Comparison of 3D Finite Element Slope Stability With 3D Limit Equilibrium Analysis

Comparaison de la stabilit des lments 3D pente finie avec l'analyse limite d'quilibre 3D
Lu H.H., Xu L.M., Fredlund M.D.

SoilVision Systems Ltd., Saskatoon, SK., Canada

Fredlund D.G.

Golder Associates Ltd., Saskatoon, SK., Canada

ABSTRACT: The two-dimensional (2D) limit equilibrium analysis is widely used in geotechnical engineering for slope stability
analysis. Three-dimensional (3D) slope stability analysis is rarely performed although all slope failures are 3D in reality. The 3D
shear strength reduction (SSR) technique using finite element method (FEM) analysis and 3D limit equilibrium method (LEM) based
on columns to predict a factor of safety for slopes have been in existence for decades. Recent software tools allow the improved
analysis of 3D slope stability through LEM and SSR techniques. The purpose of this paper is primarily to compare 3D FEM-SSR
analysis with 3D LEM analysis through the examination of benchmark slope stability analysis examples. The results indicate that
there is reasonable agreement between these two methods.
RSUM : L'analyse d'quilibre limite en deux dimensions (2D) est la mthode d'analyse de stabilit des pentes la plus couramment
utilise en gotechnique. L'analyse de stabilit des pentes en trois dimensions (3D) est rarement effectue bien que tous les
glissements de terrain sont en ralit tridimensionnels. La technique de la rduction de rsistance au cisaillement (SSR) en 3D
utilisant la mthode des lments finis (FEM), ainsi que l'analyse d'quilibre limite (LEM) en 3D base sur des colonnes pour prdire
un facteur de sret pour des pentes, existent depuis des dcennies. Les codes de calcul rcents permettent d'amliorer l'analyse de
stabilit des pentes en 3D l'aide des techniques de LEM et de SSR.
Le but de cet article est principalement de comparer l'analyse de FEM-SSR en 3D avec l'analyse de LEM en 3D travers des
exemples de rfrence d'analyse de stabilit. Les rsultats indiquent qu'il y a un accord raisonnable entre ces deux mthodes.
KEYWORDS: 3D Slope Stability Analysis, Shear Strength Reduction, Limit Equilibrium Method, Finite Element Method.
1

INTRODUCTION

The 2D LEM is widely used in geotechnical engineering for

slope stability analysis. However all slope failures are 3D in
reality. The 2D approach is generally considered to be
conservative in that 3D influences of geometry are not
accounted for in a 2D analysis. Furthermore, the assumption
that 2D analyses lead to conservative factors of safety is correct
only when the critical pessimistic section of the 3D model is
selected for the 2D analyses. It is time consuming to ensure that
the 2D section model is the critical pessimistic 2D section for
some general slopes. The use of 3D slope stability analysis is
important to model real world problems, to make the designs
more economic, and to provide a guide for 2D designs. It is
useful, for example, to know exactly what percentage the 3D
FOS is higher than the 2D analysis. The most common methods
for 3D slope stability analysis are 3D LEM based on columns
and 3D SSR based on FEM analysis.
1.1

3D LEM slope stability analysis

3D LEM slope stability analysis is traditionally based on an

extension of 2D LEM analysis. Many researchers have done
work on 3D LEM analysis (Hovland 1977, Hungr, Zhang 1988,
Salgado and Byrne 1989, Lam and Fredlund 1993, Cheng, etc.
2005). The slicing method in 2D analyses has been extended
into 3D analysis with columns by various authors due to the
popularity of 2D LEM slicing methods. Some of the benefits of
the 2D slicing method include its ability to accommodate
complex geometries, variable soils, water pressure conditions
and different reinforcement systems, etc..
The majority of the 3D LEMs are based on the assumption
that the failure direction is pre-defined in order to derive the
FOS equations, i.e. the failure sliding direction is not part of the
slope stability analysis solution. Location of the critical failure
surface and its direction is a tough global optimization problem.

759

Jiang (1997), Yamagami and Jiang (1997) provided a

optimization-minimization procedure (OMP) for their Dynamic
Programming (DP) (Baker, 1980) and random number
generation technique to find the critical slip surface and
corresponding sliding direction. Cheng and Yip (2003) derived
3D asymmetric slope stability analysis equations based on
extensions of simplified Bishop, simplified Janbu and
Morgenstern-Price methods, and the direction of slide can be
determined from 3D force/moment equilibrium equations. Their
formulation is equivalent to Yamagami and Jiang's OMP.
SoilVision Systems Ltd. (SVS) has incorporated all the
popular 3D LEMs into its commercial 3D slope stability
analysis software - SVSLOPE 3D. Recently SVS has added a
new feature to search for the critical slip surface sliding
direction in 3D, which is similar to Jiang (1997)'s procedure.
This feature enables the modeling of 3D slopes by the LEM at
any angle and is applicable for municipal designs including
calculation of setback distrances as well as the stability of open
pits in the mining industry as well as other applications.
1.2

3D FEM-SSR slope stability analysis

The finite element method (FEM) has been extensively used to

analyze various geotechnical problems. To perform slope
stability analysis with the FEM, the SSR technique dictatates
that the soil shear-strength is gradually reduced until failure
conditions occur. The factor of safety (FOS) for a SSR analysis
is defined as the ratio of the shear strength of the soil to the
shear stress developed along the critical failure surface. This
relationship is presented in the following equations.
where and are the cohesion and angle of internal friction
for the Mohr-Coulomb shear strength parameters. and
are
factored shear strength parameters. SRF is called the strength
reduction factor. In order to reach to the state of limiting

factored shear strength parameters. SRF is called the strength

reduction factor. In order to reach to the state of limiting
equilibrium, the SRF is gradually increased. This means that the
soil shear strength becomes weaker, until it is no longer possible
Proceedings
the 18thtoInternational
Conference
on Soil
for
the FE modelofanalysis
reach convergence.
At this stage,
it Mechanics and Geotechnical Engineering, Paris 2013
can be said that failure of the slope occurs and the FOS equals
the SRF. Non-convergence within a specified number of
Figure 2. Contour of total displacement of 3D FEM-SSR analysis for
iterations and tolerance is an indicator of slope failure because
example 1 at the final stage.
of the absence of force equilibrium (i.e. stress and displacement
distributions
that
satisfy
the
equations
of
equilibrium
cannot
be
equilibrium, the SRF is gradually increased. This means that the
established based on the factored set of shear strength
soil shear
strength becomes weaker, until it is no longer possible
parameters).
for the FE
model analysis to reach convergence. At this stage, it
The FEM-SSR analysis has been shown to be a powerful and
can bea useful
said that
failure of
the slope occurs
thestability
FOS equals
alternative
to conventional
LEMand
slope
analysis
the SRF.
Non-convergence
specified
of
technique,
Griffiths et alwithin
(1999,a 2007),
Wei,number
etc. (2009).
iterations
and tolerance
anhas
indicator
of slope
becauseinto
SoilVision
Systems is
Ltd.
incorporated
the failure
3D FEM-SSR
of theits
absence
of force
equilibrium
(i.e. stress
and displacement
commercial
package
- SVSOLID
3D. Consequently,
the 3D
distributions
thatissatisfy
the equations
be
FEM-SSR
now readily
availableoftoequilibrium
geotechnicalcannot
engineering
established
practice.based on the factored set of shear strength
parameters).
The
analysis has been shown to be a powerful and
2 FEM-SSR
EXAMPLES
a useful alternative to conventional LEM slope stability analysis
Figure 3. Y-section view of the contour of the total
2.1 Example
- 3Dalslope
with 2007),
external Wei,
load etc. (2009).
technique,
Griffiths 1 et
(1999,
Figure 3. Y-section
view of1the contour of the total displacement for
displacement
for example
SoilVision Systems Ltd. has incorporated the 3D FEM-SSR into
example 1
Wei, Cheng and Li (2009) considered a slope with a
its commercial
package
- SVSOLID
3D.
Consequently,
theshow
3D a
rectangle area
of vertical
external
loading
in order to
FEM-SSR
is 3D
nowfailure
readily
available
to geotechnical
distinct
surface.
As shown
in Figure engineering
1, the vertical
practice.
distribution loading length is 8m and width is 2m, while in Wei,
etc. (2009) many different combinations of length and width
2 EXAMPLES
were considered. The edge of the loading is 1m away from the
crest of the slope. The magnitude of the distribution load q is
2.1 equal
Example
- 3DThe
slope
to 1001kPa.
soilwith
shearexternal
strengthload
properties are cohesion
of 20 kPa, an angle of internal friction of 20 degrees and a unit
Wei,
Cheng
(2009)
considered
a slope
with
a
weight
of 20and
kPa.Li Table
1 shows
the results
of the
various
rectangle
area of
vertical
external
ordertheto distinct
show a3D
analyses.
Figure
2 and
Figure loading
3 clearlyinshow
distinct
3D failure
shown in Figure
the vertical
failure
surface surface.
based onAs
a FEM-SSR
analysis.1,Figure
4 shows
distribution
loading
8m and
is 2m,analysis.
while inItWei,
the critical
sliplength
surfaceisbased
on width
a 3D LEM
can be
etc. (2009)
many
combinations
length
seen that
bothdifferent
the failure
slip surfaceofshape
andand
FOSwidth
values
were considered.
The edge
of the
1m away
from the FEM-SSR
result
andloading
the 3DisLEM
result from
matchthe
well
crest of
thetheslope.
of the distribution load q is
with
Wei etThe
al. magnitude
(2009) result.
equal to 100 kPa. The soil shear strength properties are cohesion
Figure
4. Critical
slip mass
the 3DofLEM
with explosive
Figure
4. Critical
slipofmass
theanalysis
3D LEM
analysis view
with
of 3D FOS
for the of
slope
Exampleand
1 a unit
for example
1 for example 1
of 20 Table
kPa, 1.
anComparison
angle of internal
friction
20indegrees
explosive
view
weight of 20 kPa. Table 1 shows the results of the various
SVSLOPE3D
etc( 2009)
analyses. Figure
2 and Figure SVSOLID3D
3 clearly show the Wei,
distinct
3D
(SSR)
failure surface(LEM)
based on a FEM-SSR
analysis. Figure 4(SSR)
shows
2.22.2
Example
2 - A2 nonsymmetrical
slope
with
corners
Example
- A nonsymmetrical
slope
with
corners
the critical slip1.359
surface based on a1.402
3D LEM analysis. It1.42
can be
One
analysis isis that
that
Oneofofthe
theadvantages
advantagesofofthe
the 3D
3D FEM-SSR
FEM-SSR analysis
seen that both the failure slip surface shape and FOS values
thethe
sliding
in advance.
advance.AA
slidingdirection
directiondoes
doesnot
notneed
need to
to be
be specified
specified in
from the FEM-SSR result and the 3D LEM result match well
limitationforfor 3D
3D column-based
column-based LEMs
LEMs is that the
limitation
the sliding
sliding
with the Wei et al. (2009) result.
directionofofthe
thecritical
criticalslip
slip surface
surface is
is another
another variable
direction
variable which
which
mustbebedetermined
determinedthrough
through aa searching
searching procedure.
procedure. A
must
A new
new
Table 1. Comparison of 3D FOS for the slope in Example 1
featurehas
hasbeen
beenadded
added inin SVSLOPE
SVSLOPE 3D to search
feature
search for
for the
the
criticalslip
slipsurface
surface direction
direction with
with optimization.
optimization. There
SVSLOPE3D
SVSOLID3D
Wei, etc( 2009)
critical
There isis
continuedusefulness
usefulnessininthe
theLEM
LEM because
because of
of its
continued
its computational
computational
(LEM)
(SSR)
(SSR)
efficiency.
Computational
times
for
FEM-SSR
efficiency. Computational times for FEM-SSR methods
methods are
are
significantlyhigher
higherthan
thanfor
for LEM
LEM analysis.
analysis. This
This efficiency
1.359
1.402
1.42
significantly
efficiency isis
particularlyuseful
useful inin performing
performing aa 3D
3D analysis
particularly
analysis when
when the
the
number of computations is significantly increased. The purpose
number of computations is significantly increased. The purpose
of this example is to test the efficiency of both 3D FEM-SSR
of slope
this example
is to test the efficiency of both 3D FEM-SSR
stability analysis and 3D LEM slope stability analysis for
slope stability analysis and 3D LEM slope stability analysis for
Figure 1. The geometry of the slope with external load in SVSLOPE 3D
general slopes without evident sliding direction information
available.
In this example, a general asymmetrical slope with inclined
corners is considered. As shown in Figure 5 and Figure 6, there
are three slopes with different inclinations, the right slope's
Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
inclination
is 1:2, the left slope's inclination is 1:1.5 and the
2
middle slope's inclination is 1:1.3 respectively. This model is
the soil to the
digitized from Jiang (1997). The soil's Mohr-Coulomb shear
surface. This
strength parameters are a cohesion of 5 kPa and an angle of
Figure
1.
The
geometry
of
the
slope
with
external
load
in
SVSLOPE
3D
s.
internal friction of 12 degrees.
Table 2 shows the comparison from different analysis
results. Jiang (1997) shows a FOS = 0.96 based on 3D
Simplified Janbu method with DP searching. The FOS from
FEM-SSR analysis is 0.941. The contour of the final
nternal friction
displacement at the final stage is shown in Figure 7. The critical
and
are
d the strength
slip mass based on 3D LEM analysis is shown in Figure 6. The
te of limiting
FOS is 0.957 and 0.977 separately for Simplified Bishop and
means that the
Spencer method. The critical slip direction also needs to be
onger possible
found. As shown in Figure 8, the critical slip surface direction is
At this stage, it
43 degree counter-clock wise from the negative x direction. It
Figure
2.
Contour
of
total
displacement
of
3D
FEM-SSR
analysis
for
he FOS equals
can be seen clearly that both the shape and slip direction are
example
1
at
the
final
stage.
d number of
Figure 2. Contour of total displacement of 3D FEM-SSR analysis for
ailure because
example 1 at the final stage.
d displacement
ium cannot be
760
hear strength

a powerful and

Simplified Janbu method with DP searching. The FOS from

FEM-SSR analysis is 0.941. The contour of the final
displacement at the final stage is shown in Figure 7. The critical
Figure 8. Plot of rotation angle vs. FOS for Example 2
slip mass based on 3D LEM analysis is shown in Figure 6. The
Technical
Committee
103
/
Comit
technique 103
FOS is 0.957 and 0.977 separately for Simplified Bishop and
Table 2. Comparison of 3D FOS for the slope in Example 2
Spencer method. The critical slip direction also needs to be
found. As shown in Figure 8, the critical slip surface direction is
SVSLOPE3D
SVSOLID3D
Jiang( 1997)
43 degree counter-clock wise from the negative x direction. It
(LEM)
(SSR)
(LEM+DP)
can be seen clearly that both the shape and slip direction are
veryclose
closebetween
betweenthetheFEM-SSR
FEM-SSRanalysis
analysisand
andthetheLEM
LEM Table 2. Comparison
of
3D
FOS
for
the
slope
in
Example
2
very
0.957 (Bishop),
0.941
0.96
analysis
seen
Figure
6 and
Figure
analysis
asas
seen
inin
Figure
6 and
Figure
7. 7.
0.977 (Spencer)
SVSLOPE3D

(LEM)

0.957
(Bishop),
CONCLUSIONS
0.977 (Spencer)

SVSOLID3D

Jiang( 1997)

(SSR)

(LEM+DP)

0.941

0.96

An actual slope failure occurs along the most critical sliding

direction that is often unknown for general 3D slopes.
3 Determination
CONCLUSIONS
of the critical slip surface and its FOS involves
the search for the criticl sliding direction. One of the advantage
An actual slope failure occurs along the most critical sliding
of FEM-SSR technique is that it does not need to specified the
direction that is often unknown for general 3D slopes.
sliding direction in advance, however it can not give the exact
Determination of the critical slip surface and its FOS involves
sliding direction angle value either. SVSLOPE 3D provides an
the search for the criticl sliding direction. One of the advantage
optimizaton technique that can find the critical sliding direction
of FEM-SSR technique is that it does not need to specified the
as part of FOS search.
sliding direction in advance, however it can not give the exact
sliding direction angle value either. SVSLOPE 3D provides an
Based on Example 1 it can be seen that both the failure slip
optimizaton technique that can find the critical sliding direction
surface shape and FOS values from the FEM-SSR result and the
as part of FOS search.
3D LEM result match well with the Wei et al. (2009) result. The
second example also demonstrates the similarity of results
Based on Example 1 it can be seen that both the failure slip
between analyzing a complex 3D slope stability geometry
surface shape and FOS values from the FEM-SSR result and the
where the direction of the slip may have an effect on the
3D LEM result match well with the Wei et al. (2009) result. The
calulated FOS.
second example also demonstrates the similarity of results
between analyzing a complex 3D slope stability geometry
These results demonstrate the usefulness of both 3D LEM and
where the direction of the slip may have an effect on the
FEM-SSR methodologies for the analysis of slope geometries
calulated FOS.
and loading conditions which are fundamentally 3D in nature.

Figure
Plan
view
elevation
contour
Example
Figure
5. 5.
Plan
view
of of
elevation
contour
of of
Example
22

These results demonstrate the usefulness of both 3D LEM and

FEM-SSR methodologies for the analysis of slope geometries
4 REFERENCES
and loading conditions which are fundamentally 3D in nature.
Baker R. 1980. Determination of the critical slip surface in slope
stability computations. Int. J. for Numerical and Analytical
4 REFERENCES
Methods in Geomechanics, 4, 333-359.
Cheng Y.M. 2003. Locations of critical failure surface and some further
Baker studies
R. 1980.
Determination
of the critical
slip and
surface
in slope30
on slope
stability analysis.
Computers
Geotechnics
stability
computations. Int. J. for Numerical and Analytical
(3), 255267.
Methods
in Geomechanics,
333-359.
Cheng
Y.M.,
Liu H.T., Wei4,W.B.
and Au S.K. 2005. Location of
Cheng critical
Y.M. 2003.
Locations of critical
failure surface
and some
three-dimensional
non-spherical
failure surface
byfurther
NURBS
studies
on slope
Computers and
30
functions
and stability
ellipsoidanalysis.
with applications
to Geotechnics
highway slopes.
(3),Computers
255267. and Geotechnics 32 (6), 387-399.
Cheng
Y.M.,
Liuand
H.T.,
Wei
W.B.
andSlope
Au S.K.
2005.
Location
of
Griffiths
D.V.
Lane
P.A.
1999,
stability
analysis
by finite
critical
three-dimensional
failure surface by NURBS
elements,
Gotechnique,non-spherical
49 (3), 387-403.
functionsD.V.
andand
ellipsoid
withR.M.
applications
to highway slopes.
Griffiths
Marquez
2007. Three-dimensional
slope
Computers
and Geotechnics
32 (6), 387-399.
stability analysis
by elasto-plastic
finite elements, Gotechnique, 57
Griffiths
and Lane P.A. 1999, Slope stability analysis by finite
(6),D.V.
537546.
elements,
49 (3), 387-403.
Hovland
H.J.Gotechnique,
1977. Three-dimensional
slope stability analysis method.
Griffiths
D.V. and
2007.
Three-dimensional slope
J Geotech
Eng Marquez
Div, ASCER.M.
103 (9),
971986.
stability
elasto-plastic
Gotechnique,
57
Hungr
O., analysis
Salgado by
F.M.
and Byrnefinite
P.M.elements,
1989. Evaluation
of a three(6),dimensional
537546. method of slope stability analysis. Canadian
Hovland
H.J. 1977. Journal,
Three-dimensional
slope stability analysis method.
Geotechnical
26 (4), 679686.
J Geotech Eng Div, ASCE 103 (9), 971986.
Hungr O., Salgado F.M. and Byrne P.M. 1989. Evaluation of a threedimensional method of slope stability analysis. Canadian
Geotechnical Journal, 26 (4), 679686.
Jiang J.C. 1997. Determination of the three-dimensional critical slip
3 surface in slope stability analysis, PhD thesis, the University of
Tokushima, Tokushima, Japan.
Lam L. and Fredlund D.G. 1993. A general limit equilibrium model for
three-dimensional slope stability analysis, Canadian Geotechnical
Journal,30 (6), 905919.
Wei W.B., Cheng Y.M. and Li L. 2009. Three-dimensional slope failure
analysis by the strength reduction and limit equilibrium methods,
Computers and Geotechnics, 36 (1-2), 7080.
Yamagami T. and Jiang J.C. 1997. A search for the critical slip surface
in three dimensional slope stability analysis. Soils and Foundation
37 (3), 16.
Zhang X. 1988. Three-dimensional stability analysis of concave slopes
in plan view, J. Geotech. Engng, ASCE, 114 (6), 658671.

Figure
Critical
slip
mass
Example
2 from
LEM
analysis
Figure
6. 6.
Critical
slip
mass
of of
Example
2 from
thethe
3D3D
LEM
analysis

Figure
7 Contour
total
displacement
from
FEM-SSR
analysis
Figure
7 Contour
of of
total
displacement
from
FEM-SSR
analysis

Figure 8. Plot of rotation angle vs. FOS for Example 2

761