Proceedings, Slope Stability 2011: International Symposium on Rock Slope Stability in Open Pit Mining and

Civil Engineering, Vancouver, Canada (September 18-21, 2011)

Limit Analysis of Seismic Slope Stability Based on Tension-Shear Failure
Mechanism
Y. Zhang Civil and Structure Engineering, Kyushu University, Fukuoka, Japan
G. Chen Civil and Structure Engineering, Kyushu University, Fukuoka, Japan
K. Zen Civil and Structure Engineering, Kyushu University, Fukuoka, Japan
K. Kasama Civil and Structure Engineering, Kyushu University, Fukuoka, Japan
S. Dong Civil and Structure Engineering, Kyushu University, Fukuoka, Japan

Abstract
Earthquakes have a significant influence on slope stability in both open pit mining and civil engineering. This
has been investigated both from the perspective of slope stability analyses as well as seismic-induced slope
failure mechanisms. A large number of investigations of landslides induced by earthquakes show that tension
failures appear in the upper part of almost all landslides. However, existing seismic slope stability analysis
approaches usually only consider shear failure ignoring extensional failure effects. This may lead to inaccuracy
in the stability analysis of seismic slope. Therefore, to investigate the significance of the tension failure effect on
seismic stability analysis, this paper reports findings from the adoption of a method that considers both tension
and shear failure during seismic loading. Calculations of the factor of safety were conducted by applying the
kinematic theorem of limit analysis; the form and position of the failure surface are reported in the paper.
Results are presented that correspond with the actual situation from the investigations. Finally, the present
results of factor of safety were compared to those obtained from existing analysis methods that do not consider
tension effects. The findings imply that the effect of tension failure on the seismic slope stability analysis is
significant and needs to be considered where earthquake loading is a possibility.
1 Introduction
In seismically active regions, earthquake is a major trigger for instability of slopes. Stability analysis of slope
bearing earthquake loading is always the focus of geotechnical engineers. The failure mechanism of slope under
earthquake loading is the premise of slope's dynamic stability analysis. Currently, earthquake-induced slope
failures have normally been evaluated based on equilibrium of forces which provides a safety factor against
initiating failure or displacement according to the Newmark model (Newmark 1965). Both of these methods
regard the failure mechanism of slope as shear failure, and take the shear sliding surface as failure surface under
dynamic loadings. However, in recent years, frequent earthquakes occur in New ZealandJ apan and China which
have led to a renewed knowledge in the mechanism of instability of slopes. A large number of post-earthquake
investigations show that seismic induced slope failure contains tension and shear segment (Huang et al. 2009, Xu
et al. 2009, Yin et al. 2009). For this issue, some geologists have made some meaningful researches. Yan et al.
(2010) and Zheng et al. (2009) have certificated the existing of tension failure region by analyzing the
mechanism of seismic slope based on FLAC program. Huang et al. (2009), Xu et al. (2009) and Yin et al. (2009)
have given much detailed description for tension segment of slope failure surface based on post-earthquake
investigations. These studies, however, most focus on the description or explanation of phenomenon, but few on
research of stability analysis subjected to the tension failure.
Therefore, to investigate how the tension failure effect on seismic stability analysis, a method considering both
the tension and shear failure is adopted in this paper which the kinematic theorem of limit analysis be applied
and the earthquake inertia force is simulated by using the pseudo-static method.
2 Mechanism of landslides induced by earthquake
Authors hold the viewpoint that instability of slopes induced by earthquake is a process from quantitative change
to qualitative change. For homogeneity slope, at the beginning of earthquake occurred, tension failure region
engendered at trailing edge of slope with the reason of effect of inertia force, and the depth of tension failure
region will be extend and new tension crack will be occur with the continuing effect of earthquake loading. At
the same time, plasticity shear yield occur in the slope toe area under earthquake loading, and the plasticity shear
yield is extended from toe upward to slope body. Development of the plasticity shear yield aggravate the depth
of the crack region, and the crack region can caused stress concentration at shear region and caused larger
shearing deformation at the potential failure surface. When the shearing failure region connects with the crack
region, the slope will collapse.
According to the crack extend theory of fracture mechanics; depth of the crack in top of slope has the functional
relationship with the slope's geometrical characteristic, strength of slope materials, intensity of earthquake and so
on. It can reach to several meters, or dozens of meters even deeper. It should be indicated that the depth of crack
region under earthquake loading is different from it in static situation. In static situation, the limit depth of crack
is deemed generally the same with depth of tension crack at behind of retaining wall bearing active earth
pressure. In dynamic situation, the crack develops easily and suddenly with the reason of inertia force, and
results of numerical simulation show that a dynamic loading with peak acceleration of 1.29m/s
2
can caused ratio
of depth of tension failure region to height of slope of 0.3 to 0.5 (Zheng et al. 2009).
3 Strength reduction technique
For slopes, the factor of safety F is traditionally defined as the ratio of the actual shear strength to the minimum
shear strength required to prevent failure (Bishop 1955). It means the factor of safety F is a shear strength
reduction factor. It simply reduces the soil shear strength until collapse occurs. The result of factor of safety is
the ratio of the slope materials actual shear strength to the reduced shear strength at failure. This is the shear
strength reduction technique that was used by Zienkiewica et al. (1975), and so on.
To perform slope stability analysis with the shear strength reduction technique, the factor of safety F is defined
according to the equation:
F =
c
c
=
tanq
tanq
[1]
where c and are the reduced strength parameters and c and are the actual strength parameters,
respectively.
4 Upper bound method of limit analysis
4.1 Failure mode
Limit analysis has been applied to slope stability analysis in the past 35 years or so since the publishing of limit
analysis and soil plasticity in 1975 presented by Chen (1975).
This theorem state that a slope will collapse if the rate of work done by external loads and body forces exceeds
the energy dissipation rate for any assumed kinematically admissible failure mechanism. Applicability of the
theorem requires that slope material will be considered to be isotropic and perfectly plastic, and to obey both the
associated flow rule and the Mohr-Coulomb yield criterion. Deformations of the slope before the limit state are
considered to be negligible.
This method is applied here to analyze the stability of slopes under seismic loading by considering both the
tension and shear failure. The problem considered here is factor of safety F and form of failure surface for a
homogeneity slope under earthquake loading in which the effect of earthquake on a potential failure slope
material mass is represented by forces acting horizontally and vertically at the centre of gravity, which are
calculated as the product of a seismic intensity coefficients k
h
, k

and the weight of the potential sliding mass

w (Seed & Martin 1966).
The modes of failure surface for limit analysis are illustrated in Figures 1 and 2. Especially, Figure 1 shows a
simple failure mechanism where the slope material mass translates as a rigid body along a planar surface with a
vertical crack. Moreover, the Figure 2 shows a rotational mechanism involving a log-spiral failure surface
passing through below the toe of the slope. The sliding mass rigid is translating along a planar surface or rotating
with respect to the centre to 0 with an angular velocity relative to the bottom rigid, and two parts are
separated by a thin layer with a discontinuous velocity : in the translational mode and :(0) in the rotational
mode. The plastic velocity : or :(0) is making an angle with sliding surface according to the associate
flow rule. In this paper, the failure surface is considered as a rigid rotation collapse be a log-spiral expressed by
the equation:
r(0) = r
0
cxp |(0 - 0
0
)tan] [2]
where r(0) is the radius of the spiral related to angle 0, and r
0
and 0
0
are the initial values (Figure 2). It is
assumed that the failure surface of tension failure region is developing vertically.
The assumed mechanism can be specified completely by three variables: 0
0
, 0
h
, [

. For the geometrical

relationship it may be shown that the ratios,
H

0
and
L

0
, can be expressed in terms of the angles 0
0
, 0
h
and depth
of tension failure region Z in the forms:
H

0
=
sIn[

sIn ([

-u)
|sin(0
h
+ o) c
(0
h
-0
0
)tanq
- sin(0
0
+ o)] +
z

0
sIn[cosu
sIn ([-u)
[3]
and:
L

0
=
sIn(0
h
-0
0
)
sIn (0
0
+u)
-
sIn(0
h
+[

)
sIn (0
h
+u)sIn ([

-u)
|sin(0
h
+o) c
(0
h
-0
0
)tanq
- sin(0
0
+ o)] -
z

0
cos[
sIn ([-u)
[4]
4.2 Calculations of rates of work and internal dissipation
4.2.1 Rate of external work
The rates of work due to the slope material in the regions OB'C', OB'A', OA'C', A'AC' and ABB'A' are w 1, w 2,
w S, w 4 and w S, respectively. The rate of external work for region ABC'C is then found by the simple algebraic
summation, w 1 -w 2 -w S -w 4 +w S. After some simplification, it is found that the rates at which work is
done by the soil weight and seismic loading in the region ABC'C are:
w

wcght
= yr
0
3
(
1
-
2
-
3
-
4
+
5
) [5]
and:
w

scsmc
= k

yr
0
3
(
1
-
2
-
3
-
4
+
5
) +k
h
yr
0
3
(
1

-
2

-
3

-
4

+
5

) [6]
where y is the unit weight of slope material, is the angular velocity of the region ABC'C, k

and k
h
are
the vertical and horizontal seismic coefficients defined as the ratio of the inertia forces k

w and k
h
w to the
slope mass weight w, the
1
,
2
,
3
,
4
,
5
,
1

,
2

,
3

,
4

and
5

are non-dimensional functions, which

depend on geometry parameters 0
0
, 0
h
, [

and and reported in the appendix (Loukidis et al. 2003, Yang et

al. 2004).

Figure 1. Translational slope failure mechanism.

Figure 2. Rotational slope failure mechanism.
4.2.2 Rate of internal dissipation
The internal dissipation of energy occurs along the discontinuity surface B'C'. The differential rate of dissipation
of energy along the surface may be found by multiplying the differential areas,
d0
cosq
, of this surface by the
cohesion c times the tangential discontinuity in velocity, :cos, across the surface. The total internal
dissipation of energy is found by integration over the whole surface

nt
= ] c(:cos)
d0
cosq
0
h
0
0
=
c
0
2
o
2tanq
|c
(0
h
-0
0
)tanq
-1] [7]
4.3 Stability analysis
Equating the summation of external rate of work, Equations [5] and [6], to the rate of internal energy dissipation,
Equation [7], gives:
w

wcght
+w

scsmc
=

nt
[8]
and take the Equations [5] [6] and [7] into the above Equation [8], safety factor of slope can be given as:
F =
cjc
(0
h
-0
0
)tcnq
-1[
2y
0
tunq|(1+k

)(]
1
-]
2
-]
3
-]
4
+]
S
)+k
h
(]
1

-]
2

-]
3

-]
4

+]
S

)]
[9]
where = aictan (tanF), and the in functions of
1
,
2
,
3
,
4
,
5
,
1

,
2

,
3

,
4

and
5

are replaced
with aictan (tanF).
According the basic theory of upper bound theorem of limit analysis, the Equation [9] gives an upper bound for
the factor of safety F. The factor of safety F is a function of the geometrical parameters 0
0
, 0
h
, [

of the
collapse mechanism, and it is useful information only if it corresponds to its minimum value. Assume the
function for The factor of safety F is g(0
0
, 0
h
, [

). Seeking the values of 0

0
, 0
h
, [

, and substituting them

to g(0
0
, 0
h
, [

) until obtained a least upper bound for factor of safety F . In present work, the seeking for 0
0
,
0
h
, [

realized by a optimize iterative program composited by authors. Actually, The factor of safety F has a
minimum value when 0
0
, 0
h
, [

satisfy the conditions:

g
0
0
= u,
g
0
h
= u and
g
[

= u. Here u < 0
0
<
n
2
,
0
0
< 0
h
< n and[

[, and when [

= [, it means failure surface through the toe of the slope. Thus the
failure surface parameters I and becomes Equation [4] and:
=
sIn([-[

)
sIn[sIn[

jE -
z

0
sIn[cosu
sIn ([-u)
[ [10]
5 Example of application
A example is given in this paper. A cut slope is to be excavated in a normally consolidation clay. The slope has
the following parameters: the slope angle is [=50, the height of the slope is E =20m, the soil unit weight is
y=20kN/m
3
, top angle of the slope is o=0, the cohesion strength of slope material is c =28kN/m
2
, the friction
angle of slope material is =25, Earthquake loadings are horizontal coefficient k
h
range from 0 to 0.3 and
vertical coefficient k

range from 0 to k
h
. The objective of this case study is to find out the value of the factor
of safety F of this slope against earthquake loading subject to the tension failure region and what about the
form of failure surface.
5.1 Program confirmation
In order to confirm the correctness of the optimizing iterative program, It takes the earthquake loading is k
h
=0
and k

=0 without considering the tension failure region. To calculate the factor of safety and search for potential
failure surface, program confirmation is applied. Results obtained by present method and others existing
methods are shown in Figure 3. From the Figure 3, the factor of safety F and the form of failure surface which
was given by proposed method are in very good agreement with them obtained by several existing solutions.
Although according to the theory of upper bound limit analysis, the factor of safety obtained is only the upper
estimation of actual value, It is believed that the limit analysis method proposed is more appropriate as an
engineering tool than other available methods, especially when it shows that the factor of safety obtained are
lower than those from other methods.
5.2 Single horizontal earthquake loading
The influences of the tension failure to the factor of safety F and the form of failure surface under single
horizontal earthquake loading are shown in Figure 4 and Figure 5. It is seen that when ratio of depth of tension
failure region to height of slope Z/H increases, the factor of safety F is decreases slowly and the ratio of L/H is
decreases markedly under various earthquake loadings. For example, when the ratio of depth of tension failure
region to height of slope Z/H is extended from 0 to 0.3 under an earthquake loading of k
h
=0.2, the factor of
safety F is change from 1.0236 to 0.9988 (drop 2.5%), and the ratio of L/H is change from 0.2395 to 0.1056
markedly (drop 56%).
Figure 6 shows the form of failure surface and the factor of safety F under different ratio of Z/H for k
h
=0.2.
The present result is consistent with the ones from actual situation and shaking table test (Lin & Wang 2006).


Figure 3. Comparison of factor of safety and form of failure surface by methods of limit analysis and others.

Figure 4. Factor of safety F as a function of Z/H under different k
h
.
0 0.05 0.1 0.15 0.2 0.25 0.3
0.96
0.98
1
1.02
1.04
1.06
1.08
1.1
1.12
Z/H
F

kh=0
kh=0.1
kh=0.2
kh=0.3

Figure 5. Form parameter of failure surface, L/H as a function of Z/H under different k
h
.

Figure 6. Form of failure surface and factor of safety under different ratio of Z/H for k
h
=0.2.
5.3 Coupling earthquake loadings
The influences of the tension failure to the factor of safety F and the form of failure surface under coupling
earthquake loadings are shown in Figure 7 and Figure 8. It is seen that when ratio of depth of tension failure
region to height of slope Z/H increases, the factor of safety decreases slowly and the ratio of L/H decreases
markedly under various earthquake loadings. For example, when the ratio of depth of tension failure zone to
height of slope, Z/H is extended from 0 to 0.3 under coupled earthquake loadings of k
h
=0.2 and k

=0.2, the
factor of safety F is change from 0.8162 to 0.7948 (drop 2.6%), and the ratio of L/H is change from 0.3157 to
0.1488 markedly (drop 53%).
Figure 9 shows the form of failure surface and the factor of safety F under different situations. The present
result is accordance with the results which were obtained from actual situation and shaking table test.
0 0.05 0.1 0.15 0.2 0.25 0.3
0.08
0.12
0.16
0.2
0.24
0.28
Z/H
L
/
H

kh=0
kh=0.1
kh=0.2
kh=0.3


Figure 7. Factor of safety F as a function of Z/H under different coupling earthquake loadings.

Figure 8. Form parameter of failure surface, L/H as a function of Z/H under different coupled earthquake
loadings.
0 0.05 0.1 0.15 0.2 0.25 0.3
0.7
0.8
0.9
1
1.1
1.2
1.3
Z/H
F

kh=0 kv=0
kh=0.1 kv=0.1
kh=0.2 kv=0.2
kh=0.3 kv=0.3
0 0.05 0.1 0.15 0.2 0.25 0.3
0.1
0.15
0.2
0.25
0.3
0.35
Z/H
L
/
H

kh=0 kv=0
kh=0.1 kv=0.1
kh=0.2 kv=0.2
kh=0.3 kv=0.3

Figure 9. Form of failure surface and factor of safety under different situations.
6 Conclusions
This paper presents a method that incorporates a tension failure region into the limit analysis method to
investigate the influence of the depth of tension failure region on the factor of safety and the form of failure
surface for slope stability analysis under earthquake loading. Based on the analyses above, the conclusions are
drawn as follows:
(1) The value of the factor of safety F and the form of failure surface obtained using the limit analysis method
based on optimize iterative program are closer to those using other existing methods.
(2) The ratio of depth of tension failure region to height of slope Z/H has a slight influence on the factor of safety
under various earthquake loadings. And the form of failure surface obtained using the proposed method is
different from the results which were obtained by applying other existing methods, moreover, is accordance with
the results which were obtained from actual situation and shaking table test.
(3) Compared to the horizontal earthquake loading, the vertical earthquake loading has little influence to slope
stability analysis subject to the effect of tension failure.
7 Acknowledgements
This study has received financial support from Grants-in-Aid for Scientific Research (Scientific Research (B),
22310113, G. Chen) from J apan Society for the Promotion of Science. And the first writer acknowledges the
support of China Scholarship Council (CSC). Part of the work reported in the paper was carried out while the
first writer was a PhD student at Central South University, supported by a program of outstanding PhD thesis
support fund.
8 References
Bishop, AW. (1955). The use of the slip circle in the stability of slopes. Gotechnique 5: 7-17.
Chen, WF. (1975). Limit Analysis and Soil Plasticity. Elsevier Science Publishers, BV, Amsterdam, the Netherlands.
Huang, RQ., Pei, XJ ., Zhang, WF., et al. (2009). Further examination on characteristic and formation mechanism of
Daguangbao landslide. Journal of Engineering Geology 17: 725-736. (in Chinese)
Lin, ML., Wang, KL. (2006). Seismic slope behavior in a large-scale shaking table model test. Engineering Geology 86:
118-133.
Loukidis, D., Bandini, P., Salgado, R. (2003). Stability of seismically loaded slopes using limit analysis. Gotechnique 53:
463-479.
Newmark, NM. (1965). Effects of earthquakes on dams and embankments. Gotechnique 15: 139-160.
Seed, HB., Martin, GR. (1966). The seismic coefficient in earth dam design. Journal of Soil Mechanics. Foundations
Division, ASCE 92: 59-83.
Xu, Q., Pei, XJ., Huang RQ., et al. (2009). 5.12 Wenchuan Earthquake: Research on Large-scale Landslides. Science Press,
Beijing, China. (in Chinese)
Yan, ZX., Zhang, S., ZHANG, X., et al. (2010). Failure mechanism and stability analysis of slope under earthquake. Journal
of Engineering Geology 18: 844-850.
Yang, XL., Li, L., Yin, J H. (2004). Seismic and static stability analysis of rock slopes by a kinematical approach.
Gotechnique 54: 543-549.
Yin, YP., Wang, FW., Sun, P.. (2009). Landslide hazards triggered by the 2008 Wenchuan Earthquake, Sichuan, China.
Landslides 6: 139-151.
Zheng, YR., Ye, H., Huang, RQ. (2009). Analysis and discussion of failure mechanism and fracture surface of slope under
earthquake. Chinese Journal of Rock Mechanics and Engineering 28: 1714-1723.
Zienkiewica, OC., Humpheson, C., Lewis, RW. (1975). associated and non-associated visco-plasticity in soil mechanics.
Gotechnique 25: 671-689.
9 Appendix

1
=
(sIn0
h
+3tanqcos0
h
)c
3(0
h
-0
0
)tanq
-3tanqcos0
0
-sIn0
0
3(1+9tan
2
q)
[A1]

=
(3tanqsIn0
h
-cos0
h
)c
3(0
h
-0
0
)tanq
-3tanqsIn0
0
+cos0
0
3(1+9tan
2
q)
[A2]

2
=
1
6
j
L

0
+
z

0
cos[
sIn([-u)
[ ]2cos0
0
- j
L

0
+
z

0
cos[
sIn([-u)
[ coso sin(0
0
+o) [A3]

=
1
6
j
L

0
+
z

0
cos[
sIn([-u)
[ ]2sin0
0
+ j
L

0
+
z

0
cos[
sIn([-u)
[ sino sin(0
0
+o) [A4]

3
=
1
6
c
(0
h
-0
0
)tanq
_sin(0
h
-0
0
) -_
I
r
0
+
Z
r
0
cos[
sin([ - o)
_ sin(0
h
+o)_
]cos0
0
- j
L

0
+
z

0
cos[
sIn([-u)
[ coso +cos0
h
c
(0
h
-0
0
)tanq
[A5]

=
1
6
c
(0
h
-0
0
)tanq
_sin(0
h
- 0
0
) - _
I
r
0
+
Z
r
0
cos[
sin([ - o)
_ sin(0
h
+ o)_
]sin0
0
+j
L

0
+
z

0
cos[
sIn([-u)
[ sino + sin0
h
c
(0
h
-0
0
)tanq
[A6]

4
= |
H

0
-
z

0
sIn[cosu
sIn ([-u)
]
2
sIn ([-[

)
2sIn[sIn [

]cos0
0
- j
L

0
+
z

0
cos[
sIn([-u)
[ coso -
1
3
j
H

0
-
z

0
sIn[cosu
sIn([-u)
[ (cot[ +cot[

) [A7]

= |
H

0
-
z

0
sIn[cosu
sIn ([-u)
]
2
sIn ([-[

)
2sIn[sIn[

]sin0
0
-j
L

0
+
z

0
cos[
sIn([-u)
[ sino -
2
3
|
H

0
-
z

0
sIn[cosu
sIn ([-u)
] [A8]

5
=
1
2
[
z

2
cosucos[
sIn([-u)
]cos0
0
- j
L

0
+
1
3
z

0
cos[
sIn([-u)
[ coso +
z

0
L

0
coso(cos0
0
-
1
2
L

0
coso) [A9]

=
1
2
[
z

2
cosucos[
sIn([-u)
]sin0
0
+
L

0
sino -
1
3
z

0
+
z

0
L

0
coso [sin0
0
+
1
2
L

0
sino -
1
2
z

0
[A10]