See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/326446517

Effect of earthquake characteristics on the permanent displacement of a

cantilever retaining wall

Conference Paper · June 2018

DOI: 10.1201/9780429446931

CITATIONS READS

0 1,177

2 authors:

Junied Bakr Syed MOHD Ahmad

The University of Manchester The University of Manchester
10 PUBLICATIONS   44 CITATIONS    37 PUBLICATIONS   361 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Seismic response of soil-nailed earth retaining walls View project

SPH‑FEM coupled simulation for seismic stability of seawalls under earthquake and tsunami forces View project

All content following this page was uploaded by Junied Bakr on 23 July 2018.

The user has requested enhancement of the downloaded file.

Effect of earthquake characteristics on permanent displacement of a
cantilever retaining wall
Junied Bakr
PhD researcher in Geotechnical Engineering, University of Manchester, UK
Syed Mohd Ahmad
Lecturer at School of Mechanical, Aerospace and Civil Engineering, University of Manchester, UK

ABSTRACT: One of the most significant components of performance-based methods in the seismic design of
retaining walls is the accurate estimation of the anticipated permanent displacement. Little attention has been
paid to predict the seismic permanent displacement of a cantilever-type retaining wall. Finite element method
is used in the current study to predict the seismic permanent displacement of a cantilever retaining wall con-
sidering many realistic aspects associated with the real seismic behaviour of the wall-soil system. This study
mainly focuses on the evaluation of the effect of the earthquake characteristics and seismic earth pressure on
the seismic permanent displacement. The results show that the Newmark sliding block method overestimates
the seismic permanent displacement. The most critical scenario, causing maximum permanent displacement,
is the one when the ground motion having maximum amplitude but a minimum frequency content. The seis-
mic earth pressure has a low impact on the permanent displacement.

more complicated geometry than a gravity retaining

wall, and it maintains its stability from the weight of
1 INTRODUCTION backfill soil above footing slab in addition to its self-
weight. In general, the Newmark sliding block
For proper seismic design of different types of re- method has assumed that the seismic earth pressure
taining walls, there has been an increased push in us- force, which already computed by using force-based
ing the performance-based design methods. One of methods like M-O method (Mononobe & Matsuo
the very important components of the performance- (1929)), as a part of total driving force causing the
based design process is an accurate prediction of the permanent displacement; however, the recent
permanent displacement of the retaining walls. experimental methods like Nakamura (2006) and
Richards & Elms (1979) proposed the first analysis (Jo et al. 2014) have shown that the force-based
method to compute the permanent displacement of a methods are quite conservative of a gravity and
gravity type retaining wall. Newmark sliding block cantilever retaining wall respectively. On the other
method, which was derived to evaluate the seismic hand, the real seismic response of a cantilever
stability of slopes (Newmark (1965)), has also been retaining wall covers by many fundamental factors
widely used to compute the accumulated permanent like the material properties of the wall, backfill and
displacement in a time history profile of a gravity foundation soil, the interface between the wall and
retaining wall. This method was based on the backfill soil as well as between the footing slab and
introducing the concept of yield acceleration in foundation soil, boundary condition effect,
which the gravity retaining wall will accumulate amplification of acceleration response and phase
permanent displacement when the earthquake difference issue. So, it is difficult for the simplified
acceleration time history exceeds the yield Newmark sliding block method to account all
acceleration level. Following that, extensive efforts abovementioned problem associated with the real re-
have been made to predict the seismic permanent sponse of a cantilever retaining wall. Some
displacement of a gravity-type retaining wall. How- researcher like (Jo et al. 2014), (Candia et al. 2016),
ever, little attention has been paid to compute the and (Jo et al. 2017) conducted a series of centrifuge
seismic permanent displacement of a cantilever re- tests to estimate the seismic earth pressure behind
taining wall. The same Newmark sliding block the stem of the wall in order to provide a safe
method has still been used to estimate the permanent seismic structural design of the stem of the wall;
displacement of a cantilever retaining wall although however, few researchers like (Green et al. 2008)
the basic difference that the cantilever retaining has and (Kloukinas et al. 2015) carried out numerical
study and shaking table test respectively to
investigate the permanent displacement of a
cantilever type retaining wall.
Despite the fact that the permanent displacement
is a focal point in the performance-based methods,
the proper design of a cantilever retaining wall
should not ignore considering other aspects that di-
rectly affect the amplitude of permanent displace-
ment like earthquake characteristics and seismic
earth pressure. Hence, little emphasis has been given
to the effect of earthquake characteristics and seis- Figure 1. Cantilever retaining wall profile
mic earth pressure on the permanent displacement of
the cantilever retaining wall. So, finite element
method is proposed in the current study to account 3 FINITE ELEMENT METHOD
all abovementioned aspects in order to allow for
more realistic estimation of seismic permanent dis- A finite element (FE) model has been developed
placement of a cantilever-type retaining wall and in current study by using the PLAXIS 2D software
then study the effect of earthquake characteristics (Brinkgreve et al. 2016) as shown in Figure 2 in
and the seismic earth pressure on the permanent dis- order to investigate the seismic performance of a
placement of a cantilever retaining wall. cantilever-type retaining wall.
As shown in Figure 2, the height of the retaining
wall is 5.4 m and sits on a 9 m thick foundation soil.
2 PROBLEM DESCRIPTION The stem member is assumed to have a fixed con-
nection with the footing slab. The backfill soil and
As shown in Figure 1 cantilever retaining wall, the foundation soil are modelled using 6-noded tri-
with a footing slab width L and total stem height H, angular elements of the PLAXIS 2D library
is constructed to provide lateral support to a horizon- (Brinkgreve et al. 2016), while the cantilever-type
tal backfill layer. For a proper seismic design of the retaining wall was modelled using plate elements.
cantilever-type retaining wall, it should be ensured The maximum height of the element is limited by
that retaining wall does not fail by excessive perma- 20% of minimum wavelength as recommended by
nent displacement. For a real treatment of the per- Kuhlemeyer & Lysmer (1973).
manent displacement of a cantilever-type retaining The interaction between the cantilever-type retaining
wall, it is required to consider the following parame- wall and backfill soil as well as between the footing
ters (Fig. 1): 1) Total increment seismic earth pres- slab and foundation layer has been modelled by us-
sure force (∆Pae=Pae - Pa), computed along the ver- ing the 6-noded interface elements, available in the
tical line passing through the heel; where, Pae =total PLAXIS 2D library (Brinkgreve et al. 2016). The
seismic earth pressure force; Pa =total static earth absorbing boundaries are applied to the vertical
pressure force; 2) Total seismic inertia force of the boundaries of the finite element model in order to
retaining wall (FW), which includes the total seismic reduce the effect of seismic wave reflection in the
inertia forces of the stem and footing slab; 3) Total finite element domain and to increase the accuracy
seismic inertia forces of backfill soil above the foot- of the analysis.
ing slab (FS); 4) Total friction resistance force be-
tween the footing slab and foundation layer (FF). 42.85 m
The cantilever-type retaining wall maintains its
stability from the weight of backfill soil above the Stem Backfill soil layer
5.4 m
Footing slab
footing slab in addition to its self-weight. So, the
seismic earth pressure is assumed to be developed Foundation
along the vertical line extended from the heel up to 9m
the backfill soil surface (Fig. 1). The results obtained
from current finite element analysis will be com-
pared with traditional methods like Newmark sliding
16m 1.15 m 2.6 m 40.25m
block method. After that, a variety of ground mo-
tions is used in order to investigate the effect of Ground input motion
earthquake characteristics like amplitude and fre-
quency content on the permanent displacement of a Figure 2. Finite element model of the wall-soil system
cantilever retaining wall. The effect of seismic earth
pressure on the permanent displacement of a cantile-
ver-type retaining wall is also investigated
4 MATERIAL BEHAVIOUR 5 GROUND SEISMIC MOTION

The backfill soil and foundation layer are simu- Real acceleration-time history of the 1952 Kern
lated by using hardening soil with small strain mod- County earthquake used in the current study to simu-
el, which is available in PLAXIS 2D (Brinkgreve et late ground seismic motion and it is applied at the
al., 2016) library. However, the cantilever retaining nodes of the base boundary of finite element model.
wall is simulated by using a linear viscoelastic con- The peak ground acceleration (PGA) of this ground
stitutive model. Table 1 shows the parameters of seismic motion is 0.24g (see Fig. 3a). A fast Fourier
hardening soil with small strain model as well as the transform analysis is conducted for the earthquake
material parameters of a cantilever retaining wall. time history in order to obtain the dominant frequen-
cy of ground input motion. A shown in Figure 3b,
Table 1. The parameters of soil and retaining wall used to run the applied ground seismic motion has three domi-
finite element model nant frequencies (fe) of (0.4 / 1.9 / 3.33 Hz). The
uniform sinusoidal ground motions are also used in
the current study to investigate the effect of earth-
Parameter Symbol Unit Value quake characteristics on the permanent displacement
Soil of the wall, and they are simulated by three groups
Relative Density Dr % 78%
according to their frequency content; group 1:
fe=0.5Hz, group 2: fe =2Hz, and group 3: fe =4Hz.
Unit weight γ kN/m3 14.23 For each group, the uniform sinusoidal ground seis-
Effective friction o mic motion is scaled by three maximum amplitudes
φ' 40
angle of the soil of 0.2g, 0.4g, and 0.6g.
Reference stiff-
ness modulus at ௥௘௙
‫ܧ‬ହ଴ MPa 46.8 (a) 0.3
50% of ultimate
soil strength
Acceleration, a(g) (m/sec2)

0.2
Reference secant
௥௘௙ 0.1
modulus of oe- ‫ܧ‬௢௘ௗ MPa 46.8
dometer test 0.0
Reference stiff-
ness modulus of ௥௘௙ -0.1
‫ܧ‬௨௥ MPa 140.4
unloading reload-
-0.2
ing
Dilatancy angle of o -0.3
ψ 10
the soil
Poisson’s ratio for -0.4
0 5 10 15 20
unloading- vur - 0.2
Time, t (sec)
reloading
Stress-level de-
(b)
pendency of the
y - 0.5 0.20
stiffness of the
Acceleration, a(g) (m/sec2)

soil
Initial shear mod- 0.15
௥௘௙
‫ܩ‬௢ MPa 113
ulus
Reference shear 0.10
strain at 70% of γ0.7 - 0.0002
௥௘௙
‫ܩ‬௢ 0.05
Reference confin- ref 2
p kN/m 100
ing pressure 0.00
Damping ratio ξ % 3
0 2 4 6 8 10 12 14 16
Failure ratio Rf - 0.9 Frequency, fe (Hz)
Retaining wall
Modulus of elas- Figure 3. a) Real acceleration-time history of the 1952 Kern
E MPa 68000
ticity County earthquake; b) Frequency domain representation of the
Moment of inertia I m4 0.00089 1952 Kern County earthquake
Poisson’s ratio v - 0.334
Unit weight γ kN/m3 26.6
Damping ratio ξ % 3
6 SEISMIC PERMANENT DISPLACEMENT 0.05

Two permanent displacement profiles were 0.04

predicted from the current finite element analysis in

Normalised permanent
displacement, d/H
order to understand the seismic deformation 0.03
mechanism of the wall-soil system. In the first 0.02
instance, the permanent displacement was calculated
between the footing slab of the retaining wall and 0.01
and a point located in the foundation layer at a depth
Newmark sliding block method
of 0.5m; while in the second instance, the permanent 0.00 Permanent displacement of the wall
Permanent displacement of the soil
displacement was computed between the centre of
-0.01
gravity of the backfill soil above footing slab and a 0 5 10 15 20
point located in the foundation layer at a depth of Time, t (sec)
0.5m below the footing slab.
It can be noted from Figure 4 that the normalised Figure 4. Seismic permanent displacement predicted by current
permanent displacement between the retaining wall finite element analysis of the wall-soil system
and the foundation soil (d/Η) attains a maximum
value of about 0.023 for the time duration 5 – 10sec.
It is also observed that the normalised permanent 7 EFFECT OF EARTHQUAKE
displacement between the backfill soil and the CHARACTERISTICS ON THE PERMANENT
foundation soil (d/H) also achieves its maximum DISPLACEMENT
value of about 0.024 for the time duration 5 – 10 sec.
After 10 sec of the time of ground motion, it is To investigate the effect of earthquake characteris-
observed that there is no change in the permanent tics on the permanent displacement of the wall-soil
displacement of both the wall and backfill soil until system, a variety of ground motions are applied at
the end of the analysis. Thus, from Figure 4 it can be the base of finite element model. Figures 5a, 6a, and
said that the retaining wall and backfill soil slide to- 7a show three groups of ground motions applied at
gether away from the backfill soil, and almost move the base of the finite element model with frequency
as a single entity. It can be noted from the Figure 4 content 0.5Hz, 2Hz, and 4Hz respectively. It can al-
that the wall-soil system is also accumulated perma- so be noted that the amplitude of ground motion in
nent displacement towards the backfill soil, but their each group is simulated by 0.2g, 0.4g, and 0.6g.
amplitudes are smaller than the amplitude of accu- Figure 5b, 6b, and 7b show the permanent displace-
mulated displacement of the wall-soil system away ment of the cantilever retaining wall for three groups
from the backfill soil. The permanent displacement of applied ground motion described above respec-
of the wall-soil system is also computed by using tively. Figure 5b, 6b, and 7b that that as the
Newmark sliding block method. The acceleration amplitude of the input motion increases from 0.2g to
predicted in current finite element analysis at mid- 0.6g the permanent displacement of the retaining
height of the wall-soil system is used in Newmark wall increases, while with an increase in the
method to compute the yield acceleration. The fric- frequency content of the input motion from f =
tion angle between the footing slab and foundation 0.5Hz to f = 4 Hz, the permanent displacement
layer is assumed 20°, and the same was used in cur- reduces. It is also interesting to note that the
rent finite element analysis. In conjunction the retaining wall slides by about d/H=0.038 at an
weight of the retaining wall and backfill soil above ground motion amplitude of 0.6g and a frequency
footing slab with the friction angle between the foot- content of 4 Hz (Fig. 7b) while it slides by about
ing slab and foundation layer as well as the total d/H=0.05 for a ground motion amplitude of 0.4g and
seismic earth pressure force computed along the ver- a frequency content of 2 Hz (Fig. 6b). This suggests
tical line passing through the heel by using M-O that the frequency content of the input motion is a
method, the maximum yield acceleration (N.g) is more dominating factor than its amplitude which
found equal to 0.18g. The comparison between the contributes to the permanent displacement of the re-
results obtained from current finite element analysis taining wall. From the above discussion, it can be
and Newmark sliding block method as shown in very safely argued that a low frequency content of
Figure 4 shows that the normalised permanent dis- the seismic ground motion and maximum amplitude
placement computed by Newmark method (d/H = cause a critical case scenario. The results also show
0.046) is remarkably larger than that predicted by fi- that the accumulated permanent displacement of the
nite element method. This is because the Newmark retaining wall is highly sensitive to the number of
sliding block method did not take into account the acceleration cycles (duration of the input motion) in
problems associated with real seismic behaviour of which the retaining wall is still accumulated perma-
wall-soil system like the realistic representation of nent displacement away from the backfill soil with
seismic earth pressure force. increasing of the number of acceleration cycle.
 (a) (a) 0.6
0.6 0.2g
0.2g

Acceleration, a(g) (m/sec )

Acceleration, a(g) (m/sec )

2
2

0.4g 0.4g
0.4 0.4 0.6g
0.6g

0.2 0.2

0.0 0.0

-0.2 -0.2

-0.4 -0.4

-0.6 -0.6

0 5 10 15 20 25 30 0 1 2 3 4
Time, t (sec) Time, t (sec)
(b) (b)
0.25 0.04
0.2g 0.2g
0.4g 0.4g

Normalised permanent
Normalised permanent

0.20 0.6g

displacement, d/H
0.6g
displacement, d/H

0.03
0.15
0.02
0.10

0.05 0.01

0.00
0.00
0 5 10 15 20 25 30 0 1 2 3 4
Time, t (sec) Time, t (sec)

Figure 5. Group of fe=0.5Hz; a) Ground input motions Figure 7. Group of fe=4Hz; a) Ground input motions
b) Permanent displacement b) Permanent displacement

(a) 0.6
0.2g
Acceleration, a(g) (m/sec )
2

0.4
0.4g
0.6g
8 TOTAL INCREMENT OF SEISMIC EARTH
PRESSURE FORCE
0.2

0.0 The total seismic earth pressure force Pae has been
-0.2 calculated along virtual line passing through the
heel. The total increment of seismic earth pressure
-0.4
force ∆Pae is estimated by subtracting the total seis-
-0.6 mic earth pressure force Pae from the total static
0 1 2 3 4 5 6 7 8 earth pressure force Pa. Figure 9 shows the total in-
Time, t (sec) crement of seismic earth pressure force, estimated at
(b)0.12 the virtual line passing through the heel (∆Pae) be-
0.2g
0.4g
tween the time 5 sec -10 sec when the maximum
0.10 permanent displacement of the wall-soil system has
Normalised permanent

0.6g
displacement, d/H

0.08 been accumulated. It can noted from the Figure 8

that when the ground acceleration applied towards
0.06
the backfill soil (for example at time 5.5 sec, 7.1 sec,
0.04 and 9 sec (Fig. 3a) the total increment of seismic
0.02 earth pressure force is close to zero, and the total
seismic earth pressure force is close to the static
0.00
earth pressure force. However, when the ground
0 2 4 6 8 acceleration changes its direction away from the
Time, t (sec) backfill soil (for example at time 6 sec, 7 sec, and
Figure 6. Group of fe=2; a) Ground input motions 7.5 sec (Fig. 3a), the maximum increment of
b) Permanent displacement seismic earth pressure force is developed along the
virtual line passing through the heel.
 150
Total increment of seismic earth

140
pressure force, ∆Pae (kN/m)

Total seismic force, (kN/m)

120 100
100
50
80
60 0
40
-50
20
0 -100 Total increment of seismic earth pressure force, ∆Pe
Total seismic inertia force of wall-soil system, F
-20
5 6 7 8 9 10 -150
Time, t (sec) 5 6 7 8 9 10
Time, t (sec)

Figure 8. Total increment of seismic earth pressure force pre- Figure 9. Total increment of seismic earth pressure force and
dicted by finite element analysis total seismic inertia force predicted by finite element analysis

To more realistic treatment of the contribution of consider more real aspects of the seismic behaviour
seismic forces causing the permanent displacement of the wall-soil system. The soil behaviour is
of the wall, Figure 9 shows the total seismic inertia simulated in the current numerical analysis by using
force of the wall-soil system F=FW+FS, and it is hardening soil with small strain model. The results
combined with the total increment of seismic earth of the current study show that the Newmark sliding
pressure force ∆Pae between the time 5sec to 10 sec. block method overestimated the permanent dis-
Reading the Figure 9 with Figure 3a and 5 together, placement of the cantilever retaining wall. A variety
it can be noted that at the time when the acceleration of ground input motions were applied at the base of
of ground motion is applied towards the backfill soil the finite element model with different amplitude
(for example at time 5.5 sec, 7.1 sec, and 9 sec), the and frequency content to investigate the effect of
wall-soil system is accumulated maximum earthquake characteristics on the permanent dis-
permanent displacement away from the backfill soil placement of the cantilever retaining wall. The re-
and the total inertia force of the wall-soil system is sults of the parametric study shows that the critical
acting away from the backfill soil while the total scenario for the design of the cantilever retaining
increment of seismic earth pressure force has the wall, predicting maximum permanent displacement,
minimum value. However, when the ground is when the ground motion is subjected to a ground
acceleration is applied away from the backfill motion having maximum amplitude but a minimum
soil (for example at time 6 sec, 7 sec, and 7.5 sec), frequency content. The parametric study also shows
the wall-soil system is accumulated permanent that the permanent displacement is highly sensitive
displacement towards the backfill soil, and the to the duration of the earthquake. The result also
seismic inertia force of wall system is acting towards show that the seismic earth pressure force has a low
the backfill soil, and the wall-soil system is resisted impact in the accumulated permanent displacement
by the maximum seismic earth pressure force. Thus, away from the backfill soil, while the same has sig-
it can be observed that the total seismic earth pres- nificant effect to resist the accumulated permanent
sure force has low impact in the accumulated per- displacement towards the backfill soil.
manent displacement away from the backfill soil,
while it has a remarkable contribution to resist the
accumulated permanent displacement towards the REFERENCES
backfill soil.
Brinkgreve, R., Engin, E. & Swolfs, W. 2016. PLAXIS 2016.
PLAXIS bv, The Netherlands.
9 CONCLUSION Candia, G., Mikola, R. G. & Sitar, N. 2016. Seismic response
of retaining walls with cohesive backfill: Centrifuge
The main aim of this paper is to use performance- model studies. Soil Dynamics and Earthquake
based method for predicting the seismic permanent Engineering 90: 411-419.
displacement of a cantilever-type retaining wall. Green, R. A., Olgun, C. G. & Cameron, W. I. 2008. Response
Emphasis has been given to investigate the effect of and modelling of cantilever retaining walls subjected
earthquake characteristics and the seismic earth to seismic motions. Computer‐Aided Civil and
pressure on the earthquake-induced permanent dis- Infrastructure Engineering 23(4): 309-322.
placement of the cantilever retaining wall. Finite el- Jo, S.-B., Ha, J.-G., Lee, J.-S. & Kim, D.-S. 2017. Evaluation
ement method is proposed in the current study to of the seismic earth pressure for inverted T-shape stiff
 retaining wall in cohesionless soils via dynamic
centrifuge. Soil Dynamics and Earthquake
Engineering 92: 345-357.
Jo, S.-B., Ha, J.-G., Yoo, M., Choo, Y. W. & Kim, D.-S. 2014.
Seismic behaviour of an inverted T-shape flexible
retaining wall via dynamic centrifuge tests. Bulletin of
earthquake engineering 12(2): 961-980.
Kloukinas, P., di Santolo, A. S., Penna, A., Dietz, M.,
Evangelista, A., Simonelli, A. L., Taylor, C. &
Mylonakis, G. 2015. Investigation of seismic response
of cantilever retaining walls: Limit analysis vs shaking
table testing. Soil Dynamics and Earthquake
Engineering 77: 432-445.
Kuhlemeyer, R. L. & Lysmer, J. 1973. Finite element method
accuracy for wave propagation problems. Journal of
Soil Mechanics & Foundations Div 99: 421-7.
Mononobe, N. & Matsuo, M. 1929. On the determination of
earth pressures during earthquakes. Proceedings,
World Engineering Congress 9: 179-187.
Nakamura, S. 2006. Reexamination of Mononobe-Okabe
Theory Of Gravity Retaining Walls Using Centrifuge
Model Tests. Soils and Foundations 46(2): 135-146.
Newmark. 1965. Effect of earthquakes on dams and
embankments. Geotechnique 15: 139–159.
Richards JR, R. & Elms, D. G. 1979. Seismic Behaviour of
Gravity Retaining Walls. Journal of Geotechnical and
Geoenvironmental. Dn., ASCE 105: 449-464.

View publication stats