Indian Geotechnical Conference – 2010, GEOtrendz

December 16–18, 2010

IGS Mumbai Chapter & IIT Bombay

Effect of Earthquake Induced Lateral Soil Movement on

Piles in a Sloping Ground

Muthukkumaran, K. Subha, I.P.

Assistant Professor M.Tech. Student
e-mail: kmk@nitt.edu e-mail: subhaip@gmail.com

Department of Civil Engineering, National Institute of Technology, Tiruchirappalli

ABSTRACT
The performance of piles in liquefying ground under earthquake loading is a complex problem due to the effects
of a progressive build-up of pore water pressures in the saturated soils. The loss of soil strength and stiffness due
to liquefaction may develop large bending moments and shear forces in piles, possibly leading to pile damage.
The significance of liquefaction-related damage to pile foundations has been clearly demonstrated by the major
earthquakes that have occurred during past years. The present investigation is to find out the effect of earthquake
induced lateral soil movement on piles in sloping ground. The present study was carried out by numerically. In
this, 1995 Kobe earthquake data (Japan) is used. Parametric study has been done on the same model by varying
slope in the soil layers and L/D ratio of the pile. The dynamic analysis was carried out for slope angle of 1V:1.5H
in with L/D=16, L/D=25 & L/D=33. In each case, bending moment and displacement variation with depth of the
pile is noticed. Based on the study, it is concluded that for a constant slope and constant depth of liquefiable layer,
lateral displacement and bending moment is significantly increased for L/D=16 when compared to higher L/D
ratio’s of 25 and 33. However, further increase in L/D ratio is not having any significant effect in the lateral
displacement.

1. INTRODUCTION (Benuzca 1990); and the damage to numerous water and

The seismic waves generated due to earthquake cause the
ground to vibrate and create severe natural disasters.
Various types of ground deformations occur during
vibration and such deformations leading to failure can be
recognized as ground failures. The dramatic response of
saturated loose sand deposit due to earthquake can be
hazardous to constructed facilities in the seismic regions
where such soil deposits exist. In liquefiable soils,
earthquake excitation shall induce the development of
excess pore pressure and when the excess pore pressure
reaches the initial overburden effective stress, the soil is
liquefied. Lateral spreading past bridge foundations was
observed in the 1999 Chi-Chi earthquake in Taiwan. During
the 1999 Chi-Chi, Taiwan earthquake, many sand boiling
phenomena were observed in central Taiwan, which caused
severe ground settlement and structure damages. Examples
of damage to deep foundations and lifeline utilities due to
liquefaction-induced lateral spreading are; the railway
bridge foundations during the 1991 Limon earthquake
(Youd et al., 1992); the batter piles supporting the 7th Street
Terminal Wharf in the 1989 Loma Prieta earthquake
gas lines in the 1906 San Francisco earthquake (Bartlett
and Youd, 1992). The extent of damage and affected areas
in Bhuj earthquake (26th January 2001) has provided a
unique opportunity to evaluate a wide range of geotechnical
issues. A large area in the Rann of Kutch experienced
massive liquefaction resulting in ground subsidence and
lateral flow. A large number of dams in the Kutch district
suffered moderate to severe damages. Many buildings were
damaged and collapsed in the city of Ahmedabad situated
on the bank of the Sabarmati River. Dynamic FEM analyses
have the potential to better represent certain aspects of pile
foundation behaviour during earthquake loading and
liquefaction-induced deformations. FEM analyses can
explicitly include the effects of progressive liquefaction on
seismic response, ground deformations, and foundation
stiffness.
2. NUMERICAL MODELLING
Governing Factors
Numerical models involving FEM can offer several
approximations to predict true solutions. The accuracy of
210 K. Muthukkumaran and I.P. Subha

these approximations depends on the modeller’s ability to Table 1: Soil Properties (Adopted from Abdoum
portray what is happening in the field. Often the problem and Dobry, 2002)
being modelled is complex and has to be simplified to obtain Nevada Sand Slightly Cemented
Description
a solution. Two of the major factors which have a vast (Dr=40%) Sand
impact on both the real and model piles are; (1) the Density (kN/m3) 25 28
constitutive properties of the soils and (2) the soil– structure Φ 32 35
interaction at the interface over the structural surface. Ψ 2 5
k (m/s) 6.05 × 10-5 3.197 × 10-5
Constitutive Models µ 0.25 0.3
Finite element method has become more popular as a soil E (kN/m2) 38000 49000
response prediction tool. This has led to increased pressure
(flexural rigidity) and axial stiffness EA (where E is the
on researchers to develop more comprehensive descriptions
Young’s modulus, I is moment of inertia and A is the cross
for soil behaviour, which in turn leads to more complex
sectional area of the pile) are the input values. The values
constitutive relationship. Prevost and Popescu state that
of pile properties are presented in Table 2.
for a constitutive model to be satisfactory it must be able
to: (1) Define the material behaviour for all stress and strain Table 2: Pile Property (Adopted from Abdoum and
paths; (2) identify model parameters by means of standard Dobry, 2002)
material tests; and (3) Physically represent the material Material Concrete
response to changes in applied stress or strain. Previous Diameter (m) 0.6
EA (kN) 3.56 × 105
studies have explored constitutive models and found that
EI (kNm2) 8000
the use of isotropic models such as elasto-plastic Mohr–
Coulomb and Drucker-Prager models are sufficiently Boundary Conditions
accurate. In the past, linear elastic constitutive models have An absorbent boundary is aimed to absorb the increments
been commonly used in developing pile design methods. of stresses on the boundaries caused by dynamic loading,
that otherwise would be reflected inside the soil body. In
3. PLANE STRAIN ANALYSIS USING PLAXIS 2D
PLAXIS absorbent boundaries are generated by selecting
Description of Approach the standard absorbent boundaries. For plane strain models,
For this study, the model tests are analyzed using a plane the standard absorbent boundaries are generated at the left,
strain finite element approach. Plane strain analysis is the the right and the bottom boundary. The applied load is the
most straightforward of the finite element approaches product of the input value and the corresponding load
described above, and allows good representation of the pile multiplier.
group configuration and geometry, without being unduly
Calculations
complicated. In this study, a two dimensional finite element
program PLAXIS 2D has been used to model the soil layers The calculation involves two phases. The first one is a
and the pile using the concept of plane strain condition. plastic calculation in which the pile is activated and the
The model created is shown in Figure 1. The analyses are soil above the slope is deactivated. The prescribed
conducted with three layers of sand, represented by Mohr- displacement is activated while defining the calculation
Coulomb model. The Mohr-Coulomb model is used for the phases. The second is a dynamic analysis in which the
earthquake is simulated. In order to analysis the effect of
the earthquake in detail the displacements are reset to zero
at the beginning of this phase. Earthquakes are usually
0.04

0.03

0.02
Acceleration (m/sec )
2

0.01

0
0 2 4 6 8 10 12 14 16
-0.01

Fig. 1: Model Created -0.02

proposed (linear elastic-plastic) model, with plastic flow -0.03

Time (sec.)
governed by an associated flow rule. The values of soil -0.04

properties are presented in Table 1. In the present analysis, -0.05

a 15-node triangular is selected. Bending stiffness EI

Fig. 2: Kobe Earthquake Acceleration Data
Effect of Earthquake Induced Lateral Soil Movement on Piles in a Sloping Ground
applied by means of prescribed horizontal displacements.
In the present analysis, the prescribed horizontal
displacement is given as 0.01m. In the present earthquake
analysis, the Kobe earthquake acceleration data (Figure 2)
is used. A time interval of 25 sec is used for the calculation.
Then parametric study is done on the model by providing
a slope of 1V:1.5H in the first layer alone. Bending moment
and displacement variation of the pile is obtained for all
the cases. Again the same procedure is followed for L/D
ratio 16, 25 and 33.
4. RESULTS AND DISCUSSION
For validation, the bending moment for a pile obtained
using FEM model is compared with the results of a similar
study which is done by Abdoun and Dobry (2002), using
centrifuge modelling. A model with three layers of soil and
a pile is created. Slightly cemented sand is used in top and
bottom layers. Middle layer is Nevada sand (Dr=40%)
which is the liquefiable layer. The model created for the
analysis, is validated by Abdoun and Dobry (2002).
Fig. 3: Deformed Mesh
The results obtained using FEM modelling is compared
with the results given by Abdoun and Dobry (2002). Figure
3 shows the deformed mesh. Graphs are plotted with
bending moment on x-axis and depth of the pile on y-axis.
Figure 4 shows the variation of bending moment with the
depth of the pile. The results are captured after 25sec of
the application of earthquake loading, which are
corresponding to maximum values.
Fig. 4: Depth of the Pile Vs Bending Moment (Validation)
The pile bending moment in the top 2 m is first
increased with time of shaking and then decreased after
211
passive failure of the top nonliquefiable layer against the
pile, while the bending moments near the bottom increased
monotonically and never decreased, as the bottom
nonliquefiable layer did not fail.
The shapes of the bending moment profiles indicate
that the deformed shape of the pile had a double curvature
caused by the top and bottom soil layers loading the pile in
opposite directions. The double curvature shape indicates
that the nonliquefied shallow layer pushed the pile laterally
resisting this bending action. Both the passive failure of
the top layer and the moment concentrations at the top and
bottom boundaries of the liquefied layer indicated by the
figures are consistent with the experience from earthquake
case histories. Another interesting aspect is that the bending
moments vary linearly within the liquefied layer, suggesting
that they are essentially controlled by the loading of the
top and bottom layers, with the pressure of the liquefied
soil being negligible.
Effect of L/D Ratio on Lateral Pile Behaviour Under
Earthquake
Figure 5 shows the variation of displacement with the L/D
ratio when a slope of 1V:3H is provided in the first layer
(slightly cemented sand) alone. Displacement reduces with
increase in L/D ratio. When L/D ratio is increased from 16
to 33, the decrease in displacement is by 70.98%. In the
case of L/D=16, it can be seen that the displacement is
high when compared to other cases. This is because only
bottom 2m of the pile is in the slightly cemented sand and
hence the anchorage of the pile in the non-liquefiable layer
is less when compared to the other cases.
Fig. 5: Depth of the Pile Vs Displacement
Figure 6 shows the variation of bending moment with
the L/D ratio when a slope of 1V:3H is provided in the first
layer (slightly cemented sand) alone. Bending moment
decreases with increase in L/D ratio. When L/D ratio is
increased from 16 to 33, the decrease in bending moment
is by 277.32%. The point where there is a shift from
negative bending moment to positive bending moment in
212
the cases of L/D=16 and L/D=25 is near to the boundary
between top slightly cemented sand layer and the liquefiable
Nevada sand layer whereas in the case of L/D=33, the point
where this shift occurs is at the boundary between
liquefiable Nevada sand layer and bottom slightly cemented
sand layer.
Fig. 6: Depth of the Pile Vs Bending Moment
Effect of Slope and L/D Ratio on Maximum
Displacement and Maximum Bending Moment Under
Earthquake
Figure 7 shows the variation of maximum displacement
obtained for different L/D ratios with the depth of the pile
when slope is introduced in the first layer alone (slightly
cemented sand). When L/D=16, rate of increase in
displacement is more when compared to higher L/D ratio’s.
In all cases, maximum displacement decreases as L/D ratio
increases. Maximum displacement occur for the slope
1V:1.5H for all the three cases. When the slope is increased
from 1V:3H to 1V:2H for L/D=16, increase is in
displacement is by 133.3%. It got reduced to 5.8% when
L/D=33.
Fig. 7: Max. Displacement Vs Depth of the Pile
Figure 8 shows the variation of maximum bending
moment obtained for different L/D ratios with the depth of
K. Muthukkumaran and I.P. Subha
the pile when slope is introduced in the first layer alone
(slightly cemented sand). The maximum bending moment
reduces with increase in L/D ratio. In all the cases,
maximum bending moment occur for the slope 1V:1.5H.
When L/D ratio is increased from 16 to 33, the bending
moment has got reduced by 277.3% for the slope 1V:3H.
When L/D ratio is increased from 16 to 33, the bending
moment has got reduced by 263.19% for the slope 1V:1.5H.
Fig. 8: Max. Bending Moment Vs Depth of the Pile
5. CONCLUSIONS
On the basis of the results and discussion presented above,
the following general conclusions can be drawn.
1. A finite element model with three layers of soil
and a pile is created and dynamic analysis has
performed. The developed FE model has been
validated with the similar study which was done
by Abodoun and Dobry (2002). The obtained
results from the developed FE model are having
very good agreement with their centrifuge model
results.
2. The increase in L/D ratio is significantly affects
the failure mode of the pile. The failure mechanism
of L/D=16 pile is like a short rigid pile though
the pile is long flexible pile as per soil-pile relative
stiffness calculation. However, the increase in L/
D ratios (L/D=25 and 33) changes the mode of
failure of the piles more towards flexible in nature.
The increase in L/D ratio beyond 33 is not much
significantly affects the bending moment and
lateral deflection of the pile.
REFERENCES
Abdoun, T., and Dobry, R. (2002). Evaluation of pile
foundation response to lateral spreading. Soil Dynamics
and Earthquake Engineering, 1051–1058.
Youd, T.L., Rollins, K.M., Salazar, A.F. and Wallace, R.M.
(1992). Bridge damaged caused by liquefaction during
the 22 April 1991 Costa Rica earthquake, Madrid. 10th
World Conference on Earthquake Engineering.