Lecture 6 - Seismic Performance of Slopes STU

Evaluating Seismic Performance of
Dams/Slopes
Tezera Firew Azmatch, PhD, PEng

Addis Ababa University
Addis Ababa Institute of Technology
Faculty of Civil and Environmental Engineering
Contents
 Introduction
 Steps in seismic evaluation
 Selecting the design earthquake
 Liquefaction susceptibility
 Post-liquefaction stability analysis
 Pseudo static stability analysis
 Seismic deformation analyses
 Summary
 Some useful references
2
Introduction
 Earthquakes affect the stability of slopes in two ways:

(1) The acceleration produced by the seismic ground motion during an earthquake
subjects the soil to cyclically varying forces, and
(2) the cyclic strains induced by the earthquake loads may cause reduction in the
shear strength of the soil.
 If the strength of the soil is reduced less than 15% by cyclic loading, pseudostatic
analyses of the earthquake loading can be used.
 In pseudostatic analyses, the effect of the earthquake is represented crudely by applying a
static horizontal force to the potential sliding mass.
 This type of analysis, provides a semi-empirical means of determining whether
deformations due to an earthquake will be acceptably small.
 If the strength of the soil is reduced more than 15% as a result of cyclic loading,
dynamic analysis are needed to estimate the deformations that would result from
earthquakes.
 Some engineers perform this type of analysis for all slopes, even if the strength reduction
due to earthquake loading is less than 15%.
 These more complex analyses are highly specialized and are beyond the scope of this
course
Introduction
4
Introduction
Introduction
6
Two Critical Design Issues in Evaluating
Seismic Performance
 Are there materials that will lose significant strength as a result of cyclic loading (e.g.
soil liquefaction)?
 Potential for large displacements/flow slides
 If not, will the structure undergo significant deformations that may jeopardize
satisfactory performance?
 Seismically induced permanent displacements
Steps in Seismic Slope Evaluation
 Determine the design earthquake
 Determine liquefaction susceptibility
 Carry out an appropriate analysis

 Start with simplistic techniques
 Proceed to complex methods only if they are warranted
8
Seismic Analysis Procedure
Determine Design Earthquake Ground

Motion (EDGM)
Is soil expected to liquefy under the

design earthquake?
Yes
No (i.e. strength loss due to

cyclic loading is < 20%) Carry out Post Earthquake
Stability Analysis
Pseudo Static Stability If stability acceptable

Screening Analysis
Calculate Permanent Calculate Permanent
Deformations Deformations
Simplified Numerical
Techniques Modeling
Ref: ICOLD (2011), Duncan and Wright (2005) 9
Quantifying Earthquake Ground Motion
 Acceleration-time history provides a complete

definition of one of many possible earthquake
ground motions at a site
 Simplified parameters that define components

of an acceleration-time history:
 Peak Ground Acceleration (PGA)
 Mean Period (Tm) Acceleration-time History
 Significant Duration (D5-95)
 Spectral acceleration (Sa)

 Used to characterize an equivalent seismic
loading on a structure from the earthquake
ground motion
 Spectral acceleration is specified with an
associated period, e.g. Sa(0.2)
Acceleration Response Spectrum
10
10
Selecting the Design Earthquake
 CDA (2007, 2013) dam classification governs the return period or the annual
exceedance probability (AEP) to be used in selecting the design earthquake ground
motion
 A seismic hazard evaluation allows estimation of design earthquakes possible for

various AEPs and Site Classes
 Design earthquake applicable to the design is selected from the seismic hazard
evaluation using the AEP and Site Class applicable
 For Fort McMurray region, one such seismic hazard evaluation exists:
 Seismic Hazard Working Group Report (KCB, 2006)
 An industry study sponsored by Syncrude, Suncor, Albian and CNRL
11
11
CDA Dam Classification Scheme
12
12
Local Site Classification
13
13
Local Site Classification
14
14
(Contd…)
CDA (2007, 2013) Guidance for Selecting Earthquake Hazards
Sample Results from the SHS Report for Fort McMurray Region
15
15

(Contd…)
16
16
(Contd…)
Link for global seismic

hazard map
https://maps.openquake.org/map/global-seismic-hazard-map/#5/11.213/41.106
17
17

(Contd…)
18
18

Motion (EDGM)

design earthquake?
Yes
No (i.e. strength loss due to cyclic loading is < 20%)

Carry out Post Earthquake
Stability Analysis
If stability acceptable
Pseudo Static Stability
Screening Analysis
Techniques Modeling
19
Liquefaction Susceptibility
 Two types of liquefaction:
 Statically induced liquefaction (not the focus of this presentation)

– Caused by
– Rapid loading
– Foundation movements
 Seismically induced liquefaction
20
20
Static Liquefaction Susceptibility
 Semi-empirical
 Compare N1,60 or qc,1 against case

records
 Result:
 Contractive (potentially liquefiable), or
 Dilative (non-liquefiable)
21
21
Seismic Liquefaction Susceptibility
 Semi-empirical
 Based on comparing Cyclic Stress Ratio to Cyclic CRRM=7.5
Resistance Ratio (Youd. et al. 2001)
 Cyclic Stress Ratio (CSR) defines the earthquake
loading or seismic shear stress
CSR or CRR
 Cyclic Resistance Ratio (CRR) is the seismic shear

stress required to cause liquefaction
 Correlations with field test results based on case histories CSRM=7.5
 Field tests:
– CPT, SPT, Vs, and for gravelly sites, Becker
penetration tests
 If CSR>=CRRM, soil will liquefy
Ref: Seed et. al. (2003)

Note: A similar graph exists for qc1
22
22

Motion (EDGM)

design earthquake?
Yes

Stability Analysis
Screening Analysis
Techniques Modeling
23
Pseudo Static Stability Analysis
 One of the earliest procedures of analysis for seismic stability
 Limited to soils that do not lose strength by more than 20%

during the earthquake (not applicable to liquefied soils)
 Earthquake loading is represented by a static force

 Static force = Soil Weight (W) * Seismic Coefficient (k)
 Can be easily performed using computer programs such as

Slope/W
24
24
 The critical component is the selection of the seismic

coefficient
 Earlier days k = 0.05 to 0.25 (empirical)
 More recent:
– k = 0.5*PGA (Hynes-Griffin and Franklin, 1984)
– Validated by comparing with detailed seismic analyses
– Assumes an acceptable deformation of 1 m
– k = f(allowable displacements, earthquake magnitude and

spectral acceleration)
– Bray and Travasarou, 2009
 Pseudo-static stability analysis is mainly for screening

analyses
25
25
Ref: Duncan, J.M. and Wright, S.G. (2005) 26
26

Motion (EDGM)

design earthquake?
Yes

Stability Analysis
Screening Analysis
Techniques Modeling
27
Post Earthquake Stability Analysis
For soils in which liquefaction is expected to occur:

 Shear strengths of soils could be
lower after an earthquake
 Reductions in shear strengths
are generally treated differently
depending on whether or not
liquefaction occurs
 Use reduced shear strengths in
a conventional stability analysis
without a seismic coefficient
 Acceptable Factor of Safety ~
1.2-1.3 (CDA, 2007)
28
28
For soils in which liquefaction is expected to occur:
 Shear strengths of soils could be

lower after an earthquake
 Reductions in shear strengths
are generally treated differently
depending on whether or not
liquefaction occurs
 Use reduced shear strengths in
a conventional stability analysis
without a seismic coefficient
 Acceptable Factor of Safety ~
1.2-1.3 (CDA, 2007)
29
29
Liquefaction does not occur:
 Saturated soils which are potentially

liquefiable, but which have not been
subject to sufficient cyclic shear stress
to reach a liquefaction condition, will
develop positive pore pressures during
the cyclic loading of the earthquake. The
amount of positive pore pressure will
depend on many factors relating to the
soil and the cyclic loading. These pore
pressures should be allowed for in post
earthquake stability analyses.
Factor of safety against liquefaction
30
30
For soils which are not expected to liquify

 Saturated soils which are potentially
liquefiable, but which have not been
subject to sufficient cyclic shear stress
to reach a liquefaction condition, will
develop positive pore pressures during
the cyclic loading of the earthquake. The
amount of positive pore pressure will
depend on many factors relating to the
soil and the cyclic loading. These pore
pressures should be allowed for in post
earthquake stability analyses.
Factor of safety against liquefaction
31
31

Motion (EDGM)

design earthquake?
Yes

Stability Analysis
Screening Analysis
Techniques Modeling
32
Permanent Deformations
 Seismic displacements (discussed in upcoming slides)
 Post liquefaction vertical settlements (if soils liquefy)
33
33
Seismic Deformation Analyses
 Simplified Methods
 Bridges the gap between overly simplistic pseudo-
static and complex stress-deformation analysis
 Based on a box sliding on a failure plane
 Stress-Deformation Modeling
 Finite element or finite difference techniques
 Complex analyses
34
34
Parameters Affecting Seismic
Displacements
 Ground motion’s spectral acceleration (at

the degraded period of the structure)
 The most important
 The structure’s yield coefficient, ky

– This is the pseudo static seismic
coefficient that results in FOS=1
– If a slope stability analysis method that
satisfied full equilibrium is used, the
estimates of ky are fairly consistent
(Duncan 1996 as cited by Bray 2009).
 The structure’s fundamental period (Ts)
35
35
Seismic Deformation Analysis –

Simplified Techniques
 Only applicable to materials that do not lose
strength significantly
 Newmark’s Rigid Sliding Block Analysis (1965)

 Models slope movement as a rigid block that
slides on an inclined plane
 Makdisi and Seed (1978) Rigid Block
 Flexible block
 Dynamic and sliding responses are decoupled
 Calibrated against a very limited number of
ground motions
 Bray and Travasarou (2007)
 Uses a nonlinear soil response, fully coupled
sliding-block model
 Flexible block
 Calibrated against 688 measured ground motions
 Can be used for probabilistic and/or deterministic
framework
 Rathje and Antonakos (2011)
36
36
Newmark(1984)
 Newmark’s method is based on a
number of simplifying assumptions:
(1) the existence of a well-defined
slip surface, (2) a rigid, perfect
plastic slide material, (3) permanent
strains occur only if the dynamic
stress exceeds the shear resistance,
and (4) the displacements are
presumed to occur in the downslope
direction only, thus implying infinite
dynamic shear resistance in the
upslope direction.
 The most important factor in a
traditional Newmark analysis is the
selection of the design
accelerograms for modeling the
seismic motions of the rigid block.
37
37

Newmark(1984)
 The effect of the elastic response of the embankment on the acceleration of

the slide mass is not taken into account so the response of the structure is
modeled only by ky.
 This simplifying assumption causes the Newmark approach to be most
appropriate for stiff structures whose response can be approximately represented
by an appropriate outcrop acceleration record.
 The yield acceleration can be readily determined using conventional limit
equilibrium methods by calculating the inertial forces required to lower the
factor of safety against block sliding to 1.0.
 It is typical to evaluate several failure surfaces in addition to that which
produces the lowest static factor of safety. The permanent displacement is
calculated by double integration of those portions of the accelerogram that
exceed the yield acceleration for the selected failure surface.
 The method usually assumes there is negligible loss of shear strength
during shaking, although this can be approximately considered by making
the yield acceleration a function of time or earthquake-induced
displacement.
38
38
Newmark(1984)
 This procedure is illustrated in Fig. 5. No
displacements occur until time t1 when the
induced acceleration reaches the yield
acceleration for the first cycle, ky1. The relative
velocity between the slide mass and underlying
material will increase until time t2 when the
acceleration drops below the yield value. The
variation in relative velocity is computed by
integration of the acceleration history over the
shaded area.
 The relative velocity reduces to zero between
time t2 and t3 since the mobilized strength on
the slide plane exceeds the stress being
induced by the earthquake motion. Subsequent
relative velocity pulses are estimated in a like
manner, and the displacement history is
computed by integration of the relative velocity
versus time relationship.
39
39

Hynes-Griffin and Franklin (1984)
 Hynes-Griffin and Franklin (1984) refined this method to account for likely
deformations based on Newmark sliding block analyses of 349 horizontal
components of natural earthquakes and 6 synthetic records for a range of
yield accelerations. From this permanent displacements u in cm were
plotted versus N/A where N is the yield acceleration and A is the peak
value of the earthquake acceleration at the base of the dam.
 Hynes-Griffin and Franklin (1984) then used an arbitrary limit of 100 cm
(one metre) displacement as representing tolerable displacements, and
from Figure 12.44, used the upper bound plot of displacements, which
gave N/A=0.17. After allowing for amplification factor of 3 between the
base acceleration and that at the crest of the dam, they concluded that a
factor of safety of 1.0 with a seismic coefficient of one half of the peak
ground acceleration at the base of the dam, would assure that
deformations would not exceed 1 metre. Their suggested method is:
Note: N is the yield acceleration Ky
40
40
Hynes-Griffin and Franklin (1984)
 Their suggested method is:

 Carry out a conventional pseudo
static stability analysis using a
seismic coefficient equal to one-
half the predicted peak ground
acceleration.
 Use a composite S-R strength
envelope (Effective stress
strength at low stresses,
undrained strength at high
stresses) for pervious soils and
R undrained strength for clays,
multiplying the strength in either
case by 0.8.
 Use a minimum factor of safety
of 1.0. Note: N is the yield acceleration Ky
41
41

Bray and Travasarou (2007)
 Two calculations possible

 Probability of zero displacement
 Estimated displacements (deterministic or probabilistic based on a given probability of exceedance)
 Required parameters:
 Shear wave velocity, vs
 Earthquake ground motion (as a spectral acceleration at a specified period)
 Magnitude (Mw) of the design earthquake (Moment Magnitude)
 Yield coefficient (ky) of the slope
– ky was assumed constant – meaning the earth material do not undergo severe
strength loss as a result of earthquake shaking (e.g. no liquefaction)
42
42
Bray and Travasarou (2007) Equations
Probability of zero displacement
P (D = 0) = 1 – Ф (-1.76 – 3.22 ln (ky) – 0.484 Ts ln (ky) + 3.52 ln (Sa(1.5Ts)))
Displacement estimate
ln(D) = -1.10 – 2.83 ln(ky) – 0.333 (ln(ky))2 + 0.566 ln(ky)ln(Sa(1.5Ts)) +3.04

ln(Sa(1.5Ts)) – 0.244(ln(Sa(1.5Ts)))2 + 1.50Ts + 0.278(M-7) ± ɛ Eq. (5)
ɛ – normally distributed random variable with zero mean and standard deviation σ=0.66.
ɛ =0 for the median estimate of D (i.e. deterministic calculation).
D – seismic displacement (cm)
Ф – standard normal cumulative distribution
function (i.e., NORMSDIST in Excel)
Sa (x) – spectral acceleration at period x (in g units)
M – Earthquake magnitude
ky – Yield coefficient (pseudo-static seismic
coefficient for FOS=1)
Ts – Refer to the figure
A calculation spreadsheet
available for download from
www.geoengineer.org
43
43

Stress-Deformation Modeling
 Finite-element or finite difference

 Models response due to full acceleration-time record
 Earthquake-time history is applied at the base of the model
 Can incorporate liquefaction strength loss into the shaking model
 Can predict zones that will liquefy
 These zones can then be modeled in Slope/W to run post liquefaction stability analysis
 Can provide seismic displacements
 1D – PROSHAKE, FLAC
 2D – FLAC, QUAKE/W
 Predictions are only as good as the quality of input data
Inputs: Outputs:
Earthquake-time history • Identify if any zones would liquefy
Soil properties (dynamic and static) • Seismic Displacements
44
44
Simplified Procedure for Evaluating Liquefaction
Potential – Cohesionless Soils Example
45 IITK-GSDMA GUIDELINES for SEISMIC DESIGN of EARTH DAMS AND EMBANKMENTS, Provisions with Commentary and Explanatory Examples
45

46
47

48
49

50
51
Summary
 All seismic analysis techniques are approximate
 Start with simple methods… and then proceed to detailed only if required
 The key question required answering prior to selecting the analysis method:
 Will soils liquefy?
 Pay attention to the seismic coefficient in a pseudo-static stability analysis
 Seismic deformations calculations appear to be gaining popularity over pseudo static

stability analyses
52
52
Evaluation of Liquefaction Potential
Tezera Firew Azmatch, Ph.D., P.Eng.

Addis Ababa Science and Technology Institute
February 14, 2020

53
53
53
Evaluation of Liquefaction Potential
Tezera Firew Azmatch, Ph.D., P.Eng.

Addis Ababa Science and Technology Institute
February 14, 2020

54
54
54
Pells & Fell (2003) empirical method for estimating settlement, damage &
cracking
 Pells and Fell (2002, 2003) gathered data from 305 dams, 95 of which
reported cracking, and classified these for damage according to the
system shown in Table 12.7.
 Figures 12.46 and 12.47 show plots of damage contours versus
earthquake magnitude and peak ground acceleration for earthfill and earth
and rockfill dams with respect to crest settlement and cracking.
 These figures can be used to estimate the magnitude of crest settlement based
on the range of relative crest settlements given for each damage class given in
Table 12.7.
55
55

cracking
56
56
cracking
57
57

cracking
 In terms of the potential for earthquake induced cracking, Pells and Fell
(2002, 2003) concluded:
 Earthquakes cause settlement, lateral spreading and cracking of
embankment dams. Slope instability may occur but it is not common.
Longitudinal cracks are more common than transverse cracks and are
mostly in the upper part of the dam, more likely on the upstream face than
the downstream.
 There are a number of mechanisms that can lead to formation of either of
these types of cracks. Many of these mechanisms are common to both
seismic loading and normal operating loading. There are seen to be more
mechanisms that may lead to the formation of longitudinal cracks that
transverse cracks, largely because lateral displacement is more readily
achieved in the upstream-downstream directions.
58
58
cracking
 For visible longitudinal cracks to occur the dam needs to experience a
Magnitude 6.5 or greater earthquake, and a peak ground acceleration
greater than about 0.15 g for earthfill dams and 0.3 g for earth and rockfill
dams. Alternatively it needs an earthquake of Magnitude 7.0 or greater,
and a PGA of 0.05 g for earthfill and, say, 0.15 g for earth and rockfill
dams. For hydraulic fill dams, visible cracking may occur for Magnitude 6
earthquakes and 0.05 PGA.
 Dams which experience damage of Class 2 (relative settlement of 0.2% to
0.5% and/or longitudinal cracks 30–80mm wide) or greater are highly likely
to experience transverse cracking. However, transverse cracking has been
observed at under M6.0 to M6.5 earthquakes, at PGA as low as 0.1 to
0.15 g.
 There is evidence to show that higher seismic loads will result in the
formation of larger and deeper cracks and are more likely to cause
transverse cracking due to the greater differential settlements across the
valley.
59
59

cracking
 At low seismic loading, there is evidence to show that only one type of
cracking is likely to develop. This may be either longitudinal (more
common) or transverse, depending on factors within the embankment that
make it predisposed to a particular form. It is the nature of the
embankment, zoning and the foundation geometry and presence of
compressible materials and not the seismic loading, that differentiate
between which form of cracking develops and in particular determine
whether transverse cracking occurs at low seismic intensities.
 The susceptibility of an embankment to a particular type of crack could not
be related to dam type, but a weak relationship has been developed with
dam shape – dams in steeper
60
60
Some Useful References
 Jibson, R. B. (2011), “Methods for assessing the stability of slopes during earthquakes – a
retrospective”, Engineering Geology, v. 122, p. 43-50.
 ICOLD (2011), Improving Tailings Dam Safety, Technical Bulletin 139.
 Duncan, J. M. and Wright, S. G. (2005), Soil Strength and Slope Stability, Wiley, 297 pp.
 Hynes-Griffin, M.E. and Franklin, A.G. (1984), Rationalizing the Seismic Coefficient Method, US
Army Corps of Engineers Waterways Experiment Station, Miscellaneous Paper GL-84-13, 37 pp.
 Bray, J. D. (2007), “Chapter 14: Simplified Seismic Slope Displacement Procedures”, Earthquake
Geotechnical Engineering, 4th International Conference on Earthquake Geotechnical Engineering –
Invited Lectures in Geotechnical, Geological and Earthquake Engineering Series, Pitilakis, K. D.,
Ed., Springer, Vol. 6, 327-353.
 Bray, J. D., “Simplified Seismic Displacement Procedures”, a presentation available at
http://www.ucqc.canterbury.ac.nz/Resources/Bray-UCanterbury-Seismic%20Displ.pdf.
 Bray, J. D. and Travasarou, T. (2007), Simplified procedure for estimating earthquake-induced
deviatoric slope displacements, Journal of Geotechnical and Geoenvironmental Engineering, ASCE,
133: 381-392.
 Bray, J. D. and Travasarou, T. (2009), Pseudostatic coefficient for use in simplified seismic slope
stability evaluation, Journal of Geotechnical and Geoenvironmental Engineering, ASCE, 135: 1336-
1340.
 Rathje, E.M. and Antonakos, G. (2011), A unified model for predicting earthquake-induced sliding
displacements of rigid and flexible slopes, Engineering Geology, 122:51-60.
61
61

Lecture 6 - Seismic Performance of Slopes STU

Uploaded by

Copyright:

Available Formats

Lecture 6 - Seismic Performance of Slopes STU

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lecture 6 - Seismic Performance of Slopes STU

Uploaded by

Copyright:

Available Formats

Evaluating Seismic Performance of

Tezera Firew Azmatch, PhD, PEng

 Earthquakes affect the stability of slopes in two ways:

Steps in Seismic Slope Evaluation

 Determine the design earthquake

 Determine liquefaction susceptibility

 Carry out an appropriate analysis

Determine Design Earthquake Ground

Is soil expected to liquefy under the

No (i.e. strength loss due to

Pseudo Static Stability If stability acceptable

Ref: ICOLD (2011), Duncan and Wright (2005) 9

Quantifying Earthquake Ground Motion

 Acceleration-time history provides a complete

 Simplified parameters that define components

 Spectral acceleration (Sa)

Acceleration Response Spectrum

 A seismic hazard evaluation allows estimation of design earthquakes possible for

CDA Dam Classification Scheme

Local Site Classification

Selecting the Design Earthquake

Link for global seismic

Selecting the Design Earthquake

Determine Design Earthquake Ground

Is soil expected to liquefy under the

No (i.e. strength loss due to cyclic loading is < 20%)

Ref: ICOLD (2011), Duncan and Wright (2005) 19

 Two types of liquefaction:

 Statically induced liquefaction (not the focus of this presentation)

 Seismically induced liquefaction

 Compare N1,60 or qc,1 against case

Seismic Liquefaction Susceptibility

 Cyclic Resistance Ratio (CRR) is the seismic shear

Ref: Seed et. al. (2003)

Determine Design Earthquake Ground

Is soil expected to liquefy under the

No (i.e. strength loss due to cyclic loading is < 20%)

Ref: ICOLD (2011), Duncan and Wright (2005) 23

Pseudo Static Stability Analysis

 One of the earliest procedures of analysis for seismic stability

 Limited to soils that do not lose strength by more than 20%

 Earthquake loading is represented by a static force

 Can be easily performed using computer programs such as

 The critical component is the selection of the seismic

 Earlier days k = 0.05 to 0.25 (empirical)

– k = f(allowable displacements, earthquake magnitude and

 Pseudo-static stability analysis is mainly for screening

Pseudo Static Stability Analysis

Ref: Duncan, J.M. and Wright, S.G. (2005) 26

Determine Design Earthquake Ground

Is soil expected to liquefy under the

No (i.e. strength loss due to cyclic loading is < 20%)

Ref: ICOLD (2011), Duncan and Wright (2005) 27

Post Earthquake Stability Analysis

For soils in which liquefaction is expected to occur:

For soils in which liquefaction is expected to occur:

 Shear strengths of soils could be

Post Earthquake Stability Analysis

Liquefaction does not occur:

 Saturated soils which are potentially