effect of earthquake on gravity retaining walls

University of Kufa

Faculty of Engineering
Geotechnical Engineering

Department

Earth structures Report

Effect of earthquake on gravity

retaining walls
Teacher By :

Asst. Prof. Dr. Mohammed Shakir Mahmood

MSC .Student : Sheymaa Mahmood Badr

 effect of earthquake on gravity retaining walls

INTRODUCTION:
Retaining walls are one of geotechnical structures mostly used in trenches and
slopes to stabilize the earth in order to construct highways, buildings, and other
structures. Construction of these structures is especially necessary in the case
that there is limitation in usage of land. In order to calculate the induced forces
on retaining walls, theories of soil earth pressure were presented by Colomb and
Rankin (Bowles1982). In order to consider the effect of earthquake on the
behavior of retaining walls, different methods were developed based on limited
displacement concept, but the most common method is the method based on
Mononobe-Okabe theory which is based on modification of the Colomb theory
by Mononobe-Okabe (Prakash 1965). This method is based on classic
equilibrium limit theory which is applied on soil wedge and involves a static
force for considering inertial effects of earthquake loading on soil wedge.

Analyzing the behavior of earth retaining structures under seismic conditions has
been very important issue due to their wide applications in several infrastructural
applications and other structures. The problem of instability of walls is mainly
related to earth pressure distribution on the wall and the response of wall against
the earth pressure, especially, under dynamic loading condition. Soil – wall
interaction is an important property which governs the dynamic behavior of the
wall. Even after a large number of studies, the dynamic behavior of soil-wall
system is still not completely elucidated. There have been several damages to
retaining walls due to earthquakes in many parts of the world. Damages to
retaining walls used for slope stability, road and railway embankments, bridge
abetments and in nuclear power reactors are highly sensible as the damages to
these structures are related to a large number of casualties. For many decades,
the seismic analysis of retaining walls has been based on the simple extension of
Coulomb’s limit equilibrium analysis, also known as M-O method, by treating
the dynamic forces from a seismic event as quasi-static forces acting along the
failure plane. Dynamic lateral earth pressures behind a soil retaining wall
subjected to an intensive ground motion can be significant and cannot be
determined accurately using pseudo-static approach. These additional earth
pressures can result in damage to or collapse of the structure.
Dynamic earth pressures depend on a large number of parameters such as
backfill density, angle of internal friction of soil, structural design of wall,
ground motion parameters like peak ground acceleration, duration of strong
motion and predominant frequency of the earthquake.
 effect of earthquake on gravity retaining walls

1-Behavior of cantilever retaining walls under seismic

condition
study the behavior of an L-shaped precast cantilever retaining wall with dry backfill
subjected to several ground motions having different predominant frequency. The
physical model of cantilever retaining wall tested by Watanabe et al. (2003) was
simulated in finite element model and the results were compared to validate the model.
The finite element software package Abaqus® v6.8 was used to carry out the
numerical experiments. The program is widely being used for both linear and non-
linear numerical simulations in various engineering fields as it is quite fast in non-
linear analysis.

1-1 NUMERICAL MODEL

Physical model of cantilever retaining wall tested by Watanabe et al. (2003), shown in
Fig. 1-(a), was adopted as the typical geometry for the generation of 2D retaining wall
model. Total height of the wall was 530 mm and the bottom width at the base of the
cantilever retaining wall was 230 mm. The subsoil and backfill layers were made of
air-dried Toyoura sand (Gs = 2.648, emax = 0.977 and emin = 0.609) with average
relative density of 90%. A surcharge of 1 kPa was applied on the surface of the soil
backfill.
 effect of earthquake on gravity retaining walls

1-2 Results of parametric studies

Lateral earth pressures (lateral normal stress on wall) after different seismic events
have been compared in Fig. 4. The very first thing to observe is the earth pressure
profile obtained along the height of the wall was non linear. As compared to ‘at rest’
earth pressure, the post earthquake earth pressures have lower values. During a
seismic event, earth pressure increases due to ground acceleration which tends to shift
or translate and topple the wall from its rest position. As wall was allowed to tilt, this
excess pressure was mobilized and hence earth pressure reduces from actual value and
wall gets a permanent deformation after earthquake.

The changes in earth pressures (stress), from the initial rest position, due to dynamic
event have been shown in Fig. 5. The change in earth pressures due to backfill were
 effect of earthquake on gravity retaining walls

measured by observing the lateral normal stress on wall and have been presented in
Fig. 5. These values are negative and show decrement in earth pressures as the excess
earth pressures have been mobilized due to displacement and tilt of wall. Starting from
bottom of wall, stresses developed after each earthquake have almost similar profile
up to height 0.2 m and varied from 600 Pa to 800 Pa at this height. But from 0.2 m to
top of the wall, normal stresses varied and high predominant frequency seismic event
generated lower stress values after the earthquake. Again up to a height of 0.2 m, this
change is having similar pattern for all events but above this height the change in
stresses is very abrupt.
effect of earthquake on gravity retaining walls
 effect of earthquake on gravity retaining walls

2- Effect of earthquake characteristics on the permanent displacement

of a cantilever retaining wall.

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 considering 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 seismic earth pressure has a
low impact on the permanent displacement.

2-1 PROBLEM DESCRIPTION

As shown in Figure 1 cantilever retaining wall, with a footing slab width L and total
stem height H, is constructed to provide lateral support to a horizontal backfill layer.
For a proper seismic design of the cantilever-type retaining wall, it should be ensured
that retaining wall does not fail by excessive permanent displacement. For a real
treatment of the permanent displacement of a cantilever-type retaining wall, it is
required to consider the following parameters (Fig. 1): 1) Total increment seismic
earth pressure force (Pae=Pae - Pa), computed along the vertical line passing
through the heel; where, Pae =total seismic earth pressure force; Pa =total static earth
pressure force; 2) Total seismic inertia force of the retaining wall (FW), which
includes the total seismic inertia forces of the stem and footing slab; 3) Total seismic
inertia forces of backfill soil above the footing slab (FS); 4) Total friction resistance
force between the footing slab and foundation layer (FF). The cantilever-type
retaining wall maintains its stability from the weight of backfill soil above the footing
slab in addition to its self-weight. So, the seismic earth pressure is assumed to be
developed along the vertical line extended from the heel up to the backfill soil surface
(Fig. 1). The results obtained from current finite element analysis will be compared
with traditional methods like Newmark sliding block method. After that, a variety of
ground motions is used in order to investigate the effect of earthquake characteristics
like amplitude and frequency content on the permanent displacement of a cantilever
retaining wall.
 effect of earthquake on gravity retaining walls

2-2 FINITE ELEMENT METHOD

A finite element (FE) model has been developed in current study by using the
PLAXIS 2D software (Brinkgreve et al. 2016) as shown in Figure 2 in order to
investigate the seismic performance of a 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. The stem member is assumed to have a fixed connection with the footing slab.
The backfill soil and the foundation soil are modelled using 6-noded triangular
elements of the PLAXIS 2D library (Brinkgreve et al. 2016), while the cantilever-type
retaining wall was modelled using plate elements. The maximum height of the element
is limited by 20% of minimum wavelength as recommended by Kuhlemeyer &
Lysmer (1973). The interaction between the cantilever-type retaining wall and backfill
soil as well as between the footing slab and foundation layer has been modelled by
using the 6-noded interface elements, available in the PLAXIS 2D library (Brinkgreve
et al. 2016). The absorbing boundaries are applied to the vertical boundaries of the
finite element model in order to reduce the effect of seismic wave reflection in the
finite element domain and to increase the accuracy of the analysis.
 effect of earthquake on gravity retaining walls

2-3 MATERIAL BEHAVIOUR

The backfill soil and foundation layer are simulated by using hardening soil with small
strain model,which is available in PLAXIS 2D (Brinkgreve et al., 2016) library.
However, the cantilever retaining wall is simulated by using a linear viscoelastic
constitutive model. Table 1 shows the parameters of hardening soil with small strain
model as well as the material parameters of a cantilever retaining wall.
 effect of earthquake on gravity retaining walls

2-4 TOTAL INCREMENT OF SEISMIC EARTH PRESSURE FORCE

The total seismic earth pressure force Pae has been calculated along virtual line
passing through the heel. The total increment of seismic earth pressure force Pae is
estimated by subtracting the total seismic earth pressure force Pae from the total static
earth pressure force Pa. Figure 9 shows the total increment of seismic earth pressure
force, estimated at the virtual line passing through the heel (Pae) between the time 5
sec -10 sec when the maximum permanent displacement of the wall-soil system has
been accumulated. It can noted from the Figure 8 that when the ground acceleration
applied towards the backfill soil (for example at time 5.5 sec, 7.1 sec, and 9 sec
(Fig. 3a) the total increment of seismic earth pressure force is close to zero, and the
total seismic earth pressure force is close to the static earth pressure force. However,
when the ground acceleration changes its direction away from the backfill soil (for
example at time 6 sec, 7 sec, and 7.5 sec (Fig. 3a), the maximum increment of seismic
earth pressure force is developed along the virtual line passing through the heel.
 effect of earthquake on gravity retaining walls

3- Effect of vertical earthquake shaking on displacement of earth retaining

structures
3-1 Numerical Modelling
Numerical modelling was chosen for this study as this would be
independent of the previous analytical research, and would enable better
representation of the soil and earthquake shaking characteristics.
The finite difference program, FLAC (Fast Lagrangian Analyses of
Continua) was chosen because of its capability to accommodate large
displacements, its successful use by other researchers for assessing
earthquake behaviour of reinforced soil retaining walls and its availability
and familiarity. FLAC is a two-dimensional explicit finite difference code,
which simulates the behaviour of structures built of materials that may
undergo plastic flow when their yield limit is reached (Itasca Consulting
Group, 1993). Version 3.23 of FLAC was used with the associated dynamic
analysis module. The model used consisted of a 7.5 m high reinforced earth
 effect of earthquake on gravity retaining walls

wall with 7.5 m long steel strips as shown in Figure 1. To overcome

difficulties experienced during modelling due to different lateral boundary
heights, a slope was modelled well away from the front face of the wall.
The model was built in stages to represent the construction sequence.
Gravel backfill and a dense gravel foundation were modelled to minimise
foundation deformations, so that the wall displacement could be studied.
A reinforced earth retaining wall was chosen because the reinforced earth
block will displace mainly by sliding (no tilting), and its common use in
New Zealand, where walls of significant heights (greater than 3 m) are
required to support important structures.

3-2 Earthquake inputs

Earthquake time histories with significant peak vertical and horizontal
accelerations were chosen from records within 20 km of the epicentre of
earthquakes with a magnitude greater than 6.5. The characteristics of the
four Californian earthquake records chosen are presented in Table 1.
 effect of earthquake on gravity retaining walls

3-3- Analyses
The model was analysed by running FLAC with the earthquake input at the base of the
model. Displacements were measured at various points in the model during the
analyses, particularly at the top, middle and bottom of the wall face. The net
displacements were calculated as the difference between the displacements recorded at
selected points and the displacement recorded at a node within the foundation well
below the wall base. The earthquake parameters used in the assessment were derived
from accelerations recorded at the base of the wall during the numerical analyses.

3-4- Results
Pull-out of the upper strips was observed and affected the results of the displacement
at the top of the wall. The pull out is considered to be due to the lack of confining
stress at the top of the wall. In practice, the upper strips are bent down to a lower level
to attract a higher confinement stress, but this complication was avoided in the
numerical model. To minimise the pull-out of the strips at the top of the wall, longer
strips were instead modelled as shown in Figure 1. The displacements of the lower
wall under all combinations of horizontal and vertical motions were :
* less than 25 mm in the Group I earthquake analyses,
* less than 200 mm in the Group II earthquake analyses.
The maximum displacement of the lower half of the wall was only 23 mm when only
horizontal shaking was applied for both Group I and Group II earthquakes. This
illustrates the important effect of vertical earthquake motions.

3-5 APPRAISAL OF THE RESULTS

3-5 A- Energy and Frequency Content
The power spectral density (PSD) of an earthquake record is a representation of the
energy against frequency in an earthquake. It is calculated from the square of the
Fourier amplitude spectrum. The power spectral density for the input motions at the
base of the walls was calculated to represent the frequency and power content of the
earthquake inputs, and is presented in Figure 2. By summing the power in all
requency bins of the earthquake record for each component, a representation of the
total energy content of the component was obtained .
 effect of earthquake on gravity retaining walls

3-5 B- Horizontal shaking alone

The displacements for the top of the wall from the model analyses with the application
of horizontal motions alone are shown on Figure 3.

3-5-C- Horizontal and vertical shaking

The displacement ratio (DR) is the ratio of displacement with vertical and horizontal
shaking to that with horizontal shaking alone.
 effect of earthquake on gravity retaining walls

CONCLUSION
For Behavior of cantilever retaining walls under seismic condition The
wall-soil system tested by Watanabe et al. (2003) was simulated into finite
element model to study dynamic earth pressures developed during different
seismic events. Primary results produced by the FE model were comparable
to the physical model results and for static conditions the results were
approximately matching with the theoretical values. After discussing the
results - The lateral earth pressure profile for different predominant
frequencies was similar up to a certain height and then varied abruptly. This
can be explained by the formation of the sliding soil wedge at the backfill.
- The vertical stresses at wall base increased on the toe side and
consequently reduced towards the heel side due to tilting of wall base
caused by lateral seismic loads on wall stem.
- The residual wall top displacement was high for very low frequency
(0.847 Hz) seismic loads and low for high frequency (12.5 Hz) seismic
loads and showed no significant variation for middle ranged (2.0, 3.846 and
6.25 Hz) frequency earthquakes
- Stresses developed at the stem-base joint of the wall varied significantly
with the predominant frequency of the seismic events. Due to cantilever
action of the wall, earth pressures acting on wall base and stem caused the
 effect of earthquake on gravity retaining walls

wall stem-base joint to develop tensile stresses. The variation of this stress
followed the residual wall top displacement profile. For very high
predominant frequency (12.5 Hz) seismic loads, the wall top displacement
and the stresses developed in the wall stem-base joint were high and for
very low frequency (0.847) seismic loads both the parameter were quite
low. The results presented in this paper are based on the finite element
simulations on the small physical model size retaining wall model. Further
simulations are needed to be performed on full size model wall with more
realistic stress levels.

as for the second study Effect of earthquake characteristics on the

permanent displacement of a cantilever retaining wall: use performance
based method for predicting the seismic permanent displacement of a
cantilever-type retaining wall. Emphasis has been given to investigate the
effect of earthquake characteristics and the seismic earth pressure on the
earthquake-induced permanent displacement of the cantilever retaining
wall. Finite element method is proposed in the current study to consider
more real aspects of the seismic behavior of the wall-soil system. The soil
behaviour is simulated in the current numerical analysis by using hardening
soil with small strain model. The results of the current study show that the
Newmark sliding block method overestimated the permanent displacement
of the cantilever retaining wall. A variety of ground input motions were
applied at the base of the finite element model with different amplitude and
frequency content to investigate the effect of earthquake characteristics on
the permanent displacement of the cantilever retaining wall. The results of
the parametric study shows that the critical scenario for the design of the
cantilever retaining wall, predicting maximum permanent displacement, is
when the ground motion is subjected to a ground motion having maximum
amplitude but a minimum frequency content.

For the Effect of vertical earthquake shaking on displacement of earth

retaining structures:
* Peak ground acceleration reflects the amplitude of a single pulse, and
does not represent strong motion duration, energy and frequency content of
earthquake shaking. It is shown to be a poor parameter for the assessment
of the displacement of retaining structures.
 effect of earthquake on gravity retaining walls

* Power spectral density represents the energy content of earthquake

shaking, and is a better parameter for the assessment of wall displacement
performance during earthquakes.
* Vertical shaking of earthquakes having a significant energy content had a
significant effect on the displacement of the model retaining wall. The
power spectral density of the vertical motions was a better parameter to
assess wall displacements than peak vertical acceleration.
* Vertical shaking generally had a small effect on wall displacements when
the earthquake shaking only had modest energy and peak ground
acceleration, and when the energy content of the earthquake at a frequency
similar to the first-mode period of response of the wall was low.
* Retaining walls can undergo large displacements due to resonance, when
exposed to earthquakes with significant energy content at frequencies
similar to their first-mode of response, even when the total energy content
of the earthquake is not large.
* It is hypothesised that vertical shaking increases the flexibility of
reinforced earth retaining structures and hence its period, and if the
earthquake has a large energy content at this modified frequency, then
resonance and significantly greater displacements could result.
* Design methods should be modified to take into consideration the energy,
frequency content and the vertical shaking of earthquakes, particularly for
walls that support important structures.

Reference
 Brabhaharan P, Fairless G, Chapman H Effect of vertical earthquake shaking on
displacement of earth retaining structures. In: Proceedings of the Pacific
conference on earthquake engineering, New Zealand Society for Earthquake
Engineering, Christchurch, 2003 .
 Puri VK, Prakash S, Widanarti R (2004) Retaining walls under seismic loading .
 Anurag UPADHYAY AMKaKDS (2011) BEHAVIOR OF CANTILEVER
RETAINING WALLS UNDER SEISMIC CONDITIONS. Paper presented at
the 5th international conference on earthquake geotechnical engineering, Chile ,

 Junied Bakr SMA (2018) Effect of earthquake characteristics on permanent

displacement of a cantilever retaining wall .