Scientific Research and Essay Vol.4 (9), pp.

842-852, September, 2009

Available online at http://www.academicjournals.org/SRE
ISSN 1992-2248 2009 Academic Journals

Review

Stability analysis of pile-slope system

Seyhan Frat
Sakarya University, Technical Education Faculty, Department of Structure, Geotechnical Division, Esentepe Campus,
54187, Sakarya, Turkey. E -mail: sefirat@sakarya.edu.tr. Tel.: +90 264 2956473, Fax: +90 264 2956424.
Accepted 5 August, 2009

An accurate estimation of the lateral force is an important key point for the stability analysis, because
the lateral force influences both pile and slope stabilities. Therefore, overestimation of the lateral forces
on the pile will naturally lead to conservative pile design and un-conservative slope stability design and
vice versa. This is because the estimated lateral loads are used as extra resisting force(s) on slope
stability calculations due to the pile reaction. In this paper, 2 different methods of lateral force
calculations namely plastic deformation and visco-plastic flow are obtained on a pile that installed into
slope which is susceptible to failure. First, the case of a slope without piles is considered and the
methods are then used to calculate the safety factor for the slope. Then, the stability of slopes
reinforced with piles is analyzed. Expressions are derived allowing the force needed to increase the
safety factor to a desired value and the most suitable location of piles within the slope to be evaluated.

Key words: Lateral soil movement, soil-pile interaction, finite difference, slope stability, safety factor.

INTRODUCTION

Failures of mass movements often result in extensive
property damage and loss of human life. It is recognized
that ensuring the stability of both natural and man-made
slopes continues to be a fundamental issue in geotechni-
cal engineering. There is no universally accepted method
for the prevention or correction of landslides. Each slide
is unique and should be considered on the basis of
unique inherent characteristics.
Stabilization of a slope may depend on a number of
factors such as; its geometry, surface and groundwater
conditions, strength of materials and the reason for stabi-
lization. A number of techniques have been developed to
stabilize slopes considering the above mentioned condi-
tions (Abramson et al., 1996).
Piles have been used successfully in many situations in
order to stabilize slopes or to improve slope stability (Ito
and Matsui, 1975; Anagnastopoulos et al., 1991;
Popescu, 1991; Poulos, 1995; Chen and Poulos, 1997
and Won et al., 2005) and numerous methods have been
developed for the analysis of piled slopes.
The limit equilibrium method was used by Ito et al.
(1979) to deal with the problem of the stability of slopes
containing piles. In this approach the safety factor of the
piled slope was defined as the ratio of the resisting mo-
ment to the overturning moment acting on the potentially
unstable soil mass. The driving moment and the resisting
moment due to soil shearing resistance were obtained
applying the ordinary slice method. To calculate the resis-
ting moment due to the piles, Ito et al. (1979) proposed
the use of the theoretical equation, derived previously by
Ito and Matsui (1975), to evaluate the lateral force acting
on a row of piles due to soil movement.
A similar approach was developed by Lee et al. (1995)
in which Bishop's simplified method (Bishop, 1955) was
employed to find the critical sliding surface for the slope
as well as the driving moment and resisting moment due
to soil shearing resistance. These forces were calculated
using a procedure based on the boundary element
method.
Hassiotis et al. (1997) have extended the friction circle
method to incorporate the pile reaction in slope stability
analysis. The Ito and Matsui equation (1975) has been
used to evaluate the lateral force that the failing soil mass
exerts on a row of piles.
Ausilio et al. (2001) have used the kinematic approach
of limit analysis for the stability of slopes that are rein-
forced with piles. The case of a slope without piles is first
considered, and a solution is proposed to determine the
safety factor of the slope, which is defined as a reduction
coefficient for the strength parameters of the soil. Then,
the stability of a slope containing piles is analyzed. To
account for the presence of the piles, it is assumed that a
lateral force and a moment are applied at the depth of the
potential sliding surface. For simplicity, the effect of pore
 Firat 843

water pressure on slope stability is not considered. result of strain-softening caused by the movement in a
The influence of the one row of pile groups on the sta- potential landslide. For this reason it is assumed that the
bility of the weathered slope was investigated by Jeong et state of plastic equilibrium occurs only in the soil just
al. (2003) based on an analytical study and a numerical around the piles (Figure 1), satisfying the Mohr-Coulomb
analysis. An uncoupled analysis, in which the pile yield criterion. Therefore, the lateral force acting on the
response and slope stability are considered separately, pile may be estimated neglecting the change of equili-
was performed by quantifying the load transfer of passive brium condition of the whole slope. When the stress
pile groups subjected to lateral soil movements in slope. distribution in the soil AEBBEA is considered, the fric-
Recently, Martin and Chen (2005) have used displace- tional forces on surfaces AEB and AEB are neglected.
ment method to evaluate the response of piles caused by Therefore, a theoretical equation for the lateral force p
an embankment slope in a translational failure mode, per unit length may be given as follows:
induced by a weak layer or a liquefied layer beneath the
embankment. The analysis includes the kinematic p = Ac
1
exp
D1 D2
N tan tan

+ 2 N
12
tan 1
loading acting on the bridge piles caused by lateral soil N tan D2 8 4
movements and the effects of spatial variation of soil
displacement on the response of piles. 2 tan + 2 N 1 2 + N 1 2 2 tan + 2 N 1 2 + N 1 2 1 2
The interaction behaviour between pile and soil is a + 12
c D1 12
2 D 2 N
N tan + N 1 N tan + N 1
complicated phenomenon due to its 3-dimensional nature
and can be influenced by many factors, such as the cha-
racteristics of deformation and the strength parameters of (N 12
tan + N 1)
z

D1 D1 D2
+ D1 exp N tan tan + D2
both pile and soil. In addition, movement of the failure N D2 D2 8 4
mass (landslide) involves a complicated mechanism. It is
recognized that the pile design for slope stabilization is (1)
too simplified for the complicated event. Nevertheless,
there is no precise design method for stabilizing piles so Where N =tan 2 [( 4)+( 2)] and
as to utilize effectively the pile effect in slope stabilization. (N
12
tan + N 1 )
In fact, many slopes failed or conti-nuously moved even D1
A = D1
after the installation of stabilizing piles (Ito et al., 1982). D2
In this paper three different methods of lateral force
estimations have been identified. 2 of these methods are Hence, the total corresponding lateral force induced per
coded to provide predictive numerical values for the unit length of the pile due to the plastically deforming soil
lateral forces on piles installed for stabilizing slopes. Al- layer will be obtained by the integration of Equation (1)
though historically there are different methods for lateral along the soil depth layer (or until a critical slip surface).
force estimations, only the developments due to Ito and In the case of a purely frictional soil (putting the
Matsui (1975) and De Beer and Carpentier (1977) and cohesion c = 0 in Equation (1) the lateral force, p, per unit
current study are presented. Soil reinforcement by means length may be obtained by integrating Equation (2) along
of piles bored through the sliding mass to the stable the depth, z:
underlying soil is often used as a countermeasure to sta-
bilize landslides. With this case the following design steps (N 12
tan + N 1 )
z

applied; Evaluating the total shear force needed to D1 D1 D2
p= D1 exp N tan tan + D2
increase the safety factor of the slope by the desired N D2 D2 8 4
value and calculating the maximum lateral force that each (2)
pile can receive from the sliding soil and transmits to the
stable underlying soil. In the case of a cohesive soil, the angle of friction may
be neglected that is, = 0. Therefore, the lateral force p
Forces on piles in soil undergoing lateral movement\ per unit length may be obtained as the function of depth
Ito and Matsui method as follows:

Ito and Matsui (1975) and Ito et al. (1979, 1982) have D1 D1 D2
p =c D1 3 log + tan 2 ( D1 D2 ) + z ( D1 D2 ) -
developed an analysis to show that piles placed in D2 D2 8
plastically deforming ground, such as landslide mass (3)
movement, can prevent further plastic deformations.
The theory proposed by the above authors, called the
theory of plastic deformation, estimates the lateral force De Beer and Carpentier method
between the 2 extremes of zero and large landslide mass
movement. It is assumed that no reduction in shear The theory of plastic deformation was first proposed by
resistance along the sliding surface has taken place as a Ito and Matsui (1975) and later discussed by De Beer and
Sci. Res. Essays 844

Pile
[(/4)(/2)]
F B
D
[(/8)+(/4)]
C
=[(/4)+(/2)]
E A
y
D1 D2
0 x
Direction of
deformation
E' A'
C'

D'
F' B'
Pile

Figure 1. Plastic state of soil just around piles (Ito and Matsui, 1975).

Carpentier (1977). The latter authors developed compa- F1 ( ) D1 D2

F2 ( )
rable equations by modifying the Ito and Matsuis method z sin D1 D2
p= 1+ N D1 e D2 (5)
and gave the following equations as another way of N 2 D2
obtaining the lateral force p per unit length induced on the
piles. For the case of cohesive soil, putting = 0 and by inte-
grating along the depth of the pile, the lateral force per
z sin D1
F1 ( ) D1 D2
F2 ( ) unit length may be obtained as:
D2
p= 1+ N D1 e D2 + c cot g
N 2 D2
D1 D D2 3
(D1 D2 )F1 ( ) sin
D1 D2
F2 ( ) 1+
sin
N p = c D1 2 ln + 1 tan (D1 D2 ) + (D1
D1 1+ N e D2
D1 D2 2 + D2 D2 D2 8 2
N 2 N
(6)
(4)
Visco-plastic flow
Where;
N The visco-plastic flow, past a row of piles, is presented in
F1 ( ) = (1 sin ) tan + N (1 sin ) 1 following sections. The assumptions for which may be

tan + summarized as follows (Figure 2);
4 2
and 1) The visco-plastic flow is two-dimensional and is
2 assumed to be uniform in the direction of depth.
1 sin
F2 ( ) = 2
tan tan + 2) Flow past a cylinder is stationary, steady and
1 + sin 8 4 symmetrical.
3) The fluid is deemed incompressible.
4) Coupled non-linear pair of partial differential equations
The lateral force, p, acting on the piles is obtained by was used to obtain a stream function and vorticity
integrating Equation (4) along the depth of the soil layer. using finite difference solution techniques.
For the case of cohesionless soil c = 0 in Equation (4), 5) In order to analyze the flow past a row of piles, a
by integrating along the depth the lateral force per unit purely viscous fluid model has been used.
length may be obtained 6) The soil layer is represented by a Bingham plastic mo-
 Firat 845

, j
Far field
Ro

a
Flow velocity y
0
=0 V
(V) x
r

Figure 2. Visco-plastic flow past a cylindrical pile. 0 , i

Figure 3. Finite difference meshes

del approximation together with yield stress y and plastic structure.
viscosity p.
7) The force acting in a row of piles is perpendicular to
the pile shaft (Frat, 1999).
( ) (
C i j = h 2 i2j +1 + i2j + 0.5 Re i j h i +1 j i 1 j ) (12)

Numerical model of visco-plastic flow ( )

Di j = h 2 i2j + i2j 1 0.5 Re i j h i +1 j i 1 j( ) (13)

Due to rectangular nature of the flow domain, it is con-

Further terms used are:
venient to use a finite difference scheme to solve the
coupled equations numerically. A mesh is scribed on the [ ( ) (
Fi j = 2 d hTrr i j i j +1 i j 1 Tr i j i +1 j i 1 j )]
flow domain in the and directions where the interval
between grid lines in the direction is d = 0 M and in + 2 Tr i j [ i +1 j 2 i j + i 1 j h 2 ( i j +1 2 i j + i j 1 )]
the directions d = N , for positive integers M and N Trr i j h ( i +1 j +1 + i 1 j 1 i +1 j 1 i 1 j +1 )
(Figure 3). On each grid point (id, jd) a dependent (14)
variable Z can be written as Z ij . Due to the scaling me-

( )[
&rr i j = e 2id 4dd i +1 j +1 + i 1 j 1 i 1 j +1 i +1 j 1 2d ( i j +1 i j 1 ) ]
chanism r = e , the r grid lines are compressed near the (15)
cylinder allowing greater field definition.
Conventional second order accurate finite differences ( )[ (
&r i j = e 2id 2 d 2 h 2 ( i j +1 + i j 1 ) (1 d ) i +1 j (1 + d ) i 1 j + 2 i j 1 h 2 )]
are adopted to approximate the partial derivations (16)
involved. The second order Woods formulation which
couples and to set the missing boundary value on Trr i j = 2 i j &rr i j (17)
the cylinder = 0 is used as:
Tr i j = 2i j &r i j (18)
(
0 j = 2 2 )0j [(
= 6 1 j d 2 1 j ) (2d 2
(1 + d ))] (7)
i j = (q i j ) (19)
Rewrite, in finite difference approximated form, the partial
differential equations for and respectively as:
Where qij is given by:
[
i j = i +1 j + i 1 j + h ( i j +1 + i j 1 ) d e
2 2 2i d
i j ] [2(1 + h )]
2

(8) q i j = 2 &rr2 i j + &r2 i j (20)

and
i j = (Ai j i +1 j + Bi j i 1 j + C i j i j +1 + Di j i j 1 Fi j ) (Ai j + Bi j + Ci j + Di j )
There are special provisions for the use of the latter
(9) formulae at the boundaries of the flow domain and at the
domain corners. Since the equations are non-linear, an
Where h = d d and iterative approach will be adopted that requires initial
(
Ai j = i2+1 j + i2j 0.5 Re i j h i j +1 i j 1 ) (10) guess starting values for i j and i j . Although it is usual
to adopt zero starting values, the solution for the case of
Bi j = i2j + i21 j (
+ 0.5 Re i j h i j +1 i j 1 ) (11)
a Newtonian zero Re-flow situation leads to a simple
solution that is adopted for the general case, and leads to
Sci. Res. Essays 846

significant reduction in program execution times. N 1

Therefore: p = 2 d { 0 j0 j }
[( )( )]0 j sin( jd ) (27)

[ ( ) ( )]
j =1
= A e 3 4e e + B (1 2 )e e sin (21)
Where the derivative term was calculated using the
= 4 [2 A (e
e ) B e ]sin (1 (

0 )) (22) second order forward difference approximation for a
quantity Z as:
Where the condition = 0 at = 0 has been enforced;
here
( Z (
)0 j = 4 Z 1 j Z 2 j 3 Z 0 j 2 d ) (28)

[
A = 1 2 (1 0 ) X 02 (1 + 0 ) ] (23) Noting that at = 0 and = ; and are both
zero, simplifies the calculation of force on the pile, p,
( )
B = X 02 1 A , X 0 = e 0 (24)
above.

Also the viscosities i j are set to a constant value appro-
priate to the viscous model used (Frat et al., 2006).
Since an iterative approach needs to be adopted, the
following sequential steps were used:
a) Set i0j and i0j as the initial guess values for i j
and i j .
b) Then calculate interior i1 j using Equation (8), where
the values are the updated currently available and
values, in a single sweep of a Gauss-Seidel procedure.
c) Subsequently 01 j values are calculated using
Equation (9).
d) Then calculate updated i j values for the whole flow
domain.
e) Test for convergence between iterations k and k+1 for
a quantity Z i j as:
Z ikj+1 Z ikj (1 + Z )< TolZ
k
ij
Comparison of forces for benchmark results
Three different lateral force estimations are shown in
Figure 4 which illustrates the variation of the lateral force
on the pile versus the ratio D2/D1. It can be seen that the
visco-plastic flow displays a trend where lateral load
predictions lie between the data given by Ito and Matsui
(1975), De Beer and Carpentier (1977). It is seen that in
addition to parametric variation of inter-pile distance,
D2/D1, the mechanical properties of the flowing soil mass
2
has also been varied from y = 10 kN/m to 50 and then
100. KONS (KONS = Vp) is a product of flow velocity
and plastic viscosity. The following numbers represent
the relevant research work and can be found in the
legend of Figure 4;
1) Represents De Beer and Carpentier (1977).
2) Represents Ito and Matsui in terms of passive earth
pressure (1978).
3) Represents Ito and Matsui in terms of active pressure
(1975).
(25) 4) Represents current study.

It is also noted that lateral forces converge in all four

For all i and j with Tol and Tol (convergence tolerance
values) both set as appropriate small values. This
convergence test gives a combination both of absolute
and relative comparisons of successive iteration values.
f) Finally a smoothing strategy is adopted by setting;
k +1
Z i j = S Z ikj + (1-S) Z ikj+1
where S is a smoothing parameter (0 < S < 1) and the
k +1
Z i j values are used in subsequent calculations. This
helps convergence in difficult cases.
Once satisfactory convergence has been achieved the
dimensionless force per unit length on the cylinder may
be calculated using a numerical procedure. Since the
methods adopted thus far have been second order, the
Trapezium second order integration rule was used to
yield:
methods as shear stress y tends to zero.
According to the results obtained by the present visco-
plastic flow, the Ito and Matsui (1975) and the De Beer
and Carpentier (1977) methods the lateral force on the
pile becomes infinite when D2 = 0, that is, when the inter-
spaces between the piles becomes zero (Figure 1). This
(26) does not correspond to reality because the forces gene-
rated by the piles cannot be larger than those needed for
the equilibrium of the soil mass which is located
upstream of the pile row (in the event of a potential
landslide). For this reason it is suggested that the gap
(D2) between two piles should be 2/3 a , where a is the
diameter of the pile used.
Stability analysis of slopes reinforced with piles
When the safety factor for a slope is considered to be
inadequate, slope stability may be improved installing a
 Firat 847

500
1=100 kN/m2
1=50 kN/m2
1=10 kN/m2
2=100 kN/m2
450 2=50 kN/m2
2=10 kN/m2
3=100 kN/m2
3=50 kN/m2
3=10 kN/m2
4=100 kN/m2
400 4=50 kN/m2
4=10 kN/m2

350

300
Force on the pile (kN/m)

250

200

150

100

50
3
GAM () = 20 kN/m
D1-D2= 0.30 m
z=5m
KONS = 5 kN/m

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
D2/D1

Figure 4. Comparison of lateral load calculations for different theories.

retaining structure such as a row of piles (Figure 5). The
piles should be designed to provide the stabilizing force
needed to increase the safety factor to a selected value.
For the slope stability problem containing piles in a row,
two separate analyses have been carried out in terms of
slope and pile stability. In practical applications, the study
of a slope reinforced with piles is usually carried out by
extending the methods commonly used for the stability
analysis of slopes to incorporate the reaction force exer-
ted on the unstable soil mass by the piles. To date, the
limit equilibrium method is the most widely used appro-
ach to analyze slope stability due to its simplicity of use.
Moreover, this method allows for the effect of seepage,
loading and general soil conditions without requiring
additional computational efforts. Major criticism of the
limit equilibrium method is that it is generally based on
simplified assumptions and the results obtained from this
method are, in the light of limit analysis, neither upper
bounds nor lower bounds on the true solution (Ausilio et
al., 2001).
Sci. Res. Essays 848

C
x

B
Slice boundary dx

Ground surface Slip surface

y=z(x) y=y(x)
D
y Line of thrust
y=y't(x)
A
Line of internal water force
y=h(x)

Piles a) Slope stability

dx
B

X+dX

Sliding
surface
E'+dE'
A
X Pw+dPw

Piles b) Pile stability (y+dy)-(y't+dy't)

Pw dW (y+dy)-(h+dh)

E'
Figure 5. Slope stability analysis containing piles in a row. (y-h)
(y-y't) dy

The slope instability can be analyzed by dividing resis-
ting moments/forces (either together or separately, since
it depends on method of analysis used) and disturbing
moments/forces acting on the soil mass DBCAD, shown
in Figure 5. Due to pile installation, the extra resisting
force provided by piles at the plane AB is added to resis-
ting moments/forces within the parameters of normal
slope stability calculations. Pile stability may be analyzed
using forces acting on a single pile at the plane AB, as
shown in Figure 5. This force is used as an extra resisting
force for the slope stability but reactively it is also used as
a design force to calculate pile integrity and stability.
When a row of piles is installed into the slope, the
factor of safety (FOS) changes due to the additional
resisting force, Pp , provided by the piles. To evaluate this
force per unit width of the failure mass, the total force
may be integrated along the depth of the pile (until
reaching the slip surface). Then, the result, p, is divided
by centre to centre distance, D1 where Pp = p D1 .
According to Ito and Matsui (1975), the force acting on
the slope is equal to Pp regardless of the state of the
equilibrium of the slope. This force is added into the FOS
calculations. Nevertheless, an overestimation of the force
Pp may lead to un-conservative results in the design of
the slope. To remain on the safe margin, Hassiotis et al.
(1997) suggested the use of mobilized lateral force, Pm
so that:
dS dPb
g
dN'
Figure 6. Forces acting on a slice (Morgenstern and Price, 1965).
Pm = Pp co --------------------------------- (29)
Where co is greater than 1.0. In this research, Pp is
scaled by the un-reinforced FOS of the slope where it is
then reinforced by a row of piles (i.e. co = FOS). This mo-
bilized force, Pm is used to evaluate stabilized FOS, but
the total lateral force per unit length, p is still used to
design the piles.
Two distinct methods of lateral load estimations above
were used to evaluate lateral loads on the pile in a row.
With the piles in place and with the restraining forces of
the piles against the sliding soil mass (Figure 5), a
second analysis was performed to find the new stabilized
FOS against sliding without changing the failure surface.
The reinforced FOS values are given in Tables 1 and 2.
The case study was examined by two different mass
divisions namely method of slices (MOS) and Gauss
quadrature (GQ).
Method of slices (MOS)
In this method, a possible failure slip surface is divided by
vertical or inclined planes into a series of slices (Figure 6).
 Firat 849

Table 1. Un-reinforced FOS values.

Methods of analysis for FOS calculations (un-reinforced)

Programs Bishop Janbu Janbu Morg. & Sarma Fredlund
Fellenius Spencer Fellenius
simple simple general price (Kc=0) & Krahn
Reese et al. - - - - - 1.060 - - -
SLOPE
1.073 1.073 1.095 1.073 * 1.073 * * 1.073
program
This MOS 1.101 1.101 1.112 1.122 1.126 1.110 1.099 1.110 1.101
study GQ 1.085 1.085 1.110 1.120 - 1.110 1.100 1.110 1.085

Table 2. Reinforced FOS values by using two different lateral load estimations

Methods Methods of analysis for FOS calculations (reinforced)

Lateral load Pile
of mass Fellenius Janbu Janbu Morg. & Fredlund
caculations locations Bishop Spencer Sarma
division simpli simpli general price & Krahn
1 1.397 1.397 1.348 1.428 1.409 1.401 1.403 1.401
Method of
2 1.680 1.680 1.601 1.682 1.671 1.680 1.665 1.680
slices
Plastic 3 1.571 1.571 1.510 1.582 1.585 1.575 1.569 1.575
deformation 1 1.385 1.385 1.351 1.421 - 1.398 1.408 1.398
Gauss
2 1.682 1.682 1.605 1.685 - 1.680 1.669 1.680
quadrature
3 1.575 1.575 1.521 1.584 - 1.580 1.577 1.580
1 1.401 1.401 1.354 1.434 1.441 1.409 1.432 1.409
Method of
2 1.560 1.560 1.521 1.573 1.561 1.560 1.558 1.560
slices
Visco-plastic 3 1.450 1.450 1.405 1.469 1.465 1.453 1.451 1.453
flow 1 1.410 1.410 1.351 1.429 - 1.412 1.438 1.412
Gauss
2 1.558 1.558 1.525 1.568 - 1.563 1.545 1.563
quadrature
3 1.453 1.453 1.401 1.472 - 1.455 1.442 1.455
Reese et al. - - - - - 1.820 - -

The width between slices can be equal or unequal and of
necessity each slice has a variable height, measured
along the centre line. Using a sufficiently large number of
slices the base and top of each slice can be approxima-
ted to be a straight line. The potential slip surface is
divided into slices only for ease of analysis. The factor of
safety is taken to be the same for each slice. This implies
that interslice normal and shear forces must act between
each slice.
Gauss quadrature (GQ)
to be evaluated numerically from a given number of
Gauss points. Gauss considered the problem of deter-
mining the values of should be chosen in order to get
the greatest possible accuracy. In other words, how
should one subdivided the interval (-1, +1) so as to get
the best possible results? Gauss found that the points in
the interval should not be equally spaced but should be
symmetrically placed with respect to the mid-point of the
interval (Figure 7). The forces acting on a single Gauss
slice are identical with the forces considered with the
method of slices (Frat, 1999).

One of the most accurate numerical methods in ordinary

use for integrating polynomials is the Gauss quadrature
formulae. The great mathematician Karl Friederich Gauss
(1777-1855) discovered that by a special placement of
the nodes, the accuracy of the numerical integration
process could be greatly increased. Consider the definite
integral:
+1
I= f ( ) d
1 analyzed the sudden drawdown case taking the undrained
Practical example
This particular problem has been examined by Reese et
al. (1992) shown in Figure 8. The slope exists along the
bank of a river where sudden drawdown is possible.
According to the above authors, slides had been
observed along the river at numerous places and it was
found necessary to stabilize the slope to allow a bridge to
(30) be constructed. Using the Spencers method (1967), they
Sci. Res. Essays 850

XL X XR

-1 -0.906 -0.538 0 0.538 0.906 +1

Figure 7. Application of Gauss quadrature to slope

stability.

PILE Elavation (m)

a = 0.915 m
D2 / D1 = 0.6 y
(83.3;86)
EI = 19.5 MN.m2
Fill (clay) K s = 0.0
Silt, Ks = 8.15 MN/m3
Clay, K s = 12.0 MN/m3
Sand, Ks = 64.0 MN/m3

(91.5;76.2)
(85.2;75.3) 76.2
Fill (clay)
c = 47.9 kN/m2
= 19.6 kN/m3
(80;70.7)

(74;69.2)
(91.5;68.96)
Water table (67.7;67.1) 68.96

(60;66.4)
Silt
c = 23.9 kN/m2 (101.3;64)
= 17.3 kN/m3
(54.9;64.9)

(101.3;62.5) 63.0
(54.9;53.1)
Clay
c = 36.3 kN/m2 (91.5;60)
= 19.6 kN/m3 61.0

(91.5;57.3)
Sand
c=0 3
2
= 19.6 kN/m3
= 30o-40o 1

Trial pile locations

S = 80.0 m
x
S = 90.0 m
S = 95.0 m

Figure 8. Stabilized failure surface by a row of piles.


analysis and the FOS was given to be 1.060. They stated
that the value is in reasonable agreement with
observations.
Reese et al. (1992) used drilled shafts which were
0.915 m in diameter and penetrated below the sliding sur-
face. The tops of the drilled shafts were restrained with
anchors in stable soil.
The example slope, shown in Fig. 8, was reanalyzed
using eight different methods namely as Fellenius (1936),
Bishop simplified (1955), Janbu simplified and gene-
ralized (1954), Morgenstern and Price (1965), Spencer
(1967), Sarma (1973) and Fredlund and Krahn (1977)
and available comparisons were made in Tables 1 and 2
in terms of un-reinforced and reinforced FOS values. Also
SLOPE (1993) program used for un-reinforced case to
compare the results.
Three different pile locations were examined to find out
the most suitable place(s) to obtain higher stabilized FOS
values.
Conclusion
There are now bounded solutions for the lateral forces on
the pile depending on the theory and circumstances
used. This is to be welcomed, since an engineer has a
choice of examining a number of solutions. The
resistance that the piles can provide is assessed via an
analysis of the response of piles to lateral soil movement.
For the slope stability problem containing piles in a row,
two separate analyses have been carried out in terms of
slope and pile stability. In practical applications, the study
of a slope reinforced with piles is usually carried out by
extending the methods commonly used for the stability
analysis of slopes to incorporate the reaction force exer-
ted on the unstable soil mass by the piles. To date, the
limit equilibrium method is the most widely used appro-
ach to analyze slope stability due to its simplicity of use.
From the findings of this study, the following conclusions
are drawn:
1) The present visco-plastic flow displays a trend where
lateral load predictions lie between the data given by Ito
and Matsui (1975) De Beer and Carpentier (1977).
2) Similar FOS values are obtained by commercially
available SLOPE (1993) program and this study for un-
reinforced case.
3) Eight different methods of analysis were carried out to
calculate FOS values of a slope reinforced by a row of
piles. Relationships have been observed between
stabilized FOS and the location of the pile row.
4) For slopes containing piles, analytical expressions
have been derived that allow the force needed to
increase the safety factor to a desired value and the most
suitable pile location within the slope to be evaluated.
The calculations carried out using the methods obtained
show that installing a row of piles is an effective remedy
to improve slope stability especially when the sliding sur-
Firat 851
face for the un-reinforced slope is relatively shallow. Piles
appear to be very effective when they are installed in the
region around the middle of the slope.
Nomenclature
a ; Pile (cylinder) diameter, c; so-called cohesion for soil, D1;
centre to centre distance between two piles, D2; clear distance
between two piles, EI; pile stiffness, K; modulus of subgrade
reaction, M; number of grid points in the (r ) direction, N;
number of grid points in the direction, p ; lateral force on the
pile, Re; Reynolds number, Ro; distance from a pile, S;
smoothing parameter, Trr , T , Tr ; extra stress tensors, TolZ;
tolerance value for iterations, Tol ; tolerance value for
convergence of vorticity function, Tol ; tolerance value for
convergence of stream function, V; flow velocity, v; velocity
component in the direction, z; depth from ground surface, ;
unit weight of soil, &
rr , &
&
, r ; rates of strains in polar co-
ordinates, ; apparent viscosity, 0 ; Newtonian flow viscosity,
p ; Bingham plastic viscosity, ; angle of internal friction of
soil, y ; yield stress, , r; polar co-ordinates, ; vorticity
function, 0 ; effective distance in the vicinity of a pile, ;
scaling factor for r and ; stream function.
REFERENCES
Abramson LW, Lee TS, Sharma S, Boyce GM (1996). Slope Stability
and Stabilisation Methods, John Wiley and Sons.
Anagnastopoulos C, Hada M, Fukuoka M (1991). Piles as Landslides
Countermeasures-Model Study, Landslides, (Bell, D.H. ed.), A.A.
Balkema, Rotterdam.
Ausilio E, Conte E, Dente G (2001). Stability Analysis of Slopes with
Piles, Comput. Geotech. 28: 591-611.
Bishop AW (1955). The Use of the Slip Circle in the Stability Analysis of
Earth Dams, Geotechnique 5(1): 7-17.
Chen L, Poulos HG (1997). Pile Subjected to Lateral Soil Movements, J.
Geotech. Geoenviron. Eng. ASCE 123(9): 802-811.
De Beer E, Carpentier R (1977). Discussions: Methods to Estimate
Lateral Force Acting on Stabilizing Piles, Soils Found. 17(1): 68-82.
Fellenius W (1936). Calculation of the Stability of Earth Dams, Pro-
ceedings of the Second Congress on Large Dams, Washington 4:
445-459.
Frat S (1999). Critical Assessment of Existing Slope Stability Formulae
Analysis and Application to Slope Stabilisation, Ph.D Thesis,
University of Glamorgan.
Frat S, Sarbyk M, elebi E (2006). Lateral Load Estimation from
Visco-Plastic Mud-Flow Around Cylindrical Row of Piles, J. Appl.
Math. Comput. 73(2): 803-821.
Fredlund DG, Krahn J (1977). Comparison of Slope Stability Methods of
Analysis, Can. Geotech. J. 14(3): 429-439.
Hassiotis S, Chameau JL, Gunaratne M (1997). Design Method for Sta-
bilization of Slopes with Piles, J. Geotech. Geoenviron. Eng. ASCE
123(4): 314-323.
Ito T, Matsui T (1978). Discussions: Methods to Estimate Lateral Force
Acting on Stabilising Piles, Soils Found. 18(2): 41-44.
Ito T, Matsui T (1975). Methods to Estimate Lateral Force Acting on
Stabilizing Piles, Soils Found. 15(4): 43-59.
Ito T, Matsui T, Hong WY (1979). Design Method for the Stability
Analysis of the Slope with Landing Pier, Soils Found. 19 (4): 43-57.
Ito T, Matsui T, Hong WY (1982). Extended Design Method for Multi-
Row Stabilizing Piles Against Landslide, Soils Found. 22(1): 1-13.
Sci. Res. Essays 852
Janbu N (1954). Application of Composite Slip Surfaces for Stability
Analysis, Proceedings of the European Conference on Stability of
Earth Slope 3: 43-49.
Jeong S, Kim B, Won J, Lee J (2003). Uncoupled Analysis of Stabilizing
Pile in Weathered Slopes, Computers and Geotechnics 30: 671-682.
Lee CY, Hull TS, Poulos HG (1995). Simplified Pile-Slope Stability
Analysis, Comput. Geotech. 17: 1-16.
Martin GR, Chen CY (2005). Response of Piles due to Lateral Slope
Movement, Comput. Structure 83: 588-598.
Morgenstern NR, Price VE (1965). The Analysis of the Stability of
General Slip Surfaces, Geotech. 15(1): 79-93.
Popescu ME (1991). Landslide Control by Means of a Row of Piles,
Keynote paper, Proceeding International Conference in Slope
Stability Engineering, Thomas Telford, London pp. 389-394.
Poulos HG (1995). Design of Reinforcing Piles to Increase Slope
Stability, Can. Geotech. J. 32: 808-818.
Reese LC, Wang ST, Fouse JL (1992). Use of Drilled Shafts in Sta-
bilizing a Slope, Stability and Performance of Slopes and Embank-
ments-II, (Seed, R.B. and Boulanger R.W. ed.), Geotechnical Special
Publication, ASCE 31(2): 1318-1332.
Sarma SK (1973). Stability Analysis of Embankments of Slopes,
Geotech. 23(3): 423-433.
SLOPE (1993). Slope Stability and Reinforced Soil Analysis Program,
by D.L. Borin, Distributed by Geosolve.
Spencer E (1967). A Method of Analysis of the Stability of Embank-
ments Assuming Parallel Inter-Slice Forces, Geotech. 17(1): 11-26.
Won J, You K, Jeong S, Kim S (2005). Coupled Effects in Stability
Analysis of Pile-Soil Systems, Comput. Geotech. 32: 304-315.