SEISMIC SLOPE STABILITY

Jorge F. Meneses, PhD, PE, GE, D.GE, F.ASCE

Methods
 ANALYSIS PROCEDURES
 Detailed, comprehensive analyses
 Pseudostatic analyses
 Sliding block analyses
 PSEUDOSTATIC SCREENING ANALYSES
 DETERMINING PEAK ACCELERATION
 SHEAR STRENGTH FOR PSEUDOSTATIC ANALYSES
 Earthquakes immediately after construction
 Earthquakes after the slope has reached consolidation
equilibrium
 Effects of rapid load application
 POSTEARTHQUAKE STABILITY ANALYSES
ANALYSIS PROCEDURES
Detailed, comprehensive analyses
 Generally used for any large embankment or any slope
or embankment where the consequences of failure are
high or significant soil strength losses occur
 The general approach may include the following steps:
 Step1: Determine the cross section of the slope and
underlying foundation
 Step 2: Determine the anticipated acceleration time history
(GMPEs)
 Step 3: Determine the static and dynamic stress-strain
properties of the natural soils and fill materials within and
beneath the slope
 Step 4: Estimate the initial static stresses in the slope or
embankment prior to the earthquake
Detailed, comprehensive analyses (2)

 Step 5: Perform a dynamic finite element analysis to

compute the stresses and strains induced in the
embankment
 Step 6: Estimate the reductions in shear strength and
increases in PWP that will result from the earthquake
 Step 7: Compute the stability of the slope using
conventional limit equilibrium procedures with the
reduced shear strengths. This may require analyses
using both undrained and drained shear strengths to
determine which strengths are most critical
Detailed, comprehensive analyses (3)

 Step 8: If the analyses indicate that the slope will be

stable after the earthquake, compute the permanent
displacements
 Ifstrength losses due to cyclic loading are small, a
Newmark-type sliding block analysis may be used
 However, if strength losses are significant, other methods
should be used (e.g., strain potential)
 A complete nonlinear finite element analysis should be able
to calculate any permanent displacements in a slope,
however such analyses are very complex, involve
considerable uncertainties, and are seldom performed in
practice
Pseudostatic analyses
 Earthquake loading is represented by a static force,
equal to the soil weight multiplied by a seismic
coefficient, k
 The pseudostatic force is used in a conventional limit
equilibrium slope stability analysis
 The pseudostatic force is treated as a static force
and acts in only one direction, whereas the
earthquake accelerations act for only a short time
and change direction, tending at certain instances in
time to stabilize rather than destabilize the soil
Pseudostatic analyses (2)
 The vertical components of the earthquake
accelerations are usually neglected, and the seismic
coefficient usually represents a horizontal force
 The pseudostatic force is assumed to be known
force and is included in the various equilibrium
equations
Pseudostatic analyses (3)

Derivation of the equation for the FS of an infinite slope with a seismic force (kW)-
total stress analyses
Pseudostatic analyses (4)
 An issue is the location of the pseudostatic force
 Terzaghi (1950) suggested that the pseudostatic
force should act through the center of gravity of
each slice or the entire sliding mass
 Thiswould be true only if the accelerations were
constant over the entire soil mass, which they probably
are not
 Seed (1979) showed that the location can have a
small but noticeable effect on the computed FS
Pseudostatic analyses (5)
 Dynamic analyses of the response of many dams to
earthquakes (Makdisi and Seed 1978) indicate that the
peak accelerations increase (amplify) from the bottom
to the top of a dam
 Thus, the location of the resultant seismic force would be
expected to be above the center of gravity of the slice
 In the case of circular slip surfaces, this would reduce
the moment about the center of the circle due to the
seismic forces, in comparison to applying the force at
the center of gravity of the slice, and the FS would be
expected to increase
Pseudostatic analyses (6)
 Consistent with results from evaluation of Sheffield
Dam, which showed that the FS decreased when the
seismic force was located below the center of
gravity of the slice
 In the case of circular slip surfaces, this would
reduce the moment about the center of the circle
due to the seismic forces, in comparison to applying
the force at the center of gravity of the slice, and
the FS would be expected to increase
Pseudostatic analyses (7)
 Assuming that the pseudostatic force acts through the
center of gravity of the slice is probably slightly
conservative for most dams
 Thus, it appears that Terzaghi’s suggestion is reasonable
 For most pseudostatic analyses the pseudostatic force is
assumed to act through the center of gravity of each
slice
 If a force equilibrium (only) procedure is used, the location
of the pseudostatic force ha no effect on the FS computed
Pseudostatic analyses (8)
 For many years, seismic coefficients were estimated
based on empirical guidelines and codes. Typical values
for seismic coefficients used ranged from about 0.05 to
about 0.25
 However, with the development of more sophisticated
analyses, particularly displacement analyses,
correlations can be made between the seismic
coefficient, the expected earthquake accelerations, and
the probable displacements
 Most seismic coefficients used today are based on
experience and results from deformation analyses
Sliding Block Analyses
 Newmark (1965) first suggested a relatively simple
deformation analysis based on a rigid sliding block
 In this approach the displacement of a mass of soil
above a slip surface is modeled as a rigid block of
soil sliding on a plane surface
 When the acceleration of the block exceeds a yield
acceleration, ay, the block begins to slip along the
plane
(a) Actual slope; (b) sliding block representation used to
compute permanent soil displacements in a slope
subjected to earthquake shaking
Double integration of acceleration-time history to
compute permanent displacements
Seismic yield coefficient
 Limit equilibrium slope stability analyses are used to
compute the values of yield acceleration, ay, used
in sliding block analyses
 The yield acceleration is usually expressed as a
seismic yield coefficient, ky=ay/g
 The seismic yield coefficient is the seismic coefficient
that produces a FS of unity in a psudostatic slope
stability analysis
Seismic yield coefficient (2)
 The value of ky is determined using conventional
slope stability analysis procedures
 However, rather than searching for the slip surface
that gives the minimum FS, searches are conducted
to find the slip surface that gives the minimum value
of ky
 The slip surface giving the minimum value of ky is
usually different from the one giving the minimum FS
for static conditions
Pseudostatic screening analyses
 Several simple screening criteria have been
developed for evaluating seismic stability using
pseudostatic analysis procedures
 The screening criteria differ in the reference seismic
acceleration, acceleration multiplier, strength
reduction factor, acceptable FS, and tolerable
displacement criterion used
 Components of pseudostatic screening
analyses
 A reference peak acceleration, aref
 Either the peak acceleration in bedrock beneath the
slope, or the peak soil acceleration at the top of the
slope
 Peak bedrock acceleration is easier to use, because
determining peak acceleration at the top of the slope
requires a dynamic response analysis
 Components of pseudostatic screening
analyses (2)
 Acceleration multiplier
 The seismic coefficient used in the pseudostatic analysis
is equal to aref/g multiplied by an acceleration
multiplier, a/aref
k= (aref/g)(a/aref)
 Values of acceleration multiplier ranging from 0.17 to
0.75 have been recommended
 Components of pseudostatic screening
analyses (3)
 Shear strength reduction factor
 Most authorities recommend using reduced shear
strength in pseudostatic analyses
 The strength most often recommended is 80% of the
static shear strength (Makdisi and Seed 1977)
Components of pseudostatic screening
analyses (3)

 Minimum FS
 Values are either 1.0 or 1.15
 Components of pseudostatic screening
analyses (4)
 Tolerable permanent deformation
 Certain amount of earthquake-induced deformation is
tolerable
 The magnitudes of deformation judged to be tolerable
vary from 0.15m in the case of landfill base liners to
1.0 for dams
Suggested methods for performing
pseudostatic screening analyses
Shear strength for pseudostatic analyses

 The shear strength appropriate for use in a

psudostatic analysis depend on whether the analysis
is being performed for short-term (end-of-
construction) conditions or for a slope that has
been in existence for many years
 Pseudostatic analyses may need to be performed
for both short- and long-term conditions depending
on the particular slope
Shear strength for pseudostatic analyses (2)

 Because seismic loading is of short duration, it is

reasonable to assume that except for some coarse
gravels and cobbles, the soil will not drain
appreciably during the period of earthquake
shaking
 Thus, undrained shear strengths are used for most
pseudostatic analyses (with the exception of soils
that tend to dilate when sheared and may lose
strength after the earthquake as they drain)
Earthquakes immediately after construction

 Pseudostatic analyses for short-term stability are

only appropriate for new slopes
 Undrained shear strength can be evaluated using
conventional UU testing procedures and samples
identical to the ones that would be tested to
determine the shear strength for static conditions
 The analyses are performed using shear strengths
expressed in terms of total stresses
 Earthquakes after the slope has
reached consolidated equilibrium
 All slopes that will be subjected to earthquakes
should be evaluated for long-term stability using
values of undrained shear strength that reflect the
eventual long-term conditions, including
consolidation or swell after the slope is constructed
 The manner in which the undrained shear strength is
determined for this condition depends on whether
we are dealing with an existing slope or a slope
that is yet to be built
Earthquakes after the slope has
reached consolidated equilibrium (2)
 Existing slopes
 If a slope has reached consolidated equilibrium, the shear
strength can be determined by taking representative
samples of the soil and performing tests using UU testing
procedures
 The stability analysis is then performed much like a short-
term stability analysis, using shear strength parameters
expressed in terms of total stresses
Earthquakes after the slope has
reached consolidated equilibrium (2)
 New slopes
 Itis necessary to simulate the effects of future consolidation
and swell in the laboratory using CU testing procedures
 The soil is first consolidated to a state of effective stress and
then sheared with no drainage
 The undrained loading for an earthquake is caused by
seismic forces
 Then the slope stability computations are performed using a
two-stage analysis procedure
Earthquakes after the slope has
reached consolidated equilibrium (3)
 New slopes
A first-stage analysis is performed for conditions prior to the
earthquake (no seismic coefficient) to compute the
consolidation stresses, σfc and tfc
 These stresses are then used to estimate the undrained shear
strength for seismic loading
 The undrained shear strength is then used in the second-
stage computations (with seismic coefficient) to compute the
pseudostatic FS for the slope
 Drainage after the earthquake could adversely affect
stability and should be considered
Simplified procedure (R envelope and
single-stage analysis)
 Although the two-stage analysis procedure is the proper
way to perform a pseudostatic analysis for a slope,
analyses are sometimes performed using a simple
single-stage procedure
 In the single-stage procedure the R shear strength
envelope is used
 The R envelope is obtained by plotting results from CU
triaxial tests
 FS computed by single-stage procedure is less than the
FS computed by the more rigourous two-stage
procedure
Simplified procedure (R envelope and
single-stage analysis) (2)
Post-earthquake stability analyses
 Following an earthquake, the stability of a slope
may be diminished because cyclic loading has
reduced the shear strength of the soil
 The reductions in shear strength are generally
treated differently depending on whether or not
liquefaction occurs
 Stability following an earthquake can be evaluated
using a three-step process
Post-earthquake stability analyses (2)

 Step 1. Estimate if liquefaction will occur

 Step 2. Estimate reduced undrained shear strengths
 Step 3. Compute slope stability
Post-earthquake stability analyses (3)

 Step 1. Estimate if liquefaction will occur

 Youd et al. (2001)
 Idriss and Boulanger (2008)

 Boulanger and Idriss (2014)

 Cetin et al. (2005)

 Field tests: SPT, CPT, Vs measurements and BPT

 FS=CRR/CSR
Post-earthquake stability analyses (4)

 Step 2. Estimate reduced undrained shear strength

 Seed and Harder (1990)
 Poulos et al. (1985)

 Stark and Mesri (1992)

 Olson and Stark (2002)

 Idriss and Boulanger (2008)

Relationship between corrected clean sand blowcounts and
undrained residual strength from case studies (Seed and Harder
1990)
Relationship between undrained critical strength ratio and
equivalent clean sand blow count (Stark and Mesri 1992)
Liquefied strength ratio relationship based on
normalized CPT tip resistance (Olson and Stark 2002)
Post-earthquake stability analyses (5)

 Step 3. Compute slope stability

 Once the postearthquake shear strengths have been
determined, a conventional static slope stability
analysis is performed