Available online at www.eccomasproceedia.

org
Eccomas Proceedia COMPDYN (2017) 3671-3681

COMPDYN 2017
6th ECCOMAS Thematic Conference on
Computational Methods in Structural Dynamics and Earthquake Engineering
M. Papadrakakis, M. Fragiadakis (eds.)
Rhodes Island, Greece, 15–17 June 2017

NUMERICAL SIMULATIONS WITH CUSHIONING MATERIAL

INSTALLED BEHIND THE RETAINING WALL FOR RESILIENT
STRUCTURES
Kazuhiro Kaneda 1, Hiroyuki Yamazaki 2, and Satoru Ohtsuka 3
1
Takenaka Corporation
1-5-1, Ohtsuka, Inzai, Chiba, Japan, 270-1395
kaneda.kazuhiro@takenaka.co.jp
2
Port and Airport Research institute
3-1-1, Nagase, Yokosuka, Kanagawa, Japan, 239-0826
yamazaki@pari.go.jp
3
Nagaoka University of Technology
1603-1, Kamitomioka, Nagaoka, Niigata, Japan, 940-2188
ohtsuka@nagaokaut.ac.jp

Keywords: Retaining wall, Cushioning material, Earth pressure, Dynamic simulation, Quasi-
active state.

Abstract. Following the Hyogoken Nanbu earthquake in 1995, the design earthquake wave
was larger. After the Tohoku chihou Taiheiyouoki earthquake in 2011, the tsunami was fo-
cused after earthquake. In the future large earthquakes are expected to occur in Japan. We
are required to design more resilient structures that can withstand these events.
We conducted a numerical simulation of the reduction of earth pressure in a sandy backfill.
When a cushioning material was placed behind the retaining wall, the earth pressure was re-
duced, because the backfill sand moved to the active earth pressure state. It is possible that
the design section of the gravity caisson became smaller. In the dynamic approach, shaking
table tests were performed and the earth pressure acting on the retaining wall was seen to
reduce. However, the mechanism of earth pressure reduction was not clear in the dynamic
tests. Therefore in this study, the mechanism of earth pressure reduction during an earth-
quake was clarified using a numerical simulation. From the simulation results, the residual
horizontal displacement in the case of sand is larger than that in the case of cushioning mate-
rial. It is cleared that in the cushioning material the inertial force of backfill sand was ab-
sorbed for the friction force toward to the foundation ground beside the effect of quasi-active
state.

© 2017 The Authors. Published by Eccomas Proceedia.

Peer-review under responsibility of the organizing committee of COMPDYN 2017.
doi: 10.7712/120117.5674.18382

3671
 Kaneda Kazuhiro, Yamazaki Hiroyuki and Ohtsuka Satoru

1 INTRODUCTION
The mechanism of earth pressure reduction in a backfill with cushioning material behind
the retaining wall is herein discussed. The influences of Poisson’s ratio and the elastic modu-
lus of the cushioning material on the reduction of earth pressure against a retaining wall were
examined by assuming the material to be an elastic body. Numerical simulations revealed that
the attainment of the active state in the sandy backfill was due mainly to the lower Young’s
modulus and Poisson’s ratio of the cushioning material in the static condition. Under static
loading, the soil in the vicinity of a wall enters the active state, or what has also been called a
“quasi-active state,” in which it is deformed by the cushioning material buffer. The concept of
the quasi-active state was first introduced by Hazarika and Okuzono (2004), in which EPS
was used as a compressible inclusion to reduce the dynamic earth pressure against the retain-
ing wall. Hazarika et al. (2008) extended the cushioning technique to reduce the dynamic
earth pressure against the retaining wall and validated the approach by conducting shaking
table tests on gravity type retaining structures. However, the mechanism of earth pressure re-
duction was not clear in the dynamic case. Therefore, in this study, the mechanism of earth
pressure reduction during an earthquake was clarified using a numerical analysis. First, the
mechanism of earth pressure reduction was explained for the static case. Subsequently, the
deformation of the retaining wall with earth pressure reduction was discussed for the dynamic
case.

2 STATIC CASE

2.1 Simulation condition

Numerical simulations were performed in order to simulate the static and dynamic cases.
The FEM program code MuDIAN, developed by Takenaka Corporation (Shiomi et al., 1993)
was used in the simulation. The numerical mesh in the case of the cushioning material and
boundary conditions are indicated in Figure 1. A sandy foundation ground of 2 m and a sandy
backfill 2 m high and 10 m wide was constructed in the plane strain condition. The bottom of
foundation ground was kept fixed. Two cases were simulated. One case was simulated with
sand at the backfill, while the other had a 0.5 m thick layer of cushioning material at the back-
fill. The Mohr Coulomb failure criterion of the soil was used as the constitutive model in this
study. Table 1 shows the material constants of sand and the cushioning material used in the
simulation. The cushioning material and the retaining wall were considered elastic bodies.
The friction angle of soil was set at 40°. The stiffness of the foundation ground was greater
than that of the backfill sand. The cushioning material was assumed to be tier chips (Kaneda.
et.al. 2011); its elastic modulus and Poisson’s ratio were small compared with those of the
soil. The retaining wall was assumed to be composed of concrete. The slip between the retain-
ing wall and soil or cushioning material was not considered in this simulation. Initial calcula-
tions were performed using the self-weight of the soils. Then, an additional load of 50 kN/m2
was applied at the top (red part in Fig.1).

3672
 Kaneda Kazuhiro, Yamazaki Hiroyuki and Ohtsuka Satoru

50 kN/m2
4m 1 m 0.5 m Vertical load 4.5m

Retaining
wall Backfill sand 2m
Cushioning
material

Foundation ground
z 2m

x
Figure 1: Boundary conditions.

Foundation Cushioning
Backfill sand Retaining wall
ground material
Elastic modulus
112000 56000 24400000 200
(kN/m2)
Poisson ratio 0.3 0.3 0.2 0.2
Unit weight (kN/m3) 16.62 16.62 23.03 8.82
Cohesion (kN/m2) 1 1
Friction angle (Degree) 40 40

Table 1: Material constants.

2.2 Simulation results and considerations

Figure 2 shows the stress distribution in the vertical direction (compressive stress) acting
on the retaining wall from the sand or the cushioning material after loading. In the case of
sand, the compressive stress from the sandy backfill was transmitted to the retaining wall. In
the case of the cushioning material, the compressive stress is smaller than that in the case of
sand at the total. Figure 3 shows the equivalent shear strain distributions after loading. The
shear strain in the cushioning material case is larger than that in the sand case. Figure 4 shows
the mechanism of earth pressure reduction in the static condition (Kaneda. et.al. 2011). When
cushioning material is present in the backfill and the backfill sand underwent deformation, the
cushioning material entered the passive state and the stress increased. On the other hand, the
backfill sand entered the active state and the stress decreased with deformation. The equilibri-
um state is the intersection between the passive cushioning material and the active backfill
sand. Thus, earth pressure reduction is the difference between the stress values at the point of
intersection and the initial force. See the references in detail. (Kaneda. et.al. 2011).

3673
 Kaneda Kazuhiro, Yamazaki Hiroyuki and Ohtsuka Satoru

Sand 50cm
4.0
3.8

Height (m)
3.6
3.4
3.2
3.0
2.8
2.6
2.4
2.2
2.0
-20 0 20
Earth Pressure (kPa)

Figure 2: Stress distributions in the vertical direction acting on the retaining wall.

（％）
4.0

0.0

Sand case

（％）
4.0

0.0

Cushioning material case

Figure 3: Equivalent shear strain distributions.

3674
 Kaneda Kazuhiro, Yamazaki Hiroyuki and Ohtsuka Satoru

Stress σs
Earth pressure reduction σt

σt
σs

Cushioning material：Passive state

Backfill sand：Active state

Strain

Figure 4: Mechanism of earth pressure reduction.

3 DYNAMIC CASE

3.1 Simulation condition

Figure 5 shows the numerical mesh in the case of the cushioning material and boundary
condition in dynamic case. A sandy foundation ground of 8 m, a sandy backfill of 2 m high
and 40 m wide is constructed in plane strain condition. There are three simulation cases, one
with only sand, and the other two with 0.2 m and 0.5 m cushioning material. The bottom of
the mesh was set to be a viscos boundary and both the sides were connected to the free field,
which has a huge mass. The soil constitutive equation was obtained from the Yoshida mode.
See the reference for details. (Yoshida. et. al. 1993)
0.2 m,
1m 0.5 m
20 m 19.5 m

Retaining Viscos boundary

Viscos boundary wall
Backfill sand 2m

Cushioning
Free Free
material
Field Field 8 m
z
Foundation ground

Viscos boundary
x
Figure 5: Boundary conditions.

The dynamic deformation characteristic (G/G0) and damping ratio h are calculated from
equations (1) and (2). For the soil ground, hmax is assumed to be 21.0%.
G 1
(1)
G0 1
r

3675
 Kaneda Kazuhiro, Yamazaki Hiroyuki and Ohtsuka Satoru

G
h hmax 1 (2)
G0

where G: shear stiffness, G0: initial shear stiffness, r : reference strain, h: dumping, hmax:
maximum dumping
Table 2 lists the material constants. The hyperbolic model at r 0.1 was used for sand.
The material constants of both the Retaining wall and Cushioning material are the same as the
static case. Figure 6 shows the G/G0, h~γ curve in sand in this research.

Foundation Retaining Cushioning

Backfill sand
ground wall material
Elastic modulus(kN/m2) 112000 56000 24400000 200
Poisson ratio 0.3 0.3 0.2 0.2
Unit weight (kN/m3) 16.62 16.62 23.03 8.82
Reference strain r 0.1 0.1
Maximum dumping hmax(%) 21 21
Table 2: Material constants.

1 100

0.8 80

0.6 60

h (%)
G/G0

0.4 40

0.2 20

0 0
0.0001 0.01 1 100
(%)

Figure 6: G/G0 , h~γ curve.

Similar to the static case, the self-gravity analysis was performed first, followed by the dy-
namic analysis. Rayleigh damping was adopted for both structures and soils. Table 3 shows
the Rayleigh damping coefficient. The α and β were determined by f1=1Hz, f2=5Hz and
h=2%, respectively. Figure 7 shows the input motion. The bottom had an input of 2E with the
viscos boundary. Table 4 shows the viscos parameters. The slip between the retaining wall
and soil or cushioning material was also not considered in this simulation, as in the static case.

α β
damping 0.209440 0.001061

Table 3: Rayleigh damping.

3676
 Kaneda Kazuhiro, Yamazaki Hiroyuki and Ohtsuka Satoru

3
Imput motion

Acceleration (m/s2)
2
1
0
-1
-2
-3
0 10 20 30 40 50 60 70 80 90 100
Time (sec)

Time history
12 2.5
Imput motion Imput motion
Acceleration (m/s2)

10

Amplitude (m/s2)
2
8
1.5
6
1
4
2 0.5

0 0
0.01 0.1 1.0 10.0 0 5 10 15 20
Period (s) Frequency (Hz)

Response spectrum (h=5%) Fourier spectrum

Figure 7: Input motion.

Unit weight (kN/m3) 17.64

Vp(m/s) 2000
Vs(m/s) 340

Table 4: Viscos parameters.

3.2 Eigen value analysis

Table 5 shows the results of the eigen value analysis. Figure 8 shows the first eigen value
mode. The horizontal mode was observed as the first eigen value mode.

Mode Frequency (Hz) Period (s)

1 3.210 0.312 Horizontal mode
Sand 2 4.390 0.228 Horizontal mode
3 5.880 0.170 Vertical mode
1 3.200 0.312 Horizontal mode
20 cm
2 4.380 0.228 Horizontal mode
cushioning material
3 5.880 0.170 Vertical mode
1 3.200 0.312 Horizontal mode
50 cm
2 4.380 0.228 Horizontal mode
cushioning material
3 5.880 0.170 Vertical mode

Table 5: Eigen value.

3677
 Kaneda Kazuhiro, Yamazaki Hiroyuki and Ohtsuka Satoru

Sand 50 cm cushioning material

Figure 8: First eigen value mode.

3.3 Simulation results

Figure 9 shows the x direction time history from the retaining wall point (red circle). All
cases move to the left after 30 s. In the sand case, the deformation was the largest on the left
side. In the 0.5 m cushioning material case, it was the smallest. It was shown that displace-
ment is smaller when cushioning material is attached. Figure10 shows the earth pressure at the
retaining wall at both 29.1 s and 100 s (final). In the sand case, the earth pressure occurred at
the retaining wall. However, in the case of 50 cm cushioning material case, the earth pressure
did not occur. At the trend in the final case is also same. Figure 11 shows the equivalent shear
strain distributions at both 29.1 s and 100 s (final). At the 29.1 s mark, in the sand case, the
shear strain of the foundation ground in front of retaining wall occurred for the forward de-
formation. In the 0.5 m cushioning material case, there was shear strain on the lower backfill
sand. In the final case, there was more shear strain of the foundation ground in front of retain-
ing wall in the sand case. In the 0.5m cushioning material case, the shear strain occurred for
both, the lower backfill and its underlying foundation ground. In the sand case, for the inertial
force of the backfill sand, the retaining wall was pushed with the lower left of the retaining
wall acting as a fulcrum. On the other hand, in the 0.5 m cushioning material case, while there
was shear stain of the foundation ground in front of retaining wall, for the cushioning, the in-
ertial force of backfill sand was absorbed for the friction force toward to the foundation
ground as shown by the shear strain of foundation ground. As a result, it is considered that the
earth pressure at the retaining wall was small. Like the static, it is also considered that the ef-
fect of the quasi-active state existed for the reduction of earth pressure.

1.5
sand
1 20cm
Displacement (cm)

0.5 50cm
0
-0.5
-1
-1.5
-2
x 0 20 40
Time (sec)
60 80 100

Figure 9: X direction time history.

3678
 Kaneda Kazuhiro, Yamazaki Hiroyuki and Ohtsuka Satoru

Sand 50cm
4.0
3.8
Height (m)

3.6
3.4
3.2
3.0
2.8 2

Displacement (cm)
2.6 1
2.4
0
2.2
-1
2.0
-20 -10 0 10 20 30 -2
25 27 29 31 33 35
Earth Pressure (kN/m2) Time (sec)

29.10 sec

Sand 50cm
4.0
3.8
Height (m)

3.6
3.4
3.2
3.0
2.8 1
Displacement (cm)

2.6 0.5
2.4 0
-0.5
2.2 -1
2.0 -1.5
-20 -10 0 10 20 30 -2
0 20 40 60 80 100
Earth Pressure (kN/m2) Time (sec)

100 sec(final)

Figure 10: Earth pressure at retaining wall.

3679
 Kaneda Kazuhiro, Yamazaki Hiroyuki and Ohtsuka Satoru

（％）
2.0

0.0

Sand 50 cm cushioning material

（％）
2.0 29.10 s

0.0

Sand 50 cm cushioning material

100 s(final)
Figure 11: Equivalent shear strain distributions.

4 CONCLUSIONS
Earth pressure reduction of the retaining wall with cushioning material was conducted us-
ing numerical simulations. In the static case, further reduction of earth pressure was observed
in the cushioning material as compared to the sand case. This is likely due to the backfill sand
turning into the active state in the cushioning material case (quasi-active state). In the dynam-
ic case, a smaller deformation of the retaining wall was observed in the cushioning material as
compared to the sand case. It is considered that in the cushioning material, the inertial force of
backfill sand was absorbed by the friction force toward to the foundation ground in addition
to the effect of the quasi-active state.
Finally, in this simulation, the slip between the retaining wall and the cushioning material
or sand backfill was not considered. However, the effect of slip is important for the defor-
mation in the dynamic analysis. In the future, we would like to include the slip effect in nu-
merical simulations.

REFERENCES
[1] Hazarika, H. and Okuzono, S., Modeling the Behavior of a Hybrid Interactive System
Involving Soil, Structure and EPS Geofoam, Soils and Foundations, Japanese Geotech-
nical Society, Vol. 44, No. 5, 149-162., 2004.
[2] Hazarika, H., K.ohama, E. and Sugano, T., Underwater Shaking Table Tests on Water-
front Structures Protected with Tire chips Cushion, Journal of Geotechnical and Geo
Environmental Engineering, ASCE, Vol. 134, No. 12, 1706-1719., 2008

3680
 Kaneda Kazuhiro, Yamazaki Hiroyuki and Ohtsuka Satoru

[3] Shiomi, T., Shigeno, Y. and Zienkiewicz, O. C., Numerical prediction for model No.1,
Verification of Numerical Procedures for the Analysis of Soil Liquefaction Problems
(eds. By Arulanandan and Scott), Balkema, pp.213-219., 1993
[4] Kaneda, K, Hazarika, H. and Yamazaki, H, Mechanism of Earth Pressure Reduction Us-
ing Tire Chips in Sand Backfill, The 14th Asian Regional Conf. on Soil Mechanics and
Geotechnical Engineering, Hong Kong, CD-ROM., 2011
[5] Yoshida, N. and S. Tsujino, "A simplified practical stress-strain model for the multi-
dimensional analysis under repeated loading." The 28th Japan national conference on
soil mechanics and foundation engineering., 1993

3681