Soil Dynamics and Earthquake Engineering 102 (2017) 86–98

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering

journal homepage: www.elsevier.com/locate/soildyn

Dynamic characteristics of expanded polystyrene composite soil under MARK

traffic loadings considering initial consolidation state
⁎
Hongmei Gaoa,b, Chunyao Bua,b, Zhihua Wanga,b, , Yanqing Shena,b, Guoxing Chena,b
a
Urban Underground Space Research Center, Nanjing Tech University, 200 North Zhongshan Road, Nanjing 210009, China
b
Institute of Geotechnical Engineering, Nanjing Tech University, 200 North Zhongshan Road, Nanjing 210009, China

A R T I C L E I N F O A B S T R A C T

Keywords: As a type of artificial filling material, the shear modulus and damping ratio of expanded polystyrene (EPS)
EPS composite soil composite soil are two key parameters used to analyze the dynamic stability of an embankment. Nineteen
Traffic loadings combined axial–torsional tests are conducted on hollow cylinder specimens of EPS composite soil to study its
Initial consolidation state dynamic characteristics under the complex stress path induced by the simulated traffic loadings. The char-
Mixing ratio
acteristics of skeleton curve, dynamic shear modulus and damping ratio for EPS composite soil are analyzed. It is
Structural damping
found that EPS composite soil is characterized by typical dynamic nonlinearity, which is influenced by the
mixing ratio and the initial stress state. The increasing cement content can effectively improve the dynamic
strength of EPS composite soil. EPS bead content has a slight influence on the initial shear modulus of EPS
composite soil as well as the cyclic stress-strain curve in the linear elastic stage. However, the increasing EPS
bead content obviously reduces the dynamic strength. The initial shear modulus increases with increasing initial
minor principal stress for the isotropic and anisotropic consolidated specimens. The characteristics of modulus
attenuation are significantly influenced by the initial minor principal stress, EPS bead content and initial rotation
angle of the major principal stress axis. The “structural damping” effect induced by the weak interface formed
between EPS beads and cemented soil is an important component of the damping mechanism for the EPS
composite soil. Based on the experimental results, this paper provides the empirical models to describe the
skeleton curve, modulus attenuation and damping growth characteristics for EPS composite soil.

1. Introduction material.
Over the past twenty years, numerous tests have been conducted on
Thanks to the rapid development of geotechnical materials and the EPS composite soil to study its physical and mechanical characteristics.
requirements regarding the protection of the environment since the Conventional testing methods are adopted, including unconfined com-
middle of the past century, a new type of filling material—namely, EPS pression tests, uniaxial compression tests, direct shear tests, and triaxial
composite soil—is proposed and utilized in practice in the 1960s. EPS compression tests [10–12]. The mixing ratio and its influence on the
composite soil is a mixture that is usually mixed with EPS (expanded strength, deformation and failure modes have been investigated and
polystyrene) beads, soil, cement and water. EPS composite soil has been discussed [13–15].
widely used in many fields such as soft soil treatment, embankments When used on land, in the coast, or underwater, EPS composite soil
and bridge abutments, expressways, and underground pipelines [1–6]. is likely to undertake cyclic loadings such as seismic loading, traffic
As a type of artificial geotechnical material, EPS composite soil is loading, or wave loading. Therefore, the dynamic characteristics of EPS
characterized by controllable strength and deformation, and its me- composite soil are also a concern for engineering. At present, the dy-
chanical properties can be adjusted by changing the mixing ratio to namic characteristic studies on EPS composite soil are focused on the
satisfy the requirement [7–9]. In addition, it is a type of en- cyclic stress–strain relationship, modulus, damping ratio, dynamic
vironmentally friendly material [5,8]. For instance, some wastes such strength and so on. Some factors affecting dynamic characteristics are
as dredged mud and waste packaging foam can be used to prepare EPS discussed such as the mixing ratio, initial loading condition (the mag-
composite soil. In terms of convenience of construction, EPS composite nitude and frequency of cyclic loading, initial consolidation condition)
soil possesses high fluidity and good workability, which can meet the and curing days. The conventional resonance column test and cyclic
requirement of uniformity and compactness as a bulk mass filling triaxial shear test are the most commonly used methods.

⁎
Corresponding author at: Urban Underground Space Research Center, Nanjing Tech University, 200 North Zhongshan Road, Nanjing 210009, China.
E-mail address: [email protected] (Z. Wang).

http://dx.doi.org/10.1016/j.soildyn.2017.08.012
Received 20 September 2016; Received in revised form 9 June 2017; Accepted 23 August 2017
0267-7261/ © 2017 Published by Elsevier Ltd.
H. Gao et al. Soil Dynamics and Earthquake Engineering 102 (2017) 86–98

When used as the filling material of an embankment, EPS composite

soil is in a complex stress state, which can be reflected in the followings:
1) the initial stress conditions differ for the soil at different positions;
not only is the initial consolidation stress ratio different but also the
principal stress axis rotates more obviously for the soil close to the slope
shoulder; and 2) under the traffic loadings, the principal stress axis in
the soil continuously rotates and the stress path displays a typical
‘heart’ shape. In fact, the existing studies on sand, soft soil or mixed soil
have shown that the influence of the rotation of the principal stress axis
on the strength and deformation of soils cannot be neglected [16–18].
The cyclic torsional shear test on hollow cylinder specimen is the
most effective approach to study dynamic behavior of soil under com-
plex stress conditions [19,20]. The hollow cylinder torsional apparatus Fig. 1. EPS beads used in the study.
can independently apply the axial force, torque, inner and outer con-
fining pressure to the specimen, which makes it possible to simulate the
2.2. Specimen preparation
complex initial stress condition and multiple loading stress paths. The
torsional shear tests on the hollow specimens have been employed to
The specimen preparation process for EPS composite soil is as fol-
study cyclic behavior of sandy and clayey soils under drained or un-
lows: 1) the soil is oven-dried and then crushed into powder with a
drained condition, including cyclic shear strength (liquefaction re-
rubber hammer; 2) according to the desired mixing ratio, the dried soil
sistance) and residual deformation and so on [21–24]. In the meantime,
and Portland cement are weighed and mixed; 3) the desired water is
it is widely used to evaluate the dynamic performance of soil under
added into the mixture and stirred for approximately 10 min until
cyclic loading conditions such as traffic loading, seismic loading, or
uniform slurry is formed; 4) the desired amount of EPS beads are added
wave loading [25,26]. However, the present studies have not paid
into the slurry and fully stirred again for approximately 10 min until a
sufficient attention to the influence of the initial stress condition and
uniform mixture is formed; and 5) the mixture is placed into a sampling
complex stress path induced by traffic loadings on the dynamic char-
device and cured for the required number of days.
acteristics of EPS composite soil.
The hollow cylinder specimen of EPS composite soil in this study is
In this paper, a series of combined axial–torsional tests on EPS
60 mm in inner diameter, 100 mm in outer diameter and 200 mm in
composite soil were conducted to investigate its dynamic characteristics
height. To prepare and cure the specimen, a sampling device is in-
under different initial stress conditions and a ‘heart-shaped’ stress path.
dependently developed in Nanjing Tech University, China. The sam-
The characteristics of the cyclic stress–strain relationship, shear mod-
pling device mainly comprises a three-section inner mould, three-sec-
ulus and damping ratio for EPS composite soil are discussed, con-
tion outer mould, top plate and base, as shown in Fig. 2. After the
sidering the change of cement content, EPS bead content and curing
prepared mixture is placed into the sampling device, a small amount of
days. The influences of initial stress state parameters including the in-
water is apt to accumulate at the bottom. The accumulated water is
itial mean effective principal stress, consolidation stress ratio, inter-
drained by the drainage slot designed at the bottom shown in Fig. 2(d)
mediate principal stress coefficient and rotation angle of the principal
to avoid its influence on the specimen formed during the curing process.
stress axis are discussed in detail.
Generally, the specimen can be released from the sampling device
after curing for 24 h. The demolding process was separated into three
steps as follows: 1) the smaller part of the inner mould was pulled out
2. Experimental program
by a steel hook passing through a hole in it, and then the other two
larger parts were removed as shown in Fig. 2(a); 2) the outer mould
2.1. Material properties
with the hollow cylinder specimen was placed on the top plate, and
then the specimen was slowly pushed down by pressing the porous
The EPS composite soil used in this paper is a type of manmade
plate onto it as shown in Figs. 2(b) and (c); and 3) the specimen for EPS
mixture that is mixed with soil, EPS beads, Portland cement and water.
composite soil was obtained as shown in Fig. 2(d). After demolding, the
The soil sample is muddy and silty clay and taken from Yangzi Jiang
specimen was cured inside the curing box with a constant temperature
River belonging to floodplain phase deposits. The physical and me-
of 20 ± 2 °C and relative humidity at 100% until the testing date.
chanical properties of the muddy and silty clay are summarized in
Compared with the conventional sampling method, the approach pro-
Table 1.
posed herein can reduce the disturbance to the hollow specimen during
The EPS beads shown in Fig. 1 are produced from expandable
the demolding process to ensure the integrity of the specimen.
polystyrene resin, which contains microscopic cells filled with a
blowing agent (usually pentanes or butanes). The resin is exposed to
steam under controlled pressure, which softens the cell walls. The beads
2.3. Testing program and loading method
are formed after the blowing agent expands and causes individual resin
beads to increase in volume by up to 40 times. The particle size of the
2.3.1. Testing program
round EPS beads ranges from 2 to 3 mm. The bulk unit weight of EPS
The purpose of this paper is to discuss dynamic characteristics of
beads is 0.159 kN/m3. Portland cement (P.O 32.5) is used as a binding
EPS composite soil under complex stress conditions. Two factors af-
material and water is used to carry out the hydration reaction and fa-
fecting the dynamic characteristic of EPS composite soil are discussed:
cilitate the mixing process.
the mixing ratio and initial stress condition. Regarding the mixing ratio,

Table 1
Physical and mechanical properties of the muddy and silty clay.

Water content Natural unit Saturation degree Void Specific Liquid limit Plastic limit Compression coefficient Modulus of compressibility
w (%) weight γ (kN/m3) Sa (%) ratio e gravity ds wL (%) wp (%) a1-2 (MPa−1) Es (MPa)

48.1 16.8 96.0 1.37 2.74 39.2 22.6 0.98 2.42

87
H. Gao et al. Soil Dynamics and Earthquake Engineering 102 (2017) 86–98

Outer mould

Inner mould

Drainage slot
(a) Top plate (b) (c) (d)
Fig. 2. The demolding process of an EPS composite soil specimen: (a)Remove the inner mould; (b) Press down the porous plate; (c) Push out the specimen; and (d) Specimen after
demolding.

three factors are discussed: cement content, a w ; EPS bead content, λVE ; be realized by adjusting the applied W , MT , pi and pe .
and curing days, CD. Of these, a w denotes the mass of cement per vo- This paper mainly considers the influence of the initial rotation
lume of the EPS composite soil, and λVE represents the volume ratio of angle of the major principal stress axis, α o , initial intermediate principal
EPS beads to the dried soil. The weight ratio of water to dried soil for all stress coefficient, bo , initial consolidation stress ratio, R o , and mean
specimens is 0.8. The curing days considered are 7 days, 14 days and 28 effective principal stress, po′. The above factors are expressed as in Eq.
days, respectively. The mixing ratios of specimens are shown in Table 2. (1).
The stress condition of the hollow cylinder specimen under axial
loading coupling with torsional shear loading is shown in Fig. 3(a). W , 1
po′ = ′ + σ2(o)
(σ1(o) ′ + σ3(o)
′ )
3 (1a)
MT , pi and pe denote applied axial force, torsional moment, inner con-
fining pressure and outer confining pressure, respectively. ri and ro are
the internal radius and external radius of the hollow specimen, re- ′
σ1(o)
Ro =
spectively. ui and uo are the radial displacement of the inner wall and ′
σ3(o) (1b)
outer wall of the hollow specimen, respectively. The stress component
of the hollow cylinder thin-walled element is presented in Fig. 3(b), of
′ − σ3(o)
σ2(o) ′
which σz is the mean vertical normal stress, σr is the mean radial normal bo =
′ − σ3(o)
σ1(o) ′ (1c)
stress, σθ is the mean tangential stress, and τzθ is the mean torsional
shear stress on the transverse section.
The principal stress of the soil element is shown in Fig. 3(c), where 1 2τzθ (o)
αo = arctan
σ1, σ2 and σ3 are the major, intermediate, and minor principal stress, 2 σz (o) − σθ (o) (1d)
respectively; and α is the angle between the major principal stress di-
rection and the normal direction of the acting surface—namely, the The meanings of the physical quantities in Eq. (1) are the same
rotation angle of the major principal stress axis. Through the torsional interpreted as above, where the subscript “(o)” represents the initial
shear apparatus, the simulation of different initial stress conditions can stress state and the subscript “′” indicates effective stress. Based on the
above factors, the testing scheme is designed as shown in Table 2.

Fig. 3. Loading scheme and stress state on the thin-walled hollow cylinder specimen: (a) Loading scheme; (b) Stress component; and (c) Principal stress.

88
H. Gao et al. Soil Dynamics and Earthquake Engineering 102 (2017) 86–98

Table 2
Test scheme for the combine axial–torsional tests.

No. Mixing ratio and curing day Initial consolidation state Initial loading parameter

aw (kg/m3) λVE Curing day (d) bo αο (°) Ro po (kPa) pi,o (kPa) pe,o (kPa) Wo (kN) MT,o (N·m)

1 32.6 1.43 7 0.5 0 1.5 100 106.67 96 0.151 0

2 48.8 1.43 7 0.5 0 1.5 100 106.67 96 0.151 0
3 65.1 1.43 7 0.5 0 1.5 100 106.67 96 0.151 0
4 32.6 0.954 7 0.5 0 1.5 100 106.67 96 0.151 0
5 32.6 1.907 7 0.5 0 1.5 100 106.67 96 0.151 0
6 32.6 2.384 7 0.5 0 1.5 100 106.67 96 0.151 0
7 32.6 1.43 7 1 0 1.5 100 125 105 0.0942 0
8 32.6 1.43 7 0 0 1.5 100 85.71 85.71 0.2154 0
9 32.6 1.43 7 0.5 30 1.5 100 103.33 98 0.0754 3.555
10 32.6 1.43 7 0.5 60 1.5 100 96.67 102 −0.0754 3.555
11 32.6 1.43 7 0.5 0 2.5 100 114.29 91.43 0.3231 0
12 32.6 1.43 7 0.5 0 1.5 50 53.33 48 0.0754 0
13 32.6 1.43 7 0.5 0 1.5 150 160 144 0.2262 0
14 32.6 1.43 14 0.5 0 1.5 100 106.67 96 0.151 0
15 32.6 1.43 28 0.5 0 1.5 100 106.67 96 0.151 0
16 32.6 1.43 7 Isotropic 0 1.0 50 50 50 0 0
17 32.6 1.43 7 Isotropic 0 1.0 100 100 100 0 0
18 32.6 1.43 7 Isotropic 0 1.0 150 150 150 0 0
19 32.6 1.43 7 Isotropic 0 1.0 300 300 300 0 0

2.3.2. Realization of the initial stress state

Simple transformation to Eq. (1) results in Eq. (2):
3po′ R o
′ =
σ1(o)
R o + bo R o − bo + 2 (2a)

3po′ (bo R o − bo + 1)
′ =
σ2(o)
R o + bo R o − bo + 2 (2b)

3po′
′ =
σ3(o)
R o + bo R o − bo + 2 (2c)
According to Eqs. (1) and (2), the initial stress components of the
thin-walled soil element can be calculated as:
′ + σ3(o)
σ1(o) ′ ′ − σ3(o)
σ1(o) ′
σz,o = + cos 2α o
2 2 (3a)

′
σr,o = σ2(o) (3b)

′ + σ3(o)
σ1(o) ′ ′ − σ3(o)
σ1(o) ′ Fig. 4. Stress state in the soils under moving uniform traffic loading.
σθ,o = − cos 2α o
2 2 (3c)

′ − σ3(o)
σ1(o) ′ Fig. 4. The deviatoric stress and shear stress generated by the moving
τzθ,o = sin 2α o uniform loading in the soil can be obtained by:
2 (3d)
When the stress state requirement in Eq. (3) is satisfied, the initial σz − σθ q r+1 r−1
= o⎡ − ⎤
loadings applied to the hollow cylinder specimen can be determined as 2 ⎢ 1 + (r + 1)2
π⎣ 1 + (r − 1)2 ⎦
⎥ (5a)
[19]:
Wo = π [σz,o (ro 2 − ri 2) − (pe ro 2 − pi ri 2)] qo ⎡ 1 1 ⎤
(4a) τzθ = −
π⎢ 1 + (r − 1) 2 1 + (r + 1)2 ⎥
⎣ ⎦ (5b)
2π
MT ,o = τzθ,o (ro3 − ri3)
3 (4b) where qo is the magnitude of moving uniform loading (kPa) and
r = x / a . As shown in Fig. 4, a is the half width of qo , and x is the co-
(ro + ri ) σr ,o − (ro − ri ) σθ,o ordinate value of the center of qo .
pi,o =
2ri (4c) The deviatoric stress and shear stress in Eq. (5) should be corrected
because of the existence of the inner membrane and outer membrane in
(ro + ri ) σr ,o + (ro − ri ) σθ,o
pe,o = the cyclic torsional test [21,22]. EPS composite soil in this study pos-
2ro (4d)
sesses the cementation strength, and the shear strain of any specimen
Based on Eq. (4), the required initial loadings Wo , MT,o , pi,o and pe,o under each loading condition does not exceed 5%. Based on the em-
for 19 tests conducted in this study are summarized in Table 2. pirical equation given by Chiaro et al. [21], the apparent shear stress is
quite small (about 0.3 kPa) when the shear strain is 5%. Therefore, the
2.3.3. Realization of stress path induced by traffic loadings deviator stress and shear stress presented in this study are not corrected
The dynamic characteristics of soil under traffic loadings have been for the effects of membrane force.
studied by some scholars [26–28]. Chen and Pan [29] analyzed the For the hollow cylinder specimen in the isotropic consolidation
stress state of soil under moving uniform traffic loading as shown in state, combining Eq. (4) with Eq. (5), the required axial force and

89
H. Gao et al. Soil Dynamics and Earthquake Engineering 102 (2017) 86–98

The axial force and torsional moment can be obtained by expressing

r as a time variable and substituting it into Eq. (6). In this study, all 19
tests utilized the same loading modes; i.e., the axial force and torsional
moment were applied by stages, each stage corresponding to a qo . The
magnitude of qo for the first stage was 15.63 kPa, and it increased by
15.63 kPa at each stage. For each stage, 5 cycles of stress were per-
formed under stress-controlled condition and frequency of 0.5 Hz
[26,30]. The time histories of the applied axial force and torsional
moment are shown in Fig. 5(a) and (b) for test No.1. Correspondingly,
the stress path of the specimen under the simulated traffic loading is
realized as shown in Fig. 5(c) for test No.1.

3. Results and discussions

3.1. Cyclic shear stress–strain relationship and skeleton curves

In the combined axial–torsional loading tests, the cyclic shear and

axial stress–strain response can be obtained. Herein, the dynamic
nonlinear characteristics of EPS composite soil are discussed using only
the cyclic shear stress–strain relationship obtained from the experi-
ments. The cyclic shear stress-strain curves of specimens 1 and 17 are
shown in Fig. 6, which represent anisotropic and isotropic consolida-
tion states, respectively. As the number of cycle and shear stress am-
plitude increase, the shear modulus for two specimens gradually at-
tenuates and the accumulation of plastic deformation appears which
reflects that EPS composites soil possesses typical viscoelastic and
plastic characteristics. What is different, however, is specimen 1 under
anisotropic consolidation state generates larger plastic deformation and
the modulus attenuation is more obvious compared to specimen 17. The
maximum shear strain exceeds 4% for specimen 1 while less than 2%
for specimen 17. Obviously, this difference in shear strain results from
the difference in initial stress state among two specimens. Specimen 1
will generate greater shear strain when subjected to cyclic shear stress
due to the existence of initial shear stress caused by the anisotropic
consolidation.
In order to more conveniently analyze the influence of mixing ratio
and initial consolidation state, the cyclic shear stress-strain skeleton
curves constructed according to Fig. 7 are used to depict the dynamic
nonlinear development process of EPS composite soil under cyclic
loading. The hyperbolic model given by Hardin and Drnevich [31] is
widely used to depict the cyclic shear stress (τd )–shear strain (γd ) ske-
leton curves of soils, which is expressed as:
γd
τd =
at + bt γd (7)

at = 1/ Go (8)

bt = Go/ γr (9)

where Go is the initial or maximum shear modulus of soil; γr is the re-

ference shear strain (γr = τmax / Go ) which means the ratio of maximum
shear stress corresponding to the asymptote of skeleton curve to initial
(maximum) shear modulus; and γd is the shear strain.
For the cyclic combined axial–torsional test, the axial force and
torsional moment simultaneously yield the shear strain. Herein, γd is
expressed as:
Fig. 5. Cyclic loading mode used in the test of No.1: (a) Time history of axial force, W ; (b)
2
Time history of torsional moment, MT ; and (c) Heart-shaped stress path. γd = (ε1 − ε2)2 + (ε3 − ε1)2 + (ε2 − ε3)2
3 (10)

torsional moment applied to the specimen can be calculated as: where ε1, ε2 , and ε3 are the major, intermediate and minor principal
strain of the element of hollow cylinder specimen, respectively.
r+1 r−1
W = 2qo (ro 2 − ri 2) ⎡ 2
− ⎤ It is found from the data of the 19 tests that the shear stress–strain
⎢
⎣ 1 + (r + 1) 1 + (r − 1)2 ⎥
⎦ (6a) skeleton curves can be well fitted by Eq. (7) for EPS composite soil
regardless of isotropic or anisotropic consolidation state. Moreover, the
2qo 3 1 1
MT = (ro − ri3) ⎡ − ⎤ R-square of each fitted curve is greater than 0.98. Table 3 gives the
3 ⎢ 1 + (r − 1) 2 1 + (r + 1)2 ⎥
⎣ ⎦ (6b) fitting parameter and R-square value for each loading condition. It is

90
H. Gao et al. Soil Dynamics and Earthquake Engineering 102 (2017) 86–98

Fig. 6. Cyclic shear stress vs shear strain: (a) Specimen 1 in Table 2; and (b) Specimen 17 in Table 2.

Fig. 8. Influence of cement content and curing days on the skeleton curves of EPS com-
posite soil.

Fig. 7. Idealized sketches of shear stress–strain skeleton curve.

3.1.1. Influence of mixing ratio on the skeleton curve
The test scheme designed in Table 2 can be used to analyze the
Table 3
Fitting parameters for cyclic stress-strain skeleton curve using Hardin and Drnevich
influence of some factors on the cyclic stress–strain response of EPS
model. composite soil. Fig. 8 shows the influence of cement content (a w ) and
curing days (CD) on the skeleton curves of EPS composite soil with the
No. at (10−3) bt (10−2) Go (MPa) γr (10−3) R-square same λVE of 1.43. Obviously, the larger the cement content, the greater
1 8.46 3.99 118.2 2.96 0.993
the required stress amplitude which leads to the same strain response.
2 5.83 3.37 171.4 5.09 0.996 This conclusion agrees with the dynamic response of most cemented
3 4.97 1.38 201.2 14.60 0.997 soils [32,33]. It shows that the cementation strength formed by the
4 8.25 3.19 121.2 3.80 0.992 cement hydration products is an important component of dynamic
5 8.49 4.41 117.8 2.67 0.987
strength for EPS composite soil. That is why skeleton curves of 3 spe-
6 8.55 5.09 116.9 2.30 0.984
7 7.56 4.68 132.2 2.82 0.998 cimens are different with curing days as shown in Fig. 8.
8 5.64 3.49 177.4 5.09 0.998 As an important component of EPS composite soil, EPS beads en-
9 9.52 4.08 105.0 2.58 0.987 capsulated by the cemented soil also play an important role in the
10 11.74 4.41 85.2 1.93 0.990 skeleton curve as shown in Fig. 9. It is worth noting that after the strain
11 8.97 4.20 111.5 2.65 0.999
12 9.95 3.84 100.5 2.62 0.993
reaches approximately 0.25%, an obvious discrepancy appears re-
13 7.49 3.72 133.6 3.60 0.999 garding the skeleton curve of the specimens with four different λVE
14 7.89 3.61 126.7 3.51 0.980 values. The larger the EPS bead content, the smaller the required stress
15 6.67 2.99 149.9 5.01 0.987 to generate the same strain. That is to say, as the strain or stress in-
16 8.35 3.22 119.7 3.72 0.997
creases, EPS beads will significantly affect the nonlinear development
17 8.39 3.06 124.4 4.07 0.996
18 7.56 3.36 132.2 3.94 0.993 process of EPS composite soil. Moreover, the influence of EPS beads on
19 6.51 2.14 153.5 7.17 0.998 the strength of EPS composite soil is negative. The negative influence is
manifested in two ways: 1) The nonlinearity of stress-strain relationship
for EPS beads. The cyclic triaxial testing on EPS block conducted by
shown that the hyperbolic model given by Hardin and Drnevich [31] is Trandafir et al. [34] showed that when the strain exceeded 0.54% the
suitable for describing the dynamic nonlinear characteristics of EPS plastic deformation accumulation and stiffness degradation phenomena
composite soil even under the complex stress path. appeared. Although EPS beads differ from EPS block in the shape, they
have similar material properties. Apparently, the plastic yielding

91
H. Gao et al. Soil Dynamics and Earthquake Engineering 102 (2017) 86–98

Fig. 9. Influence of EPS bead content on the skeleton curves of EPS composite soil.

property of EPS material has a negative influence on the dynamic

strength of EPS composite soil. 2) The material properties of EPS beads
are greatly different from surrounding cemented soil. The interface
between EPS beads and cemented soil is apt to generate tensile de-
formation and shear sliding. Thus, the structural interfaces in EPS
composite soil are weak mechanical planes. More EPS beads leading to
more weak interfaces result in a lower strength.
It is also mentioned that the specimens in Fig. 9 with different EPS
bead contents have the same cement content of 32.6 kg/m3. There is no
obvious discrepancy in the skeleton curve when the strain is smaller
than 0.25%, and the stress–strain relationship for four specimens with
different λVE values can be approximately fitted by a straight line. That
is, the linear elastic rigidity of EPS composite soil within a small strain
range mainly depends on the skeleton constructed by the cemented soil.
As a type of filling material, EPS composite soil with the advantage
of light weight can remarkably reduce the settlement of compressible
soft soil strata. In practice, the phenomena reflected in Fig. 9 should be
considered. For example, for an embankment filled with EPS composite
soil, the zoning filling method can be adopted because the stress and
strain induced by the traffic loading gradually decrease from the top to
bottom of the embankment. For the lower part of the embankment, EPS
composite soil with higher EPS bead content may be used; and for the
upper part, fewer EPS beads may be used. The zoning filling method
can not only ensure the global stability of the embankment under traffic
loadings but also further reduce the weight of the filling material and
result in a smaller subgrade settlement.

3.1.2. Influence of initial consolidation state on the skeleton curves

For the embankment, the stress anisotropy and the principal stress
rotation phenomena really exist. When EPS composite soil is used as the
filling material of the embankment, it is essential to consider the effects
of the initial consolidation state on the dynamic characteristics of EPS
composite soil. As shown in Fig. 10, the skeleton curves of EPS com-
posite soil are illustrated under various consolidation stress ratios, R o , Fig. 10. Influence of initial consolidation state on the skeleton curve of EPS composite
intermediate principal stress coefficients, bo , and rotation angles of the soil: (a) Consolidation stress ratio, R o ; (b) Intermediate principal stress coefficient, bo ; and
principal stress axis, α o . (c) Rotation angle of major principal stress axis, α o .
It is generally recognized that when the consolidation stress ratio is
not equal to 1, the skeleton of natural soil formed by the soil particles effective confining pressure constant. Conversely, for the tests shown in
may be in a more stable state owing to the existence of the initial static Fig. 10(a), the initial mean effective principal stress, po′, are all equal to
shear stress. Thus, the dynamic strength increases with increasing 100 kPa, but the initial minor principal stresses, σ3(o) ′ , are 100 kPa,
consolidation stress ratio when the soil is in the anisotropic con- 80 kPa and 57.1 kPa, corresponding to the consolidation stress ratios of
solidation state [35,36]. However, the results in Fig. 10(a) show that 1.0, 1.5, and 2.5. Accordingly, the results from Fig. 10(a) seem to be
the anisotropic consolidation state seems to weaken the strength of EPS induced by the difference of σ3(o)′ . In fact, assuming the strain of 1% as
composite soil. This is because the conventional dynamic triaxial tests ′ , are 0.257, 0.263
the failure strain, the stress ratios at failure, τd/ σ3(o)
mostly focus on the influence of the consolidation stress ratio by and 0.335 when the consolidation stress ratios are 1.0, 1.5, and 2.5,
changing the axial stress (major principal stress) and keeping the

92
H. Gao et al. Soil Dynamics and Earthquake Engineering 102 (2017) 86–98

respectively. From this viewpoint, the effect of consolidation stress ratio

on the dynamic strength of EPS composite soil is consistent with that of
natural soil [30]. Thus, it is not necessary to analyze the dynamic
strength of the soil under triaxial stress state considering only the
consolidation stress ratio excluding the influence of σ3(o)′ .
It is shown from Fig. 10(b) that the intermediate principal stress
coefficient, bo , also has an obvious influence on the skeleton curves for
EPS composite soil. The larger the intermediate principal stress coeffi-
cient, the smaller the stress required to induce the same strain. Ac-
′ , is equal to the
cording to Eq. (3), the intermediate principal stress, σ2(o)
radial normal stress, σr . During the initial consolidation process, the
radial stress can be simulated by adjusting the inner and outer confining
′ > σ3(o)
pressures. Consequently, if bo > 0 , i. e., σ2(o) ′ , the inner confining
pressure should be larger than the outer pressure. At this moment, a
tensile force from inside to outside the hollow specimen wall is gen-
erated in the initial consolidation process. Obviously, this initial stress
state can weaken the capacity to bear the following torsional shear ′ on Go under anisotropic and isotropic consolidation conditions.
Fig. 12. Influence of σ3(o)
force coupling with axial loading.
Similar to natural soil [37], the influence of the rotation angle of the
confining pressure has a significant influence on the initial shear
major principal stress axis on the strength of the EPS composite soil is
modulus. As depicted in Fig. 12, for EPS composite soil under the iso-
also negative, which is shown in Fig. 10(c). This negative influence is
′ = po′), initial shear modulus has a good
tropic consolidation state (σ3(o)
more prominent as the rotation angle of the principal stress axis in-
′ . For the anisotropic consolidation state,
linear relationship with σ3(o)
creases. The influence of the rotation angle of the major principal stress
the initial shear modulus also increases with increasing value of σ3(o) ′ ,
axis on the strength of the EPS composite soil is mainly due to the
but a power function can better depict the relationship between them. It
existence of initial torsional shear stress in the specimen.
should be noted that σ3(o)′ ≠ po′ for the anisotropic consolidation state.
There are 4 cases under the anisotropic consolidation state as shown in
3.2. Dynamic shear modulus properties Fig. 12: test No. 1 and No. 11 (see Table 2) are under the same mean
effective principal stress ( po′) of 100 kPa but have consolidation stress
Owing to the lack of a resonance column test in the strain range of ratios of 1.5 and 2.5, respectively; test No. 12 and No. 13 (see Table 2)
10−6 for the EPS composite soil, the initial shear modulus, Go , is ob- are under the same consolidation stress ratio of 1.5 but have po′ of
tained by the fitted parameter, at , as expressed in Eq. (8). The initial 50 kPa and 150 kPa, respectively. The 4 cases have the same inter-
shear modulus of specimens under each loading condition is summar- mediate principal stress coefficient of 0.5 without the rotation of the
ized in Table 3. The initial shear modulus will be subsequently used to major principal stress axis. Consequently, both po′ and consolidation
analyze the modulus attenuation properties of EPS composite soil. stress ratio, R o , have no clear relationship with Go for EPS composite
soil under the anisotropic consolidation state. Instead, the value of Go
3.2.1. Initial shear modulus, Go for EPS composite soil is closely dependent on σ3(o) ′ .
Fig. 11 reflects the influence of mixing ratio and curing days of EPS In addition, the fitting line under the isotropic consolidation con-
composite soil on Go . It is found that increasing cement content (a w ) dition intersects with that under the anisotropic consolidation condition
results in a larger value of Go . For example, when a w is 65.1 kg/m3 and as shown in Fig. 12. It is concluded that when σ3(o)′ is smaller, Go under
32.6 kg/m3, Go is 201.2 MPa and 118.2 MPa, respectively. The differ- the isotropic consolidation condition is relative larger than that under
ence between them is approximately 41%. Go for EPS composite soil the anisotropic consolidation condition; when σ3(o) ′ is larger, the oppo-
cured for 28 days is 21% larger than that cured for 7 days. However, site occurs.
EPS bead content, λVE , has a slight influence on Go . The largest differ- Fig. 13 compares the influences of the intermediate principal stress
ence on Go for EPS composite soil with four different values of λVE is coefficient, bo , and rotation angle of the major principal stress axis, α o ,
only 0.6%. It can be concluded that the cementation strength plays a ′ , Go decreases as the rotation angle, α o , in-
on Go . Under the same σ3(o)
decisive role on the initial stiffness of EPS composite soil. Conversely, creases. The maximum value of Go is achieved when bo = 0 and σ3(o) ′ =
the influence of EPS beads on the initial stiffness is slight. 85.7 kPa, followed when bo = 1.0 and σ3(o)′ = 75 kPa, and the minimum
It is widely accepted that the effective consolidation stress or

Fig. 11. Influence of cement content and curing days on Go . Fig. 13. Influence of bo and α o on Go .

93
H. Gao et al. Soil Dynamics and Earthquake Engineering 102 (2017) 86–98

value is attained when bo = 0.5 and σ3(o) ′ = 80 kPa. The difference obtained when a certain factor changes, which is shown in Fig. 15.
between maximum and minimum Go is approximately 60 MPa, even The Df ～log γd curves regarding the mixing ratio (a w and λVE ) and
though the difference of the corresponding σ3(o)′ is only approximately ′ , α o and bo ) are shown in Fig. 15. The
initial consolidation state (σ3(o)
5 kPa. Go decreases by approximately 33% as bo increases from 0 to 0.5. following conclusions can be obtained.
The relationship between Go and σ3(o)′ does not conform to a power
function as shown in Fig. 12 even though the specimens in Fig. 13 are (1) The influence of each factor on the G / Go ～log γd relationship con-
all under the anisotropic consolidation state. Thus, the influence of bo centrates in the strain range of 0.01–1%. When γd is approximately
′ . In fact, a large number of existing static
on Go is independent of σ3(o) 0.1%～0.4%, Df approaches the maximum value.
and dynamic tests on natural soil and artificial filling material have (2) Under the anisotropic consolidation state, the influence of σ3(o)′ on
proved that the intermediate principal stress coefficient has an obvious the G / Go ～log γd curve is the most remarkable, followed by the in-
effect on their strength and deformation [38,39]. However, it is ne- fluence of α o , and that of bo is the weakest. The influence of σ3(o)
′ on
cessary to conduct more tests to further understand the influence of bo the G / Go ～log γd curve under the isotropic consolidation state is
on Go for EPS composite soil. relatively smaller than that under the anisotropic consolidation
state.
3.2.2. Attenuation characteristics of shear modulus (3) As mentioned above, due to the nonlinearity of EPS beads material
One typical characteristic of the nonlinearity of soils lies in the fact and the existence of weak interface between EPS beads and ce-
that shear modulus gradually attenuates with increasing strain during mented soil, EPS bead content significantly affects the dynamic
the loading process. At present, massive attenuation formulas regarding nonlinearity development of specimens under cyclic loadings.
the dynamic modulus have been proposed for various types of soils, Therefore, compared with cement content, a w , EPS bead content,
such as the hyperbolic model given by Hardin and Drnevich [31], the λVE , has a more obvious influence on the modulus attenuation
Darendeli model [40] and the model suitable for gravels proposed by properties of EPS composite soil as shown in Fig. 15.
Rollins et al. [41]. Based on the full comparison of the existing models,
the Darendeli model is considered to be suitable to describe the at- In general, the attenuation characteristics of dynamic shear mod-
tenuation characteristic of the shear modulus for EPS composite soil. ulus for EPS composite soil are similar to those for natural soil or other
The Darendeli model is expressed as: artificial filling materials. In particular, to more accurately simulate the
nonlinear dynamic characteristics of EPS composite soil located in
G 1
= different zones when used as the embankment material, the deviatoric
Go 1 + (aγd )b (11) stress state and the rotation effect of the major principal stress axis
where a = 1/ γr , and b is the fitting parameter. The relationship between should be considered. Instead, the influence of intermediate principal
the shear modulus and shear strain for EPS composite soil in all cases stress coefficient is not apparent. In addition, when EPS composite soil
will be fitted by Eq. (11) in the following section, and R-square is all used as a filling material is evaluated in a nonlinear working state, it is
greater than 0.98. The value of a ranges from 2.1 to 4.6, and b ranges necessary to consider the weakening effect on its strength induced by
from 0.9 to 1.4. EPS beads. When EPS composite soil enters the nonlinear deformation
The testing results and fitting curves for the shear modulus at- stage, adding excessive EPS beads may induce a significant reduction in
tenuation of EPS composite soil considering the change of mixing ratio stiffness.
and consolidation condition are shown in Fig. 14. It is found that ce-
′ under both isotropic
ment content, EPS bead content, bo , α o , and σ3(o) 3.3. Damping ratio properties
and anisotropic consolidation conditions have different effects on the
G / Go ～log γd curves. The following phenomena can be observed: 3.3.1. Damping mechanism and damping ratio characteristics
Herein, the damping ratio related to the cyclic shear strain is de-
(1) As shown in Fig. 14(a), under the same initial consolidation state, fined as the ratio of energy dissipation to the elastic strain energy of
the larger λVE value gives a smaller G / Go value when γd ranges from soils within one cycle of loading. The damping ratio is calculated as:
0.01% to 3%. The testing data of G / Go ～log γd for EPS composite soil 1 A (l)
with various cement contents are concentrated in a small range, and D= ⋅
4π A (t ) (13)
the dispersion is relatively small as shown in Fig. 14(b).
(2) The influence of σ3(o)′ on the G / Go ～log γd curve is consistent for where A(l) is the area of the hysteresis loop in one cycle of loading, and
isotropic and anisotropic consolidation conditions; that is, at the A(t) is the area of the shaded triangle as shown in Fig. 7.
same shear strain, G / Go decreases with decreasing σ3(o)′ , which is EPS composite soil is made from soil, EPS beads, cement and water.
shown in Fig. 14(c) and (d). Taking the cemented soil part from EPS composite soil as one type of
(3) bo has no obvious effect on the G / Go ～log γd curve. However, the material, the damping mechanism of EPS composite soil under dynamic
fitting curves for the G / Go ～log γd relationship intersect with each loadings includes three factors: (1) material damping of cemented soil;
other under different bo values, which is shown in Fig. 14(e). The γd (2) material damping of EPS beads; and (3) damping of the interface
value corresponding to the intersection point is approximately structure formed between cemented soil and EPS beads, which is called
0.5%. “structural damping” here. Therefore, EPS composite soil differs from
(4) When γd < 0.7%, a larger α o value generated a smaller G / Go value as natural soil in the damping mechanism. On one hand, affected by the
shown in Fig. 14(f). When γd exceeds 0.7%, the influence of α o on cementation strength, the failure mode of EPS composite soil may be
the G / Go ～log γd curve is not obvious. flexible deformation or brittle crack damage [7]. The difference in
failure mode indicates the diversity of the damping growth mechanism.
To quantitatively investigate the influence of various factors on On the other hand, the interface structure between EPS beads and ce-
G / Go ～log γd relationships for EPS composite soil, the index, Df , is de- mented soil is weak. The energy consumed by the shear or tensile be-
fined based on the fitting curves in Fig. 15: havior in the interface between two materials does not exist in most
natural soils.
Df = (G / Go)max − (G / Go)min (12)
In Fig. 16, the damping ratio growth curve of EPS composite soil
where Df is the difference of the modulus ratio indicating the influence under the isotropic consolidation state is compared with that of several
extent of a certain factor on the G / Go ～log γd relationship; (G / Go)max and typical soils. The damping ratio growth curve of EPS composite soil is
(G / Go)min are the maximum and minimum modulus ratios, respectively, borrowed from that proposed by Rollins et al. [41] for gravels, which is

94
H. Gao et al. Soil Dynamics and Earthquake Engineering 102 (2017) 86–98

′ under isotropic consolidation state; (d) σ3(o)

Fig. 14. Shear modulus attenuation curves with the influence factors for EPS composite soil: (a) λVE ; (b) aw ; (c) σ3(o) ′ under anisotropic
consolidation state; (e) bo ; and (f) α o .

Fig. 15. The Df ～log γd curves regarding different influence factors.

Fig. 16. Damping ratio growth curves of several typical soils [41–43].

95
H. Gao et al. Soil Dynamics and Earthquake Engineering 102 (2017) 86–98

Fig. 17. Cracks in the specimen after testing (specimen 10).

expressed as:

(
D = Dmin + Dmax 1 + kγdn1 )n 2
(14)

where Dmin is the minimum damping ratio; Dmax is the maximum

damping ratio; and k , n1 and n2 are fitting parameters. For all cases in
this study, Dmin = 1% is appropriate. Dmax ranges from 18% to 27%,
which is related to the mixing ratio and initial consolidation state.
The damping ratio growth curves for the 4 types of soils in Fig. 16
indicate that material damping plays an important role in the formation
of damping. In the whole process of strain growth, the damping ratio of
EPS composite soil is smaller than that of both sand and sand–rubber
mixture. The cementation enhances the strength and reduces the
damping effect of EPS composite soil. When γd < 0.5%, the damping
ratio of EPS composite soil is smaller than that of gravels; when γd > Fig. 18. Influence of mixing ratio on the growth curve of damping ratio for EPS composite
0.5%, the case is the opposite. It is noted that during the strain growth soil: (a) aw ; and (b) λVE .
process, the compaction effect of the particles limits the growth of the
damping ratio for gravels [44]. In contrast, for EPS composite soil, as strain range of 1%, the damping ratio of EPS geofoam with different
the strain increases, in addition to the energy consumption of global densities exhibits a poor correlation with the strain amplitude [34].
deformation, some behavior will consume some energy such as internal Accordingly, it is easy to understand why the damping ratio of EPS
original micro-crack spreading and the shear sliding even cracking of composite soil changes irregularly with different λVE when the shear
the interface between EPS beads and cemented soil (see Fig. 17). In strain is smaller than 0.5%. When the shear strain exceeds 0.5%, the
summary, it is understandable that when the strain reaches a certain regular difference in damping ratio maybe resulted from two factors: 1)
extent, the damping ratio of EPS composite soil will be larger than that the influence of EPS bead content is more prominent after material
of gravels; this can be considered as the effect of the “structural damping changes when EPS beads experience plastic yielding; and 2)
damping” of EPS composite soil. larger EPS bead content generates more weak interfaces between EPS
beads and cemented soil. As the shear strain increases, the weakening
3.3.2. Influence of mixing ratio on damping ratio behavior of the interface between EPS beads and cemented soil will
Fig. 18(a) gives the damping ratio growth curves of EPS composite consume more energy when the EPS bead content increases. Appar-
soil with different cement contents. The R-square of the mean line ca- ently, it will lead to a larger damping ratio. Hence, the influence of EPS
librated by Eq. (14) using all test data of 3 cement contents reaches beads on the damping ratio growth curve is reflected not only in the
0.985, which reflects that the cement content has no obvious influence “material damping” but also in the “structural damping”.
on the damping ratio of the tested EPS composite soil. Until γd reaches
5%, the difference between the maximum and minimum damping ratios 3.3.3. Influence of initial consolidation state on damping ratio
under different cement contents is only 2.8%. The influence of cement Fig. 19(a) shows the test results and fitting curve of D ～log γd for
content on the damping ratio growth curve is mainly reflected in the EPS composite soil with various σ3(o)′ under the anisotropic consolida-
material damping. tion state. The R-square value of the calibrated mean line is 0.991,
In contrast to the cement content, EPS bead content has a more which indicates that the influence of σ3(o)′ on the damping ratio of EPS
obvious influence on the damping ratio as shown in Fig. 18(b). When composite soil under the anisotropic consolidation state is not obvious.
the shear strain is less than 0.5%, the damping ratio of EPS composite This conclusion is similar with that under the isotropic consolidation
soil differs with EPS bead content, but no obvious regularity is found. state as shown in Fig. 16, in which R-square value of the calibrated
Conversely, when shear strain exceeds 0.5%, the larger EPS bead con- mean line is 0.989. Moreover, a slight difference exists in the mean lines
tent, the larger the damping ratio. Moreover, as the shear strain in- between the anisotropic and isotropic consolidation states. When
creases, this regularity is more noticeable. When γd= 5%, the difference γd < 0.2%, the damping ratio under the isotropic consolidation state is
between the maximum and minimum damping ratios reaches 6.1%. It is slightly larger than that under the anisotropic consolidation state. When
reported in the related studies that EPS geofoam is still in the elastic γd > 0.2%, the opposite occurs.
deformation stage until its strain reaches 0.5–1% [34,45]. Within the In contrast, as shown in Fig. 19(b) and (c), bo and α o have a more

96
H. Gao et al. Soil Dynamics and Earthquake Engineering 102 (2017) 86–98

effect on the original micro-crack and the interface in EPS composite

soil. Once the initial consolidation is completed, the microstructural
difference exists inside EPS composite soil. During the following
loading, the energy dissipation is essentially different as the wave
propagates through the soil. This is why a larger difference appears in
the damping ratio even if the strain is very small.

4. Conclusions

In this paper, a series of combined axial–torsional tests are con-

ducted on EPS composite soil to investigate its dynamic characteristics
under complex stress states. The simulation method regarding the
complex stress path induced by the traffic loadings is introduced in
detail through the coupling cyclic loadings of axial force and torsional
moment. The combined axial–torsional tests comprising 19 cases are
carried out on EPS composite soil considering the influence of mixing
ratio and initial consolidation state. The following conclusions are
reached.

(1) The dynamic nonlinearity of EPS composite soil is obvious, and its
shear stress–strain relationship is consistent with the hyperbolic
function. The initial shear modulus and the strength will be im-
proved as the cement content and curing days increase. The influ-
ence of EPS bead content on the initial shear modulus and the shear
stress-strain relationship in the linear elastic stage can be neglected.
However, EPS bead content plays an important role in the skeleton
curves of EPS composite soil in the nonlinear deformation stage.
(2) Whether under the isotropic or anisotropic consolidation state, the
initial shear modulus will increase with increasing initial minor
principal stress. The relationship between them is approximately
linear for the isotropic consolidation state; however, the relation-
ship is more suitable to be described by a power function for the
anisotropic consolidation state. The larger the rotation angle of the
major principal stress axis, the smaller the initial shear modulus.
The influence of the initial intermediate principal stress coefficient
on the initial shear modulus is slight.
(3) It is proved that the attenuation characteristic of dynamic shear
modulus for EPS composite soil can be described by the Darendeli
model. When estimating the modulus attenuation of EPS composite
soil under the anisotropic consolidation state, several factors should
be emphasized including the initial minor principal stress, EPS bead
content and rotation angle of the major principal stress axis. In this
study, for EPS composite soil under the isotropic consolidation
state, the initial minor principal stress, cement content and initial
intermediate principal stress coefficient have no obvious influence
on the attenuation characteristic of shear modulus.
(4) An empirical model is proposed to depict the damping ratio growth
characteristics of EPS composite soil. The damping mechanism can
be explained by the coupling effect from the material damping and
the “structural damping” of the weak interface between the EPS
beads and cemented soil. The “structural damping” plays its role
Fig. 19. Influence of initial consolidation state on the growth curves of damping ratio for
′ under anisotropic consolidation state; (b) bo ; and (c) α o .
during the whole process of cyclic loading, even if EPS composite
EPS composite soil: (a) σ3(o)
soil is in a very low strain stage. This is an important behavior of
EPS composite soil as a filling material differing from natural soil.
obvious influence on the damping ratio growth curves of EPS composite
soil. As shown in Fig. 19(b), the R-square values of the calibrated mean Acknowledgements
lines are 0.973. The influence of bo on the damping ratio begins to
appear when γd reaches 0.02%. As the strain increases, the growth trend This work is supported by the National Natural Science Foundation
of the damping ratio for EPS composite soil with different values of bo is of China (grant number 51578286) and China Postdoctoral Science
consistent. Foundation (grant number 2013T60529). The contributions of anon-
In contrast to bo , the influence of α o on the damping ratio begins to ymous reviewers and editors are also acknowledged.
appear when γd only is 0.001% as shown in Fig. 19(c). When γd = 0.04%,
the maximum damping ratio obtained at α o = 60° is larger than the References
minimum damping ratio obtained at α o = 0° by 3.8%. This possibly
results from the fact that the torsional moment, which is applied to [1] Satoh T, Tsuchida T, Mitsukuri K, Hong ZS. Field placing test of lightweight treated
achieve the rotation of the major principal stress axis, has a negative soil under seawater in Kumamoto port. Soils Found 2001;41(5):145–54.

97
H. Gao et al. Soil Dynamics and Earthquake Engineering 102 (2017) 86–98

[2] Tsuchida T, Kang MS. Use of lightweight treated soil method in seaport and airport [23] Cai Y, Sun Q, Guo L, Juang CH, Wang J. Permanent deformation characteristics of
construction project. In Proceedings of the international workshop on lightweight saturated sand under cyclic loading. Can Geotech J 2016;52(6):795–807.
geo-materials. Tokyo, Japan; 2002. p. 215-228. [24] Qian JG, Wang YG, Yin ZY. Experimental identification of plastic shakedown be-
[3] Yoonz GL, Jeon SS, Kim BT. Mechanical characteristics of light-weighted soils using havior of saturated clay subjected to traffic loading with principal stress rotation.
dredged materials. Mar Georesour Geotechnol 2004;22(4):215–29. Eng Geol 2016;214:29–42.
[4] Ma SD. Stability and settlement of embankment filled with stabilized light soil. Rock [25] Stewart HE. Permanent strains from cyclic variable-amplitude loadings. J Geotech
Soil Mech 2003;24(3):331–4. Eng 1986;112(6):646–60.
[5] Miao LC, Wang F, Han J, Lv W, Li J. Properties and applications of cement-treated [26] Xiao JH, Juang CH, Wei K, Xu SQ. Effects of principal stress rotation on the cu-
sand-expanded polystyrene bead lightweight fill. J Mater Civ Eng mulative deformation of normally consolidated soft clay under subway traffic
2013;25(1):86–93. loading. J Geotech Geoenviron Eng 2014;140(4):80–90.
[6] Tsuchida T, Tang YX, Watabe Y. Mechanical properties of lightweight treated soil [27] Grabe PJ, Clayton CRI. Effects of principal stress rotation on permanent deforma-
cured in water pressure. Soils Found 2007;47(4):731–48. tion in rail track foundations. J Geotech Geoenviron Eng 2009;135(4):555–65.
[7] Liu HL, Deng A, Chu J. Effect of different mixing ratios of polystyrene pre-puff beads [28] Wichtmann T, Niemunis A, Triantafyllidis T. Steady Strain accumulation in sand
and cement on the mechanical behaviour of lightweight fill. Geotext Geomembr due to cyclic loading: drained triaxial tests. Soil Dyn Earthq Eng
2006;24(6):331–8. 2005;25(12):967–79.
[8] Gao HM, Liu JY, Liu HL. Geotechnical properties of EPS composite soil. Int J [29] Chen GX, Pan H. The characteristics and laboratory test simulation of stress path
Geotech Eng 2011;5(1):69–77. induced by traffic loading. China Civ Eng J 2010;43. [Suppl: 340-Suppl: 345].
[9] Illuri HK. Development of soil-EPS mixes for geotechnical applications [30] Estelle DR, Irini DM, Richard C, Daniel G. Shear modulus and damping ratio of
[Dissertation]. Australia: Queensland University of Technology; 2007. grouted sand. Soil Dyn Earthq Eng 2004;24(6):461–71.
[10] Deng A, Xiao Y. Shear behavior of sand-expanded polystyrene beads lightweight [31] Hardin BO, Drnevich VP. Shear modulus and damping in soils: measurement and
fills. J Cent South Univ Technol 2008;15. [Suppl: 174-Suppl: 179]. parameter effect. J Soil Mech Found Div 1972;98(SM6):603–24.
[11] Wang F, Miao LC. A proposed lightweight fill for embankments using cement- [32] Maher MH, Ro KS, Welsh JP. High strain dynamic modulus and damping of che-
treated Yangzi River sand and expanded polystyrene (EPS) beads. Bull Eng Geol mically grouted sand. Soil Dyn Earthq Eng 1994;13(2):131–8.
Environ 2009;68(4):517–24. [33] Gao YF, Wang SM, Chen CB. A united deformation-strength framework for
[12] Tsuchida T, Porbaha A, Yamane N. Development of a geomaterial from dredged bay Lightweight Sand–EPS Beads Soil (LSES) under cyclic loading. Soil Dyn Earthq Eng
mud. J Mater Civ Eng 2001;13(2):152–60. 2011;31(8):1144–53.
[13] Gao HM, Chen YM, Liu HL, Liu JY, Chu J. Creep behavior of EPS composite soil. Sci [34] Trandafir AC, Erickson BA. Stiffness degradation and yielding of EPS geofoam under
China Technol Sci 2012;55(11):3070–80. cyclic loading. J Mater Civ Eng 2012;4(1):119–24.
[14] Kim YT, Ahn J, Han WJ, Gabr MA. Experimental evaluation of strength char- [35] Hyodo M, Aramaki N, Itoh M, Hyde AFL. Cyclic strength and deformation of
acteristics of stabilized dredged soil. J Mater Civ Eng 2010;22(5):539–44. crushable carbonate sand. Soil Dyn Earthq Eng 1996;15(5):331–6.
[15] Minegishi K, Makiuchi K, Takahashi R. Strength-deformation characteristics of EPS [36] Flora A, Lirer S, Silvestri F. Undrained cyclic resistance of undisturbed gravelly
beads-mixed lightweight geomaterial subjected to cyclic loadings. Proceedings of soils. Soil Dyn Earthq Eng 2012;43:366–79.
the international workshop on lightweight geo-materials. Tokyo, Japan; 2002. p. [37] Yoshimine M, Ishihara K, Vargas W. Effects of principal stress direction and inter-
119-125. mediate principal stress on undrained shear behavior of sand. Soils Found
[16] Lin H, Penumadu D. Experimental investigation on principal stress rotation in 1998;38(3):179–88.
Kaolin clay. J Geotech Geoenviron Eng 2005;131(5):633–42. [38] Lade PV, Rodriguez NM, Dyck EJV. Effects of principal stress directions on 3D
[17] Rodriguez NM, Lade PV. Effects of principal stress directions and mean normal failure conditions in cross-anisotropic sand. J Geotech Geoenviron Eng
stress on failure criterion for cross-anisotropic sand. J Eng Mech 2014;40(2):362–3.
2014;139(11):1592–601. [39] Inam A, Ishikawa T, Miura S. Effect of principal stress axis rotation on cyclic plastic
[18] Shen Y, Yan J, Liu HL. Critical behavior of high-density silt under cyclic principal deformation characteristics of unsaturated base course material. Soils Found
stress rotation after different initial consolidation ratio. International symposium on 2012;52(3):465–80.
geomechanics and geotechnics: from micro to macro. Boca Raton, USA; 2010. p. [40] Darendeli MB. Development of a new family of normalized modulus reduction and
117-124. material damping curves [Dissertation]. Austin: University of Texas; 2001.
[19] Hight DW, Gens A, Symes MJ. The development of a new hollow cylinder apparatus [41] Rollins KM, Evans MD, Diehl NB, Daily WD. Shear modulus and damping re-
for investigating the effects of principal stress rotation in soils. Geotechnique lationships for gravels. J Geotech Geoeviron Eng 1998;124(5):396–405.
1983;33(4):355–83. [42] Senetakis K, Anastasiadis A, Pitilakis K. Dynamic properties of dry sand/rubber
[20] Vaid YP, Sayao A, Hou E, Negussey D. Generalized stress-path-dependent soil be- (SRM) and gravel/rubber (GRM) mixtures in a wide range of shearing strain am-
haviour with a new hollow cylinder torsional apparatus. Can Geotech J plitudes. Soil Dyn Earthq Eng 2012;33(1):38–53.
1990;27(5):601–16. [43] Seed HB, Wong RT, Idriss IM, Tokimatsu K. Moduli and damping factors for dy-
[21] Chiaro G, Koseki J, Sato T. Effects of initial static shear on liquefaction and large namic analyses of cohesionless soils. J Geotech Eng Div 1986;112(11):1016–32.
deformation properties of loose saturated Toyoura sand in undrained cyclic tor- [44] Lin SY, Lin PS, Luo HS, Juang CH. Shear modulus and damping ratio characteristics
sional shear tests. Soils Found 2012;52(3):498–510. of gravelly deposits. Can Geotech J 2000;37(3):638–51.
[22] Chiaro G, Kiyota T, Koseki J. Strain localization characteristics of loose saturated [45] Zarnani S, Bathurst RJ. Influence of constitutive model on numerical simulation of
Toyoura sand in undrained cyclic torsional shear tests with initial static shear. Soils EPS seismic buffer shaking table tests. Geotext Geomembr 2009;27(4):308–12.
Found 2013;53(1):23–34.

98