Professional Documents
Culture Documents
net/publication/290467867
CITATIONS READS
555 5,144
2 authors, including:
D. V. Griffiths
Colorado School of Mines
266 PUBLICATIONS 6,411 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by D. V. Griffiths on 09 December 2016.
D. V. G R I F F I T H S a n d P. A . L A N E {
The majority of slope stability analyses per- En grande majoriteÂ, les analyses de stabilite de
formed in practice still use traditional limit pente meneÂes dans la pratique continuent aÁ
equilibrium approaches involving methods of utiliser les meÂthodes traditionnelles d'eÂquilibre
slices that have remained essentially unchanged limite et des systeÁmes de tranches qui n'ont
for decades. This was not the outcome envisaged pratiquement pas change depuis des dizaines
when Whitman & Bailey (1967) set criteria for d'anneÂes. Ce n'eÂtait pas le reÂsultat envisageÂ
the then emerging methods to become readily quand Whitman et Bailey (1967) ont eÂtabli des
accessible to all engineers. The ®nite element criteÁres pour que ces meÂthodes alors eÂmer-
method represents a powerful alternative ap- geantes puissent devenir facilement accessibles aÁ
proach for slope stability analysis which is accu- tous les ingeÂnieurs. La meÂthode d'eÂleÂments ®nis
rate, versatile and requires fewer a priori qui repreÂsente une alternative puissante pour les
assumptions, especially, regarding the failure analyses de stabilite de pente, est exacte, poly-
mechanism. Slope failure in the ®nite element valente et demande moins d'hypotheÁses `a
model occurs `naturally' through the zones in priori', surtout en ce qui concerne les meÂca-
which the shear strength of the soil is insuf®- nismes de rupture. La rupture de pente dans le
cient to resist the shear stresses. The paper modeÁle aÁ eÂleÂments ®nis se produit `naturelle-
describes several examples of ®nite element ment' aÁ travers des zones dans lesquelles la
slope stability analysis with comparison against reÂsistance au cisaillement du sol est insuf®sante
other solution methods, including the in¯uence pour reÂsister aux contraintes tangentielles. Cet
of a free surface on slope and dam stability. expose deÂcrit plusieurs exemples d'analyses de
Graphical output is included to illustrate defor- stabilite de pente utilisant les eÂleÂments ®nis et
mations and mechanisms of failure. It is argued eÂtablit des comparaisons avec d'autres meÂth-
that the ®nite element method of slope stability odes, comme l'in¯uence d'une surface libre sur
analysis is a more powerful alternative to tradi- la stabilite d'une pente et d'une digue. Nous
tional limit equilibrium methods and its wide- joignons une repreÂsentation graphique pour il-
spread use should now be standard in lustrer les deÂformations et meÂcanismes de rup-
geotechnical practice. ture. Nous avancËons que la meÂthode d'eÂleÂments
®nis pour analyser la stabilite des pentes consti-
tue une alternative plus puissante aux meÂthodes
traditionnelles d'eÂquilibre limite et que son utili-
KEYWORDS: dams; limit equilibrium methods; num- sation devrait maintenant devenir une pratique
erical modelling; plasticity; slopes. standard en geÂotechnique.
387
388 GRIFFITHS AND LANE
graphical methods will often be adequate for rou- Lambe & Silva (1995), who maintained that the
tine problems but the FE approach may be valuable ordinary method of slices had an undeservedly bad
if awkward geometries or material variations are reputation.
encountered which are not covered by traditional A dif®culty with all the equilibrium methods is
chart solutions. that they are based on the assumption that the
The use of nonlinear analysis in routine geotech- failing soil mass can be divided into slices. This in
nical practice is harder to justify, because there is turn necessitates futher assumptions relating to side
usually a signi®cant increase in complexity which force directions between slices, with consequent
is more likely to require the help of a modelling implications for equilibrium. The assumption made
specialist. Nonlinear analyses are inherently itera- about the side forces is one of the main character-
tive in nature, because the material properties and/ istics that distinguishes one limit equilibrium meth-
or the geometry of the problem are themselves a od from another, and yet is itself an entirely
function of the `solution'. Objections to nonlinear arti®cial distinction.
analyses on the grounds that they require excessive
computational power, however, have been largely
overtaken by developments in, and falling costs of, FINITE ELEMENT METHOD FOR SLOPE STABILITY
computer hardware. A desktop computer with a ANALYSIS
standard processor is now capable of performing Duncan's review of FE analysis of slopes con-
nonlinear analyses such as those described in this centrated mainly on deformation rather than stabi-
paper in a reasonable time spanÐminutes rather lity analysis of slopes; however, attention was
than hours or days. drawn to some important early papers in which
Slope stability represents an area of geotechni- elasto-plastic soil models were used to assess stabi-
cal analysis in which a nonlinear FE approach lity. Smith & Hobbs (1974) reported results of
offers real bene®ts over existing methods. As this öu 0 slopes and obtained reasonable agreement
paper will show, slope stability analysis by elasto- with Taylor's (1937) charts. Zienkiewicz et al.
plastic ®nite elements is accurate, robust and (1975) considered a c9, ö9 slope and obtained good
simple enough for routine use by practising en- agreement with slip circle solutions. Grif®ths
gineers. Perception of the FE method as complex (1980) extended this work to show reliable slope
and potentially misleading is unwarranted and stability results over a wide range of soil properties
ignores the real possibility that misleading results and geometries as compared with charts of Bishop
can be obtained with conventional `slip circle' & Morgenstern (1960). Subsequent use of the FE
approaches. The graphical capabilities of FE pro- method in slope stability analysis has added further
grams also allow better understanding of the con®dence in the method (e.g. Grif®ths, 1989;
mechanisms of failure, simplifying the output Potts et al., 1990; Matsui & San, 1992). Duncan
from reams of paper to manageable graphs and mentions the potential for improved graphical re-
plots of displacements. sults and reporting utilizing FE, but cautions
against arti®cial accuracy being assumed when the
input parameters themselves are so variable.
TRADITIONAL METHODS OF SLOPE STABILITY Wong (1984) gives a useful summary of poten-
ANALYSIS tial sources of error in the FE modelling of slope
Most textbooks on soil mechanics or geotech- stability, although recent results, including those
nical engineering will include reference to several presented in this paper, indicate that better accu-
alternative methods of slope stability analysis. In a racy is now possible.
survey of equilibrium methods of slope stability
analysis reported by Duncan (1996), the character-
istics of a large number of methods were sum- Advantages of the ®nite element method
marized, including the ordinary method of slices The advantages of a FE approach to slope
(Fellenius, 1936), Bishop's Modi®ed Method stability analysis over traditional limit equilibrium
(Bishop, 1955), force equilibrium methods (e.g. methods can be summarized as follows:
Lowe & Kara®ath, 1960), Janbu's generalized pro-
cedure of slices (Janbu, 1968), Morgenstern and (a) No assumption needs to be made in advance
Price's method (Morgenstern & Price, 1965) and about the shape or location of the failure
Spencer's method (Spencer, 1967). surface. Failure occurs `naturally' through the
Although there seems to be some consensus that zones within the soil mass in which the soil
Spencer's method is one of the most reliable, text- shear strength is unable to sustain the applied
books continue to describe the others in some shear stresses.
detail, and the wide selection of available methods (b) Since there is no concept of slices in the FE
is at best confusing to the potential user. For approach, there is no need for assumptions
example, the controversy was recently revisited by about slice side forces. The FE method
SLOPE STABILITY ANALYSIS BY FINITE ELEMENTS 389
preserves global equilibrium until `failure' is Table 1. Six-parameter soil model
reached. ö9 Friction angle
(c) If realistic soil compressibility data are avail- c9 Cohesion
able, the FE solutions will give information ø Dilation angle
about deformations at working stress levels. E9 Young's modulus
(d) The FE method is able to monitor progressive í9 Poisson's ratio
failure up to and including overall shear ã Unit weight
failure.
1.2 H 2H
Rollers
Fixed
Fig. 1. Example 1: Mesh for a homogeneous slope with a slope angle of 26´578 (2:1), ö9 208, c9=ã H 0´05
392 GRIFFITHS AND LANE
E ′δmax /γH 2
1´40 1´476 1000
0.8
1
When FOS 1:4, there is a sudden increase in the 1 increment
2 equal increments
dimensionless displacement E9ämax =ã H 2 , and the 1.2 3 equal increments
5 equal increments
algorithm is unable to converge within the iteration
limit. Fig. 2 shows a plot of the data from Table 2, and 1.4
indicates close agreement between the FE result and
the factor of safety given for the same problem by the 1.6
0 0.2 0.4 0.6 0.8 1 1.2
charts of Bishop & Morgenstern (1960).
Gravity factor
Figure 3 shows the in¯uence of gravity loading
increment size on displacements in example 1. Fig. 3. In¯uence of gravity increment size on maxi-
With a `failure' factor of safety of FOS 1:4 mum displacement at failure (FOS 1´4) from exam-
applied to the soil properties, the four graphs ple 1
correspond to the maximum displacement obtained
when gravity was applied in a single increment as
compared with that obtained with two, three or ®ve
equal increments. The ®gure demonstrates that the to the unconverged situation with FOS 1:4. The
displacement obtained with full gravity loading is deformed mesh corresponding to this unconverged
barely affected by the increment size. solution gives a rather diffuse indication of the
Figures 4(a) and 4(b) give the nodal displace- failure mechanism. This is due to the relatively
ment vectors and the deformed mesh corresponding crude FE mesh, which must remain continuous
even at `failure'. Conventional FE analysis is un-
able to model gross discontinuities along potential
failure surfaces, although techniques have been
Bishop & Morgenstern (1960) described for enhancing the visualization of the
FOS 5 1.380 failure surfaces (e.g. Grif®ths & Kidger, 1995).
More advanced FE methods for modelling shear
0.25 bands in conjunction with adaptive mesh re®ne-
0.5 ment techniques have been described by Loret &
0.75 Prevost (1991) and Zienkiewicz et al. (1995).
1
1.3
1.5 Example 2: Homogeneous slope with a foundation
1.8 layer ( D 1:5)
Figure 5 shows that a foundation layer of thick-
E ′δmax /γH 2
2
2.3
ness H=2 has now been added to the base of the
2.5
slope of example 1, with all other properties and
geometry remaining the same.
2.8 D 5 1.0 (Example 1)
D 5 1.5 (Example 2)
The initial mesh and the deformed mesh at
3
failure are shown in Figs 5(a) and 5(b) respectively.
3.3
It is clear from Fig. 5(b) that a mechanism of the
3.5
`toe failure' type has been obtained. Fig. 2 indi-
3.8 cates that the critical factor of safety is essentially
4
0.8 1 1.2 1.4 1.6 unchanged from example 1 at FOS 1:4, although
the displacements are increased due to the greater
FOS
volume of compressible soil.
Fig. 2. Examples 1 and 2: FOS versus dimensionless
This FE result con®rms that the addition of the
displacement. The rapid increase in displacement and foundation layer has not led to any perceptible
the lack of convergence when FOS 1´4 indicates change in the factor of safety of the slope. Bishop
slope failure & Morgenstern (1960) give FOS 1:752 as one
SLOPE STABILITY ANALYSIS BY FINITE ELEMENTS 393
(a)
(b)
DH
(a)
(b)
Fig. 5. Example 2: Homogeneous slope with a foundation layer. Slope angle 26´578 (2:1), ö9 208, c9=ã H 0´05,
D 1´5: (a) undeformed mesh; (b) mesh corresponding to unconverged solution with FOS 1´4
394 GRIFFITHS AND LANE
possible solution for this example (D 1:5, c9= be performed with respect to effective stresses by
(ã H) 0:05, ö9 208, 2:1 slope), although it is adding a large apparent ¯uid bulk modulus to the
important to check the alternative solution corre- soil constitutive matrix (Naylor, 1974). In this case,
sponding to D 1:0 to verify which gives the however, a total stress analysis using a Tresca
lower FOS. The charts of Cousins (1978) essen- failure criterion (öu 0) is presented.
tially agree with the FE result and indicate that, Figure 6 shows a slope on a foundation layer
with a foundation layer, the critical circular me- (D 2) of undrained clay. The slope includes a
chanism at its lowest point passes fractionally thin layer of weaker material which initially runs
below the base of the slope and gives a slightly parallel to the slope, then horizontally in the
lower factor of safety than when there is no foundation and ®nally outcrops at an angle of
foundation layer present. 458 beyond the toe. Although this example may
Solving this example using a proprietary slip seem contrived, it is not unlike the situation of a
circle program also found the possible `result' of thin, weak liner within a land®ll system. The
FOS 1:7 when a failure circle tangent to the factor of safety of the slope was estimated by FE
base of the foundation was assumed. It was neces- analysis for a range of values of the undrained
sary to force the slip circle to pass through the toe shear strength of the thin layer (cu2 ) while main-
to obtain the `correct' FOS of 1´4. taining the strength of the surrounding soil at
This example demonstrates one of the main cu1 =ã H 0:25.
advantages of FE slope stability analysis over con- The FE results shown in Fig. 7 give the com-
ventional methods. The FE approach requires no a puted factor of safety expressed to the nearest
priori assumption of the location or shape of the 0´05. For a homogeneous slope (cu2 =cu1 1), the
critical surface. Failure occurs `naturally' within computed factor of safety was close to the Taylor
the zones of the soil mass where the shear strength solution (Taylor, 1937) of FOS 1:47 and gave
of the soil is insuf®cient to resist the shear stresses. the expected circular base failure mechanism. As
The use of a limit equilibirum method requires, at the strength of the thin layer was gradually re-
the very least, some experience and care on the duced, a distinct change in the nature of the results
part of the user in order to initiate appropriate was observed when cu2 =cu1 0:6.
search procedures which avoid the possibility of Also shown on this ®gure are limit equilibrium
homing in on the wrong `critical' circle. solutions obtained using Janbu's method assuming
both circular (base failure) and three-line wedge
mechanisms following the path of the weak layer.
Example 3: An undrained clay slope with a thin The discontinuity when cu2 =cu1 0:6 clearly re-
weak layer presents the transition between the circular me-
The next example demonstrates a stability analy- chanism and the non-circular mechanism governed
sis of a slope of undrained clay. Effective stress by the weak layer. For cu2 =cu1 . 0:6, the (circular)
analysis of an undrained slope of this type could base failure mechanism governs the behaviour, and
φu 5 0
2H 2H cu2 , cu1
0.6H 1.2H
2
cu2 1
H 2H
2
1.2H 0.6H
1 cu1
cu1 0.4H
1
H 1
0.4H
Fig. 6. Example 3: Undrained clay slope with a foundation layer including a thin weak layer (D 2,
cu1 =ã H 0´25)
SLOPE STABILITY ANALYSIS BY FINITE ELEMENTS 395
2 the factor of safety is essentially unaffected by the
1.8 strength of the weaker thin layer. For cu2 =cu1 , 0:6,
cu1/γH 5 0.25 Taylor (1937) the (non-circular) thin layer mechanism takes over
1.6 FOS 5 1.47 and the factor of safety falls linearly.
1.4 This behaviour is explained more clearly in
Fig. 8, which shows the deformed mesh at failure
1.2 for three different values of the ratio cu2 =cu1 . Fig.
8(a), corresponding to the homogeneous case
FOS
1
(cu2 =cu1 1), indicates an essentially circular fail-
0.8 ure mechanism tangent to the ®rm base as pre-
Finite elements
0.6
Slope program—wedges
Slope program—circles dicted by Taylor. Fig. 8(c), in which the strength
of the thin layer is only 20% of that of the
0.4
surrounding soil (cu2 =cu1 0:2), indicates a highly
0.2 concentrated non-circular mechanism closely fol-
lowing the path of the thin weak layer. Fig. 8(b),
0
0 0.2 0.4 0.6 0.8 1 in which the strength of the thin layer is 60% of
cu2/cu1 that of the surrounding soil (cu2 =cu1 0:6), indi-
cates considerable complexity and ambiguity. At
Fig. 7. Example 3: Computed factor of safety (FOS) least two con¯icting mechanisms are apparent.
for different values of cu2 =cu1 First, there is a base failure mechanism merging
(a)
(b)
(c)
with the weak layer beyond the toe of the slope, FOS
and second, there is a mechanism running along 2.6
the weak layer parallel to the slope and outcrop- Taylor
2.4 cu1/γH 5 0.25
ping at the toe. (cu2 .. cu1)
Without prior knowledge of the two alternative 2.2 FOS 5 2.10
mechanisms, a traditional limit equilibrium search 2.0
could seriously overestimate the factor of safety.
This is illustrated in Fig. 7 where, for example, a 1.8 Taylor
circular mechanism with cu2 =cu1 0:2 would in- 1.6
(cu2 5 cu1)
dicate FOS 1:3, when the correct factor of safety FOS 5 1.47
is closer to 0´6. 1.4
φu 5 0
2H 2H
2
1
H 2H
cu1
cu2
H
(a)
(b)
(c)
INFLUENCE OF FREE SURFACE AND RESERVOIR remaining parts of the algorithm relating to the
LOADING ON SLOPE STABILITY assessment of Mohr±Coulomb yield and elasto-plas-
We now consider the in¯uence of a free surface tic stress redistribution. Note that the gravity loads
within an earth slope and reservoir loading on the are computed using total unit weights of the soil.
outside of a slope as shown in Fig. 12. The external loading due to the reservoir is
Regarding the role of the free surface, a rigor- modelled by applying a normal stress to the face
ous approach would ®rst involve obtaining a of the slope equal to the water pressure. Thus, as
good-quality ¯ow net for free surface ¯ow through shown in Fig. 13, the applied stress increases
the slope, enabling pore pressures to be accurately linearly with water depth and remains constant
estimated at any point within the ¯ow region. For along the horizontal foundation level. These stres-
the purposes of slope stability analysis, however, ses are converted into equivalent nodal loads on
it is usually considered suf®ciently accurate and the FE mesh (e.g. Smith & Grif®ths, 1998) and
conservative to estimate pore pressure at a point added to the initial gravity loading.
as the product of the unit weight of water (ãw )
and the vertical distance of the point beneath the
free surface. In Fig. 12 the pore pressures at two Example 5: Homogeneous slope with horizontal
locations, A and B, have been calculated using free surface
this assumption. Figure 14 shows a similar slope to that analysed
In the context of FE analysis, the pore pressures in example 1, but with a horizontal free surface at a
are computed at all submerged (Gauss) points as depth L below the crest. Using the method described
described above, and subtracted from the total nor- above, the factor of safety of the slope has been
mal stresses computed at the same locations follow- computed for several different values of the draw-
ing the application of surface and gravity loads. The down ratio (L= H), which has been varied from ÿ0:2
resulting effective stresses are then used in the (slope completely submerged with the water level
398 GRIFFITHS AND LANE
Free surface
Reservoir Level
Embankment
hA
hW
hB
A uA 5 hAγW
B uB 5 hBγW
Reservoir Level
Linearly increasing
Free surface normal stress from
zero to h WγW
hW
Embankment
Constant normal
stress 5 h WγW
Finite elements
Limit equilibrium
FOS
1.6
Bishop &
Morgenstern
(1960)
1.4
FOS 5 1.4
1.2
20.2 20.1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
L/H
ever, the increased frictional strength starts to have a for `rapid drawdown' problems has also been
greater in¯uence than the increased weight, and the demonstrated for a range of slopes using a similar
factor of safety rises. Other results of this type have approach (Lane & Grif®ths, 1997).
been reported by Lane & Grif®ths (1997) for a slope
which was stable (FOS . 1) when `dry' or fully
submerged, but became unstable (FOS , 1) at a Example 6: Two-sided earth embankment
critical value of the drawdown ratio L= H. It should The example given in Fig. 16 is of an actual
also be pointed out from the horizontal part of the earth dam cross-section including a free surface
graph in Fig. 15, corresponding to L= H < 0, that which slopes from the reservoir level to foundation
the factor of safety for a fully submerged slope is level on the downstream side (Torres & Coffman,
unaffected by the depth of water above the crest. 1997). For the purposes of this example, the materi-
Excellent agreement with Morgenstern (1963) al properties have been made homogeneous. Fig. 17
Dimensions in metres
7.3
Reservoir level
Free
17.1 surfa
ce H 5 21.3
18° 23°
7.3
Fig. 16. Example 6: Two-sided earth embankment with a sloping free surface. Homogeneous dam, ö9 378,
c9 13´8 kN=m2 , ã 18´2 kN=m3 (above and below WT)
E ′δmax /γH 2
A second analysis was also performed with no
free surface corresponding to the embankment be- 20
fore the reservoir was ®lled. FE slope stability
analysis led to the results shown in Fig. 18. Both
25
cases were also solved using a conventional limit Limit
equilibrium approach which gave FOS 1:90 with equilibrium
No free surface
a free surface and FOS 2:42 without a free sur- 30 With a free surface solutions
face. The limit equilibrium and FE factors of safety
values were in close agreement in both cases. FOS 5 1.90 FOS 5 2.42
Regarding the critical mechanisms of failure, 35
1 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6
Figs 19 and 20 show the deformed mesh cor-
FOS
responding to the unconverged FE solution as
compared with the slip circle that gave the Fig. 18. Example 6: FOS versus dimensionless dis-
lowest factors of safety from the limit equilibrium placement
approach. As expected, the lowest factor of safety
occurs on the steeper, downstream side of the
embankment in both cases. It should also be noted free surface (Fig. 20). Fig. 21 shows the corre-
that both the FE and limit equilibrium results sponding displacement vectors from the FE solu-
indicate a toe failure for the case with no free tions. Reasonably good agreement between the
surface (Fig. 19), and a deeper mechanism extend- locations of the failure mechanisms by both types
ing into the foundation layer for the case with a of analysis is indicated.
(a)
5.4m
R 5 62.7m
62.4m
(b)
Fig. 19. Example 6 with no free surface: (a) deformed mesh corresponding to the unconverged solution by
®nite elements; (b) the critical slip circle by limit equilibrium. Both methods give FOS 2´4
SLOPE STABILITY ANALYSIS BY FINITE ELEMENTS 401
(a)
11.5m
R 547.8m
43.9m
(b)
Fig. 20. Example 6 with a free surface: (a) deformed mesh corresponding to the unconverged solution by
®nite elements; (b) the critical slip circle by limit equilibrium. Both methods give FOS 1´9
THE CRITERIA FOR COMPUTER-AIDED ANALYSIS for the parameters being input. FE analysis meets
Whitman & Bailey (1967) looked forward to the these criteria with a degree of accuracy decided by
future of computer-aided analysis for engineers and the engineer in designing the model.
set criteria by which it could be judged. Their It should be possible, in a realistic timescale, to
comments were originally addressed to the automa- do suf®cient trials to examine all the key modes of
tion of limit equilibrium methods, but they also behaviour; to consider different times in the life of
commented on the then emerging numerical analy- the structure and to vary parameters during design
sis techniques. to test options for cost and ef®ciency. All this is
They judged that the system must be suf®ciently now possible with FE methods.
accurate for con®dence in its use and appropriate Finally, the method of human±machine commu-
FOS 5 2.4
(a)
FOS 5 1.9
(b)
Fig. 21. Example 6: Displacement vectors corresponding to the unconverged solution by ®nite elements:
(a) no free surface; (b) with a free surface. Only those displacement vectors that have a magnitude . 10% of
the maximum are shown
402 GRIFFITHS AND LANE
nication must be user-friendly and readily accessi- of embankments and slopes. J. Geotech. Engng, ASCE
ble. This is partly a matter of program design but 107, 691±693:
easily achieved. Graphical output greatly enhances Clough, R. W. & Woodward, R. J. (1967). Analysis of
the process of design and analysis over and above embankment stresses and deformations. J. Soil. Mech.
Found. Div., ASCE 93, SM4, 529±549.
that from the numerical results.
Cousins, B. F. (1978). Stability charts for simple earth
Similarly, Chowdhury (1981), in his discussion slopes. J. Geotech. Engng, ASCE 104, No. 2,
of Sarma (1979), commented on the perceived 267±279:
reluctance to develop alternatives to limit equili- Duncan, J. M. (1996). State of the art: limit equilibrium
brium methods for practice when the tools to do so and ®nite-element analysis of slopes. J. Geotech.
were already available. Since then, numerous appli- Engng, ASCE 122, No. 7, 577±596:
cations and experience have veri®ed the possibili- Duncan, J. M. & Dunlop, P. (1969). Slopes in stiff
ties offered by ®nite elements. ®ssured clays and soils. J. Soil Mech. Found. Div.,
The key issues of cost and turnaround time have ASCE 95, SM5, 467±492:
been overtaken by the falling cost of powerful Fellenius, W. (1936). Calculation of the stability of earth
dams. Proc. 2nd congr. large dams, Washington DC 4.
hardware and processor speeds which now make Grif®ths, D. V. (1980). Finite element analyses of walls,
the FE method available to engineers at less than footings and slopes. PhD thesis, University of Man-
the cost of their CAD systems. What remains is chester.
the concern of powerful tools used wrongly. That Grif®ths, D. V. (1982). Computation of bearing capacity
is no more true of ®nite elements after years of factors using ®nite elements. GeÂotechnique 32, No. 3,
application than of limit equilibrium methods, 195±202.
which can themselves produce seriously misleading Grif®ths, D. V. (1989). Computation of collapse loads in
results. Engineering judgement is still essential, geomechanics by ®nite elements. Ing Arch 59, 237±244:
Grif®ths, D. V. (1990). Failure criterion interpretation
whichever method is being used.
based on Mohr±Coulomb friction. J. Geotech. Engng,
ASCE 116, GT6, 986±999:
Grif®ths, D. V. & Kidger, D. J. (1995). Enhanced visuali-
CONCLUDING REMARKS zation of failure mechanisms in ®nite elements. Com-
The FE method in conjunction with an elastic± put. Struc. 56, No. 2, 265±269:
perfectly plastic (Mohr±Coulomb) stress±strain Hicks, M. A. & Boughrarou, R. (1998). Finite element
method has been shown to be a reliable and robust analysis of the Nelerk underwater berm failures. GeÂo-
method for assessing the factor of safety of slopes. technique 48, No. 2, 169±185:
One of the main advantages of the FE approach is Hicks, M. A. & Wong, S. W. (1988). Static liquefaction
that the factor of safety emerges naturally from the of loose slopes. Proc. 6th Int. Conf. Numer. Methods
Geomech., 1361±1368:
analysis without the user having to commit to any Janbu, N. (1968). Slope stability computations. Soil Mech.
particular form of the mechanism a priori. Found. Engng Report. Trondheim: Technical Univer-
The FE approach for determining the factor of sity of Norway.
safety of slopes has satis®ed the criteria for effec- Lambe, T. W. & Silva, F. (1995). The ordinary method of
tive computer-aided analysis. The widespread use slices revisited. Geotech. News 13, No. 3, 49±53.
of this method should now be seriously considered Lambe, T. W. & Whitman, R. V. (1969). Soil mechanics.
by geotechnical practitioners as a more powerful New York: Wiley.
alternative to traditional limit equilibrium methods. Lane, P. A. & Grif®ths, D. V. (1997). Finite element slope
stabilityÐwhy are engineers still drawing circles?
Proc. 6th Int. Symp. Numer. Models Geomech.
589±593:
ACKNOWLEDGEMENTS Loret, B. & Prevost, J. H. (1991). Dynamic strain locali-
The writers acknowledge the support of NSF sation in ¯uid-saturated porous media. J. Eng. Mech.,
Grant No. CMS 97-13442. Dr Lane was supported ASCE 117, No. 4, 907.
by the Peter Allen Scholarship Fund of UMIST. Lowe, J. & Kara®ath, L. (1960). Stability of earth dams
upon drawdown. Proc. 1st Pan-Am. Conf. Soil Mech.
Found. Engng, 537±552:
REFERENCES Matsui, T. & San, K.-C. (1992). Finite element slope
Abramson, L. W., Lee, T. S., Sharma, S. & Boyce, G. M. stability analysis by shear strength reduction tech-
(1995). Slope stability and stabilisation methods. Chi- nique. Soils Found. 32, No. 1, 59±70:
chester: Wiley. Molenkamp, F. (1981). Elastic-plastic double hardening
Bishop, A. W. (1955). The use of the slip circle in the sta- model Monot. Technical report. Delft: Delft Geotechnics.
bility analysis of slopes. GeÂotechnique 5, No. 1, 7±17: Morgenstern, N. R. (1963). Stability charts for earth
Bishop, A. W. & Morgenstern, N. R. (1960). Stability coef®- slopes during rapid drawdown. GeÂotechnique 13,
cients for earth slopes. GeÂotechnique 10, 129±150: 121±131:
Borja, R. I., Lee, S. R. & Seed, R. B. (1989). Numerical Morgenstern, N. R. & Price, V. E. (1965). The analysis of
simulation of excavation in elasto-plastic soils. Int. J. the stability of general slip surfaces. GeÂotechnique 15,
Numer. Anal. Methods Geomech. 13, No. 3, 231±249. No. 1, 79±93:
Chowdhury, R. N. (1981). Discussion on stability analysis Naylor, D. J. (1974). Stresses in nearly incompressible
SLOPE STABILITY ANALYSIS BY FINITE ELEMENTS 403
materials by ®nite elements with application to the Taylor, D. W. (1937). Stability of earth slopes. J. Boston
calculation of excess pore pressure. Int. J. Numer. Soc. Civ. Eng. 24, 197±246:
Methods Engng 8, 443±460: Torres, R. & Coffman, T. (1997). Liquefaction analysis
Naylor, D. J. & Pande, G. N. (1981). Finite elements in for CAS: Willow Creek Dam. Technical Report WL-
geotechnical engineering. Swansea: Pineridge Press. 8312-2. Denver: US Department of the Interior, Bu-
Paice, G. M. (1997). Finite element analysis of stochastic reau of Reclamation.
soils. PhD thesis, University of Manchester. Whitman, R. V. & Bailey, W. A. (1967). Use of compu-
Perzyna, P. (1966). Fundamental problems in viscoplasti- ters for slope stability analysis. J. Soil Mech. Found.
city. Adv. Appl. Mechanics 9, 243±377: Div., ASCE 93, SM4, 475±498:
Potts, D. M., Dounias, G. T. & Vaughan, P. R. (1990). Wong, F. S. (1984). Uncertainties in FE modeling of
Finite element analysis of progressive failure of Car- slope stability. Comput. Struct. 19, 777±791:
sington embankment. GeÂotechnique 40, No. 1, 79±102: Cormeau, I. C. (1974). Viscoplasticity, plasticity and
Sarma, S. (1979). Stability analysis of embankments and creep in elastic solids. A uni®ed approach. Int. J.
slopes. J. Geotech. Engng, ASCE 105, 1511±1524: Numer. Methods Engng 8, 821±845:
Smith, I. M. & Grif®ths, D. V. (1998). Programming the Zienkiewicz, O. C., Humpheson, C. & Lewis, R. W.
®nite element method, 3rd edn. Chichester: Wiley. (1975). Associated and non-associated viscoplasticity
Smith, I. M. & Hobbs, R. (1974). Finite element analysis and plasticity in soil mechanics. GeÂotechnique 25,
of centrifuged and built-up slopes. GeÂotechnique 24, 671±689:
No. 4, 531±559: Zienkiewicz, O. C., Huang, M. & Pastor, M. (1995).
Snitbhan, N. & Chen, W. F. (1976). Elastic±plastic large Localization problems in plasticity using ®nite ele-
deformation analysis of soil slopes. Comput. Struct. 9, ments with adaptive remeshing. Int. J. Numer. Anal.
567±577: Methods Geomech. 19, No. 2, 127±148:
Spencer, E. (1967). A method of analysis of the stability Zienkiewicz, O. C. & Taylor, R. L. (1989). The ®nite
of embankments assuming parallel interslice forces. element method, Vol. 1, 4th edn. London, New York:
GeÂotechnique 17, No. 1, 11±26: McGraw-Hill.