You are on page 1of 8

Shear Strength Reduction Approach for Slope Stability Analyses

Mark S. Diederichs

Geo-Engineering Centre at Queens University and Royal Military College, Kingston, Ontario

Matt Lato

Department of Geological Sciences and Geological Engineering, Queens University, Kingston, Ontario.

Reginald Hammah

Rocscience Inc. Toronto, Ontario

Pete Quinn

Geo-Engineering Centre at Queens University and Royal Military College, Kingston, Ontario

ABSTRACT: This paper is intended to illustrate the applicability of Shear Strength Reduction (SSR) as a
general technique for obtaining Factor of Safety estimates for slopes in variable geology, progressive or locally brittle yield behaviour and with ground-structure interaction. Comparisons are made with Limit Equilibrium solutions (LEM). Implications of the assumptions required to ensure this correlation are discussed including uniform stiffness, rigid-plastic behaviour, instantaneous interaction between geological units, and
instantaneous generation of support loads. The paper uses Finite Element Modelling (FEM) as a vehicle for
demonstration although Finite Difference solutions are equally valid. General applicability of the method is
demonstrated and the limitations explored using Discrete Element simulation of multi-block slope failure.
1 INTRODUCTION
The Shear Strength Reduction (SSR) technique (Matsui & San 1992, Dawson et al 1999; Griffiths &
Lane 1999; Cala & Flisiak 2003, Hammah et al
2005a, 2006) enables finite element (or finite difference) techniques to be used to calculate factors of
safety for slopes, providing an alternative to limit
equilibrium calculations and a potentially more reliable analysis of slopes with heterogeneous stiffness,
strain-softening and passive structure-ground interaction. The methodology is general and can be applied to other non-linear problems such as multiblock discrete element simulations.
Geotechnical engineers primarily conduct slope
design based on calculated factor of safety values.
Limit Equilibrium (LEM) techniques that compare
resisting forces to driving forces (or moments) are
ideally suited to the generation of a nominal safety
factor (Krahn 2003). Instantaneous slide surface
mobilization and consideration of stresses and forces
independent of pre-failure movement are inherent in
these techniques, and may result in inadequate representation of the systems actual stability state.
Non-linear modelling using the Shear Strength
Reduction (SSR) technique can also used to determine factors of safety. The approach offers a number
of advantages over LEM (Griffiths & Lane, 1999)
including the elimination of a priori assumptions on
the shape and location of failure surfaces, the elimination of assumptions regarding the inclinations and
locations of interslice forces, the capability to model
progressive failure, the calculation of deformations

and the incorporation of displacement controlled


ground-structure interaction. The general approach
is valid for a wide range of applications.
The automated inclusion into specific analysis
tools mandates the need for some accepted stability
criteria. Maximum unbalanced force tolerance,
maximum displacement limits or other criteria have
been incorporated into commercially available codes
although these are not fundamentally related to the
methodology of SSR. Examples of alternative stability indicators are proposed in this paper. Stability
indicators can be tailored for the problem at hand
through manual application of the SSR approach.
Some key issues, nevertheless, require further
resolution including the questions: should factored
shear strength parameters represent the dominant
variables in the definition of factor of safety for
more complex failure modes (shear plus toppling,
extension, etc); should cohesion and friction be subjected to the same reduction factors; should all capacity related parameters, including support, be included in the SSR process and should support be
factored by the same Strength Reduction Factor as
the soil/rock shear strength ?
Duncan (1996) provides some insight into the latter issue although more discussion is needed on this
and other details of SSR application.
1.1 Limit Equilibrium Analysis
The general limit equilibrium approach developed
by Fredlund and Krahn (1977) calculates a composite factor of safety, FS, for moment equilibrium as:

FS M =

(c' R + ( N u )R tan ')


(Wx Nf ) Dd

(1)

and for force equilibrium as:


(c' cos + ( N u ) tan ' cos )
FS M =
( N sin ) D cos

(2)

where the slice base normal force is:


c' sin + u sin tan '
W + (X R X L )
FS
(3)
N=
sin tan '
cos +
FS
where u is the pore pressure, W is the slice weight, D
is an applied line load, is the base inclination and
, R, x, f, d and are geometric parameters.
The significant assumptions in this formulation
concern the inter-slice forces (shear and normal).
Early techniques (Fellenius 1936) ignored these
forces or made significant simplifying assumptions
(Janbu 1954, Bishop 1955). Morgenstern and Price
(1965) and Spencer (1967) proposed a solution for
this problem used in the general formulation by
Fredlund and Krahn (1977). Krahn (2003) summarizes and compares a number of approaches for
LEM analysis. These differ primarily in the individual or combined consideration of moment or force
equilibrium and the consideration of inter-slice normal and shear forces and the assumed or calculated
inclination of resultant inter-slice forces as illustrated in Figure 1 and summarized in Table 1.

Figure 1: Slice and interslice forces for a composite limit equilibrium approach to slope stability.

The basal stress state in the original LEM formulations are ultimately a function of gravitational
loading on the individual slices. This does not consider the influence of topographical and stiffness

based variations in the underground stress state.


A number of authors, including Zou et al. (1995)
and Krahn (2003) have proposed the use of linear
elastic finite element analysis to provide the initial
stress map for a subsequent LEM calculation and the
integration of slice forces. Stianson et al. (2004) investigate the use of non-linear plastic analysis (for
stable problems) to compute the basal stress state
and slice forces. These methods collectively represent an important advance over conventional techniques. The use of elastic FEM stress calculations
takes into account the stress variability due to stiffness contrasts and topography (as in GeoSlope 2001,
for example). The use of plastic analysis to compute
stresses for an LEM calculation may account for
progressive yield and stress redistribution although
the additional complexity of the integrated solution
procedure is of dubious added value.
Another assumption involves the consideration of
line loads in the formulation. Support loading can be
considered by adjusting resisting forces or driving
forces within the factor of safety equation. One approach to handle this problem is to include so-called
active support (e.g. pre-tensioned anchors) as a reduction to the driving forces while passive anchors
(e.g. grouted dowels) are incorporated as an increase
in resisting forces. This selection, while logical, is
somewhat arbitrary and this challenge is discussed,
without the provision of a unique solution, by Krahn
(2003). The displacement dependency of resisting
loads provided by some passive support systems are
not considered in this formulation.
Finally, LEM techniques must be performed on an
assumed continuous slip surface. This surface was
originally circular in early formulations but most
techniques after Fellenius can be applied to noncircular and very general surfaces (Figure 2). The
surface that generates the minimum factor of safety
is located using a grid or optimized search (center
and radius), block search (relocating vertices of an
arbitrary slip polygon) or other statistical techniques. For simple geometries, modern search
methods yield robust results although for problems
including extreme geometries (very thin weak layers, for example), search techniques can often miss
the true global minimum (FS) geometry.

Table 1: Comparison of equilibrium considerations and interslice force assumptions (Krahn 2003)
Method
Fellenius
Bishop Simplified
Janbu Simplified
Spencer
Morgenstern-Price
Corps of Engin. 1
Corps of Engin. 2
Lowe-Karafiath

Moment
Equilibrium

Horiz. Force
Equilibrium

Interslice
Normal Force

Interslice
Shear Force

YES
YES
NO
YES
YES
NO
NO
NO

NO
NO
YES
YES
YES
YES
YES
YES

NO
YES
YES
YES
YES
YES
YES
YES

NO
NO
NO
YES
YES
YES
YES
YES

Inclination of
Interslice Force
No Force
Horizontal
Horizontal
Constant
Variable
= Crest to Toe Average Dip
= Slice Ground Surface Dip
= Average of Surface and Base Dip

loadings, or shear strain localization (as in Matsui &


San 1992) as desired. Figure 4 illustrates the convergence approach for the same slope in Figure 2.

Figure 2: a) LEM grid search using SLIDE (RocScience


2005a) for circular slip surface and b) block search (optimized
vertices). FS<2 surfaces for slope with weak layer below toe.

1.2 Shear Strength Reduction (SSR) Techniques


The SSR technique involves the use of factored
strength parameters in a non-linear numerical model.
The shear strength properties of the model material
are scaled until the stability limit is reached. The ratio between the actual strength and the model
strength at the stability limit is the Factor of Safety.
The continuum approach to SSR considers a discretized zone in a finite element or finite difference
continuum model and uses predefined displacement
limits at points of interest or using model convergence as an indicator of equilibrium. For similar
element shape functions, the set of algebraic equations solved in finite element analysis is identical to
the finite difference method (Dawson et al 1999).
In the case of FLAC (Itasca 2005) the set of equations is solved using dynamic relaxation. Equilibrium is assumed when the unbalanced forces reach a
minimum tolerance. Figure 3 illustrates an example
of the SSR solution procedure (as in Cala & Flisiak
2003, Lorig & Varona 2004) using unbalanced force
as an indicator of slope stability. Alternatively, critical displacements could be used as a stability limit.

Figure 3: Finite difference analysis result using total unbalanced force magnitude as a stability indicator. FS=1.025 for
this slope (after Dawson et al. 1999).

SSR analyses in FEM codes such as PHASE2


(RocScience 2005b) conventionally rely on overall
model solution convergence as an indicator of stability for automated determination of SSR. Manual iteration about a critical SSR can also involve consideration of critical slope displacements, support

Figure 4: SSR solution using FEM (PHASE2) for the same


slope in Figure 2. Maximum displacement is plotted two
right most points represent non-converging models. Notice that
the critical surface geometry defined in Figure 2.b. is naturally
defined in the SSR analysis without a priori specification and
that the FS (=SRF) is similar (1.49 above vs 1.46 for Figure 2).

For Mohr Coulomb strength parameters the modification is as follows (Dawson et al. 1999):
1
c'trial =
c'measured
(4)
FStrial

1
tan

FStrial

'trial = tan 1

(5)

'(i )
'(i )
max = ctrial
+ n tan trial
for the ith material

(6)

where, for SSR analyses, the same strength reduction factor (SRF = FStrial) is used for all components of strength and for all materials within the stability problem. The SSR approach for the HoekBrown criterion (Hoek et al. 2002) involves a more
involved but philosophically similar factoring process described by Hammah et al (2005b).
This factoring of strength parameters allows reinforcement and other external effects to be modeled
without modification to determine stability. This is
analogous to the factored load approach for returning structural load demand from ground behaviour.
The support capacities can have a different factor of
safety is desired (as suggested by Duncan 1996), although there is some merit in considering reinforcement within the SSR factoring process as well.
The approach allows for the reinforcement and the
rockmass itself (if heterogenous) to develop modified internal loads and stresses as a function of preyield displacements and strains. Most importantly,
the critical failure surface develops naturally during
the course of the non-linear solution and does not
need to be pre-defined or determined through optimization algorithms. Brittle or strain softening behaviour and dilation effects are also accommodated.

2 SSR EXAMPLES USING FEM ANALYSIS


This section describes four slope examples, which
were employed to verify the SSR methods performance. Some of the examples were taken from published papers that report factors of safety calculated
with limit equilibrium methods. The slope examples
were also computed with a commercial limit equilibrium program (RocScience 2005a) for the purposes
of obtaining failure mechanisms (slip surfaces) in
addition to factors of safety. The performance of the
SSR method can be verified by comparing its factor
of safety results and predicted shear failure mechanism to the limit equilibrium answers.
2.1 Example 1: Simple Homogeneous Slope
The slope example (Hammah et al. 2005) shown in
Figure 5 comprises a simple, homogeneous slope in
material that has Mohr-Coulomb strength.

Figure 6. The Bishop critical failure surface superimposed on


the contours of maximum shear strain of Slope 1 at failure (as
predicted by the SSR finite element method).

The slope, which is shown on Figure 7, includes a


thin seam of weak material. The Mohr-Coulomb
properties of the materials are given in Table 3.

Soil 2

Figure 7. Geometry of slope in Example 2.


Table 3. Strength parameters of the materials in Example 2.
Figure 5. Geometry of slope in Example 1.

The cohesion of the material is 10.5 kPa and the


friction angle equal to 35 degrees. Bishops method
of limit equilibrium analysis produced a factor of
safety of 1.21. Table 2 provides SSR factors of
safety for different assumed values of Youngs
modulus and Poissons ratio for the slope material.
The results compare very well to the LEM answer.
Table 2: Factors of Safety for Example 1 for Different Assumptions on Deformation Parameters (Hammah et al. 2005)
No.
1
2
3
4
5
6

Homogeneoeus Slope Example


E (kPa)

SRF=FS
2,000
2,000
20,000
20,000
200,000
200,000

0.2
0.48
0.2
0.48
0.2
0.48

1.210
1.180
1.210
1.180
1.210
1.180

c (kN/m2)

(deg.)

(kN/m3)

28.5
0

20.0
10.0

18.84
18.84

Soil #1
Soil #2

Table 4 outlines the factors of safety calculated


by the SSR method and various limit equilibrium
approaches. Figure 8 compares the Spencer noncircular failure surface against that (as described by
contours of maximum shear strain at failure) predicted by the SSR method. The SSR results agree
very well with limit equilibrium predictions.
Table 4. Factor of safety results for Example 2.
Method
Factor of Safety
SSR (FS=SRF)
Spencer
GLE
Janbu Corrected
Published results

1.270
1.258
1.246
1.275
1.240 1.270

Figure 6 superimposes the critical failure surface


obtained from Bishops method on the contours of
maximum shear strain at failure in the slope produced from the SSR finite element method.
2.2 Example 2: Slope with thin weak seam
Example 2 deals with a verification problem described as Example 3a in the five basic slope stability problems distributed in a survey sponsored by
the Australian Association for Computer Aided Design and summarized by Giam & Donald (1989).

Figure 8. The Spencer critical failure surface superimposed on


the contours of maximum shear strain of Slope 2 at failure.

2.3 Example 3: Slope with pore water pressures


The slope in Example 3 is the fifth example in the
ACADS set of slope problems (Giam & Donald
1989). The geometry is shown on Figure 9, and the
soil properties given in Table 5.
Table 5. Strength parameters of the materials in Example 3.
c (kN/m2)

(deg.)

(kN/m3)

11.0

28.0

20.00

This slope was excavated at a slope of 1:2 below


an initially horizontal ground surface. The pore
pressure grid shown on Figure 10 was derived from
the flow net provided in the ACADS publication.
Table 6. Factor of safety results for Example 3.
Method

Factor of Safety

SSR (FS=SRF)
Bishop
Spencer
GLE
Janbu Corrected
Published results

1.475
1.498
1.501
1.500
1.457
1.530

Figure 9. Geometry of slope in Example 3, including pore


pressure grid and ponded water.

Table 6 and Figure 10 show that the results of the


SSR method are in agreement with those determined
from limit equilibrium analysis methods.

Figure 11: LEM results for embankment with no support and


with passive and active support. Minimum FS values are
given for selected solution methods.

The support system consists of 4 frictional dowels


spaced 2m in plane and 1.5 m out of plane. In limit
equilibrium, the full capacity of the dowels (100kN)
is assumed to act upon the sliding mass. The bolts
contribute to the normal pressure on the slice base
(increasing frictional resistance). The tangential
component (parallel to the base) can act in two
ways. SLIDE (RocScience 2005a) imposes active
support as a reduction of the driving forces while
passive support acts to increase the resisting force.
This determination brackets the range of support action such that active support gives a slightly higher
FS than passive a realistic ballpark assumption.
It is more significant to note that passive reinforcement is assumed to act with full nominal capacity in the calculation where in reality the development of resistance is displacement or strain
controlled. The unsupported minimum FS surfaces
as well as the results of supported analysis are
shown in Figure 11.
The FEM-SSR results are shown in Figure 12. In
this analysis, the dowels are modelled as frictional
steel supports (19mm diameter and 200GPa stiffness) with the same limiting interface shear strength
(50kN/m) as in the LEM model. Figure 12 shows
that the unsupported slip surface and factor of safety
corresponds well to the LEM solution.

Figure 10. The Spencer critical failure surface superimposed


on the contours of maximum shear strain of Slope 3 at failure.

2.4 Example 4: Passive Support


The example in Figure 11 is an 25 m high embankment on a strong base. The upper soil has a cohesion
of 50 kPa and a friction angle of 30 degrees.

Figure 12: FEM-SSR analysis with and without support.


Lower right image also shows relative bolt loads (axial) for the
SRF being considered. Bolt loads are not factored.

In the supported case, however, the results differ


in terms of the location of the minimum FS ( =
SRF), although the Factor of Safety for this new
failure zone falls within the range calculated for this
zone in LEM (FS=1.5 to 1.6). The main difference is
that the LEM still locates the minimum surface
(FS=1.37-1.46) through the support as in Figure 11.
In addition, the lower right plot in Figure 12 illustrates that the generation of axial load in the supports is not constant and does not reach the maximum capacity. The axial (and shear) load develops
as the slope deforms. This is more physically accurate than the assumption of static full capacity (or
even factored nominal capacity) in LEM analysis.
In this analysis, however, the capacity of the bolts
are not factored along with soil shear strength in the
SSR process. There is, at present, no consensus on
whether they should be (i.e. apply the same SSR to
bolt capacity) or how the actual loads in the bolts
should be compared with the bolt capacities to give
an FS for the bolts. This is a discussion that is
needed in order to adopt a standard approach for
supported slopes. It is possible to use the SSR approach in a different way and discretely select limiting SRF (for short and long term stability for example) according to accepted practice. This would be
compatible with the factored load approach for
merging structural and geotechnical engineering.
The bolts could then be scaled to provide stability
for the strength-reduced slip surface within a structural (steel code) safety margin. This is just one suggestion. It is clear, however, that more discussion is
warranted on the incorporation of support in SSR
models. Accepting that SSR is mechanically correct,
the accounting process for support needs to be further discussed in the learned community.
2.5 Example 5: Stiffness and Strain Softening
The examples in the previous section were selected
to demonstrate the compatibility of LEM and FEMSSR techniques for simple slope problems. In each
case the conditions are fixed and the units are of uniform and high stiffness (the correct approach to obtain direct comparison between the two methods).
The following example explores the effect of multiple materials, stiffness, strain softening and sensitivities to convergence tolerance and mesh density.
The layer properties (according to Hoek et al. 2002)
for the example in Figure 13 are given in Table 7.

Figure 13 shows the minimum FS calculated for


three selected LEM solutions using peak and residual strength parameters for each layer. In this example, the residual strength is specified by reducing the
normalizing UCS. The slope crest is 205m above the
lake level.

Figure 13: LEM analysis (FS values shown) for a stratified


rock slope. Left analysis uses peak strength from Table 7;
Right analysis uses resid strength for the Sandstone.

Figure 14 compares the FEM-SSR results for


models with elastic-plastic response (using peak
parameters from Table 7) using identical elastic
properties and variable (true) elastic properties. The
effect of stiffness contrasts in this case is not particularly significant although there is no reason not
to include stiffness variation in an FEM-SSR analysis. The results in either case compare best to the
non-circular Bishop analysis in Figure 13.

Figure 14: FEM-SSR analysis of slope in Fig. 13 and Table 7.


Shear strain contours and displacement vectors (x25) plotted.

Figure 15 shows the SSR results for an analysis


with strain weakening (brittle) behaviour prescribed
for the sandstone. Other lithologies are plastic. SSR
is applied to both peak and residual parameters in
this case.

Table 7. Strength parameters of the materials in Example 5.


UCS
MPa

UCS
resid

mi

GSI

mb

Erm
GPa

SSt
60
12
21 70
7.2
0.036
16.5
Sh1
15
15
8
60
1.9
0.012
3.0
Silt
25
25
10 65
2.9
0.020
9.4
Sh2
10
10
7
50
1.2
0.004
2.0
SSt =Sandstone, Silt =Siltstone, Sh1=shale, Sh2 =weak shale

Figure 15: FEM-SSR analysis for Example 5 with strain weakening behaviour for the sandstone (zero dilation). Water table
is solid line (1). Shear strain (contours) and displacement vectors (x25) are plotted. Note bimodal failure (2 blocks).

Interestingly, the FS (=SRF=1.85) for this analysis is within the range of the LEM analyses for residual parameters. In addition, close examination of
Figure 15 shows that both the non-circular, deepseated slip surface as well as the shallower circular
surface suggested by alternate LEM analyses (right
side of Figure 13) are both simulated in the FEM
analysis (two distinct zones of elevated shear strain).
The results shown are for zero-dilation. Failure geometry is similar for dilation (mdil = 0.25mb) with an
elevated FS=SRF=1.92. Table 8 summarizes the impact of mesh density and convergence tolerance.
Beyond a reasonable minimum element density, the
results are not very sensitive to mesh configuration
although poor results are obtained with an elevated
tolerance or the use of 3-noded triangular elements
(as opposed to 6-noded linear strain elements).
Table 8: Effects of tolerance, maximum iterations per cycle
and mesh density for Example 5 with brittle sandstone (Figure
15) and zero dilation.
Mesh
Tolerance Max Iterations
FS=SRF
(*6 nodes)
*Standard
0.001
500
1.85
*Standard
0.01
250
2.05
*Standard
0.0001
1000
1.82
*Coarse (x 0.5)
0.001
500
1.86
*Fine (x 2)
0.001
500
1.82
3 node (stand)
0.001
500
2.03

3 DISCRETE / DISCONTINUOUS MODELS


The application of SSR is not restricted to continuum models such as finite element or finite difference tools. This section presents two examples of
application in discontinuum analysis.
3.1 SSR in Discrete Element Analysis
While the equilibrium calculation is based on a
group of slices, limit equilibrium techniques still
consider the behaviour of a contiguous block of rock
or soil in a state of limiting sliding instability. In
multi-block applications, the development of several
unstable blocks and limited movement does not
immediately indicate failure in an operational sense.
The example in Figure 16 illustrates the use of SSR
for a blocky rockmass. The model is created using
3DEC and the characterization technique described
by Kalenchuk et al. (2006). A vertical slope is used
for this simple example. In this model the actual
joint shear strength properties are specified (cohesoin = 90 kPa, friction angle = 56 deg) although the
initial conditions could be variable and joint specific. The boundary conditions shown are unrealistic
and for demonstration only. Failure must be defined
and could be based on critical displacement or, as in
this case, a limit for released blocks or cumulative
released block volume as shown in Figure 17.

Figure 16: Schematic 3DEC block models used in Figure 17.

Figure 17: Results of SSR analysis for composite model in Figure 16. Vertical lines bound the range of FS (=SRF) based on
two possible failure criteria (see axes). Each point is averaged
over several stochastic simulations.

3.2 SSR in Discontinuous Deformation Analysis


One of the fundamental premises of the SSR technique is that the factor of safety for any problem can
be related to the factoring of shear strength for materials and interfaces. Numerous verification examples
illustrate that this gives the same result as an overall
force and moment comparison for sliding problems.
It is less clear whether the SSR technique is valid for
problems that do not have a single sliding surface
and which have internal separation and/or vorticity.
The example in Figure 18 demonstrates the application of SSR to a simple toppling problem.

Figure 18: Example of SSR applied to toppling analysis using


Discontinuous Deformation Analysis. Critical SRF for interblock friction and cohesion for this analysis is 1.24. The nominal friction angle is 30 degrees and the cohesion is 5kPa.

This analysis employs discontinuous deformation


analysis (Sitar et al 2001) assuming that cohesion
and friction are the only variables to control stability. The implementation appears simple and reliable
enough and reasonable answers can be obtained, although it is important to debate the universality of
relating stability in complex situations (with sliding,
extension and rotation) to shear strength alone. In
addition to the necessary factoring of tensile
strength, crucial in rock mechanics problems, the
SSR technique may not adequately capture geometric factors such as block aspect ratio and rotational
mechanics not related to shear strength. The method
is attractive for non-standard applications but the
mechanics of each case should be compatible.
4 CONCLUSIONS
For problems dominated by sliding, the SSR technique has been clearly demonstrated as a valid alternative to Limit Equilibrium Methods. The technique
can be used in non-linear finite element or finite difference continuum models or in discontinuum analysis and is not unreasonably sensitive to model setup
(although mesh density, solution tolerance and shape
function sensitivity must always be checked).
The greatest advantage to the SSR technique is
the lack of a requirement to discretely or iteratively
pre-define a sliding surface for consideration although modern LEM techniques have sophisticated
search algorithms for non-circular surfaces.
In sliding models with no tension crack or with no
support, the factor of safety can be directly and
uniquely related to the factoring of shear strength for
materials and interfaces giving similar results to an
overall force and moment comparison (LEM).
If tension is involved then tensile strength must
also be included in the SSR process for the geomaterial. It may be necessary to debate the validity of
equal factoring of cohesion and friction in the SSR
solution. Cohesion is often specifed with much less
confidence than frictional properties and yet for
shallow problems, dominates over frictional effects.
For simple cases the LEM (force and moment)
definition of FS coincides with the SSR (FS=critical
SRF) definition. The definitions deviate for more
complex geologies and material behaviour. SSR
may perform better in such cases. Caution is needed
though, as shear strength may not be the dominant
parametric control on stability in many cases.
It is also not clear how shear strength reduction
should be standardized in cases where support is involved. More discussion is needed to decide whether
support strength be fixed during the SSR iterations
or included in the over all strength reduction.
REFERENCES
Bishop AW. 1955. The use of the slip circle for the stability
analysis of slopes. Geotechnique. 5(1) 7-17.

Cala M, & Flisiak J. 2003. Complex geology slope stability


analysis by shear strength reduction. Proc. of the 3rd Int.
FLAC Symp., Sudbury. Lisse: Balkema. 99-102.
Dawson EM, Roth WH, Dresher A. 1999. Slope stability
analysis by strength reduction. Geotechnique, 49(6). 835840.
Duncan, J.M. 1996. State of the art: limit equilibrium and finite-element analysis of slopes, Journal of Geotechnical
Engineering, 122(7). 577-596.
Fellenius W. 1936. Calculation of the stability of earth dams.
2nd Congress on Large Dams. Washington. V4. p445.
Fredlund DG & Krahn J. 1977. Comparison of slope stability
methods of analysis. Can. Geotech. J. v14. 429-439.
Geo-Slope. 2001. SLOPEW v5. Slope stability analysis.
Giam PSK & Donald IB. 1989. Example problems for testing
soil slope stability programs. Civil Engineering Research
Report No. 8/1989. Monash University.
Griffiths DV, Lane PA. 1999. Slope stability analysis by finite
elements. Geotechnique. 49(3). 387-403
Hammah RE, Yacoub TE, & Curran JH. 2006. Investigating
the performance of the shear strength reduction (SSR)
method on the analysis of reinforced slopes. In Proceedings
of the 59th Canadian Geotechnical Conference, Vancouver.
Hammah RE, Yacoub TE, Corkum B & Curran JH. 2005a. A
comparison of finite element slope stability analysis with
conventional limit-equilibrium investigation. In Proceedings of the 58th Can. Geotech. Conference, Saskatoon.
Hammah RE, Yacoub TE, Corku, B & Curran J. 2005b. The
Shear Strength Reduction Method for the Generalized
Hoek-Brown Criterion. Alaska Rocks: 40th U.S. Symposium on Rock Mechanics. Paper 05-810. 6pgs.
Hoek E, Carranza-Torres CT, Corkum B. 2002. Hoek-Brown
failure criterion-2002 edition. In: 5th North American rock
mechanics symposium, Toronto, Canada, vol. 1. 26773.
Itasca. 3DEC v3. 200a 3-Dimensional distinct element code.
Itasca. FLAC/FLAC-SLOPE. 2003b. Fast Lagrangian Analysis
of Continua Slope modelling package.
Janbu N. 1954. Application of composite slip surfaces for stability analysis. Proc. European Conf. on the Stability of
Earth Slopes. V3. p39-43.
Kalenchuk, K. , Diederichs, M.S. and McKinnon, S. 2006.
Characterizing block geometry in jointed rockmasses. Int
Jour of Rock Mech. & Min. Sc.. 43(8) , pg 1212-1225.
Krahn J. 2003. The 2001 R.M. Hardy Lecture: The limits of
limit equilibrium analysis. Can. Geotech. J. v40. 643-660.
Lorig L., & Varona P. 2004. Numerical Analysis. in Rock
Slope Engineering, pp. 218-244, D. C. Wyllie and C. W.
Mah, Eds. London: Spon Press.
Matsui T & San KC. 1992. Finite element slope stability analysis by shear strength reduction technique. Soils and Foundations. 32(1). 59-70.
Morgenstern NR & Price VE. 1965. The analysis of the stability of general slip surfaces. Geotechnique 15(1). 79-93.
Rocscience Inc. 2005a. SLIDE v5 Software for 2D LimitEqulibrium Analysis for soil and rock slopes.
Rocscience Inc. 2005b PHASE2 v6 Elasto-plastic finite element stress analysis software.
Sitar N, MacLaughlin MM, Doolin DM & Abbot T. 2001. Investigation of slope stability kinematics using discontinuous deformation analysis. Int. Jour. of Rock Mech. and
Min. Sci. v. 38, pp. 753-762,
Spencer, E. 1967. A method of analysis of embankments assuming parallel interslice forces. Geotechnique.17(1) 11-26
Stianson JR, Chan D & Fredlund DG. 2004. Comparing slope
stability analysis based on linear elastic or elasto-plastic
stresses using dynamic programming techniques. 57th Can.
Geotech Conf. Session 7C. p23-30.
Zou JZ, Williams DJ and Xiong WL. 1995. Search for critical
slip surfaces based on finite element method. Can. Geotech.
J. 32(2) 233-246.

You might also like