See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/281278364

SEISMIC RESPONSE OF EARTH DAMS

Research · August 2015

DOI: 10.13140/RG.2.1.3162.2887

CITATION READS

1 2,582

1 author:

Madhavi Latha Gali

Indian Institute of Science
144 PUBLICATIONS   3,305 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Stability Analysis of Slopes Supporting Chenab Bridge View project

Advanced Technologies for Post-Disasters Reconnaissance, Forensic and Environmental Impact Studies-Geotechnical View project

All content following this page was uploaded by Madhavi Latha Gali on 27 August 2015.

The user has requested enhancement of the downloaded file.

 SEISMIC RESPONSE OF EARTH DAMS

Dr. G. Madhavi Latha, Indian Institute of Science

1. INTRODUCTION

Earth dams are water-retaining structures built for the purpose of irrigation, hydropower
generation or recreation. They can be more than 100m in height, having a water holding
capacity of several million cubic meters and be located just miles upstream of populated
areas. Earth dams are a class of structures, which, if breached during an earthquake can
produce serious consequences ranging from direct loss of life to indirect economic losses.
The seismic vulnerability of earth dams and the potential consequences of their failures
necessitate detailed investigation of their seismic response.

Most common source of distress to earth dams during earthquakes is slope failure, piping and
displacement or settlement resulting from shearing of weak soil and/or liquefaction of loose,
saturated sands within the embankment or foundation. Possible damaging effects of
earthquakes on earth dams include:

 Fault displacement in the foundation.

 Overtopping of dam caused by settlement or by earthquake generated water waves in
the reservoir
 Sliding or settlement of weak or liquefiable foundation soils.
 Slide failure of an embankment composed of weak or liquefiable soils.
 Piping through cracks in the dam core resulting from settlement or displacement

Earthquake loading may trigger any one of the above failure modes or their combinations.

2. PAST EXPERIENCES

Over the years, many researchers have studied the performance of earth dams during
earthquakes. One of the oldest and most frequently studied cases is the Sheffield Dam failure,
as described in detail by Seed et al. (1969). The 7.6 m high dam was subjected to a M=6.3
(Richter’s magnitude), R=11 km (R is the distance from epicenter) earthquake. The peak
acceleration was estimated to be 0.2g. Castro et al. (1985) and Gu et al. (1993) have
summarized the failure of the Lower San Fernando Dam where the 43-m high dam failed
when subjected to the M=6.5, R=8km, 1971 San Fernando earthquake. Again the dam failed
due to liquefaction of the upstream foundation soils. Holzer (1998) reported the damage to
earth dams in California due to the 1989 Loma Prieta earthquake.

Tani (1996, 2000) summarized the seismic behavior and earthquake damage of earth dams in
Japan. Sitharam and Govindaraju (2004) have summarized the behavior of earth dams
subjected to 2001 Bhuj, India earthquake (M= 6.9). Total 18 large irrigation dams; including
Rudramata dam and Tapar dam were severely damaged mainly due to liquefaction. Reviews
of methods for estimating the seismic performance of earth dams can be found in Gazetas
(1987) and Gazetas and Dakoulas (1992). Seed et al. (1978) undertook a comprehensive
analysis of the performance of earth dams during earthquakes. They concluded that earth

1
dams have performed both well and poorly when subjected to earthquakes. They observed
that hydraulic fill dams in particular have been found to be vulnerable. They also emphasize
that dams constructed of clay soils on clay or rock foundation have withstood strong shaking
but embankments made of saturated cohesionless soil have failed when subjected to strong
shaking, primarily due to built up of pore water pressure.

3. SEISMIC STABILITY ANALYSIS OF EARTH DAMS

Seismic stability analysis of dam can be performed by any of the following methods.

 Pseudostatic Stability Analysis

 Sliding Block Method
 Dynamic Analysis (Simplified or Rigorous)

Well-compacted embankments and dams founded on dense soils located in zones of low to
moderate seismicity (Zones II and III specified by IS: 1893 Part I, 2002) may be evaluated by
the pseudostatic method. In high seismic zones (Zones IV and V) or where foundation
liquefaction potential exists, more sophisticated methods of analyses such as Sliding Block
Analysis or Simplified Dynamic Analysis will be necessary. Site-specific seismic evaluations
should be performed for all projects located in active fault zone or where the liquefaction
potential exists both for embankment and foundation and for all large dams.

3.1 PSEUDOSTATIC STABILITY ANALYSIS

The most common approach to seismic slope stability evaluation is the pseudostatic analysis,
which is based on limit equilibrium methods (Kramer, 2003). In pseudo-static analysis, the
complex dynamic earthquake shaking is replaced by a single constant unidirectional pseudo-
static acceleration. Slope stability analysis is similar to that for static conditions except for the
application of horizontal and vertical inertia forces over every portion of the soil mass. These
inertial forces, Fh and Fv, are assumed to act through the centroid of the failure mass. The
seismic coefficient and the allowable factor of safety are specified such that the cumulative
permanent deformation in the design earthquake is small enough to be acceptable.

The design seismic inertia forces for pseudostatic analyses shall be taken as:

In the horizontal direction: Fh = (ZI/3) W

In the vertical direction: Fv =  (2/3) FH

Z is the Seismic Zone factor, I is the importance factor given in IS: 1893 (Part 1) 2002and W
is the weight of the sliding mass.

These horizontal and vertical forces are usually expressed as a product of horizontal or
vertical seismic coefficients (kh and kv) and the weight of the potential sliding mass.
Horizontal pseudostatic force decreases the factor of safety by reducing the resisting force
and increasing the driving force. The vertical pseudostatic force typically has less influence
on the factor of safety since it has similar effects on both driving and resisting forces. For this
reason, vertical acceleration is often ignored in pseudostatic analyses. The limit equilibrium
factor of safety of a slope strongly depends on the value of the horizontal seismic coefficient.

2
Typical seismic coefficients based on spectral curves and factors of safety used in practice
can be obtained from IS: 1893(Part I), 2002.

This approach may also be used in conjunction with the method of slices for soils possessing
cohesion and friction, where the earthquake induced forces are considered by applying a
horizontal force [Sarma (1979)]. The method then involves the computation of the critical
horizontal acceleration required to bring the soil above the slip surface into a state of limiting
equilibrium, which can then be used as an index of stability.

Although the pseudostatic approach to stability analysis is simple and straight forward
producing factor of safety, it suffers from many limitations as it can not really simulate the
complex dynamic effects of earthquake shaking through a constant unidirectional
pseudostatic acceleration. These limitations were recognized by many researchers including
Terzaghi (1950), Seed (1968) etc. More specifically in case of soils that build up large pore
water pressures or have degradation in strength of more than 15% due to the earthquake
shaking the analysis can be unreliable. As shown by Seed (1979) a number of dams such as
the upper and lower San Fernando Dams, Sheffield Dam etc have in fact failed due to
earthquake shaking although the calculated factors of safety were well above 1.0.

3.2 SLIDING BLOCK METHOD

Newmark (1965) developed a displacement-based method to predict seismically induced

permanent slope deformations of slopes. Newmark made an analogy between the unstable
soil mass in a slope and a block resting on an inclined plane (Fig. 1). In the Newmark
analysis, the mass of soil located above the critical failure surface is represented as a rigid
block. A pseudostatic analysis is performed to determine the yield acceleration, a y, i.e. the
pseudostatic acceleration that produces FS = 1.0. If the inclined plane is subjected to dynamic
motion greater than the yield acceleration, the block will slide down the inclined slope. The
yield acceleration is defined to be the product of the horizontal seismic coefficient, kh and the
acceleration of gravity.

Figure 1. Analogy between potentially unstable slope and a rigid block on an inclined plane

Horizontal peak ground acceleration for sliding block analysis shall be taken as (ZIg) and
vertical peak ground acceleration as two-thirds of the horizontal peak ground acceleration,
where Z is the Zone factor, I is the importance factor and g is acceleration due to gravity.

Sliding block method assumes that the potential failure mass and the dam are rigid. Although
this assumption may be partly satisfied in the case of earth dams composed of very stiff or

3
hard soils or slopes subject to low frequency motion, this may not be true for softer soils.
This means that the resultant inertial force and the resulting permanent displacements
calculated with the rigid sliding block method are significantly overestimated. Sliding block
deformation estimates are strictly applicable only to dams not subjected to liquefaction
failure.
Makdisi and Seed (1978) adopted the Newmark approach and added a few simplistic
assumptions like dynamic yield strength of soil being equal to 80% of the undrained strength,
etc. Using various real and synthetic earthquakes, Makdisi and Seed computed the variation
in deformation of the embankment with ay/amax ratio and magnitude, which was published as
charts. Various steps involved in this method are summarized below.

 Determine Yield Acceleration (ay) from standard methods of stability analysis as the
average acceleration producing a horizontal inertia force on a potential sliding mass
so as to produce a factor of safety of unity and thus causes it to experience permanent
displacement.
 Determine Earthquake Induced Average Acceleration on the basis of one and two-
dimensional equivalent linear site response analyses of earth dams.
 Determine Permanent Seismic Vertical Deformation using the simplified seismic
deformation chart suggested by Makdisi and Seed (1978) given in Figure 2. This
figure was developed from the results of two dimensional finite element analyses of
earth dams. This figure includes the effect of amplification of seismic motions by the
earth dam and provides upper and lower bounds on the permanent deformation as a
function of magnitude.
 Check whether the calculated permanent seismic displacement is within acceptable
limits of deformation or not.

Figure 2 Permanent Displacement versus Normalized Yield Acceleration for Embankment dams
(Makdisi and Seed, 1978)

4
3.3 DYNAMIC ANALYSIS

Dynamic analyses of the response of soil systems to earthquakes have until recently been
solved by very simple numerical models. More recently equivalent linear methods have been
developed in which the nonlinear characteristics of the soil skeleton are accounted for by use
of equivalent linear soil properties, and an iterative procedure is used to obtain modulus and
damping values compatible with the effective strains computed in each zone of the soil mass.
Dynamic analysis essentially involves estimation of the deformation behavior of an earth dam
or an embankment using the finite element or finite difference method. A complete and
detailed dynamic analysis is a major undertaking that requires extensive database and
specialized skills. A rigorous dynamic analysis is recommended for large dams and dams
located over active fault zone. Small to intermediate size dams found to be unsafe by the
Simplified Dynamic Analysis may also require rigorous dynamic analysis. Seed (1979) and
Finn et al. (1986) summarize procedures for dynamic analyses of dams. These procedures
usually involve the following steps:

 Determine pre-earthquake static stress using a static finite element analysis in the
embankment for initial effective normal stress and shear stress along the potential
failure surface.
 Evaluate the dynamic soil behaviour from cyclic laboratory tests for input soil
properties required in the dynamic analyses.
 Determine the dynamic response of the embankment and foundation to base rock
motion. Using a suite of accelerograms representing earthquakes of magnitude and
peak acceleration similar to those of the design earthquake from earthquakes recorded
in a similar geologic environment, the response of the embankment is determined by
dynamic finite element or finite difference modeling, using either equivalent linear or
nonlinear procedures.
 Evaluate the stability of the embankment by comparing stresses induced by the
earthquake to those required to cause liquefaction or to exceed prescribed limits of
local strain. The locations within the embankments for which induced stresses exceed
those required to cause liquefaction or produce unacceptable strain determine the
extent of liquefaction or unacceptable performance.
 Evaluate embankment deformations on the basis of strain potential for the individual
elements, which corresponds to the strain that would be experienced if the element
were not constrained by surrounding soil.
 Calculate total embankment deformation on the basis of gravity loads and softened
material properties to determine whether they lie within the acceptable limits.

Dynamic analysis needs a set of input parameters for analyzing the seismic response of earth
dams. These input parameters can be classified as
 Physical and design parameters (e.g. dam height, storage capacity, embankment
slopes, depth of foundation, etc.)
 Embankment material data (e.g. strength and permeability of the clay core, etc.)
 Foundation material data (e.g. density and strength of foundation soil, etc)
 Material Constitutive Law (Cyclic stress-strain behavior of the soil)
 Acceleration Time Histories of expected ground motions

Dynamic analysis employing a non-linear stress-strain relationship provides a rational

framework for estimation of deformation of an earth dam. The biggest difficulty in

5
employing these models is to obtain soil stress-strain models that are representative of the soil
insitu behavior. This approach requires an accurate characterization of the stress-strain
behavior of the materials within the body of the earth dam or embankment and foundation.
Dynamic analysis of earth dams and embankments also require a suite of earthquake time
histories representing design earthquakes.

3.4 LIQUEFACTION EFFECTS

It has been found that only dams susceptible to liquefaction have failed due to earthquakes.
The pseudostatic approach and the Newmark sliding approach were unable to explain the
foundation failure of embankments during earthquake. Seed and Lee (1966) had
demonstrated that pore pressure builds up in sands when subjected to dynamic loading.
Liquefaction of saturated sands in a dam foundation may be caused due to pore water
pressure build-up, as a result of densification during earthquake shaking. Liquefaction leads
to loss of strength and increase in settlement. Loose sand is highly susceptible to liquefaction.
Because the seismic shaking occurs so quickly, the cohesionless soil is subjected to an un-
drained loading. The increase in pore water pressure causes an upward flow of water to the
ground surface, where it emerges in the form of mud spouts or sand boils. For this state of
liquefaction the effective stress is zero. In contrast, clayey soils are not subjected to such
effects and they do not really liquefy. Sensitive clays however may suffer considerable loss
of strength, as deformations of the initial pulses may tend to destroy the structure. Dense sand
on the other hand didn’t exhibit this behavior because of its dilative nature. Later, Seed et al.
(1978) performed a comprehensive review of the performance of earth dams during
earthquakes and emphasized the importance of liquefaction. Although many deposits never
attain the state of zero effective stress, they deform substantially during earthquakes leading
to the development of liquefaction-like failures. Liquefaction is therefore often functionally
defined as the state in which a double amplitude shear strain of 5% develops. If the material
is found to be liquefaction susceptible, the deformation of the embankment should be
estimated for the design ground motion accounting for liquefaction effects.

4. DESIGN MEASURES

The following defensive measures while designing new dams will reduce the failures in the
event of these dams subjected to earthquakes.

 Earth dams should be ideally located away from any potentially active fault or an area
underlain by liquefiable or sensitive soils or abutments prone to static or cyclic
instability.
 Additional dam height should be provided to avoid the possibility of overtopping by
earthquake-generated water waves or the loss of crest elevation due to deformation,
slumping, and fault displacement.
 Wider transition and filter sections are to be provided as a defense against cracking.
 Filter layers should be adequately permeable and well-graded
 Embankment should be zoned with wide impervious cores of plastic clay to minimize
saturation of materials and to accommodate deformation.
 Slopes should be stabilized to provide safety against large slides
 Liquefaction susceptible foundation soils should be removed and replaced or suitable
mitigation techniques are to be adopted.
 Dam design should include well-planned instrumentation and monitoring schemes to
minimize or manage loss that may result from failure

6
 REFECENCES

Castro, G., Poulos, S.J. and Leathers, F.D. (1985). Re-examination of slide of lower San
Fernando Dam, J. Geotechnical Engg., ASCE, 111:9, 1093–1107.
Finn, W.D.L., Yogendrakumar, M., Yoshida, M., and Yoshida, N. (1986). TARA-3: A
Program to Compute the Response of 2-D Embankments and Soil-Structure
Interaction Systems to Seismic Loadings, Department of Civil Engineering,
University of British Columbia, Vancouver.
Gazetas, G. (1987). Seismic response of earth dams: some recent developments, Soil
Dynamics and Earthquake Engg., 6, 2-47.
Gazetas, G., and Dakoulas, P. (1992). Seismic analysis and design of rockfill dams: State-of-
the-art. Soil Dynamics and Earthquake Engg., 11, 27-61.
Gu, W.H., Morgenstern, N.R. and Robertson, P.K. (1993). Progressive failure of lower San
Fernando Dam, J. Geotechnical Engrg., ASCE, 119:2, 333–349.
Holzer, T.L., (1998). The Loma Prieta, California, Earthquake of October 17, 1989- Earth
structures and engineering characterization of ground motion, U.S. Geological Survey
Professional Paper 1552-D, U.S. Government Printing Office, Washington.
Kramer, S. L. (2003). Geotechnical Earthquake Engineering, Pearson Education, India.
Makdisi, F.I. and Seed, H.B. (1978). Simplified procedure for estimating dam and
embankment earthquake-induced deformations, Journal of the Geotechnical
Engineering Division, ASCE, 104: GT7, 849-867.
Newmark, N.(1965). Effects of earthquakes on dams and embankments, Geotechnique, 15: 2,
139-160.
Sarma, S. K. (1979). Stability Analysis of Embankments and Slopes, Journal of the
Geotechnical Engineering Division, ASCE, 105: GT 12, 511-1524.
Seed, H. B. (1968). Landslides during earthquakes due to soil liquefaction, J. Geotech. Engg.
Div. ASCE, 94: SM5, 1055–1123.
Seed, H. B. (1979). Consideration in the earthquake resistant design of earth and rock-fill
dams, Geotechnique, 29: 3, 215-263.
Seed, H.B., F.I. Makdisi, and P. DeAlba, (1978). Performance of Earth Dams During
Earthquakes, Journal of the Geotechnical Division, ASCE, 104:GT7.
Sitharam, T.G. and Govindaraju, L. (2004) Geotechnical aspects and ground response studies
in Bhuj earthquake, India, Geotechnical and Geological Engineering, 22: 439–455.
Tani S. (2000) Behavior of large fill dams during earthquake and earthquake damage,
Journal of Soil Dynamics and Earthquake Engineering, 20, 223-229.
Tani, S. (1996) Damage to earth dams, Soils Foundations, 36:1, 263–272.
Terzaghi, K. (1950). Mechanisms of landslides, Engineering Geology (Berkey) Volume,
Geological Society of America, NY, 83-123.

View publication stats