Pile On Slope
Pile On Slope
Pile On Slope
SOSKO
STABILIZATION OF SEISMICALLY UNSTABLE SLOPES
USING PILES: PARAMETRIC ANALYSIS
Rallis KOURKOULIS
1
, Fani GELAGOTI
2
, Ioannis ANASTASOPOULOS
3
, George GAZETAS
4
ABSTRACT
A recently developed and validated simplified numerical model for the investigation of the response of
slope stabilizing piles is utilized to explore the parameters determining the effectiveness of such systems.
Pile diameter and spacing, depth of pile embedment, soil strength and stiffness are the key problem
parameters investigated. It is shown that a pile spacing of 4 diameters is the most cost-effective being able
to generate soil arching between the piles. For relatively small pile embedment, pile response is dominated
by rigid-body rotation, without substantial flexural distortion: The critical embedment depth to achieve
fixity conditions at the base of the pile is found to range depending on the relative strength of the unstable
ground compared to that of the stable ground (i.e. the soil below the sliding plane)
Keywords: Slope Stabilization, Piles, finite elements
INTRODUCTION
Slope stabilization using piles constitutes a widely accepted and successfully applied method [e.g.
Heyman & Boersma, 1961; Kitazima & Kishi, 1967; Leussink & Wenz, 1969; Nicu et al., 1971; De Beer
& Walleys, 1972; Ito & Matsui, 1975; D Appolonia et al., 1977]. Existing design methods are either
pressure or displacement-based [e.g. De Beer et al., 1972; to & Matsui, 1975; Poulos, 1995] or numerical
methods [e.g. Oakland & Chameau, 1984; Poulos & Chen, 1997]. Although the latter are in principle the
most rigorous since they can provide fully coupled solutions to the problem, their 3D application is
computationally expensive and time consuming.
The methodology employed herein for the design of slope stabilization piles is formulated on the basis of
the decoupled approach [Viggiani, 1981; Hull, 1993; Poulos, 1995; 1999], and combines the simplicity of
widely accepted analytical techniques with the advantages of 3D FE modeling. The method entails two
steps:
Step 1 : Conventional slope stability analysis to compute the required lateral resisting force RF needed to
increase the safety factor of the slope to the desired value, and
Step 2 : Selection of a pile configuration capable of offering the required RF (to increase the safety factor
of the slope to the desired level) for a prescribed deformation level.
Recently, Kourkoulis et al (2010) developed and validated a new approach for the calculation of RF
(second step), which is based on the decoupling of slope geometry from the computation of pile lateral
capacity, thus allowing numerical simulation of only a limited region of soil around the piles. The present
paper utilizes this decoupled analysis method to derive insights on the factors affecting the response of
piles and pile groups. The approach of Kourkoulis et al., 2010 is briefly described in the ensuing.
1
Post Doctoral Researcher, Soil Mechanics Laboratory, NTUA.
2
Doctoral Researcher, Soil Mechanics Laboratory, NTUA.
3
Post Doctoral Researcher, Soil Mechanics Laboratory, NTUA.
4
Professor, School of Civil Engineering, NTUA.
5th International Conference on Earthquake Geotechnical Engineering
January 2011, 10-13
Santiago, Chile
Stable
Soil
Unstable
Soil
Interface
H
u
L
e
L
p
u
Piles of
Diameter D
S / 2
S
5D 5D
S / 2
Assumed displacement
profile u
S
2S
interface
Stable soil
Unstable soil
x
y
z
S
S / 2
S / 2
5 D 5 D
L
p
H
u
L
e
2 S
Figure 1: Schematic illustration of the simplified decoupled methodology for estimation of
pile ultimate resistance. Instead of modeling the whole slope-soil-pile system (top sketch),
we focus on the pile(s) and a representative region of soil at its immediate vicinity (blue
box). The geometry and key parameters of the simplified model are shown at the bottom
sketch.
DECOUPLED METHODOLOGY FOR ASSESSING PILE LATERAL CAPACITY
As schematically illustrated in Figure 1, in Step 2, instead of modeling the whole slopesoilpile system,
we focus on a representative region of soil around the pile. The ultimate resistance is computed by
imposing a uniform displacement profile onto the model lateral boundary. This simplification has been
justified as reasonable by Poulos, 1999; and Kourkoulis et al., 2010, among others. Having eliminated the
detailed slope geometry, a sliding interface at depth H
u
is pre-specified in the simplified FE model; where
the piles, of diameter D and length L
p
at spacing S, are embedded into the stable soil layer for a length L
E
.
Since the zone of influence of each pile does not exceed 5D (Reese & Van Impe, 2001), the length of the
5th International Conference on Earthquake Geotechnical Engineering
January 2011, 10-13
Santiago, Chile
model can be limited to 10D. In order to model a representative soil slice, the width of the model has been
taken equal to 2S.
An elastoplastic constitutive model with Mohr-Coulomb failure criterion is used for the soil, while the pile
is modeled with 3D beam elements, circumscribed by 8-noded hexahedral continuum elements of nearly
zero stiffness. As discussed in detail in Kourkoulis et al. [2010], the nodes of the beam are rigidly
connected with the circumferential solid element nodes of the same elevation. Hence, each pile section
represents a rigid disc. The beam elements provide the pile strength and stiffness characteristics while the
circumferential solid elements are able to capture the 3D geometry effects. Analyses are conducted
assuming linear or nonlinear pile response.
PARAMETRIC ANALYSES
A plethora of parametric analyses have been performed in order to examine the behaviour of slope
stabilizing piles nailing unstable cohesionless soil layers of various depths and material properties. The
interface depth from the surface (H
u
), is varied parametrically, covering the range from a shallow (H
u
=
4m) to a quite deep (H
u
= 12 m) landslide. The factors examined are:
(a) Effect of Pile Spacing
(b) Stable Soil Layer Strength
(c). Depth of Pile Embedment depth into the Stable Layer
(d). Pile Non-linearity
Effect of Pile Spacing
This section investigates the effect of pile-to-pile spacing on the effectiveness of the latter in nailing soil
slopes. It is widely accepted that the pile spacing must ensure sufficient arching effect between the piles in
a row. In general, arching stems from the stress transfer through the mobilization of shear strength (i.e. the
transfer of stress from yielding parts of a soil mass to adjoining non-yielding or less compliant parts).
Wang and Yen (1974) studied analytically the behavior of piles in a rigid-plastic infinite soil slope with
emphasis on arching effects, and concluded that a critical pile spacing exists in both sandy and clayey
slopes, beyond which practically no arching develops. A number of pile spacings have been investigated
in order to determine the dependence of arching mechanism on pile distance by means of the proposed
simplified approach. Our analyses assumed that loading is imposed on the free field (i.e. far enough from
the piles region) on the soil nodes. After application of the load, the pile displacement has a value of u
p
,
while the soil between the piles displaces u
ip
. Soil arching is assumed to be accomplished if the ratio of
u
ip
/ u
p
, ranges between 1 and 2, i.e. the pile and neighboring soil displace almost equally. For higher
u
ip
/ u
p
ratios arching cannot be claimed achieved.
Figure 2 displays two characteristic snapshots of the FE analyses, comparing a dense (s=2D) to a loose
pile arrangement (s=7D). The unstable soil layer is considered to be a sand with = 28
, = 2 , and
c = 3 kPa . The bottom soil layer is assumed to be very hard soil with S
u
= 600 kPa. The interface
properties are = 16
, c = 3 kPa, and = 1 and its location is assumed at 4m depth from the soil free
surface. Figure 2a plots the displacements contours on the model surface for the case of piles of diameter
D = 1.2 m spaced at 2 D, i.e , 2.4 m. From the displacements contours distribution it is evident that the
soil between the piles has been restricted by the presence of the piles hence displacing almost equally with
thema clear manifestation of arching. On the contrary, in the case of piles spaced at 7 D (Fig. 2b), the
intermediate soil has not been confined by the piles and flows between them. 3 D numerical parametric
analyses have been performed to define the maximum pile spacing that ensures sufficient degree of
arching as a function of their diameter. The results are summarized in the plot of Figure 3. It is apparent
that spacings of 2, 3 and 4 times the pile diameter are able to provide soil arching. For spacings greater
than 5 diameters soil flows between the piles; such arrangements are therefore not applicable to slope
5th International Conference on Earthquake Geotechnical Engineering
January 2011, 10-13
Santiago, Chile
stabilization and will not be examined. Evidently, the most economical pile arrangement in terms of
arching is the spacing of 4 diameters. However, both the cases of 2 D and 3 D will be examined since
these correspond to the most common arrangements of single piles used for slope stabilization purposes.
As expected, increasing the pile spacing improves each piles effectiveness but reduces the total resistance
force offered per unit width. Figures 4a and b indicate the effect of pile spacing on the produced pile
lateral resistance for the case of a shallow landslide of
u
= 4m and for a relative deep landslide of
u
=
8m respectively. It is observed that in the shallow landslide case (the behaviour of all arrangements is
similar and almost independent of pile spacing. Yet the 2 D spacing ensures slightly higher RF values for
the same pile head displacement compared to that achieved by the 4 D configuration. In the deep landslide
case of
u
=8m (Fig. 4b), the discrepancies among the different pile spacings are more obvious. The
increased flexibility of the soil-pile system requires substantially increased pile deformation for the same
Resistance Force (RF) to be developed, which in turn amplifies the differences between the alternative
arrangements. For instance, a 2 D configuration may offer almost double pile resistance ( 1000 kN/m)
compared to that offered by a 4 D system ( 500 kN/m) when the pile is deformed 5 cm at its top. It is
worth noting that these results refer to elastic piles of diameter D = 1.2m. In case of the non-linear pile,
the maximum realistic moment that may be developed must not exceed the actual structural strength of the
pile. Hence, it must be pointed out that although the maximum value of the resistance force developed
(elastically) is independent of pile spacing for all landslide depths, the only acceptable ultimate RF values
are those which are achieved at acceptable levels of the bending moment.
Imposed
Force
Imposed
Force
5 D 5 D
2 D
7 D
Imposed
Force
Imposed
Force
Figure 2: Contours of horizontal displacements (a) of a dense pile configuration (distance
between piles 2D) and (b) of a sparse pile configuration (pile distance 7D).
5th International Conference on Earthquake Geotechnical Engineering
January 2011, 10-13
Santiago, Chile
1
6
11
16
21
26
31
0 0.2 0.4 0.6 0.8 1 1.2
S = 7D
S = 5D
S 4D
u
ip
/ u
p
u
p
/ u
p
max
Figure 3: Comparison of the interpile displacements calculated for different pile spacings in
sandy soil. It is obvious that for spacings S > 5D, soil flows between the piles.
1.2 m
H
u
= 4 m
Slope and Pile Characteristics
L
e
= H
u
Pile Arrangement
= 2D, 3D or 4D S
Hard
Rock
Soft silty
Sand
H
u
H
u
= 4 m
H
u
= 4m
H
u
= 8 m
H
u
= 8 m
Soft silty
Sand
Hard Rock
L
e
= H
u
H
u
S = 2D, 3D or 4D
Figure 4: (a) Resistance Force offered by the pile vs Pile head displacement diagrams and RF vs
maximum Bending Moment Diagrams for various pile spacings for a shallow landslide ( H
u
= 4m)
(b) Resistance Force offered by the pile vs Pile head displacement diagrams and RF vs maximum
Bending Moment Diagrams for various pile spacings for a deep landslide of H
u
= 8m
Effect of Stable Ground Strength
The strength and stiffness of the stable ground were investigated parametrically to model materials
ranging from relatively loose sand to a rock-type material. The idealized soils of the stable ground layer
are as follows :
i. loose silty sand : = 28
, = 2, c = 3 Kpa, G = 16 Mpa
ii. dense sand : = 38
, = 2, G = 32 Mpa
iii. soft rock : = 45