Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)

ARTICLE
Limit equilibrium slope stability analysis using the nonlinear
strength failure criterion
Dong-ping Deng, Lian-heng Zhao, and Liang Li
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15

Abstract: The limit equilibrium stability analysis of two- and three-dimensional slopes with the nonlinear failure strength
criterion uses a number of variables to determine the normal and shear stress on the slip surface. The equation for the nonlinear
strength failure criterion is expressed using a Taylor series after analyzing the stress of an elemental slice or column. Multivar-
iate linear equations are then derived to determine these variables based on the force and moment equilibrium conditions the
sliding body is subject to. The stress on the slip surface can also be obtained to calculate the slope safety factor. The validity of
the current method was verified by comparing it with established examples. Charts were produced for slope stability analysis
with the nonlinear strength failure criterion under general conditions using the current method. The results of this study show
that the slope safety factor decreases with an increase in the geotechnical material parameter m in the nonlinear strength failure
criterion. The results of the current method are in close correspondence with other traditional limit equilibrium methods and
are more reliable than the Swedish method. The charts can be used to determine slope design parameters that meet specific
requirements.

Key words: two- and three-dimensional slopes, stability analysis, limit equilibrium, nonlinear strength failure criterion, safety
factor.

Résumé : L’analyse de la stabilité d’équilibre limite de pentes bidimensionnelles ou tridimensionnelles, avec prise en compte du
For personal use only.

critère de résistance à la rupture non linéaire, fait appel à un certain nombre de variables afin de déterminer la contrainte
normale et la contrainte de cisaillement sur une surface de glissement. L’équation du critère de résistance à la rupture non
linéaire est exprimée à l’aide d’une série de Taylor après analyse des contraintes exercées au niveau d’une tranche ou colonne
élémentaire. Des équations linéaires multidimensionnelles sont ensuite dérivées pour calculer ces variables à partir des condi-
tions d’équilibre des forces et des moments s’exerçant sur l’élément glissant sur la surface. On peut également déterminer la
contrainte exercée sur la surface de glissement afin de calculer le coefficient de sécurité des pentes. On a vérifié la validité de la
méthode actuelle en la comparant avec des exemples existants. Des diagrammes d’analyse de stabilité des pentes ont été réalisés,
en tenant compte du critère de résistance à la rupture non linéaire dans des conditions générales, à l’aide de la méthode actuelle.
Les résultats de la présente étude montrent que le coefficient de sécurité des pentes diminue lorsque le paramètre m du matériau
géotechnique augmente dans le cadre du critère de résistance à la rupture non linéaire. Les résultats de la méthode actuelle
concordent bien avec ceux d’autres méthodes traditionnelles d’équilibre limite et sont plus fiables que la méthode Suédoise. Les
diagrammes peuvent être utilisés pour déterminer les paramètres de conception des pentes qui respectent les exigences
spécifiques. [Traduit par la Rédaction]

Mots-clés : pentes bidimensionnelles et tridimensionnelles, analyse de stabilité, équilibre limite, critère de résistance à la rupture
non linéaire, coefficient de sécurité.

Introduction of the sliding body. Examples include the Swedish method

In slope stability analysis, it is important that an appropriate (Fellenius 1936), simplified Bishop method (Bishop 1955), Janbu
strength failure criterion is adopted to describe the loss of the method (Janbu 1973), Spencer method (Spencer 1967), Morgen-
structural integrity of the geotechnical material. Linear and non- stern–Price (M–P) method (Morgenstern and Price 1965), Sarma
linear strength failure criterions may be used, each with a certain method (Sarma 1973), and the stress method (Bell 1968; Zhu and
scope of application. However, the linear strength failure crite- Lee 2002). Different assumptions are used for the interslice forces,
rion is only one special case of the nonlinear strength failure and the number of static equilibrium conditions is not the same
criterion (Li 2007; Li et al. 2013). A large number of experimental in each of these methods, which leads to differences in the results
data show that the shear strength of most geotechnical materials obtained. The results of the Swedish method tend towards a more
can be better represented using a nonlinear strength failure cri- conservative safety factor for slope stability, as the method ig-
terion (Jiang et al. 2003; Eid 2010). Therefore, it is more meaning- nores the effect of the interslice forces. Compared with other
ful to analyze slope stability using the nonlinear strength methods, the Swedish method can easily calculate the normal and
failure criterion. shear stress on the slip surface, resulting in a convenient way of
In traditional limit equilibrium methods, a simple equation can analyzing slope stability using the nonlinear strength failure cri-
be derived for calculating the safety factor used to analyze slope terion (Hu et al. 2006). However, the Swedish method is only suit-
stability by solving the force and moment equilibrium equations able for a circular slip surface and satisfies only a portion of static

Received 29 March 2014. Accepted 11 September 2014.

D.-p. Deng, L.-h. Zhao, and L. Li. School of Civil Engineering, Central South University, Changsha 410075, China.
Corresponding author: Lian-heng Zhao (e-mail: [email protected]).

Can. Geotech. J. 52: 1–14 (2015) dx.doi.org/10.1139/cgj-2014-0111 Published at www.nrcresearchpress.com/cgj on 16 September 2014.
 Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
2 Can. Geotech. J. Vol. 52, 2015

equilibrium conditions. In other methods, it is difficult to utilize Fig. 1. Schematic of nonlinear strength failure criterion.
the nonlinear strength failure criterion (which is usually an expo-
nential equation) to derive a simple equation for calculating the τ
slope safety factor when the interslice forces are considered. How-
m=1
ever, most of the methods are suitable for arbitrary curve slip
surfaces and satisfy all the static equilibrium conditions. There- m>1
fore, establishing a simple limit equilibrium method that can be c0
applied under general conditions and incorporates the nonlinear
strength failure criterion would be very beneficial.
Previous studies have investigated two methods of converting
the nonlinear strength failure criterion into a form that makes it -σt φ 0 σ
easier to derive equations and perform calculations. The equiva-
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15

lent linear Mohr–Coulomb (M–C) criterion can be used instead of

the nonlinear strength failure criterion (Hoek 1990; Hoek et al. criterion, where c0 is the intercept of the ␶ axis and −␴t is the
2002) and is usually adopted in limit equilibrium methods. normal stress for a shear stress of ␶ = 0. It should be noted that this
However, the equivalent linear M–C criterion has a complex cal- does not apply to soils with zero effective cohesion, i.e., c0 = 0.
culation process and does not fully represent the nonlinear
characteristics of soil. The equivalent instantaneous linear M–C Slope safety factor definition
criterion can be used instead of the nonlinear strength failure As shown in Fig. 2, the slope safety factor of a slope subject to
criterion by using the tangent method (Collins et al. 1988; shear failure can be defined as the ratio of the failure shear force
Drescher and Christopoulos 1988; Yang and Yin 2004; Zhao et al. and the shear force on the slip surface:

冕
2010). At present, this is widely used for limit analysis methods
b
and has little application on limit equilibrium methods.
␶f dl
To use the limit equilibrium method for analyzing slope stabil- a

冕
ity using the nonlinear strength failure criterion, some variables (2) Fs ⫽ b
were used to determine the normal and shear stress on the slip ␶ dl
surface. After analyzing the stress of an elemental slice or column a

in the sliding body, the nonlinear strength failure equation was

For personal use only.

expressed using a Taylor series expansion. Multivariate linear where Fs is the slope safety factor; ␶ is the shear stress on the slip
surface; 冕a ␶ dl is the sum of the shear force along the entire slip
b
equations were then derived to determine these variables accord-
ing to the limit equilibrium conditions on the sliding body. The surface; ␶f is the failure shear stress calculated using eq. (1);
normal and shear stress on the slip surface were also obtained to 冕ab ␶f dl is the sum of the failure shear force along the entire slip
establish formulas for calculating the slope safety factor. The va- surface; and a and b are the lower and upper integration limits for
lidity of the current method was verified by comparing it with the two-dimensional slip surface, respectively.
established two- and three-dimensional slope examples. Charts
suitable for analyzing slope stability using the nonlinear strength Limit equilibrium method for slope stability analysis
failure criterion under general conditions were also generated. Stability analysis of two-dimensional slope
Compared with other limit equilibrium methods, the current As shown in Fig. 3, the coordinate system is established by
method has the following advantages: (i) the equation for calcu- setting the origin at the slope toe. The outer line of the slope is
lating the slope safety factor is simple and does not require denoted as g(x) and the slip surface as s(x). The elemental slice abcd
iterative computations; (ii) it satisfies all the static equilibrium in the two-dimensional sliding body is selected for stress analysis.
conditions on the sliding body; (iii) it can analyze slope stability The portions aefd and ebcf are located above and below the water
using both the linear and nonlinear strength failure criterion; table, respectively. The width of the elemental slice is dx, ␣ is the
(iv) it can be applied to the stability analysis of two- and three- horizontal inclination angle of the tangent of the midpoint of the
dimensional slopes; and (v) it can analyze slope stability for an slip surface in the elemental slice, and dl is the length of the slip
arbitrary curve slip surface. surface in the elemental slice with dl ⫽ dx/cos ␣. Under general
circumstances, the following forces act on the elemental slice:
Nonlinear strength failure criterion and definition
w dx is the gravitational force on the elemental slice.
for slope safety factor kHw dx is the horizontal seismic force where kH is the horizontal
Nonlinear strength failure criterion seismic action coefficient.
In geotechnical engineering, the shear strength of soil is gener- kVw dx is the vertical seismic force where kV is the vertical seis-
ally represented by a nonlinear strength failure criterion, partic- mic action coefficient. kV is positive when the vertical seismic
ularly for clays. The nonlinear strength failure criterion for soil force acts downwards and negative when the force acts upwards.
(Hobbs 1966; Yang and Chi 2013) can be expressed by qx dx is the horizontal external load on the slope.
qy dx is the vertical external load on the slope.

(1) 冉
␶f ⫽ c0 1 ⫹
␴
␴t 冊 1/m ␴ dl is the normal force on the slip surface where ␴ is the nor-
mal stress.
␶ dl is the shear force on the slip surface where ␶ is the shear
stress.
where ␴ is the normal stress on the slip surface; ␶f is the shear u dl is the pore-water pressure on the slip surface.
strength under the corresponding normal stress ␴; and c0, ␴t, and
m (where m ≥ 1) are geotechnical material parameters. Ignoring the interslice forces, the x- and y-direction force equi-
As shown in Fig. 1, eq. (1) expresses the linear M–C criterion librium equations of the elemental slice can be expressed as
when m = 1, where c0 is the cohesion, and ␸ is the internal friction
angle. When m > 1, eq. (1) expresses the nonlinear failure strength (3a) kHw ⫺ ␶0 ⫺ qx ⫹ (␴0 ⫹ u)s  ⫽ 0

Published by NRC Research Press

 Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
Deng et al. 3

Fig. 2. Schematic of slope subject to shear failure. w, gravitational force. Fig. 4. Stress analysis of elemental column in three-dimensional
sliding body: (a) three-dimensional sliding body; (b) elemental
column. H, slope height; ␤, slope angle.

y Width of three-dimensional
σ sliding body (~ ) w1
w Neutral surface
x
Slip surface
0 τ Left sliding body
w2
Fig. 3. Stress analysis of elemental slice in two-dimensional sliding body.

qy
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15

B z
qx d s(x) (a)
g(x) k Vw 0 x

H
a
y f y
k Hw
e c Saturation line Elemental column
b w Right sliding body £
0 α
A x σ τ
Slip surface u
dy
(b)
Table 1. Calculation parameters in stability analysis of two-dimensional qz
D1 C1
slope.
Parameter Formula dx
qy
qx
⫺冕a ␴0s  dx
a11 b A1 B1
For personal use only.

k Vw
a12 冕ab ␶01 dx z
a13 冕ab ␶02 dx k Hw
αy(-) αy(+)
b1 冕ab 共kHw ⫺ qx ⫹ us 兲 dx w
C y
z x D
a21 冕 dx b
a ␴0
αx
a22 冕ab ␶01s  dx 0 y τ σ u
a23 冕ab ␶02s  dx A B
αy
b2 冕ab 冋共1 ⫺ kV兲w ⫹ qy ⫺ u册 dx
a31 冕ab ␴0共x ⫹ ss 兲 dx the slope. Using the definition for the slope safety factor (eq. (2)),
a32 冕ab ␶01共⫺s ⫹ xs 兲 dx the following relationship can be obtained: ␶ ⫽ ␶f/Fs. Substituting this
into eq. (1) yields a modified expression for the shear stress:
a33 冕ab ␶02共⫺s ⫹ xs 兲 dx
b3
冕ab 兵冋共1 ⫺ kV兲w ⫹ qy ⫹ u册x ⫹ us s ⫺ 21 kHw共s ⫹ g兲 ⫹ qxg其 dx (5) ␶⫽
1
冉
c 1⫹
Fs 0
␴
␴t 冊 1/m

Expanding eq. (5) in a Taylor series gives


(3b) (1 ⫺ kV)w ⫹ qy ⫺ ␶0s ⫺ (␴0 ⫹ u) ⫽ 0

where ␴0 is the normal stress on the slip surface; ␶0 is the shear

(6) ␶⫽
1
Fs 冉
c0 1 ⫹
␴0
␴t 冊 1/m
⫹
1 d␶
Fs d␴
(␴ ⫺ ␴0) ⫹ H(␴ ⫺ ␴0)

stress on the slip surface; and s= is first derivative of the slip

surface equation s(x), i.e., s= = tan␣.
The equation of the normal stress ␴0 on the slip surface in the where
d␶
⫽
d␴ m␴t
c0
1⫹
␴t冉
␴ 共1⫺m兲/m
冊
; and H(␴ − ␴0) represents the higher
absence of the interslice forces can be obtained by solving eq. (3) to order error terms.
give To calculate the normal stress ␴ on the slip surface easily and
effectively, ␴ can be related to ␴0 using the variable ␭1:
[(1 ⫺ kV)w ⫹ qy] ⫹ (⫺kHw ⫹ qx)s 
(4) ␴0 ⫽ ⫺u (7) ␴ ⫽ ␭1␴0
1 ⫹ (s )2

In the limit equilibrium slope stability analysis, the safety fac- where ␴0 is calculated in the absence of the interslice forces. A
tor of the slip surface in the elemental slice is the same as that of more accurate ␴0 obtained using other methods can also be ap-

Published by NRC Research Press

 Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
4 Can. Geotech. J. Vol. 52, 2015

Table 2. Calculation parameters in stability analysis of three-dimensional slope.

Parameter Formula
c11 A␴0nx␴ ⌬ dx dy
c12 y
A␴0nx␴ sin ␣y ⌬ dx dy
W
c13 A␶01n␶x ⌬ dx dy
c14 y
A␶01n␶x sin ␣y ⌬ dx dy
W
c15 A␶02n␶x ⌬ dx dy
c16 y
A␶02n␶x sin ␣y ⌬ dx dy
W
d1 A共kHw ⫺ qx ⫺ unx␴ ⌬兲 dx dy
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15

c21 A␴0ny␴ ⌬ dx dy
c22 y
A␴0ny␴ sin ␣y ⌬ dx dy
W
c23 A␶01n␶y ⌬ dx dy
c24 y
A␶01n␶y sin ␣y ⌬ dx dy
W
c25 A␶02n␶y ⌬ dx dy
c26 y
A␶02n␶y sin ␣y ⌬ dx dy
W
d2 ⫺A共qy ⫹ uny␴ ⌬兲 dx dy
c31 A␴0nz␴ ⌬ dx dy
c32 y
A␴0nz␴ sin ␣y ⌬ dx dy
W
c33 A␶01n␶z ⌬ dx dy
c34 y
A␶01n␶z sin ␣y ⌬ dx dy
W
For personal use only.

c35 A␶02n␶z ⌬ dx dy
c36 y
A␶02n␶z sin ␣y ⌬ dx dy
W
d3 A冋共1 ⫺ kV兲w ⫹ qz ⫺ unz␴ ⌬册 dx dy
c41 A␴0共ny␴s ⫺ nz␴ y兲 ⌬ dx dy
c42 y
A␴0共ny␴s ⫺ nz␴ y兲 sin ␣y ⌬ dx dy
W
c43 A␶01共n␶ys ⫺ n␶z y兲 ⌬ dx dy
c44 y
A␶01共n␶ys ⫺ n␶z y兲 sin ␣y ⌬ dx dy
W
c45 A␶02共n␶ys ⫺ n␶z y兲 ⌬ dx dy
c46 y
A␶02共n␶ys ⫺ n␶z y兲 sin ␣y ⌬ dx dy
W
d4 ⫺A兵冋共1 ⫺ kV兲w ⫹ qz册y ⫹ qyg ⫹ u共ny␴s ⫺ nz␴ y兲 ⌬其 dx dy
c51 A␴0共⫺nx␴s ⫹ nz␴ x兲 ⌬ dx dy
c52 y
A␴0共⫺nx␴s ⫹ nz␴ x兲 sin ␣y ⌬ dx dy
W
c53 A␶01共⫺n␶s ⫹ n␶ x兲 ⌬ dx dy
x z

c54 y
A␶01共⫺n␶xs ⫹ n␶z x兲 sin ␣y ⌬ dx dy
W
c55 A␶02共⫺n␶xs ⫹ n␶z x兲 ⌬ dx dy
c56 y
A␶02共⫺n␶xs ⫹ n␶z x兲 sin ␣y ⌬ dx dy
W
d5
A兵 1
其
冋共1 ⫺ kV兲w ⫹ qz册x ⫹ qxg ⫺ kHw共g ⫹ s兲 ⫺ u共⫺nx␴s ⫹ nz␴ x兲 ⌬ dx dy
2
c61 A␴0共nx␴ y ⫺ ny␴ x兲 ⌬ dx dy
c62 y
A␴0共nx␴y ⫺ ny␴ x兲 sin ␣y ⌬ dx dy
W
c63 A␶01共n␶xy ⫺ n␶y x兲 ⌬ dx dy
c64 y
A␶01共n␶xy ⫺ n␶y x兲 sin ␣y ⌬ dx dy
W
c65 A␶02共n␶xy ⫺ n␶y x兲 ⌬ dx dy
c66 y
A␶02共n␶xy ⫺ n␶y x兲 sin ␣y ⌬ dx dy
W
d6 A冋qyx ⫺ 共qx ⫺ kHw兲y ⫺ u共nx␴y ⫺ ny␴ x兲 ⌬册 dx dy

Published by NRC Research Press

 Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
Deng et al. 5

plied in the current method and does not change the calculation Fig. 5. Schematic of slope in example 1.
procedure.
Substituting the aforementioned relation into eq. (6) yields B

R=
30
(8a) ␶ ⫽ ␭2␶01 ⫹ ␭3␶02

m
H = 20 m
冉 冊
1.5
␴0 1/m
γ = 18.82 kN/m3
(8b) ␶01 ⫽ c0 1 ⫹
␴t y 1 c0 = 42.65 kPa
σt = 155.44 kPa

(8c) ␶02 ⫽
c0
m␴t
1⫹ 冉
␴0
␴t 冊 (1⫺m)/m
␴0 A
0
θ
x
Circular slip surface
Arbitrary curve slip surface
α1
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15

where ␭2 and ␭3 are the unknown variables. In the current

method, ␭2 and ␭3 consider the effect of H(␴ − ␴0) in eq. (6) on shear
stress. When ␴ is close to ␴0, ␭1 tends to 1 and ␭3 tends to 0, which
shows that the second term on the right side of eq. (8a) has little 冕 冉 冊
a
b
c0 1 ⫹
␴
␴t
1/m 1
cos␣
dx

冕
contribution on the shear stress ␶. If ␴ is the same as ␴0 (i.e., ␭1 = 1), (12) Fs ⫽ b
1
eq. (5) does not need be expanded in a Taylor series. In other ␶ dx
words, ␭2 ⫽ 1/Fs and ␭3 = 0 in eq. (8a). When this special situation a cos␣
appears, ␴0 is actually the normal stress ␴ that needs to be calcu-
lated; thus, the solutions obtained according to the static equilib- From the aforementioned derivation, it can be shown that the
rium equations for the sliding body would be consistent with the normal stress ␴0, used usually in the Swedish method, has been
results discussed earlier in the text. The current method therefore also assumed to be the initial normal stress to calculate the nor-
accounts for this special case. mal stress ␴ on the slip surface in the current method. However,
The x- and y-direction force equilibrium equations and the mo- all static equilibrium conditions are considered in the current
ment equilibrium equations about a point (xc, yc) for a two- method. Therefore, the normal stress ␴ in the current method is
closer to the actual normal stress, and the results of slope analysis
For personal use only.

dimensional sliding body can be expressed as

are also stricter than that in the Swedish method. When a circular

冕 冕
b b slip surface is considered in a simple slope, the current method
(9a) (⫺ ␴ s  ⫹ ␶) dx ⫺ (kHw ⫺ qx ⫹ us ) dx ⫽ 0 can be reduced to the Swedish method if ␭1 = 1, and only the
a a equations for i = 1, 2 in eq. (11) are used.

(9b) 冕 b

a
(␴ ⫹ ␶ s ) dx ⫺ 冕
b

a
[(1 ⫺ kV)w ⫹ qy ⫺ u] dx ⫽ 0
Stability analysis of a three-dimensional slope
When the sliding body is a symmetric body in a three-dimensional
slope, the approximate solution for slope stability can be obtained

冕兵 b by the two-dimensional simplified slope model. However, when

(9c) [(1 ⫺ kV)w ⫹ qy ⫹ u](x ⫺ xc) ⫹ us (yc ⫺ s) the sliding body is an asymmetric body, there is a clear difference
a between the stability of a three-dimensional slope and a two-

共 s⫹g
兲 其
dimensional simplified slope. Therefore, it is necessary to estab-
⫹ kHw yc ⫺ ⫺ qx(yc ⫺ g) dx lish a three-dimensional limit equilibrium method based on
2

冕 b the two-dimensional method discussed earlier in the text. In

⫺ [(⫺ ␴ s  ⫹ ␶)(yc ⫺ s) ⫹ (␴ ⫹ ␶ s )(x ⫺ xc)] dx ⫽ 0 the stability analysis of a three-dimensional slope, the assump-
a tions for the two-dimensional slope regarding the normal and
shear stress on the slip surface are still applicable. However,
Substituting eqs. (9a) and (9b) into eq. (9c) and simplifying yields as six limit equilibrium equations must be satisfied for a three-
dimensional sliding body (compared with three limit equilibrium

冕兵
equations for a two-dimensional sliding body), these assumptions
其
b
1
(10) [(1 ⫺ kV)w ⫹ qy ⫹ u]x ⫹ us s ⫺ k w(s ⫹ g) ⫹ qxg dx have to be extended. Compared with the two-dimensional slice
a 2 H method, the direction of the normal and shear stress on the slip

⫺ 冕
a
b
[⫺(⫺ ␴ s  ⫹ ␶)s ⫹ (␴ ⫹ ␶ s )x] dx ⫽ 0
surface in the three-dimensional column has to be determined
analytically.
As shown in Fig. 4a, the two-dimensional symmetry slip surface
or the two-dimensional main slip surface in a three-dimensional
Ternary linear equations for variables ␭1, ␭2, and ␭3 can then be sliding body is taken as the neutral surface. The coordinate system
obtained by substituting eqs. (7) and (8) into eqs. (9a), (9b), and (10): is established by setting the origin at the slope toe of the neutral
surface. The positive x direction points toward the slope, the pos-

兺
3 itive z direction is opposite to the direction of gravity, and the
(11) aij␭j ⫽ bi, i ⫽ 1, 2, 3 positive y direction is determined by the right-hand rule. The
j⫽1
three-dimensional sliding body has width W and is divided into
left and right portions by the neutral surface. The widths of the
The equations for aij and bi are presented in Table 1. left and right portions are w1 and w2, respectively. The equation of
Once the variables ␭1–␭3 are determined, the normal stress ␴ the outer surface is z = g(x, y), and the equation of the three-
and shear stress ␶ on the slip surface can be determined by sub- dimensional slip surface is z = s(x, y).
stituting them into eqs. (7) and (8). The safety factor of a two- As shown in Fig. 4b, the elemental column ABCDA1B1C1D1 is
dimensional slope can then be calculated using selected for stress analysis. The widths of the elemental column in

Published by NRC Research Press

 Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
6 Can. Geotech. J. Vol. 52, 2015

Table 3. Comparison of calculation results using linear M–C criterion in example 1.

Calculation method
Swedish Simplified Spencer Current
Type of slip surface method Bishop method method M–P method method
Circular slip surface 1.345 1.415 1.410 1.408 1.416
Arbitrary curve slip surface — — 1.425 1.429 1.442

the x and y directions are dx and dy, respectively; and ␣x and ␣y are Fig. 6. Comparison of calculation results using nonlinear strength
the horizontal inclination angles of the slip surface of the elemental failure criterion in example 1.
column in the x–z and y–z plane, respectively. ⌬ dx dy is the area of the
1.5
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15

slip surface of the elemental column, where ⌬ ⫽ 兹1⫹sx2⫹sy2, sx = tan␣x,

and sy = tan␣y. Under general conditions, the forces acting on the ele-
1—Current method
1.4
mental column ABCDA1B1C1D1 are as follows: 2—Swedish method
1.3 1
w dx dy is the gravitational force acting on the elemental
column. 1.2

Fs
kHw dx dy is the horizontal seismic force where the definition of
1.1
kH is consistent with that in the two-dimensional slope.
kVw dx dy is the vertical seismic force where the definition and 1.0
sign of kV is consistent with that for the two-dimensional slope. 2
qx dx dy is the external load on the slope in the x direction. 0.9
qy dx dy is the external load on the slope in the y direction. 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
qz dx dy is the external load on the slope in the z direction. m
␴⌬ dx dy is the normal force on the slip surface where ␴ is the
normal stress.
␶⌬ dx dy is the shear force on the slip surface where ␶ is the Fig. 7. Schematic of slope in example 2.
shear stress.
For personal use only.

u⌬ dx dy is the pore-water pressure on the slip surface.

R=
B

18
The definitions of the other parameters such as n␴x , n␶x, n␴y , n␶y, n␴z , soil 1

m
H = 10 m

and n␶z are listed in Appendix A. Based on the equations relating 2 P3

these parameters, the x-, y-, and z-direction force equilibrium con- 1
dition, and ignoring the effect of the interslice forces, the normal y
θ P2
stress ␴0 in the absence of the interslice forces in the three- P1 soil 2
P0
dimensional column can be obtained (eq. A6). The detailed deri- A x
vation of ␴0 is presented in Appendix A. 0 P4
α1 soil 3
In the stability analysis of a two-dimensional slope, assumptions Circular slip surface Arbitrary curve slip surface
are made regarding the normal and shear stress on the slip surface,
which can also be applied to the three-dimensional slope. However,
there are six limit equilibrium equations for a three-dimensional
sliding body. Therefore, we amend eqs. (7) and (8) to give (14d) A 兵(␴ ⫹ u)n ⌬(y z
␴ c ⫺ y) ⫹ (␴ ⫹ u)n␴y ⌬(s ⫺ zc)

(13a) 共
␴ ⫽ ␭1 ⫹ ␭2
y
W 兲
sin␣y ␴0 ⫹ ␶n␶y ⌬(s ⫺ zc)⫹␶n␶z ⌬( yc ⫺ y) ⫺ [(1 ⫺ kV)w ⫹ qz](yc ⫺ y)
⫺ qy(zc ⫺ g)其 dx dy ⫽ 0
␶ ⫽ 共␭ sin␣ 兲␶ ⫹ 共␭ 兲
y y
(13b) 3 ⫹␭ 4 y 01 5 ⫹ ␭6 sin␣y ␶02
A 兵(␴ ⫹ u)n ⌬(x ⫺ x ) ⫹ (␴ ⫹ u)n ⌬(z
W W z x
(14e) ␴ c ␴ c ⫺ s)
where ␭1–␭6 are the unknown variables; y is the y coordinate of ⫹ ␶n␶x ⌬(zc ⫺ s) ⫹ ␶n␶z ⌬(x ⫺ xc) ⫺ [(1 ⫺ kV)w ⫹ qz](x ⫺ xc)

冋 册其 dx dy ⫽ 0
the center of the elemental column; W is the width of three-
dimensional sliding body; and ␣y is the horizontal inclination 1
⫺ qx(g ⫺ zc) ⫺ kHw zc ⫺ (g ⫹ s)
angle of slip surface in the y–z plane. ␶01 and ␶02 are also calculated 2
using eq. (8).
The stress analysis of the elemental column shown in Fig. 4b
yields the x-, y-, and z-direction force equilibrium equations and
(14f) A 冋(␴ ⫹ u)n ⌬(y ⫺ y ) ⫹ (␴ ⫹ u)n ⌬(x
x
␴ c
y
␴ c ⫺ x)

the moment equilibrium equations around the x, y, and z axes ⫹ ␶n␶x ⌬(y ⫺ yc) ⫹ ␶n␶y ⌬(xc ⫺ x)⫺(qx ⫺ kHw)(yc ⫺ y)
⫺ qy(x ⫺ xc)册 dx dy ⫽ 0
about a point (xc, yc, zc) for a three-dimensional body:

(14a) A 冋(␴ ⫹ u)n ⌬ ⫹ ␶n ⌬ ⫺ k w ⫹ q 册 dx dy ⫽ 0

x
␴
x
␶ H x
Substituting eqs. (14a)–(14c) into eqs. (14d)–(14f) and simplifying
yields
(14b) A 冋(␴ ⫹ u)n ⌬ ⫹ ␶n ⌬ ⫹ q 册 dx dy ⫽ 0
y
␴
y
␶ y
(15a) A 兵⫺(␴ ⫹ u)n ⌬y ⫹ (␴ ⫹ u)n ⌬s ⫹ ␶n ⌬s ⫺ ␶n ⌬y
z
␴
y
␴
y
␶
z
␶

(14c) A 冋(␴ ⫹ u)n ⌬ ⫹ ␶n ⌬ ⫺ (1 ⫺ k )w ⫺ q 册 dx dy ⫽ 0

z
␴
z
␶ V z ⫹ [(1 ⫺ kV)w ⫹ qz]y ⫹ qyg其 dx dy ⫽ 0

Published by NRC Research Press

 Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
Deng et al. 7

Table 4. Comparison of calculation results using linear M–C criterion in example 2.

Calculation method
Swedish Simplified Spencer Current
Type of slip surface method Bishop method method M–P method method
Circular slip surface 1.320 1.486 1.456 1.458 1.450
Arbitrary curve slip surface — — 1.522 1.539 1.603

(15b) A 兵(␴ ⫹ u)n ⌬x ⫺ (␴ ⫹ u)n ⌬s ⫺ ␶n ⌬s ⫹ ␶n ⌬x

z
␴
x
␴
x
␶
z
␶
Fig. 8. Comparison of calculation results of different methods using
nonlinear strength failure criterion in example 2: (a) m1 = m2 = m3 = m;

⫺ [(1 ⫺ kV)w ⫹ qz]x ⫺ qxg ⫹

1
其
k w(g ⫹ s) dx dy ⫽ 0
2 H
(b) m1 = m3 = 1.0.
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15

1.5

(15c) A 冋(␴ ⫹ u)n␴x ⌬y ⫺ (␴ ⫹ u)n␴y ⌬x ⫹ ␶n␶x ⌬y ⫺ ␶n␶y ⌬x 1.4 1 1—Current method

2—Swedish method
⫹ (qx ⫺ kHw)y ⫺ qyx册 dx dy ⫽ 0 1.3
1.2

Fs
Finally, substituting eq. (13) into eqs. (14a)–(14c) and eqs. (15a)–
1.1
(15c) yields a set of equations for the six variables ␭1–␭6: 2
1.0 (a)

兺
6
(16) cij␭j ⫽ di, i ⫽ 1, 2, Ê, 6 0.9
j⫽1

1.00 1.02 1.04 1.06 1.08 1.10 1.12 1.14 1.16 1.18 1.20
where cij and di are calculated using the equations listed in Table 2. m
Once the variables ␭1–␭6 are known, the normal stress ␴ and
shear stress ␶ can be obtained by solving eq. (13). Note that in the
three-dimensional symmetric sliding body, the y-direction force
1.5
For personal use only.

equilibrium equation and the moment equilibrium equations 1 (b)

around the x and z axes about a point (xc, yc, zc) are in closed form,
1—Current method
which leaves only three equations (eqs. (14a), (14c), and (15b)) to 1.4
solve these variables. We therefore let ␭2 = 0, ␭4 = 0, and ␭6 = 0. The 2—Swedish method
variables ␭1, ␭3, and ␭5 can be solved according to eqs. (14a), (14c),
1.3
Fs

and (15b). The safety factor of a three-dimensional slope (Fs) can

then be determined using
1.2 2

(17) Fs ⫽
A 冉
c0 1 ⫹
␴
␴t 冊 1/m
⌬ dx dy
1.1
A ␶⌬ dx dy 1.0 1.1 1.2 1.3 1.4 1.5
m2
1.6 1.7 1.8 1.9 2.0

Results and comparison with classical examples

Two-dimensional slopes
AB; and ␣1 is the horizontal inclination angle of the first plotline
Example 1 (Arai and Tagyo 1985) that is positive in the clockwise direction. The following parame-
A schematic of this slope is shown in Fig. 5. The slope height is ters were used: n = 100; x0 ⫽ 共xB ⫺ xA兲/n; and ␣1 = 10°.
H = 20 m, and slope ratio is 1:1.5. The soil parameters are as fol- For m = 1.0, i.e., the linear M–C criterion with an equivalent
lows: unit weight ␥ = 18.82 kN/m3; c0 = 41.65 kPa; and ␴t = cohesion of c = 41.65 kPa and an internal friction angle of ␸ = 15°,
155.44 kPa. The coordinate system is established by setting the
the results of the Swedish method, simplified Bishop method,
origin at the slope toe. Two kinds of slip surfaces are considered
for the slope stability analysis: circular and arbitrary curve. The Spencer method, M–P method, and the current method are listed
coordinates of points A and B, the upper and lower boundary in Table 3. For the circular slip surface and m > 1.0, i.e., the non-
points of the slip surface, are (0, 0) and (38, 20), respectively, linear strength failure criterion, the results of the Swedish
expressed in metres. For the circular slip surface, the arc radius is method and the current method are shown in Fig. 6. From Table 3
R = 30 m. The arbitrary curve slip surface (Deng et al. 2011) is and Fig. 6, the following observations can be made:
formed by plotlines of n sections:
(1) When m = 1.0, the results obtained by the current method are
(i ⫺ 1)i close to those of other methods and are slightly higher than
(18a) xi ⫽ ix0 ⫺ [nx ⫺ (xB ⫺ xA)] ⫹ xA those of the Swedish method. However, in the current
(n ⫺ 1)n 0
method, the normal stress ␴ on the slip surface is obtained by
(18b) yi ⫽ (xi ⫺ xA) tan i 冋 2(n(n␪ ⫹⫹ ␣1))册 ⫹ y
1
A
linearly amending the initial stress ␴0 and satisfies all the
limit equilibrium conditions. In the Swedish method, ␴0 is
the same as the normal stress on the slip surface. Therefore,
where (xi, yi) are the coordinates of the end point in the ith plo- the results of the current method are stricter.
tline; (xA, yA) are the coordinates of point A; xB is the y-axes coor- (2) When m > 1.0, the results obtained by the current method are
dinate of point B; n is the number of plotlines; x0 is the width of also higher than those of the Swedish method. The slope
the first plotline; ␪ is the horizontal inclination angle of the line safety factor decreases with increases in m.

Published by NRC Research Press

 Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
8 Can. Geotech. J. Vol. 52, 2015

Table 5. Comparison of results of slope stability analysis with limit analysis method and limit equilibrium method.
Parameter, m 1.0 1.2 1.4 1.6 1.8 2.0
Critical slope height, Hcr (m) (Zhang and Chen 1987) 72.900 56.475 48.690 43.650 40.950 39.510
Calculation method
Zhang and Chen's method (Zhang and Chen 1987) 1.000 1.000 1.000 1.000 1.000 1.000
Swedish method 0.958 0.961 0.963 0.972 0.969 0.958
Current method 1.000 0.997 0.994 0.999 0.993 0.980

Based on this analysis, the current method is suitable for the Fig. 9. Schematic of slope in example 4.
stability analysis of slopes using the linear or nonlinear strength
O B
failure criterion and can obtain a reliable solution of the safety factor.
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15

R
Example 2 (Donald et al. 1989) =
40
A schematic of the slope is shown in Fig. 7. The slope height (H) m

H = 40 m
is 10 m, and the slope ratio is 1:2. There are three soil layers with
different parameters. For soil 1, the parameters are ␥1 = 19.5 kN/m3,
c01 = 3.0 kPa, ␴t1 = 3.84 kPa, and m1. For soil 2, the parameters are
␥2 = 19.5 kN/m3, c02 = 5.3 kPa, ␴t2 = 12.49 kPa, and m2. For soil 3, the z γ = 22 kN/m3
parameters are ␥3 = 19.5 kN/m3, c03 = 7.2 kPa, ␴t3 = 19.78 kPa, and c0 = 30 kPa
m3. The coordinate system is established by setting the origin at β σt = 51.96 kPa
the slope toe. The soil layer interface is described by the points A x
P0–P4, with coordinates (0, 0), (10, 2), (20, 4), (24, 6), and (22, –1), 0
respectively. The circular and arbitrary curve slip surfaces are
considered. The coordinates of points A and B, the upper and
lower boundary points of the slip surface, are (0, 0) and (22, 10), Fig. 10. Results of current method for example 4.
respectively, expressed in metres. For the circular slip surface, the 1.3
arc radius is R = 18 m. For the arbitrary curve slip surface, the
parameters are given as follows: n = 100; x0 ⫽ 共xB ⫺ xA兲/n; and ␣1 = 1.2
For personal use only.

12°. It should be noted that in Donald’s example, c01 = 0 kPa for 1.1
soil 1, while the nonlinear strength failure criterion in the current
1.0
method does not apply to this particular situation. We therefore
Fs

amended this material parameter of soil 1. 0.9

When m1, m2, and m3 are assumed to be equal to m and m = 1.0 0.8
(i.e., the linear M–C criterion), with an equivalent cohesion of c1 =
3.0 kPa and an internal friction angle ␸1 = 38° for soil 1, an equiv- 0.7
alent cohesion of c2 = 5.3 kPa and an internal friction angle of ␸2 = 0.6
23° for soil 2, and an equivalent cohesion of c3 = 7.2 kPa and an
1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0
internal friction angle of ␸3 = 20° for soil 3, the results of the m
current method are compared with other methods in Table 4.
When m1 = m2 = m3 = m and m > 1.0 (i.e., the nonlinear strength
failure criterion), the results of the current method are compared using the current method and Swedish method based on the cir-
with the Swedish method in Fig. 8a. When m1 = m3 = 1.0 and cular slip surface. The results calculated by Zhang and Chen’s
m2 changes from 1.0 to 2.0, the results of the current method and method, the Swedish method, and the current method are listed
Swedish method are shown in Fig. 8b. From Table 4 and Fig. 8, the in Table 5. The difference between the results of the limit analysis
following observations can be made: method and the current method is <2%. The results of the current
method are also larger and more reasonable than those of the
(1) In the stability analysis of the layered slope, the current
Swedish method with increases in m in the nonlinear strength
method produces results that are close to other methods.
failure criterion. The current method considers all conditions of
They are also higher than those of the Swedish method be-
static equilibrium that the sliding body would be subject to and
cause they satisfy all of the limit equilibrium conditions.
amends the initial stress ␴0 on the slip surface in the absence of
(2) In the stability analysis of the layered slope, as one or all soil
the interslice forces.
layers are subject to the nonlinear strength failure criterion,
the results of the current method are larger than those of the Three-dimensional slopes
Swedish method. This indicates that the current method is
also suitable for stability analysis in the layered slope using Example 4 (Zhang 1988)
the nonlinear strength failure criterion. Figure 9 is a cross-sectional schematic of a three-dimensional
slope. The slope height is H = 40 m, and the slope angle is ␤ = 45°,
Example 3 with the following soil parameters: ␥ = 22 kN/m3; c0 = 30 kPa; and
Zhang and Chen (1987) assumes a slope angle of ␤ = 45° in a ␴t = 51.96 kPa. The three-dimensional slip surface is assumed to be
homogeneous slope, with the following soil parameters: ␥ = 20 kN/m3; an ellipsoid defined by 冋共x ⫺ x0兲2册/a2 ⫹ 冋共z ⫺ z0兲2册/a2 ⫹ y2/b2 ⫽ 1
c0 = 90 kPa; and ␴t = 247.3 kPa. Zhang and Chen (1987) presented the with x0 = 0 m, z0 = 40 m, a = 40 m, and b ⫽ W/兹2, where W is the
critical slope height Hcr calculated using the limit analysis – upper width of three-dimensional sliding body. The coordinate system is
bound method when the minimum slope safety factor (Fs,min) is established by setting the origin at the slope toe of the two-
equal to 1.0. To compare the difference between the results of the dimensional main slip surface. The coordinates of points A and B,
limit analysis method and the limit equilibrium analysis method the upper and lower boundaries of the two-dimensional main slip
using the nonlinear strength failure criterion, Fs,min in the slope surface, are (0, 0, 0) and (40, 0, 40), respectively, expressed in
for a given critical slope height is searched for and calculated metres. The radius of the two-dimensional main slip surface is R =

Published by NRC Research Press

 Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
Deng et al. 9

Fig. 11. Schematic of slope in example 5: (a) two-dimensional main slip surface; (b) width of three-dimensional sliding body. hmax, maximum
depth; ␤01 and ␤02, maximum horizontal inclination angles of left and right sliding bodies in y–z plane, respectively.

H = 12.2 m
(a)
2
z 1
θ hmax
A 0 x
α1

z
w1 is respectively 20 m, 25 m, 30 m, 35 m, 40 m, 45 m, (b)
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15

w2
50 m, 55 m, 60 m, 65 m, 70 m, 75 m and 80 m

β01 β02
hmax
y
0

40 m. When W = 8H, the results of the current method with Fig. 12. Safety factor of slope in example 5.
respect to changes in m are shown in Fig. 10.
1.80
When m = 1.0 (i.e., the linear M–C criterion), with an equivalent 1—Safety factor for three-dimensional slip surface
cohesion of c = 30 kPa and an internal friction angle of ␸ = 30°, the 1.75 2—Safety factor for two-dimensional main slip surface
slope safety factor calculated by the current method is 1.171. This
1.70
result lies between the values of 1.260 and 1.154 obtained from the
improved three-dimensional safety factor method (Wang and 1.65 1
Fs

Deng 2003) and the three-dimensional ordinary slice method

1.60
(Hovland 1997), respectively. It is close to the values of 1.193, 1.142, 2
For personal use only.

1.209, 1.214, and 1.163 obtained from the three-dimensional sim- 1.55
plified Bishop method (Hungr 1987), three-dimensional simplified
1.50
Janbu method (Hungr 1989), three-dimensional Spencer method
40 50 60 70 80 90 100 110 120 130 140 150 160
(Zhang 1988), three-dimensional M–P method (Chen and Zhu W (m)
2010), and three-dimensional Sarma method (Deng and Li 2013),
respectively. This illustrates that the current method is suitable

冋 2(n(n␪ ⫹⫹ ␣1))册 ⫹ y
for analyzing the stability of three-dimensional slopes. In Fig. 10,
the safety factor of the slope decreases with an increase in m, which 1
(19e) z1i ⫽ (xi ⫺ xA) tan i A
is consistent with the trend observed in the two-dimensional slope.

z ⫽ y tan冋j 册 (left sliding body)

Example 5 (Zhang 1988) 2(␤ ⫹ ␣ ) 01 01
(19f)
A schematic of the slope is shown in Fig. 11a. The slope height is 2i
m(m ⫹ 1)
i

H = 12.2 m, with a slope ratio of 1:2. The soil parameters are ␥ =

z ⫽ ⫺y tan冋j 册 (right sliding body)
19.2 kN/m3, c0 = 29.3 kPa, ␴t = 80.50 kPa, and m = 1.5. The coordi- 2(␤ ⫹ ␣ ) 02 02
(19g)
nate system is established by setting the origin at the slope toe of 2i
m(m ⫹ 1)
i

the two-dimensional main slip surface. The slope stability analysis

is performed using a three-dimensional arbitrary curve slip sur-
where (xA, yA) are the coordinates of point A; xB is y-axes coordi-
face at specific positions. The coordinates of points A and B, the
nate of point B; x0 is the width of the first row parallel to the y axis;
upper and lower boundaries of the two-dimensional main slip
surface, are (−5.71, 0, 0) and (28.72, 0, 12.20), respectively, ex- y01 and y02 are the widths of the first column parallel to the x axis
pressed in metres. in the left and right sliding bodies, respectively; ␣1 is the horizon-
The three-dimensional arbitrary curve slip surface can be gen- tal inclination angle of the first plotline of the first column paral-
erated based on the method proposed by Li et al. (2011). The three- lel to x axis, as shown in Fig. 11a; ␣01 and ␣02 are the horizontal
dimensional sliding body is divided into n × 2m columns (as shown inclination angles of the first plotline of the first row parallel to
in Fig. 4a) where n is the number of rows parallel to the y axis, and y axis in the left and right sliding bodies, respectively; ␪ is the
2m is the number of columns parallel to the x axis. The left and horizontal inclination angle of the line AB, as shown in Fig. 11a;
right sliding bodies are both divided into m columns, and their ␤01 and ␤02 are the maximum horizontal inclination angles of the
widths are w1 and w2, respectively. The coordinates of the ith row left and right sliding bodies in the y–z plane, as shown in Fig. 11b,
and jth column are calculated as follows: respectively; ␤01 ⫽ arctan 共hmax /w1兲; ␤02 ⫽ arctan 共hmax /w2兲; and
hmax is the maximum depth between the two-dimensional main
(i ⫺ 1)i slip surface and the outline of slope. It should be noted that the
(19a) xi ⫽ ix0 ⫺ [nx ⫺ (xB ⫺ xA)] ⫹ xA
(n ⫺ 1)n 0 point calculated by eq. (19) should not exceed the outline of the
slope.
(j ⫺ 1)j
(19b) yi ⫽ jy01 ⫺ (my01 ⫺ w1) (left sliding body) In this example, the three-dimensional symmetric arbitrary
(m ⫺ 1)m
curve slip surface is considered, and the parameters are given as
(j ⫺ 1)j follows: n = 100; m = 100; x0 ⫽ 共xB ⫺ xA兲/n; y01 ⫽ w1/m; y02 ⫽
(19c) yi ⫽ ⫺jy02 ⫹ (my02 ⫺ w2) (right sliding body) w2/m; ␣1 = 18°; ␣01 = 0°; and ␣02 = 0°. The safety factor of slope for
(m ⫺ 1)m
the three-dimensional slip surface and two-dimensional main slip
(19d) zi ⫽ z1i ⫹ z2i surface calculated by the current method are shown in Fig. 12.

Published by NRC Research Press

 Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
10 Can. Geotech. J. Vol. 52, 2015

Fig. 13. Charts for slope stability analysis using nonlinear strength failure criterion.
0.21
0.19 (a)
0.17 β = 30°

H
0.15 γ‫ޔ‬c0‫ޔ‬σt‫ޔ‬m m = 1.0
£

c0/γH
0.13
0.11 Fs is respectively 1.0, 1.1,
0.09
1.2, 1.3, 1.4, and 1.5
0.07
0.05
0.03
0.07 0.12 0.17 0.22 0.27 0.32 0.37 0.42 0.47 0.52 0.57 0.62
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15

c0/σt

0.25
0.23
(b)
0.21
β = 30°

H
0.19 γ‫ޔ‬c0‫ޔ‬σt‫ޔ‬m
£ m = 1.5
0.17
c0/γH

0.15
0.13 Fs is respectively 1.0, 1.1,
0.11 1.2, 1.3, 1.4, and 1.5
0.09
0.07
0.05
0.07 0.17 0.27 0.37 0.47 0.57 0.67 0.77 0.87 0.97 1.07 1.17 1.27
For personal use only.

c0/σt

0.25
0.23 (c)
0.21 β = 30°
H

0.19 γ‫ޔ‬c0‫ޔ‬σt‫ޔ‬m
£ m = 2.0
0.17
c0/γH

0.15
Fs is respectively 1.0, 1.1,
0.13
1.2, 1.3, 1.4, and 1.5
0.11
0.09
0.07
0.07 0.27 0.47 0.67 0.87 1.07 1.27 1.47 1.67 1.87 2.07
c0/σt

From Fig. 12, it can be seen that the safety factor of the three- To verify the feasibility of these charts, two examples are cited
dimensional slope decreases and approaches that of the two- to analyze slope stability as follows:
dimensional slope with an increase in W.
Example 6
Charts for slope stability analysis using nonlinear A slope height of H = 12 m and a slope angle of ␤ = 45°. The soil
strength failure criterion parameters are ␥ = 18 kN/m3, c0 = 24 kPa, ␴t = 73.86 kPa, and
Compared with the three-dimensional slope, the two-dimensional m = 1.0.
simplified slope is generally more stable and safer. In engineer-
ing applications, the three-dimensional slope is usually simpli- Example 7
fied into a two-dimensional slope to analyze its stability. A slope height of H = 12 m and a slope angle of ␤ = 45°. The soil
Therefore, the current method is adopted to establish some parameters are ␥ = 18 kN/m3, c0 = 30 kPa, ␴t = 96.79 kPa, and
practical charts for slope stability analysis using the nonlinear m = 1.4.
strength failure criterion. For unit weight of ␥ = 17.8 kN/m3 This gives c0/␥H ⫽ 0.111 and c0/␴t ⫽ 0.325 for example 6 and
(suitable for clays), slope heights of H = 10, 15, and 20 m, slope c0/␥H ⫽ 0.139 and c0/␴t ⫽ 0.310 for example 7. The slope safety
angles of ␤ = 30°, 45°, and 60°, and m = 1.0, 1.5, and 2.0, the factor in examples 6 and 7 are 1.286 and 1.258, respectively, using
dimensionless parameters c0/␥H and c0/␴t are used to establish
the charts in Fig. 13 and the linear interpolation method. Com-
contour curves of safety factor. This forms the charts for slope
stability analysis using the nonlinear strength failure criterion paring with the calculated minimum slope safety factor (1.286
shown in Fig. 13. It should be noted that when c0/␥H is given, the for example 6, and 1.243 for example 7), there are only small
slope height has little effect on the safety factor of the slope. differences between the two methods. The charts can therefore
Therefore, the charts in Fig. 13 can be used for analyzing slope be used for slope stability analysis using the nonlinear strength
stability with different slope heights. failure criterion. Furthermore, the charts could be used to de-

Published by NRC Research Press

 Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
Deng et al. 11

Fig. 13 (continued).

0.25
c0/σt
0.23 (d)
0.21
β = 45°

H
0.19 γ‫ޔ‬c0‫ޔ‬σt‫ޔ‬m
£ m = 1.0
0.17

c0/γH
0.15
Fs is respectively 1.0, 1.1,
0.13
0.11 1.2, 1.3, 1.4, and 1.5
0.09
0.07
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15

0.05
0.07 0.17 0.27 0.37 0.47 0.57 0.67 0.77 0.87 0.97
c0/σt

0.27
0.25
(e)
0.23
β = 45°
0.21 H γ‫ޔ‬c0‫ޔ‬σt‫ޔ‬m
0.19 £ m = 1.5
c0/γH

0.17
0.15 Fs is respectively 1.0, 1.1,
0.13 1.2, 1.3, 1.4, and 1.5
0.11
0.09
0.07
0.07 0.27 0.47 0.67 0.87 1.07 1.27 1.47 1.67
c0/σt
For personal use only.

0.27
0.25
(f)
0.23
β = 45°
H

0.21 γ‫ޔ‬c0‫ޔ‬σt‫ޔ‬m
£ m = 2.0
0.19
c0/γH

0.17 Fs is respectively 1.0, 1.1,

0.15
1.2, 1.3, 1.4, and 1.5
0.13
0.11
0.09
0.07 0.27 0.47 0.67 0.87 1.07 1.27 1.47 1.67 1.87 2.07 2.27
c0/σt

sign the parameters of slopes that meet certain safety require- validity was verified by comparing it with established exam-
ments. ples. Charts suitable for slope stability analysis using the non-
linear strength failure criterion under general conditions are
Example 8 obtained with the current method. The following conclusions
The slope ratio is designed to be equal to 1:1.5, i.e., the slope
can be drawn from this study:
angle is ␤ = 33.69°. The soil parameters are ␥ = 18 kN/m3,
c0/␴t ⫽ 0.325, and m = 1.3. Let the minimum slope safety factor be (1) The current method can be used to analyze the stability of
equal to 1.2. This results in c0/␥H ⫽ 0.0916 for ␤ = 33.69°, m = 1.3, two- and three-dimensional slopes and can obtain reliable
and c0/␴t ⫽ 0.325 using the charts in Fig. 13. Likewise, if the min- solutions of safety factor.
imum slope safety factor is equal to 1.3 or 1.4, this results in
(2) In current method, the normal stress ␴0 in the absence of
c0/␥H ⫽ 0.1049 or c0/␥H ⫽ 0.1185, respectively.
interslice forces was applied as the initial stress to calculate
When c0/␥H is known, the slope height H or soil parameter c0
can be obtained, and the results are listed in Table 6. Some obser- the normal stress on the slip surface. More accurate initial
vations can be made based on the results in Table 6: (i) for a given stresses obtained by other methods are also suitable for the
c0/␥H, the slope height H has little effect on the slope safety factor; current method.
and (ii) the results obtained by the charts in Fig. 13 are close to the (3) The slope safety factor decreases with an increase of m in
calculated safety factor of slope. This illustrates that the charts the nonlinear strength failure criterion. In the homoge-
can be used to design the parameters of slopes that meet specific neous slope, the slope height H has little effect on the slope
requirements. safety factor for a given value of the dimensionless param-
eter c0/␥H.
Conclusions (4) The charts for slope stability analysis using the nonlinear
The limit equilibrium slope stability analysis method using strength failure criterion can be used to determine parame-
the nonlinear strength failure criterion was derived, and its ters of slopes that meet specific design requirements.

Published by NRC Research Press

 Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
12 Can. Geotech. J. Vol. 52, 2015

Fig. 13 (concluded).

0.29
0.27
(g)
0.25
β = 60°

H
0.23 γ‫ޔ‬c0‫ޔ‬σt‫ޔ‬m
£ m = 1.0
0.21
0.19
c0/γH 0.17 Fs is respectively 1.0, 1.1,
0.15 1.2, 1.3, 1.4, and 1.5
0.13
0.11
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15

0.09
0.07
0.07 0.17 0.27 0.37 0.47 0.57 0.67 0.77 0.87 0.97 1.07 1.17
c0/σt

0.29
0.27 (h)
0.25
β = 60°
H

0.23 γ‫ޔ‬c0‫ޔ‬σt‫ޔ‬m
£ m = 1.5
0.21
c0/γH

0.19
0.17 Fs is respectively 1.0, 1.1,
0.15 1.2, 1.3, 1.4, and 1.5
0.13
For personal use only.

0.11
0.09
0.07 0.27 0.47 0.67 0.87 1.07 1.27 1.47 1.67 1.87 2.07
c0/σt

0.31
0.29
(i)
0.27
β = 60°
H

0.25 γ‫ޔ‬c0‫ޔ‬σt‫ޔ‬m
0.23 £ m = 2.0
c0/γH

0.21
0.19 Fs is respectively 1.0, 1.1,
0.17 1.2, 1.3, 1.4, and 1.5
0.15
0.13
0.11
0.07 0.27 0.47 0.67 0.87 1.07 1.27 1.47 1.67 1.87 2.07 2.27 2.47 2.67 2.87
c0/σt

Table 6. Summary of specific design parameters and slope safety factor in example 8.
Parameter Value
Design requirement, Fs,min = 1.2
H (m) 10 12 14 16 18 20
c0 (kPa) 16.49 19.79 23.09 26.39 29.69 32.99
Calculated safety factor of slope 1.163 1.163 1.163 1.164 1.164 1.164
Design requirement, Fs,min = 1.3
H (m) 10 12 14 16 18 20
c0 (kPa) 18.89 22.66 26.44 30.22 34.00 37.77
Calculated safety factor of slope 1.262 1.262 1.262 1.262 1.262 1.262
Design requirement, Fs,min = 1.4
H (m) 10 12 14 16 18 20
c0 (kPa) 21.34 25.61 29.87 34.14 38.41 42.68
Calculated safety factor of slope 1.361 1.361 1.361 1.361 1.361 1.361
Note: Fs,min, minimum safety factor.

Published by NRC Research Press

 Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
Deng et al. 13

References stability using a nonlinear failure criterion. Computers and Geotechnics,

54(10): 185–191. doi:10.1016/j.compgeo.2013.06.007.
Arai, K., and Tagyo, K. 1985. Determination of noncircular slip surface giving the
Yang, X.-L., and Yin, J.-H. 2004. Slope stability analysis with nonlinear failure
minimum factor of safety in slope stability analysis. Soils and Foundations,
25(1): 43–51. doi:10.3208/sandf1972.25.43. criterion. Journal of Engineering Mechanics, 130(3): 267–273. doi:10.1061/
(ASCE)0733-9399(2004)130:3(267).
Bell, J.M. 1968. General slope stability analysis. Journal of the Soil Mechanics and
Foundations Division, ASCE, 94(SM6): 1253–1270. Zhang, X. 1988. Three-dimensional stability analysis of concave slopes in plan
Bishop, A.W. 1955. The use of the slip circle in the stability analysis of earth view. Journal of Geotechnical Engineering, ASCE, 114(6): 658–671. doi:10.
slopes. Géotechnique, 5(1): 7–17. doi:10.1680/geot.1955.5.1.7. 1061/(ASCE)0733-9410(1988)114:6(658).
Chen, C.F., and Zhu, J.F. 2010. A three-dimensional slope stability analysis pro- Zhang, X.J., and Chen, W.F. 1987. Stability analysis of slopes with general non-
cedure based Morgenstern-Price method. Chinese Journal of Rock Mechanics linear failure criterion. International Journal for Numerical and Analytical
and Engineering, 29(7): 1473–1480. [In Chinese.] Methods in Geomechanics, 11(1): 33–50. doi:10.1002/nag.1610110104.
Collins, I.F., Gunn, C.L.M., Pender, M.J., and Wang, Y. 1988. Slope stability anal- Zhao, L.-h., Li, L., Yang, F., Luo, Q., and Liu, X. 2010. Upper bound analysis of slope
yses for materials with a non-linear failure envelope. International Journal stability with nonlinear failure criterion based on strength reduction tech-
for Numerical and Analytical Methods in Geomechanics, 12(5): 533–555. doi: nique. Journal of Central South University of Technology, 17(4): 836–844.
10.1002/nag.1610120507. doi:10.1007/s11771-010-564-7.
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15

Deng, D.P., and Li, L. 2013. Quasi-rigorous and non-rigorous 3D limit equilibrium Zhu, D.Y., and Lee, C.F. 2002. Explicit limit equilibrium solution for slope stabil-
methods for generalized-shaped slopes. Chinese Journal of Geotechnical En- ity. International Journal for Numerical and Analytical Methods in Geome-
gineering, 35(3): 501–511. [In Chinese.] chanics, 26(15): 1573–1590. doi:10.1002/nag.260.
Deng, D.P., Li, L., and Zhao, L.H. 2011. A new method of sliding surface searching Zhu, D.Y., Ding, X.L., and Qian, Q.H. 2007. Three-dimensional limit equilibrium
for general stability of slope based on Janbu method. Rock and Soil Mechan- solution to generalized-shaped slope stability. Chinese Journal of Geotechni-
ics, 32(3): 891–898. [In Chinese.] cal Engineering, 29(10): 1460–1464. [In Chinese.]
Donald, I.B., Giam, P.S.K., and Monash University. 1989. Example problems for
testing soil slope stability programs. Department of Civil Engineering, Mo- Appendix A
nash University, Clayton, Vic. n␴x and n␶x are the x direction cosines of the normal and
Drescher, A., and Christopoulos, C. 1988. Limit analysis slope stability with
nonlinear yield condition. International Journal for Numerical and Analyti- shear stress on the slip surface in the elemental column,
cal Methods in Geomechanics, 12(3): 341–345. doi:10.1002/nag.1610120307. respectively.
Eid, H.T. 2010. Two- and three-dimensional analyses of translational slides in n␴y and n␶y are the y direction cosines of the normal and
soils with nonlinear failure envelopes. Canadian Geotechnical Journal, 47(4): shear stress on the slip surface in the elemental column,
388–399. doi:10.1139/T09-110.
Fellenius, W. 1936. Calculation of the stability of earth dams. In Proceedings of respectively.
the 2nd Congress on Large Dams, Washington, D.C., pp. 445–462. n␴z and n␶z are the z direction cosines of the normal and
Hobbs, D.W. 1966. A study of the behaviour of a broken rock under triaxial shear stress on the slip surface in the elemental column,
compression, and its application to mine roadways. International Journal of
respectively.
For personal use only.

Rock Mechanics and Mining Sciences & Geomechanics Abstracts, 3(1): 11–43.
doi:10.1016/0148-9062(66)90030-1. Based on the aforementioned definitions, the direction cosines
Hoek, E. 1990. Estimating Mohr-Coulomb friction and cohesion values from the n␴x , n␴y , and n␴z can be expressed as
Hoek-Brown failure criterion. International Journal of Rock Mechanics and
Mining Sciences & Geomechanics Abstracts, 27(3): 227–229. doi:10.1016/0148-
sx
9062(90)94333-O. (A1a) n␴x ⫽ ⫺
Hoek, E., Carranza-Torres, C., and Corkum, B. 2002. Hoek-Brown failure crite- ⌬
rion - 2002 edition. In Proceedings of the North American Rock Mechanics
Society NARMS-TAC 2002, University of Toronto Press, Toronto, Ont., sy
pp. 67–273. (A1b) n␴y ⫽ ⫺
Hovland, H.J. 1997. Three-dimensional slope stability analysis method. Journal of ⌬
Geotechnical Engineering Division, 103(9): 971–986.
1
Hu, W.D., Zhu, X.N., and Li, X.Q. 2006. Slope stability analysis with nonlinear (A1c) n␴z ⫽
failure criterion. Journal of Hunan Institute of Science and Technology (Nat- ⌬
ural Sciences), 19(3): 84–91. [In Chinese.]
Hungr, O. 1987. An extension of Bishop’s simplified method of slope stability
analysis to three dimensions. Géotechnique, 37(1): 113–117. Note that the direction cosines of the normal and shear stress
Hungr, O., Salgado, F.M., and Byrne, P.M. 1989. Evaluation of a three- on the slip surface in the elemental column obey the following
dimensional method of slope stability analysis. Canadian Geotechnical Jour- relationships:
nal, 26(4): 679–686. doi:10.1139/t89-079.
Janbu, N. 1973. Slope stability computations. In Embankment Dam Engineering:
Casagrande Memorial Volume. Edited by E. Hirschfield and S. Poulos. John (A2a) 共n␶x兲2 ⫹ 共n␶y兲2 ⫹ 共n␶z兲2 ⫽ 1
Wiley, New York. pp. 47–86.
Jiang, J.-C., Baker, R., and Yamagami, T. 2003. The effect of strength envelope
nonlinearity on slope stability computations. Canadian Geotechnical Jour- (A2b) n␶xn␴x ⫹ n␶yn␴y ⫹ n␶zn␴z ⫽ 0
nal, 40(2): 308–325. doi:10.1139/t02-111.
Li, L., Deng, D.P., and Zhao, L.H. 2011. Application of new method of arbitrary
sliding surface search to stability analysis of complex three-dimensional In stability analysis of three-dimensional slope, the direction of
slopes. Chinese Journal of Geotechnical Engineering, 33(4): 544–553. [In the shear stress in the elemental column is usually assumed to be
Chinese.] parallel to the x–z plane (Zhu 2007), therefore
Li, N.S., Tang, B., Tan, F.J., and Xie, L.H. 2013. Slope stability analysis of earth-rock
dams based on unified strength criterion by genetic algorithm. Rock and Soil
Mechanics, 34(1): 243–249. [In Chinese.] (A3) n␶y ⫽ 0
Li, X. 2007. Finite element analysis of slope stability using a nonlinear failure
criterion. Computers and Geotechnics, 34(3): 127–136. doi:10.1016/j.compgeo.
2006.11.005. Substituting eqs. (A1) and (A3) into eq. (A2), and noting that n␶x is
Morgenstern, N.R., and Price, V.E. 1965. The analysis of the stability of general always positive, n␶x and n␶z can be calculated using
slip surfaces. Géotechnique, 15(1): 79–93. doi:10.1680/geot.1965.15.1.79.
Sarma, S.K. 1973. Stability analysis of embankments and slopes. Géotechnique,
23(3): 423–433. doi:10.1680/geot.1973.23.3.423. n␴z
Spencer, E. 1967. A method of analysis of the stability of embankments assuming (A4a) n␶x ⫽
parallel inter-slice forces. Géotechnique, 17(1): 11–26. doi:10.1680/geot.1967.17.
1.11.
兹共n␴x 兲2 ⫹ 共n␴z 兲2
Wang, Y.X., and Deng, H.K. 2003. An improved method for three-dimensional
slope stability analysis. Chinese Journal of Geotechnical Engineering, 25(5): n␴x
611–614. [In Chinese.] (A4b) n␶z ⫽ ⫺
Yang, X.-g., and Chi, S.-c . 2013. Upper bound finite element analysis of slope 兹共n␴x 兲2 ⫹ 共n␴z 兲2
Published by NRC Research Press
 Pagination not final (cite DOI) / Pagination provisoire (citer le DOI)
14 Can. Geotech. J. Vol. 52, 2015

As shown in Fig. 4b and ignoring the effect of the interslice where ␴0 and ␶0 are the normal and shear stress in the absence of
forces, the x-, y-, and z-direction force equilibrium conditions can the interslice forces on the slip surface of the elemental column,
be expressed by the following set of equations: respectively.
Substituting eqs. (A1), (A3), and (A4) into eq. (A5), the normal
kHw ⫺ qx stress ␴0 on the slip surface of the elemental column can be cal-
(A5a) (␴0 ⫹ u)n␴x ⫹ ␶0n␶x ⫺ ⫽0
⌬ culated by combining eqs. (A5a) and (A5c):

qy
(A5b) (␴0 ⫹ u)n␴y ⫹ ␶0n␶y ⫹ ⫽0 [(1 ⫺ kV)w ⫹ qz]n␶x ⫺ (kHw ⫺ qx)n␶z
⌬
(A6) ␴0 ⫽ ⫺u
(1 ⫺ kV)w ⫹ qz ⌬(n␴z n␶x ⫺ n␴x n␶z)
(A5c) (␴0 ⫹ u)n␴z ⫹ ␶0n␶z ⫺ ⫽0
⌬
Can. Geotech. J. Downloaded from www.nrcresearchpress.com by San Francisco (UCSF) on 02/19/15
For personal use only.

Published by NRC Research Press