a complex
of the dam. Deformations of the dam were also esti- task as it involves many factors and lacks reliable field
mated by semi-empirical and empirical methods such data during an earthquake. A ground motion is unique to
as Seed and Makdisi’s method, Newmark’s double in- a site for a particular earthquake. When real earthquake
tegration method, Jansen’s method and Swaisgood’s data is not available, the site-specific spectral compatible
method. Results from different methods are compared time history is used. The response of a dam is affected by
to obtain a range for the value of permanent deforma-
the type of construction, water level in the reservoir,
tions of the dam. It is observed that the lateral defor-
mation obtained by Seed and Makdisi’s method is the height of the dam, etc.2. Nowadays sophisticated numeri-
highest while Jansen’s method predicts the highest cal tools like finite element method, finite difference
crest settlement. The crest settlement of the dam is method are available for assessing the dynamic response
found to vary between 11.8 mm and 17.8 mm, which is of a dam. But a realistic deformation value of a dam
within the safety limits according to IITK-GSDMA obtained from these advanced numerical tools may be
guidelines. expected only when the material parameters and numeri-
cal models accurately represent the actual conditions ex-
Keywords: Earthen dam, dynamic analysis, deforma- isting at a site. Despite several limitations, an advanced
tions, non-linear finite element analysis. numerical analysis is recommended for performing the
earthquake safety evaluation of a dam. Over the decades,
T HE main focus of this study is to predict the dynamic starting with the fifth Rankine lecture by Newmark 3, sev-
performance of Apsara dam using 2D plane strain finite eral simplified empirical, semi-empirical methods3–6 have
element (FE) analysis. Apsara dam is 18.5 m high earthen been proposed to evaluate the permanent deformation of
embankment located in Trombay, a suburb of Mumbai in a dam during an earthquake. Due to limited availability of
India. The area, being within 350 km from the Gulf of such studies, all the empirical and semi-empirical rela-
Cambay and Rann of Kutch, is known to be seismically tions are developed based on the statistical analysis of
active and has experienced several devastating earth- these limited data.
quakes like Koyna (1967) and Bhuj (2001) in the near Singh et al.7 conducted finite element analyses of sev-
past. Also, Trombay region is located in the seismic Zone eral earthen dams affected by the 2001 Bhuj (Mw = 7.6)
III (ref. 1) of India, a zone with moderate seismic hazard. earthquake which were within 50 km of the epicentre.
Apsara dam was constructed in 1950. The seismic These dams underwent free-field ground motions with
criteria used earlier for the design of nuclear research peak ground accelerations (PGAs) between 0.28 and
facilities were less rigorous than the present ones. For 0.52 g. Basudhar et al.8 conducted 2D finite element
this reason, a new study was undertaken to check whether (FEM) seismic analyses of Tehri dam located in a seismi-
the dam satisfies the present seismic design criteria for cally active region for studying displacement, velocity
and acceleration time histories near the crest and bottom
of the dam using the acceleration time history of 2001
*For correspondence. (e-mail: [email protected]) Bhuj earthquake. Their results revealed that vertical

902 CURRENT SCIENCE, VOL. 113, NO. 5, 10 SEPTEMBER 2017

 RESEARCH ARTICLES
displacement at any location of an embankment dam was For cohesive material, the shear modulus is obtained
negligibly small compared to horizontal displacement. from the relationship given by Hardin and Drenvich 10
The velocity time history shows a maximum velocity in
forward direction at the crest of the dam, while in the re- 3230(2.97  e) 2
verse direction, the same is experienced by shell and core G (OCR) k ( 0 ) 0.5 , (2)
(1  e)
of the dam supplemented by a noticeable phase differ-
ence. The accelerations obtained near the crest of a dam
will be more than those near its bottom portions. where G is in kPa, e the void ratio, 0 the mean effective
The present article provides an estimate of the perma- confining stress (in kPa), OCR the over consolidation
nent deformations of Apsara embankment dam subjected ratio, and k is a parameter which is a function of the plas-
to an earthquake of magnitude, Mw = 6.5 and a PGA of ticity index of the soil (=0.18 in this study).
0.156 g by a plane strain finite element method and com- The modulus of elasticity (E) of the soil is obtained
pares them with those obtained by some of the well- from the equation
established semi-empirical and empirical methods.
E
G , (3)
2(1   )
Description of the dam

Apsara dam is an earthen zoned dam. It has a maximum where  is the Poisson’s ratio of the soil.
height of 18.5 m at the deepest location. The crest has a
width of 2 m and spans 500 m in length. The foundation Response spectrum and ground motion
of the dam lies on top of a weathered rock of thickness
10 m overlying a solid rock base. The dam consists of a The IS : 1893 (ref. 1) code gives the seismic zoning of
central impervious core consisting of compacted clay India as per the comprehensive intensity scale of earth-
with a width of 18 m at the base and tapers to a width of quakes (MSK64). Apsara dam is located in Zone III, a
1.5 m just one metre below the crest. A 900 mm thick fil- zone with moderate seismic hazard. The intensity associ-
ter/drain layer is located on the downstream side of the ated with Zone III1 is VII which corresponds to 0.156 g
core. The central clay zone is protected by rockfills on PGA and the corresponding local magnitude (ML) is 5.67
both downstream and upstream sides. The downstream based on the relationship proposed by Gutenberg and
slope is 1(V) : 2(H) and the upstream slope is 1(V) : Richter11. The corresponding moment magnitude (Mw) is
2.5(H). The upstream slope is provided with stone revet- around 5.9 proposed by Idriss12 . For the present study, an
ment to protect against surface erosion. A cut-off trench earthquake with Mw equal to 6.5 and the zero-period ac-
is also provided underneath the clay core along the length celeration (PGA) to 0.156 g is selected. This value repre-
of the dam for seepage control. A free draining rockfill sents the maximum credible earthquake (MCE) for the
toe is provided on the downstream for additional stability. region. The site-specific data or the uniform hazard re-
A freeboard of 2 m is provided at the crest of the dam. sponse spectrum (UHRS) for this region is obtained from
The storage capacity of the reservoir is around 120 mil- the probabilistic seismic hazard analysis (PSHA) study13
lion litres. At one end of the dam, a surplus weir of and shown in Figure 2 b. The above spectrum is consid-
12.2 m in length and 2.2 m in width is provided to let the ered as a representative ground motion for the finite
surplus water overflow onto the downstream side. Figure
1 a shows a typical section of the dam.

Material properties

The shear strengths of the dam materials are obtained

from in situ tests and laboratory triaxial tests and summa-
rized in Table 1. In the present case, the following em-
pirical relationship given by Seed and Idriss9 to find the
shear modulus (G) of a cohesionless soil is utilized

G  1000 K ( 00.5 ), (1)

where G is in kPa, K is an empirical factor which is a

function of the relative density of the soil (= 13.0 in this
study) and 0 is the mean effective confining stress (in Figure 1. a, Typical cross-section and finite element discretization of
Apsara dam; b, Locations of the upstream and downstream failure sur-
kPa). faces in a typical cross-section of Apsara dam.

CURRENT SCIENCE, VOL. 113, NO. 5, 10 SEPTEMBER 2017 903

RESEARCH ARTICLES

Figure 2. a, Selected ground motion used in 2D finite element analysis; b, Comparison of Trombay response spectra and
that of Saguenay earthquake (Canada), 1988; c, Response spectra of Saguenay Earthquake (Canada), 1988 at 5% and 10%
of critical damping.

Table 1. Material strength parameters respectively. The ratio of the maximum velocity and
Zones of dam acceleration (v max/amax) is 0.054, which is less than 0.1,
confirms that it is a rock motion3. Figure 2 b compares
Clay Shell Coarse the site-specific response spectra (at 5% damping) and
Soil properties core (U/S and D/S) filter
spectra of the selected ground motion. A good match is
Bulk unit weight (kN/m3 ) 19.2 21 20 observed between the response spectra of selected motion
Saturated unit weight (kN/m3 ) 19.5 22 20 and the site-specific spectra which confirms the adequacy
Angle of internal friction 15 37 42 of the selected ground motion.
Cohesion (kPa) 30 0 0

Methodology

element analysis of Apsara dam. A number of real earth- Analysis using equivalent linear approach
quakes are reviewed and the Saguenay Earthquake of
Canada (25 November 1988), Site 7 (Mw = 6.5, NW com- The ‘shear beam’ concept serves as a basis for many of
ponent) (see Figure 2 a) was chosen based on its PGA the recently developed models. To obtain the natural fre-
and frequency content of the motion. The selected motion quencies and mode shapes of the dam, the response spec-
is a bed rock motion with a focal depth of 29 km, an epi- trum is analysed in which the dam is modelled as a
centre distance of 45.14 km with a predominant fre- triangular shear beam with variable stiffness. The re-
quency of 4 Hz. The ground motion was chosen such that sponse spectra at 5% and 10% of critical damping used in
its frequency content matches with the predominant time the shear beam analysis for Mw = 6.5, PGA = 0.156 g
period of the dam, which typically lies between 0.1 sec earthquake are shown in Figure 2 c. The dependence of
and 1 sec (ref. 14). The selected earthquake motion has a soil modulus on the confining pressure has been estab-
predominant time period, PGA, velocity and displace- lished by Ghaboussi and Wilson 15. Using the well-
ment of 0.34 sec, 153.036 cm/s2, 8.402 cm/s and 56 cm established distributions of static confining stresses in an
904 CURRENT SCIENCE, VOL. 113, NO. 5, 10 SEPTEMBER 2017
 RESEARCH ARTICLES
idealized dam section, along with the assumption that the Nonlinear dynamic analysis
element soil modulus, G(z), increases in proportion to
 00.5 , many researchers recommended the following The finite element discretization of Apsara dam is shown
dimensionless expression to be used to estimate G(z) with in Figure 1 a. The embankment dam and its foundation
the depth with a reasonable accuracy9,10,16 are modelled by 4-noded and 3-noded 2D plane strain
finite elements. The finite element mesh consists of 639
m nodes and 597 elements. The material properties of each
 z 
G  Gb   , (4) zone are specified in Table 1. For static analysis, a fixed
H boundary condition is applied at the bottom, which
restricts the movements in both horizontal and vertical
where m is the material non-homogeneity parameter, Gb directions, whereas the two sides are restricted to move in
the shear modulus at the base of the dam, z the depth the horizontal direction. In dynamic analysis, an absorb-
from crest of the dam and H is the total height of the ing boundary is applied at the two side boundaries to
dam. minimize the reflection of waves. Finite element analysis
Based on the above expressions, an inhomogeneous is performed in three stages – in the 1st stage, the
shear beam analysis is conducted for clay core of the dam entire problem is solved in the presence of gravity (or
by using the principle of an equivalent linear approach in self-weight) only. In the present study, the construction
which the strain-dependent dynamic soil properties are sequence of the dam is ignored. In the next stage, a
estimated in an iterative process for a certain value of steady state analysis is performed where the reservoir is
material non-homogenity parameter, m (= 0.5 in this impounded. The seismic response of the dam at normal
study). The curves proposed earlier9,17 are used to repre- operating condition is then analysed using acceleration
sent the strain-dependent dynamic stiffness and damping time history of the selected Saguenay earthquake (1988)
of the clayey material as shown in Figure 3. The results of PGA 0.156 g. The selected ground motion is shown in
of the shear beam analysis in terms of maximum crest Figure 2 a. The motion is applied at the bottom of the
acceleration, fundamental period of the dam, damping foundation for 15 sec. Deformation of the dam and
ratio and the corresponding shear strain are shown in change in the pore water pressures due to earthquake
Table 2. motions are obtained from the analysis.
Figure 4 shows the deformed geometry of the dam at
the end of the dynamic analysis. The predominant defor-
Table 2. Results of the shear beam analysis
mations of the dam are found to be in its downstream
Magnitude of earthquake, Mw = 6.5 shell. Vertical settlement is obtained near the dam’s crest
Maximum crest acceleration, a max 0.50 g and the lateral deformation obtained at the upstream face
Average equivalent strain 0.033%
Strain compatible damping 6.5%
as shown in Figure 5 a and b. The pore water pressure
Fundamental time period 0.218 sec variation near the bottom and mid-point of the dams’s
2nd mode time period 0.105 sec core is shown in Figure 6. The analysis predicts no lique-
3rd mode time period 0.069 sec faction for the dam and its foundation for this earthquake
event. The response of the input acceleration is measured
at the top of the dam. Figure 7 shows the amplification of
motion at the dams’s crest and accordingly, the amplifi-
cation of PGA is around 2.21 times at the crest. However,
different amplification value of 3.2 times the base motion
is obtained in the shear beam analysis of the dam. A pos-
sible reason for this deviation is due to the onset of non-
linear (damping) behaviour of the dam materials during
earthquake motions, which prevents the development of
high acceleration predicted by shear beam analysis16.
Besides the shear beam and FEM analysis, several
empirical and semi-empirical methods exist to estimate
the permanent displacement in an earthen embankment
during an earthquake.

Estimation of lateral displacements of a dam by

Seed and Makdisi

Figure 3. Dependency of soil stiffness and damping of clay core on This is a simple yet rational approach to the design of an
strain level. embankment dam under earthquake loadings. The method
CURRENT SCIENCE, VOL. 113, NO. 5, 10 SEPTEMBER 2017 905
RESEARCH ARTICLES

Figure 4. Deformed geometry of the dam at the end of the earthquake (deformation magnified 1000 times).

of safety is determined according to Bishop’s simplified

method18. The results of the factor of safety and the cor-
responding values of yield acceleration, Ky, are summa-
rized in Table 3.
Considering the approximate nature of the method, the
average relationship4 adopted for determining the maxi-
mum average acceleration for a potential sliding mass
based on the maximum crest acceleration, is considered
accurate enough for practical purposes. For design
purposes, a conservative estimate of the accelerations is
desirable. Hence, the upper limit curve is used in the pre-
sent study. The maximum average acceleration along
with the Ky/Kmax values for all cases is illustrated in Table
4 and the lateral permanent displacements U, for each of
the failure surface is calculated from the figure given
in Makdisi and Seed4. The corresponding displacements
thus obtained for Mw = 6.5 earthquake are shown in
Table 5.
This approach has two drawbacks; the peak accelera-
tion at the crest (k max) is highly variable, and the fre-
quency content of the motion is not captured in shear
beam analysis. In addition, the range of the upper and
lower bounds on the Makdisi and Seed 4 plot of k max/PGA
versus y/h and k y/k max versus displacement (U) may not be
true, because of the limited number of analysis of the
Figure 5. a, Lateral deformation at the upstream berm of the dam; b, earth structure with variable ground motion. These curves
vertical settlement at the crest of the dam. need refinement and upgradation as a larger number of
analytical results for embankments are obtained.
is based on the concept that whenever the rigid body
acceleration, Kmax for a sliding area exceeds the yield Estimation of displacements of a dam by
acceleration, Ky a permanent deformation takes place. Newmark’s method
The yield acceleration is determined from a series of
pseudostatic analyses. After obtaining the yield accelera- Newmark3 proposed a method of computing a seismically
tion, the permanent deformation U is calculated, which is induced deformation based on the sliding wedge concept.
a function of an earthquake magnitude and fundamental When the inertia of the mass exceeds the frictional
period of a dam for a given ratio of Ky/Kmax. resistance along the sliding surface, the sliding failure of
The yield acceleration Ky is defined as an acceleration the mass occurs and the corresponding relative displace-
at which a potential sliding surface would develop a fac- ments are calculated by double integration of the differ-
tor of safety of unity. Yield accelerations are obtained for ence between mass acceleration and the yield acceleration
three failure masses on the upstream and downstream K y.
faces of the dam whose locations are shown in Figure 1 b. A concept of ‘elastic response method’ has been pro-
The water level on the upstream face is assumed to be at posed by Seed and Martin 19 to determine the amplified
E l.16 m (maximum normal operating pool) and the factor absolute acceleration time history at any point in the dam.
906 CURRENT SCIENCE, VOL. 113, NO. 5, 10 SEPTEMBER 2017
 RESEARCH ARTICLES
Table 3. Yield accelerations

Zones of dam Location of surface FS in static case Yield acceleration (K y) Kmax (g) K y/Kmax

Upstream slope 1/3 height 2.207 0.28 0.455 0.615

2/3 height 2.101 0.20 0.310 0.645
Full height 2.051 0.19 0.250 0.760

Downstream Slope 1/3 height 1.444 0.13 0.455 0.28

2/3 height 1.628 0.215 0.310 0.69
Full height 1.733 0.25 0.250 1.00

The total or absolute acceleration, ug acting on the dam

at any time t, is given by

ua ( y , t )  u( y, t )  ug (t ). (6)

where y is the depth from the crest of the dam, ug the base
acceleration and ua is the total acceleration.
This may be readily determined from the values of
u(y, t) and the known values of ground acceleration.
Hence, the absolute acceleration y, at any level, in the
dam at t, is expressed by


1  y
ua ( y , t )   2n J0  n 
n 1  n J1 (  n )  H 

t
  ug e nn (t  ) sin(n (t   ))d , (7)
Figure 6. Variation of pore water pressure near the bottom and mid-
point of the core of the dam. 0

where n is the number of significant mode shapes to be

considered in the analysis. In the present analysis, three
mode shapes are considered. The above equation is
solved for 1/3rd, 2/3rd and full heights of the dam.
The actual irregular acceleration time history is con-
verted into equivalent time history of several cycles of
constant amplitude as proposed by Seed and Idriss20. The
present time history is modelled by six cycles of identical
full sinusoidal waves of average acceleration amplitude,
aavg (=2/3 of PGA) of 0.30, 0.21 and 0.104 g for
the upper 1/3rd, 2/3rd and full height of the dam respec-
tively.
The acceleration time history is represented by an
equivalent sine wave given by

 t
Figure 7. Acceleration time history at the crest of the dam. K (t )  aavg sin  2  , (8)
 T
According to this concept, damped response to the ran-
dom ground motion, ug, which might be induced by an where T is the fundamental period of the dam
earthquake, can be shown by (= 0.218 sec in the present case).
Based on the procedure by Sengupta14, the relative
y displacements of the dam at 1/3rd height are found to be

n  2J0  n  t 2 mm for the upstream side and 76 mm for the down-
 H  u e  Dnn (t  ) sin( (t   )) d
u ( y, t )   stream side. No significant deformations are observed at
) 0
g dn
n 1  dn  n J 0 (  n 2/3rd and full height of the dam from the above metho-
(5) dology.
CURRENT SCIENCE, VOL. 113, NO. 5, 10 SEPTEMBER 2017 907
RESEARCH ARTICLES
Table 4. Permanent displacements by Makdisi and Seed’s procedure

Upstream Downstream

Location of sliding surface Depth y/H K y/Kmax U/KmaxgT0 U (mm) K y/Kmax U/KmaxgT0 U (mm)

1/3 height 0.33 0.615 0.0085 8.27 0.28 0.08 77.84

2/3 height 0.66 0.645 0.006 3.97 0.69 0.0058 3.845
Full height 1.00 0.760 0.004 2.14 1.00 – –

Table 5. Comparison of deformations (lateral and vertical)

Deformations (mm) due to the Mw = 6.5 earthquake

Lateral displacement (mm)

Methods used Upstream Downstream

Seed and Makdisi’s 8.27 77.84

Double integration 2.0 76.0
2D finite element method 13.69 9.05
Crest settlement (mm)
Jansen’s 17.80
Swaisgood’s 11.80
2D finite element method 12.32

One of the basic assumptions to calculate seismic dis- Estimation of crest settlement by Jansen’s
placement according to this method is that the failure method
mass is considered to be totally rigid. This means that the
dynamic response of the slope and embankment, for Jansen’s5 method for estimating the crest settlement of a
which the seismic stability is estimated, is not considered dam is based on the data collected from one hypothetical
calculating seismic displacement. The soil slopes and and four existing dams. A relationship is developed for
embankments are flexible systems characterized by a computing crest settlement based on earthquake magni-
relatively large fundamental period. In order to consider tude, maximum crest acceleration, Km and yield accelera-
this for the estimation of seismic displacement of a slope, tion, Ky, as follows
a decoupled method is proposed by Makdisi and Seed4
 8
which consists of two steps. The first step focusses on the  M  ( K m  K y ) 
U (cm)   48.26  w  . (9)
estimation of dynamic response of the failing soil mass in  10  Ky 
terms of an equivalent acceleration time history after cal-  
culation of response of the earth structure. The time his- The value of crest amplification of Apsara dam is
tory of an equivalent horizontal acceleration is defined as obtained from Jansen’s method5. The corresponding crest
the integral of horizontal stresses respectively, along the settlement, U, at different locations is shown in Table 6.
slip surface divided by the weight of the sliding mass. In From the tabulated values, it may be concluded that the
the second step, the permanent seismic displacement is maximum range of crest settlement of Apsara dam lies
obtained from double integration of the relative accelera- between 12.6 and 17.8 mm respectively, for a Mw = 6.5
tion. Difference between the applied and critical accelera- earthquake.
tion is defined as the relative acceleration. A decoupled The above equation implies no deformation if maxi-
sliding block analysis overestimates the system’s response, mum crest acceleration Km  Ky and it also calculates
because the forces at the sliding interface are not taken crest settlement without considering the height of the
into account, when considering the dynamic response. dam.
A decoupled sliding block analysis does not correctly
model the forces at the sliding interface, which leads to
an over-prediction of the system response. Estimation of crest settlement by Swaisgood’s
Crest settlement is considered for an earthquake related method
damage and deformation, because it is often the most
quantified measurement of damage in such cases. This Swaisgood6 developed a crest settlement relationship
parameter also has a relationship with the severity of de- based on 69 case histories which takes into account the
formations and cracking in a dam during an earthquake. level of severity of deformation and cracking in the dam.
908 CURRENT SCIENCE, VOL. 113, NO. 5, 10 SEPTEMBER 2017
 RESEARCH ARTICLES
Table 6. Estimation of crest settlement by Jansen’s method

Upstream slope Downstream slope

Location of
Value of Km (g) Mw sliding surface K y (g) U (cms) K y (g) U (cms)

0.546 6.5 1/3 height 0.28 0.77 0.13 1.78

2/3 height 0.20 1.19 0.215 1.09
Full height 0.19 1.26 0.25 0.92

Here, the crest settlement depends on the earthquake that lateral displacements and crest settlements of the
magnitude and PGA at the site as given below dam during a probable earthquake are found to be within
safety limits according to prescribed guidelines.
(%)  SEF.K typ .K dh .Kat , (10)
Conclusion
where (%) is the crest settlement (in percentage) of the
dam and the alluvial thickness, SEF the site energy factor The variations in predicted displacements of Apsara dam
related to the earthquake magnitude (M) and PGA, and under an earthquake motion indicate that there is still
expressed as SEF = e(0.7168 M  6.405PGA 9.098) , Ktyp a con- some room for understanding the failure mechanism of
stant related to the type of dam construction (=1.363 for such a dam during a dynamic event. The finite element
earthfill dam), Kdh a constant related to the dam height study shows that maximum deformation occurs on the
and expressed as Kdh  9.134  DH 0.437 where DH is the upper portion of the dam, while the foundation level has
height of the dam in feet. Kat a constant related to the negligible deformations. Among analytical methods,
alluvial thickness and expressed as Kat  0.851.e(0.00368. At ) Makdisi and Seed method overestimates the deformations
where At is the alluvial thickness in feet. obtained, than those predicted by the double integration
The estimated crest settlement of the dam for the seis- method and finite element analysis. From this study (refer
mic energy and site-specific factors recommended by to Table 5), it may be concluded that the evaluation of
Swaisgood is found out to be 11.80 mm. crest settlement by Swaisgood’s method provide results
In Swaisgood method, the accuracy for predicting crest closer to FEM results than Jansen’s method.
settlement depends on the accuracy with which the site-
specific PGA is estimated from probabilistic studies. In
addition, crest deformation of the dam mainly spreads 1. USBR, Seismic design and analysis of embankment dams, Design
Standards. Chapter 13, In United States Department of the Inte-
and settles during an earthquake event. There is no dis-
rior, Bureau of Reclamation, Denver, Colorado, DS-13(13)-7,
tinct failure along a particular shear plane. 1989.
2. IS:1893, Indian Standard criteria for earthquake resistant design of
structures, Part 1 – general provisions and buildings, Bureau of
Results and discussion Indian Standards, Fifth Revision, New Delhi, 2001.
3. Newmark, N. M., Effects of earthquakes on dams and embank-
ments. Geotechnique, 1965, 15(2), 139–159.
A comparison of displacements (both horizontal and ver- 4. Makdisi, F. I. and Seed, H. B., Simplified procedure for estimating
tical) of Apsara dam predicted by various methods is dam and embankment earthquake induced deformations. J. Geo-
shown in Table 5. The results indicate a wide variation in tech. Eng. Div., 1978, 104(7), 849–867.
values obtained from different methods. This may be due 5. Jansen, R. B., Estimation of embankment dam settlement caused
by earthquake. Int. Water Power Dam Constr., 1990, 42(12), 35–
to the fact that these empirical and semi-empirical meth-
40.
ods have very little in common. For a Mw = 6.5 earth- 6. Swaisgood, J. R., Estimating deformation of embankment dams
quake, the lateral displacements of the upstream and caused by earthquakes. In ASDSO Western Regional Conference,
downstream shells are in the range of 2–13.69 mm and Montana, USA, 1995, pp. 1–7.
9.05–77.84 mm respectively. The crest settlement ob- 7. Singh, R., Roy, D. and Jain, S. K., Analysis of earth dams affected
by the 2001 Bhuj Earthquake. Eng. Geol., 2005, 80, 282–291.
tained from empirical methods varies between 11.8 and
8. Basudhar, P. K., Rao, N. S. V. K., Bhookya, M. and Dey, A., 2D
17.8 mm. From the 2D plane strain finite element analy- FEM analysis of earth and rockfill dam under seismic condition.
sis, lateral displacement is found to be around 12.32 mm In 5th International Conference on Recent Advances in Geotech-
which is in reasonable agreement with values obtained nical Earthquake Engineering and Soil Dynamics, 2010.
from analytical procedures. As per IITK-GSDMA21 9. Seed, H. B. and Idriss, I. M., Soil modulii and damping factors for
dynamic response analysis. Report EERC, 70–10, University of
guidelines, the acceptable deformation of Apsara dam
California, Berkeley, 1970.
along any failure plane should not exceed 1 m. The range 10. Hardin, B. O. and Drenvich, V. P., Shear modulus and damping in
of deformations obtained by various methods lies well soils: design equations and curves. J. Soil Mech. Found. Div.,
within this prescribed limit. Hence, it may be concluded 1972, 98(7), 667–692.

CURRENT SCIENCE, VOL. 113, NO. 5, 10 SEPTEMBER 2017 909

RESEARCH ARTICLES
11. Gutenberg, B. and Richter, C. F., Earthquake magnitude, intensity, 17. Seed, H. B. and Sun, J. H., Implication of site effects in the Mex-
energy and acceleration (second paper). Bull. Seismol. Soc. Am., ico City earthquake of 19 September 1985 for Earthquake-
1956, 46(2), 105–145. resistance-design criteria in the San Francisco Bay Area of Cali-
12. Idriss, I. M., Evaluating seismic risk in engineering practice. In fornia. In Report No. UCB/EERC-89/03, University of California,
Proceedings of the 11th International Conference on Soil Mechan- Berkeley, USA, 1989.
ics and Foundation Engineering, San Francisco, USA, 1985, vol. 18. Bishop, A. W., The use of slip circles in the stability analysis of
1, pp. 255–320. earth slopes. Geotechnique, 1955, 5(1), 7–17.
13. Ghosh, A. K., Rao, K. S. and Kushwaha, H. S., Development for 19. Seed, H. B. and Martin, G. R., The seismic coefficient in earth
UHRS for Tarapur, Trombay and Kakrapara Sites. Report dam design. J. Soil Mech. Found. Div., 1966, 92(3), 25–58.
BARC/2003/E/019, 2003. 20. Seed, H. B. and Idriss, I. M., A simplified procedure for evaluat-
14. Sengupta, A., Estimation of permanent displacements of the Tehri ing soil liquefaction potential. J. Soil Mech. Found. Div., ASCE,
dam in the Himalayas due to future strong earthquakes. Sadhana, 1971, 97(9), 1249–1274.
2010, 35(3), 373–392. 21. IIT-K, GSDMA guidelines for seismic design of earth dams and
15. Ghaboussi, J. and Wilson, E. L., Seismic analysis of earth dam embankments, 2007.
reservoir systems. J. Soil Mech. Found. Div., 1973, 99(10),
849–862.
16. Dakoulas, P. and Gazetas, G., A class of inhomogeneous shear Received 20 June 2016; revised accepted 9 March 2017
models for seismic response of dams and embankments. Soil Dyn.
Earthq. Eng., 1985, 4(4), 166–182. doi: 10.18520/cs/v113/i05/902-910

910 CURRENT SCIENCE, VOL. 113, NO. 5, 10 SEPTEMBER 2017