Soil Dynamics and Earthquake Engineering 94 (2017) 75–82

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering

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

An energy method for deformation behavior of soft clay under cyclic loads MARK
based on dynamic response analysis
⁎
Qing-Lu Deng, Xing-Wei Ren
Faculty of Engineering, China University of Geosciences, Wuhan 430074, China

A R T I C L E I N F O A BS T RAC T

Keywords: Soft clay is fragile to be disturbed and to produce deformation under cyclic loads. A simple theoretical energy
Energy method method is presented to explore the dynamic response of soft clay under cyclic loads, then to discuss its
Soft clay deformation behavior. In the formulation, equations are derived based on general thermodynamic principle and
Cyclic loads Newton's laws of motion. A series of cyclic triaxial tests are conducted for undisturbed samples of soft clay to
Dynamic response
verify the proposed method. Cumulative deformation is computed by the proposed approach. Results are
Plastic accumulative deformation
compared with the experimental data, and good agreement is achieved. This paper develops a new insight to
describe deformation behavior of soft soil under cyclic loads, and to predict the plastic accumulative
deformation.

1. Introduction [3,4,20–26]. Compared with the theoretical method, the empirical

Plastic deformation for soft soils can be produced under cyclic loads
caused by traffic vehicles, waves, operation of machinery, etc. [1–6].
This deformation could be accumulated over time, and will bring high
maintained costs and reduced design life. Taking Shanghai Metro Line
1 for example, tunnel settlement was relatively small (2–6 mm) during
the first two years before formal operation (formal operation at 1995),
but then increases year by year, reaching 60 mm in 1997, 150 mm in
1999, and up to 280 mm in 2007 [7]. Therefore, knowledge of
cumulative plastic deformation and prediction methods for soft soil
under cyclic loads are essential for design and maintenance cost
control.
Deformation behavior of soft soils under cyclic loads has been
investigated by many researchers. The existing methods can be mainly
classified into two categories: theoretical method and experimental/
empirical method. Generally, the theoretical method is firstly based on
certain constitutive models of soils, such as, yield surface model [8,9],
bounding surface model [10–14], and visco-elastic–plastic model
[15,16]. Then deformation of each cycle can be computed by the
constitutive models. The second method is empirical model, which is
based on laboratory experimental or in-site tests, to predict accumu-
lative plastic deformation by considering various influencing factors
such as soil type, soil properties and, stress state. Among numerous
empirical models, one of the most often used is a power model, which is
first proposed by Monismith et al. [17] and then modified by Li and
Selig [18], Chai and Miura [2], Huang et al. [19]. In addition, some
other empirical models or semi-empirical models have been developed
method has become more and more popular in recent years because of
its simplicity and satisfactory performance in practical engineering
applications. However, the theoretical method is a fundamental and
absolutely necessary tool for further exploring and understanding
deformation mechanism. Therefore, this paper tries to provide a simple
theoretical method in terms of energy to investigate the plastic
accumulative deformation of soils under cyclic loads.
The aim of this paper is to present a theoretical method based on
single freedom vibration equation, and to estimate the accumulated
plastic deformation of soft clay subjected to cyclic loads. Since move-
ment and deformation of the object is always accompanied by energy
changes, therefore, it is possible to use the energy change to describe
dynamic response and deformation behavior of soft soil under cyclic
loads. The rest of this paper is organized as follows. We employ laws of
thermodynamics in combination with vibration differential equation of
soil to carry out a rigorous formula derivation for calculating plastic
strain accumulation energy under cyclic loads. Then, a number of cyclic
triaxial tests are conducted, and comparison between the tested results
and computed results is made to verify the proposed model.
Limitations of the proposed model were discussed in the last part of
the paper.
2. Methodology
2.1. Literature review of energy method in this field
Compared with the existing theoretical (constitutive models) and

⁎
Corresponding author.

http://dx.doi.org/10.1016/j.soildyn.2016.12.012
Received 1 August 2016; Received in revised form 14 November 2016; Accepted 26 December 2016
Available online 16 January 2017
0267-7261/ © 2017 Elsevier Ltd. All rights reserved.
Q.-L. Deng, X.-W. Ren Soil Dynamics and Earthquake Engineering 94 (2017) 75–82

empirical methods (such as the power model proposed by Monismith

et al. [17] and its improved models) reviewed in the above, the energy

Cyclic load F (N)

approach has several advantages such as its independence of the load
waveform and the type of testing device [27]. The most important
among them is that energy is a scalar, which determines its quantity
can be directly superimposed without the need to consider its direction.
This advantage makes a much simpler for solving complex issues by
use of energy approach. Time t (s)
The study on the soil vibration problem using energy method began
from Nemat-Nasser and Shokooh [28]. Based on the idea of energy
dissipation, they studied the densification and liquefaction of cohesion-
less soils under vibration loads by analysis of the relationship between
excess pore water pressure and dissipated energy in the process of
vibration. Following the study by Nemat-Nasser and Shokooh [28], an
energy-based liquefaction evaluation method (EBM) was proposed by
Davis and Berrill [29] and applied in liquefaction assessment by several f f
other researchers [30–39]. Azeiteiro et al. [40] carried out study on
energy dissipated in undrained cyclic triaxial tests, and investigated the
correlation between the energy dissipated per unit volume and the
generated excess pore water pressure. In addition, some researchers
study response of soil deposits and of structures with using formulation
of the solutions in terms of energy carried by seismic waves [41–43].
Nevertheless, most of the existing literatures for energy approach focus
on sand liquefaction, very rare on dynamic response and cumulative
plastic deformation saturated soft clay.
Fig. 1. Schematic diagram of stress analysis of studied system.

t
2.2. Theoretical derivations Wf = ∫o fvdt
(4)
It could be learned from the laws of thermodynamics and con- In addition, kinetic energy Ek and elastic potential energy Ee can be
servation of energy that energy may neither be created nor destroyed calculated by the following formulas, respectively.
and the sum of all the energies in the system is a constant. Therefore,
1 2
the total mechanical energy produced by external cyclic loads is equal Ek = mv
2 (5)
to the sum of dissipated energy generated by damping of system,
kinetic energy, elastic potential energy and plastic deformation accu- 1 2
mulation energy, that is: Ee = kz
2 (6)
WF = Wf + Ek + Ee + EP (1) where, m is quality of studied object [kg]; v is the same as the above; k
is restoring force coefficient, also known as the generalized stiffness [N/
where, WF is the total mechanical energy produced by external cyclic m]; z is displacement of deviation from the initial position [m].
loads in the loading process of system; Wf is dissipated energy The following work is to derive the vibration differential equation
generated by damping; Ek is kinetic energy of system obtained at the and solve it. We first establish a physical model of vibration system.
time of calculating; Ee is elastic potential energy of system obtained at Assume the studied soil object is a forced vibration system of a single
the time of calculating; Ep is plastic accumulative deformation energy degree of freedom with damping in the vertical direction, whose stress
during the cyclic loading. state under cyclic loads is shown in Fig. 1.
Since the cyclic loads adopted in the most of cyclic triaxial tests are Thus, according to Newton's first law,
harmonic loads, we choose sinusoidal harmonic load as the external
excitation load F expressed as follows. ma = mz ̈ = F − f − Fp (7)

F = F0 + FA sin ωt (2) In which, a is vibration acceleration of system [m/s2],

where, F0 is a base value of cyclic loads [N]; FA is dynamic amplitude
of loads [N], and its value is smaller than the base value in order to
avoid tensile stress, namely FA < F0 ; ω is circular frequency [s−1], and
ω = 2π / T , in which T is cycle time [s]. Since the external excitation
force (the applied cyclic loads) F always exist in the whole loading
process, the total mechanical energy WF is the product of the force and
displacement z of the object s on which the force acting. The object
displacement z is a function of object velocity v and acting time t. Thus,
the WF can be expressed as follows.
t t
WF = Fz = F ∫0 vdt = ∫0 (F0 + FA sin ωt ) vdt
(3)
In which, v is the velocity of object on which the force acting on[m/
s]; t is the time [s].
Similarly, the dissipated energy Wf caused by damping f can be
expressed as
a = z ̈ = d 2z / dt 2 ; FP is restoring force of system, FP=kz; others are the
same as the above.
Damping of soil makes energy dissipate and lose in the process of
vibration, and it could generally be divided into viscous damping and
friction damping in engineering practice. However, their contribution
to the dissipation of energy is not the same. It has been proven by fluid
mechanics that viscosity resistance generated by viscous damping of
media is proportional to vibration velocity of studied object and related
to frequency; frictional resistance produced by friction damping is
independent of vibration frequency and merely subject to displacement
or strain. As the amount of energy dissipation is mainly related to
frequency of wave [30,35,40], this paper only considers the impact of
viscous damping to facilitate the calculation. Thus, the resistance f can
be calculated by formula (8).
f = −cv = −cz ̇ (8)
where, c is coefficient of viscous resistance, and also known as the

76
Q.-L. Deng, X.-W. Ren Soil Dynamics and Earthquake Engineering 94 (2017) 75–82

generalized damping coefficient [N s/m]

Thus, making use of (2), (7) and (8), vibration differential equation v = z ̇ = e−nt [( C1 λ2 − n2 − C2 n cos ) λ 2 − n2 t
of studied system could be expressed as:
mz ̈ = F0 + FA sin ωt − cz ̇ − kz (9) (
− C1 n + C2 λ2 − n2 sin ) λ 2 − n2 t ]
Formula (9) could be rewritten as: +
ωFA
⋅cos(ωt + φ)
m (λ2 − ω2)2 + 4n2ω2
c k F F
z ̈ + z ̇ + z = A sin ωt + 0
m m m m (10) (18)
Define n = c /(2m ), λ2 = k / m . Thus,
By use of the initial conditions z|t=0 = z 0 , z ̇|t=0 = v0 , we obtain
F F
z ̈ + 2nz ̇ + λ2z = A sin ωt + 0
m m (11) ⎡
1 ⎢v0 + nz 0 − nFA nF
C1 = sin φ − 2 0
Solution of non-homogeneous differential Eq. (11) is composed of λ2 − n2 ⎢⎣ m ( λ 2 − ω 2 ) 2 + 4 n 2ω 2 λ m
the general solution of its corresponding homogeneous equation and a ⎤
ωFA
special solution of it. That is − ⋅ cos φ⎥
m (λ 2 − ω 2 )2 + 4n2ω 2 ⎥⎦
z = z + z* (12)
where, z is the solution of non-homogeneous differential Eq. (11); z is
FA F0
the general solution of corresponding homogeneous equation of (11); C2 = z 0 − sin φ −
m ( λ 2 − ω 2 ) 2 + 4 n 2ω 2 λ2 m
z* is a special solution of (11)
Defining the special solution as:
To solve Eq. (1), kinetic energy Ek and elastic potential energy Ee
F0 must be given first.
z* = z 0 sin(ωt + φ) +
λ2 m (13)
1 2
With use of (11) and (13), the special solution could be obtained. Ek = mv
2
FA F0 1
z* = sin(ωt + φ) + = m {e−nt [(C1 E − C2 n )cos Et − (C1 n + C2 E )sin Et ] + Hω cos(ωt + φ)}2
m (λ2 − ω 2 )2 + 4n2ω 2 λ2 m (14) 2
2nω (19)
In which,φ = arctan 2 2 .
λ −ω
The following work is to solve the general solution of homogeneous
1 2 1 ⎡ −nt F ⎤
2
equation corresponding to (11). The homogeneous equation is Ee = kz = k ⎢e (C1 sin Et + C2 cos Et ) + H sin(ωt + φ) + 2 0 ⎥
2 2 ⎣ λ m⎦
z ̈ + 2nz ̇ + λ2z = 0 (15)
(20)
Known from some relevant knowledge of calculus, there are three
FA
different cases of roots of characteristic equation of (15). where, E = λ 2 − n2 ; H = .
m (λ2 − ω2)2 + 4n2ω2
Correspondingly, (11) and (15) have three different solutions under Thus, making use of (1), (3), (4), (5), (6), (19) and (20), plastic
different cases. deformation energy could be written as

(A) The first case of n2 − λ2 > 0 , namely the damping ratio D = n / λ ≥ 1 Ep = ∫ (FA sin ωt + F0 )⋅{e−nt [(C1 E − C2 n )cos Et − (C1 n + C2 E )sin Et ]
⎛ ⎞ ⎛ ⎞ + Hω cos(ωt + φ )} dt
⎜−n + n2 − λ2 ⎟ t ⎜−n − n2 − λ2 ⎟ t
z = A1 er1t + A2 er 2 t = A1 e⎝ ⎠ + A2 e⎝ ⎠ (16a) − ∫ c {e−nt [(C1 E − C2 n )cos Et − (C1 n + C2 E )sin Et ] + Hω cos(ωt + φ )}2 dt
1
− 2 m {e−nt [(C1 E − C2 n )cos Et − (C1 n + C2 E )sin Et ] + Hω cos(ωt + φ )}2
(B) The second case of n2 − λ2 = 0 , namely the damping ratio
1 ⎡ F0 ⎤2
D = n /λ = 1 − 2 k ⎢e−nt (C1 sin Et + C2 cos Et ) + H sin(ωt + φ ) + ⎥
⎣ λ2 m ⎦

z = (B1 + B2 t ) ert = (B1 + B2 t ) e−nt (16b) (21)

The above Eq. (21) is the formula of plastic deformation energy

(C) The third case of n2 − λ2 < 0 , namely the damping ratio under third case of damping ratio D < 1. Similarly, the solution under
D = n /λ < 1 other two cases can be obtained by use of the same derivation.
Under the first case of D > 1,
(
z = e−nt C1 sin λ2 − n2 t + C2 cos λ 2 − n2 t ) (16c)
Substituting Eq. (14) and (16) into Eq. (12) yields the solution of Ep = ∫ (FA sin ωt + F0 )⋅[(B2 − nB2 t − nB1) e−nt + Hω cos(ωt + φ)] dt
Eq. (11), which is the displacement-time function. After derivation − ∫ c [(B2 − nB2 t − nB1) e−nt + Hω cos(ωt + φ)]2 dt
calculus to this function we can get the velocity equation, then 1
substituting displacement and velocity into Eqs. (1)–(6), the plastic − 2 m [(B2 − nB2 t − nB1) e−nt + Hω cos(ωt + φ)]2

1 ⎡ ⎤2
deformation accumulation can be obtained. We take the third case that F0
is the most common in engineering practice for example to present the − 2 k ⎢ (B1 + B2 t ) e−nt + H sin(ωt + φ) + ⎥
⎣ mλ2 ⎦
detailed derivation.
(22)
Using Eqs. (14), (16c) and (12), the solution of Eq. (11) is given by
where,
(
z = z + z* = e−nt C1 sin λ2 − n2 t + C2 cos λ 2 − n2 t ) F0 ⎛
,B2 = v0 − Hω cos φ + n ⎜z 0 − H sin φ −
F0 ⎞
B1 = z 0 − H sin φ − ⎟
+
FA
sin(ωt + φ) +
F0 mλ 2 ⎝ mλ 2 ⎠
m (λ2 − ω2)2 + 4n2ω2 λ2 m (17) Under the second case of D = 1,
Thus, the velocity can be written as

77
Q.-L. Deng, X.-W. Ren Soil Dynamics and Earthquake Engineering 94 (2017) 75–82

Ep = ∫ (FA sin ωt + F0 )⋅[A1 r1 er1t + A2 r2 er 2 t + Hω⋅ cos(ωt + φ)] dt
− ∫ c [A1 r1 e r1 t + A2 r2 er 2 t + Hω⋅ cos(ωt + φ)]2
1
− 2 m [A1 r1 er1t + A2 r2 er 2 t + Hω⋅ cos(ωt + φ)]2
1 ⎡ F0 ⎤2
− 2 k ⎢A1 er1t + A2 er 2 t + H sin(ωt + φ) + ⎥ stresses. Thus the K0-consolidation method is adopted in the experi-
⎣ λ2 m ⎦ (23)
where r1 = −n + n2 − λ 2 , r 2 = − n − n2 − λ 2 ,
r2 (z 0 − H sin φ − F0 / mλ2 ) − v0 + Hω cos φ
A1 = , before K0-consolidation, and the axial pressure σ1 and confining
r2 − r1
r1 (z 0 − H sin φ − F0 / mλ2 ) − v0 + Hω cos φ
A2 = .
r1 − r2
Eqs. (21), (22) and (23) are formulas of plastic cumulative
deformation under cyclic loads. However, the system does non-vibra-
tion attenuation movement under the first two cases D ≥ 1, and has
nothing to do with the frequency, cycle, etc. [44]. Thus, we only
consider the case of D < 1 in this paper.
With regard to the parameters of Eq. (21) such as coefficient of
viscous resistance c and generalized stiffness k, they can be obtained by
back calculation method with laboratory and/or filed test data. Since
Eq. (21) represents the energy dependent on the deformation gener-
ated by cyclic loads, these parameters can be fitted out by Least Squares
Method as long as the deformation at a given time is obtained by tests.
An example for back calculation will be presented in detail in the
Section 4. By use of the proposed energy approach, problems of
dynamic response and deformation properties of soils subjected to
cyclic loads could be explored well in combination with some labora-
tory experimental and field test..
3. Laboratory GDS test
To verify the reasonability and applicability of the proposed
approach, a number of cyclic triaxial tests are conducted for soft clay.
3.1. Studied soil
Tested soil samples are soft clay drilled from the fourth soil stratum
of Shanghai at the depth of 13.5 m. The studied mucky clay is coastal-
shallow sedimentary of Q24 with the properties of high moisture
content, large void ratio, high compressibility, low sheer strength and
permeability and high sensitivity. A summary of its properties is
presented in Table 1.
3.2. Testing procedure
Several factors are known to affect the dynamic response of
cohesive soils. Our study considers the dynamic stress amplitude,
Table 1
Physical properties of the studied clays.
Soil properties
confining stress, initial stress and frequency of loading in the develop-
ment of plastic accumulative deformation of clay. The equipment used
dt
in this work is the Global Digital Systems (GDS), which has been
described in detail by Tang et al. [45]. Non-hydrostatic conditions
existing in natural soil result in the occurrence of initial static shear
ment to reproduce the original stress state. The value of K0 is 0.7
according to the physical and mechanical properties of soil sample and
engineering experience. Back pressure σb saturating is conducted
pressure σ3 could be determined by use of following formula.
σ1 = σ ′1 + σb
σ3 = σ2 = σ ′3 + σb
σ ′1 = γ′h
σ ′3 = K0 σ ′1 (24)
where, σb is the back pressure, which is 100kPa in our experiments; γ′ is
the average saturated weight of soil covered above the sampling
location; it is 20kPa in this test; h is the depth of sampling, which is
13.5 m. Thus, σ1 = 235 kPa, σ3 = σ2 = 195 kPa.
Besides, the soil samples are still subjected to additional stress
caused by subway tunnel system, which ranges from 20–40 kPa [46],
and 30 kPa is adopted in the test. Therefore, the initial basic axial
pressure of cyclic loading set on the sample is 265 kPa, which contains
the axial pressure of K0-consolidation (235 kPa) and the additional
stress (30 kPa). Sine waveform cyclic loading is chosen and frequency
of the load is 1.5 Hz. The scheme and parameters of test are shown in
Table 2. The specific testing procedure has been described in detail
[45,47].
3.3. Test results
Plastic accumulative strain under cyclic loading is observed, and
typical results are presented in Fig. 2. It could be seen that strain of
saturated soft clay under cyclic loading presents the characteristics of
both recoverable elasticity and unrecoverable plasticity in one cycle,
and strain gradually increases with time. Fig. 3 shows the unrecover-
able axial strain over time under different cyclic deviator stress
amplitude: 2 kPa, 10 kPa, 25 kPa respectively. As can be observed
from these curves, the deformation rate of all of three samples changes
from fast to slow, and has the trends to be zero. Its deformation is
increasing, but not up to failure. These observations indicate that the
applied cyclic deviator stress is less than the critical cyclic stress, which
is defined as the stress below which the soil will not suffer failure
regardless of the number of repeated load applications [48,49]. The
development of plastic cumulative deformation of soils under the stress
below the cyclic critical stress could be approximately divided into
three stages. In the first stage, deformation rate is relatively largest, and
strain sharply increases up to 50–65% of the total in a very short period
of time; after this level is reached, the deformation rate rapidly
decreases, but strain accumulation grows gently up to 80–90%. This
duration time is longer than that of the first stage. Then, deformation
rate continues to fall until it is close to zero, and the final about 10% of
Values

Depth (m) 13.5 Table 2

Natural moisture content (%) 49–58 Scheme and parameters of test *.
Void ratio 1.44–1.82
Initial density (g/cm3) 1.76 Parameters/Sample A B C D
Specific gravity 2.75
Degree of saturation (%) 92–100 Waveform of cyclic load sine sine sine sine
Plastic limit (%) 15–24 Frequency (Hz) 1.5 1.5 1.5 1.5
Liquid limit (%) 42–47 Back pressure (kPa) 100 100 100 100
Plasticity index 18–32 Confining pressure (kPa) 195 195 195 195
Permeability coefficient (cm/s) 3.00×10−7 Basic axial pressure (kPa) 265 265 265 265
Coefficient of pressure at rest, K0 0.7 Cyclic stress amplitude (kPa) 2 10 25 40
Sensitivity 6.2
*
Compression modulus (MPa) 2.03–3.01 Three tests are conducted under each condition, namely a total of 12 experiments
were done in this study.

78
Q.-L. Deng, X.-W. Ren Soil Dynamics and Earthquake Engineering 94 (2017) 75–82

Fig. 4. Attenuation curve of displacement produced by free vibration.

FA F0
Fig. 2. Typical cyclic plastic strain accumulation of soft clay (Frequency f=1.5 Hz; cyclic z ≈ zforced = sin(ωt + φ) +
m (λ2 − ω 2 )2 + 4n2ω 2 λ2 m (27)
stress amplitude FA=2 kPa; basic pressure F0=265 kPa).

Thus, the Eq. (21) is correspondingly simplified to

Ep = ∫ (FA sin ωt + F0 )⋅Hω cos(ωt + φ) dt − ∫ c [Hω cos(ωt + φ )]2 dt

1 1 ⎡ F0 ⎤2
− 2 m [Hω cos(ωt + φ)]2 − 2 k ⎢H sin(ωt + φ) + ⎥
⎣ λ2 m ⎦

(28)
In the light of the form of mathematical expressions of kinetic
energy Ek (Eq.4) and elastic potential energy Ee (Eq. (5)), we assume
that the plastic deformation energy is expressed as follows:
Ep = ax b (29)
In which, a, b are plastic deformation coefficients; x is plastic
cumulative deformation [m].
Thus, substituting Eq. (28) into Eq. (29), the plastic deformation
accumulation of soil dynamic response is written as

Fig. 3. Test results of plastic accumulative deformation of soft clay. ⎧ 1 (F sin ωt + F )⋅Hω cos(ωt + φ) dt ⎫1/ b
⎪a ∫ A 0 ⎪
⎪
⎪ − 1 ∫ c [Hω cos(ωt + φ)]2 dt ⎪
⎪
total strain amount is completed in this period. This stage lasts the x=⎨ a ⎬
⎪ ⎡ 2⎪
F0 ⎤
longest time. The deformation rate and duration time of each stage
⎪− m [Hω cos(ωt + φ)]2 − k ⎢H sin(ωt + φ) + ⎥ ⎪
relate to soil type, soil properties, cyclic stress, load frequency, ⎪ 2a 2a ⎣ λ2 m ⎦ ⎪
⎩ ⎭ (30)
confining pressure, etc. [50]. Different parameters, different deforma-
tion rate and duration time of each stage. The Least Squares Method was used to obtain the parameters in the
Eq. (30) by fitting the data set of plastic cumulative strain εp and time t.
With the fitted parameters, plastic cumulative strain was computed by
the Eq. (30), and made a comparison with the tested values. The fitted
4. Applications and discussions
values of parameters are shown in Table 3.
Fig. 5 demonstrates the calculation results of axial deformation
4.1. Comparison test results with calculation
under dynamic loading in the first few cycles. We can see that the
specific unrecoverable plastic strain and recoverable elastic strain in
It can be learned from Eq. (17) that displacement of soil dynamic
each load cycle are clearly described, which is consistent with the
response is composed of two parts produced free vibration and forced
corresponding test results shown in Fig. 2. Fig. 6 presents the
vibration respectively.
comparison test results with calculation results of plastic accumulative
strain under different cyclic deviator stress. It can be seen that most of
(
zfree = e−nt C1 sin λ2 − n2 t + C2 cos λ 2 − n2 t ) (25) the computed results match very well with test results in all four cases.
Thus, it indicates that the energy method proposed in this paper is
FA F0
zforced = sin(ωt + φ) +
m ( λ 2 − ω 2 ) 2 + 4 n 2ω 2 λ2 m (26) Table 3
Calculation coefficients of accumulative plastic deformationa.
However, displacement zfree sharply attenuates to zero in a very
Case/ Coefficient c (N·s/m) m (kg) k (N/m) a b
short period of time (Fig. 4). By use of the test results, the necessary
parameters of Eq. (21) were obtained through back calculation method, FA=2 kPa 1.577×106 2.769×105 1.294×108 4.008×105 2.724
which are c=3.26×106 N s/m, m=8.64×104kg, k=8.28×108 N/m, FA=10 kPa 4.017×106 3.242×105 1.098×108 1.132×106 3.801
FA=25 kPa 1.26×106 8.927×105 5.489×108 1.722×104 3.734
FA=2 kPa, F0=265 kPa. Substituting these parameters into the formula
FA=40 kPa 6.126×106 4.051×105 1.117×108 8.259×104 4.372
(25), displacement-time curve could be obtained. As can be seen from
it the axial displacement declines from 4.6 mm to 0 in about 0.3 s. a
The basic pressure and cyclic loading frequency are same in all the four cases,
Thus, the effect of free vibration on displacement and plastic strain F0=265 kPa, f=1.5 Hz respectively.
could be omitted to simplify the calculation. That is Eq. (17) could be
simplified to:

79
Q.-L. Deng, X.-W. Ren Soil Dynamics and Earthquake Engineering 94 (2017) 75–82

0.18
tical model based on single freedom vibration, which is one-dimen-
0.16 sional. Since lateral direction is completely constrained in one-dimen-
sional system, its vertical deformation should be theoretically smaller
0.14
than the deformation produced in a three-dimensional system. This
0.12 results in a “born” and inevitable gap between the test results and the
computed ones. Moreover, the greater the deformation, the greater the
Axial strain (%)

0.1
gap is. This is why the residual error in the case shown in Fig. 7d
0.08 (whose axial strain reaches up to 12% shown in Fig. 6d)is much greater
0.06
than the errors in the other three cases shown in Fig. 7a, b, and c (the
largest strain in these three cases is no more than 3%).
0.04

0.02

0
0 1 2 3 4 5
Time (s)
Fig. 5. Calculation of axial strain in the first few cycles of load (Frequency f=1.5 Hz;
Cyclic stress amplitude FA=2 kPa; Basic pressure F0=265 kPa).
valid for analyzing dynamic response of soil under cyclic loading, and
for predicting its plastic accumulative strain.
A histogram plot of the residual error is shown in Fig. 7. All of the
residual errors between the test results and the computed results are
less than 0.5%, and most of them are less than 0.1%. These figures
demonstrate the good quality of our equations for plastic accumulative
strain, which is also able to meet the engineering requirements in this
case. In addition, it is worth noting that most of the computed values
are smaller than the test results, especially when time is over 1000 s (as
shown in Fig. 6). This observation also can be seen from Fig. 7 that
most of the residual errors are positive. These errors may be caused for
two reasons. (1) In the derivation process, we ignore the effect of free
vibration on displacement and plastic strain, which probably result in
the computed results are smaller than expected. And the error caused
by this effect will accumulate over time. (2) The test results are
obtained by triaxial cycle test of GDS, which is in three-dimensional;
however, the computed results are calculated by the proposed theore-
6 7 8 9 10
4.2. Limitations
To make sure the proposed method is of practical value for
geotechnical engineers involved in analysis of deformation under traffic
loads, particular attention is paid to simplify the calculation process in
this paper. Some assumptions and simplifications were taken as
follows. (1) Assume the soil-body as a black box with properties of
viscosity, elasticity and plasticity to bypass the extremely complicated
constitutive relation of soil internal. (2) Omit the effect of free vibration
on displacement and plastic strain energy. (3) Simplify a complicated
vibration system as a dumped linear single degree of freedom oscillator
to determine the displacement and corresponding velocity of the
system, and then to determine the cumulative plastic energy of the
system using energy conservation.
Since the real case of soil vibration system is complicated and
highly nonlinear, these assumptions and simplifications will certainly
affect the accuracy and rationality of the proposed methodology.
Nevertheless, based on satisfactorily reproducing the measured re-
sponse, this energy method herein can still be regarded as a good
attempt for discussing the dynamic response and deformation mechan-
ism of soft soils subjected to cyclic loads.

Fa=2kPa,Fo=265kPa,w=3pi Fa=10kPa,Fo=265kPa,w=3pi
a 1 b 1.2

0.8 1
A x ia l s tra in (% )

A x ia l s tra in (% )

0.8
0.6 Calculation Calculation
Test 0.6 Test
0.4
0.4

0.2 0.2

0 0
0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000
Time (s) Time (s)

c Fa=25kPa,Fo=265kPa,w=3pi d Fa=40kPa,Fo=265kPa,w=3pi
2.5 12

2 10
Calculation
A x ia l s tra in (% )

A x ia l s tra in (% )

Calculation
8 Test
1.5 Test

6
1
4
0.5
2

0 0
0 1000 2000 3000 4000 5000 6000 0 1000 2000 3000 4000 5000 6000
Time (s) Time (s)

Fig. 6. Comparison between the computed results and tested results of plastic accumulative strain under different cyclic stress amplitude of 2 kPa (a), 10 kPa (b), 25 kPa (c), and 40 kPa
(d).

80
Q.-L. Deng, X.-W. Ren
a 0.04
Residual (%) 0.02
0
-0.02
-0.04
0 1000 2000 3000 4000
Time (s)
c 0.1
0.05
Residual (%)
0
-0.05
-0.1
0 1000 2000 3000 4000
Time (s)
Fig. 7. Histogram of residual errors of calculation for plastic accumulative strain.
5. Summary and conclusion
Soft clay is fragile to be disturbed and to produce deformation
under cyclic loads. Many geotechnical and geology engineering pro-
blems are involved in this soft soil area subjected to traffic loads,
mechanical vibration loads, waves, and other long-term cyclic loads. It
is necessary and important to investigate dynamic response and
deformation mechanism of soft soil under these loads. To do this, a
simple theoretical energy method was proposed by assuming soil to be
a one-dimensional vibration system. In the formulation, equations are
derived based on general thermodynamic principle and Newton's laws
of motion.
To verify the proposed model, a number of cyclic triaxial tests were
conducted for undisturbed soft clay. Plastic cumulative deformation
was computed by the proposed approach without considering the effect
of free vibration, which has been proved to have little influence on
displacement and plastic strain in this paper. Results were compared
with the experimental data, and good agreement was achieved.
Residual errors and their causes were also analyzed. The comparison
result reveals the capability of the model to describe the development
of deformation behavior of soft soil under cyclic loads, as well as its
flexibility to predict the plastic accumulative deformation. In addition,
limitations of the proposed model were discussed in the last part of the
paper.
Since this paper bypasses the extremely complicated constitutive
relation of soil, the proposed model does not need to compute plastic
accumulative strain cycle-by-cycle, which successfully avoids excessive
complexity in calculation especially for the huge number of cyclic load
applications. This is a major advantage distinguishing from other
existing theoretical methods. Furthermore, this paper develops a new
insight to describe dynamic response and to predict plastic accumula-
tive deformation of soft soils subjected to cyclic loads. On the basis of
this work, future work will focus on the investigation of model
performance in engineering practice, and improvements will possibly
be introduced to make the model with few readily measurable para-
meters.
Soil Dynamics and Earthquake Engineering 94 (2017) 75–82
b 0.03
0.02
Residual (%)
0.01
0
5000 6000 0 1000 2000 3000 4000 5000 6000
Time (s)
d 0.5
Residual (%)
0
-0.5
5000 6000 0 1000 2000 3000 4000 5000 6000
Time (s)
Acknowledgements
This study was supported by the Major State Basic Research
Development Program of China (No. 2011CB710601) and the
National High Technology Research and Development Program (No.
2012AA121303). The Project was also supported by the Fundamental
Research Funds for the Central Universities (No. CUG160701 and No.
CUG160813), China University of Geosciences(Wuhan).
References
[1] Li LL, Dan HB, Wang LZ. Undrained behavior of natural marine clay under cyclic
loading. Ocean Eng 2011;38(16):1792–805.
[2] Chai JC, Miura N. Traffic-load-induced permanent deformation of road on soft
subsoil. J Geotech Geoenviron Eng 2002;10:907–16.
[3] Ren XW, Tang YQ, Li J, Yang Q. A prediction method using grey model for
cumulative plastic deformation under cyclic loads. Nat Hazards 2012;64:441–57.
[4] Wang J, Guo L, Cai YQ, Bian XC, Gu C. Strain and pore pressure development on
soft marine clay in triaxial tests with a large number of cycles. Ocean Eng
2013;74:125–32.
[5] Tang LS, Chen HK, Sang HT, Zhang SY, Zhang JY. Determination of traffic-load-
influenced depths in clayey subsoil based on the shakedown concept. Soil Dyn
Earthq Eng 2015;77:182–91.
[6] Guo L, Chen J, Wang J, Cai Y, Deng P. Influences of stress magnitude and loading
frequency on cyclic behavior of K0-consolidated marine clay involving principal
stress rotation. Soil Dyn Earthq Eng 2016;84:94–107.
[7] Ng CWW, Liu GB, Li Q. Investigation of the long-term tunnel settlement
mechanisms of the first metro line in Shanghai. Can Geotech J 2013;50:674–84.
[8] Prevost JH. Anisotropic undrained stress-strain behavior of clay. J Geotech Eng Div
ASCE 1978;104(8):1075–90.
[9] Papadimitriou GA, Bouckovalas GD. Plasticity model for sand under small and large
cyclic strains: a multiaxial formulation. Soil Dyn Earthq Eng 2002;22:191–204.
[10] Krieg RD. A practical two-surface plasticity theory. J Appl Mech ASME
1975;42:641–6.
[11] Dafalias YF. Bounding surface plasticity (I): mathematical formulation and
hypoplasticity. J Engng Mech ASCE 1986;112(9):966–87.
[12] Dafalias YF, Herrmann LR. Bounding surface plasticity (II):application to isotropic
cohesive soils. J Eng Mech ASCE 1986;112(12):1263–91.
[13] Cheng Z, Jeremic B. Numerical modeling and simulation of pile in liquefiable soil.
Soil Dyn Earthq Eng 2009;29:1405–16.
[14] Tasiopoulou P, Gerolymos N. Constitutive modeling of sand: formulation of a new
plasticity approach. Soil Dyn Earthq Eng 2016;82:205–21.
[15] Zienkiewicz OC, Leung KH, Pastor M. Simple model for transient soil loading in
earthquake analysis (I): basic model and its application. Int J Numer Anal Methods
Geomech 1985;9:453–76.
[16] Pisano F, Jeremic B. Simulating stiffness degradation and damping in soils via a
simple visco-elastic–plastic model. Soil Dyn Earthq Eng 2014;63:98–109.
81
Q.-L. Deng, X.-W. Ren
[17] Monismith CL, Ogaw N, Freeme CR. Permanent Deformation Characteristics of
Subgrade Soils due to Repeated Loading, TRR 537, Washington D.C., TRB 1975, p.
1–17
[18] Li D, Selig ET. Cumulative plastic deformation for fine-grained subgrade soils. J
Geotech Eng 1996;122(12):1006–13.
[19] Huang MS, Li JJ, Li XZ. Cumulative deformation behavior of soft clay in cyclic
undrained tests. Chn J Geotech Eng 2006;28(7):891–5.
[20] Seed HB, Chan CK, Monismith CL. Effects of repeated loading on the strength and
deformation of compacted clay. Proc Hwy Res Rec 1955;3:541–58.
[21] Seed HB, Chan CK. Effects of stress history and frequency of stress application on
deformation of clay subgrades under repeated loading. HRB Proc 1958;37:555–75.
[22] Andersen KH, Pool JH, Brown SF, Rosenbrand WF. Cyclic and static laboratory
tests on Drammen clay. J Geotech Eng Div ASCE 1980;106(5):499–529.
[23] Yasuhara K. Cyclic strength and deformation of normally consolidated clay. Soils
Found 1982;22(3):77–91.
[24] Moses GG, Rao SN, Rao PN. Undrained strength behaviour of a cemented marine
clay under monotonic and cyclic loading. Ocean Eng 2003;30:1765–89.
[25] Sun L, Gu C, Wang P. Effects of cyclic confining pressure on the deformation
characteristics of natural soft clay. Soil Dyn Earthq Eng 2015;78:99–109.
[26] Kaya Z, Erken A. Cyclic and post-cyclic monotonic behavior of Adapazari soils. Soil
Dyn Earthq Eng 2015;77:83–96.
[27] Voznesensky EA, Nordal S. Dynamic instability of clays: an energy approach. Soil
Dyn Earthq Eng 1999;18:125–33.
[28] Nemat-Nasser S, Shokooh A. A unified approach to densification and liquefaction of
cohesionless sand in cyclic shearing. Can Geotech J 1979;16(4):659–78.
[29] Davis RO, Berrill JB. Energy dissipation and seismic liquefaction in sands. Earthq
Eng Struct D 1982;10(1):59–68.
[30] Law KT, Cao YL, He GN. An energy approach for assessing seismic liquefaction
potential. Can Geotech J 1990;27(3):320–9.
[31] Cao YL, Law KT. Energy dissipation and dynamic behaviour of clay under cyclic
loading. Can Geotech J 1992;29:103–11.
[32] Igarashi S. Dislocation energy of liquefaction. J Geotech Eng
1993;481(25):175–86.
[33] Okada N, Nemat-Nasser S. Energy dissipation in inelastic flow of saturated
cohesion less granular media. Geotechnigue 1994;44(1):1–19.
[34] Figueroa JL, Saada AS, Liang L, Dahisaria NM. Evaluation of soil liquefaction by
energy principles. J Geotech Eng 1994;120(9):1554–69.
[35] Liang L, Figueroa JL. Liquefaction under random loading: unit energy approach. J
Geotech Eng 1995;121(11):776–81.
[36] Baziar MH, Jafarian Y. Assessment of liquefaction triggering using strain energy
82
Soil Dynamics and Earthquake Engineering 94 (2017) 75–82
concept and ANN model: capacity energy. Soil Dyn Earthq Eng
2007;27(12):1056–72.
[37] Kokusho T. Liquefaction potential evaluations: energy-based method versus stress-
based method. Can Geotech J 2013;50:1088–99.
[38] Jafarian Y, Vakili R, Abdollahi AS, Baziar MH. Simplified soil liquefaction
assessment based on cumulative kinetic energy density: attenuation law and
probabilistic analysis. Int J Geomech ASCE 2014;14(2):267–81.
[39] Trifunac MD. Empirical criteria for liquefaction in sands via standard penetration
tests and seismic wave energy. Soil Dyn Earthq Eng 1985;14:419–26.
[40] Azeiteiro RJN, PALF Coelho, DMG Taborda, Grazina JC. Dissipated Energy in
Undrained Cyclic Triaxial Tests. 6th International Conference on Earthquake
Geotechnical Engineering, 1–4 November, Christchurch, New Zealand, 2015.
[41] McKevitt WE, Anderson DL, Nathan ND, Cherry S. Toward simple energy method
for seismic design of structures [Second U.S. national conference on Earthquake
Engineering]. Berkeley, CA: Earthquake Engineering Research Institute; 1979. p.
383–92.
[42] Bertero RD, Bertero VV, Teran-Gilmore A. Performance-based earthquake resistant
design based on comprehensive design philosophy and energy concepts. In:
Proceedings of the Eleventh world conference on earthquake engineering, Disc2,
Paper no.611. Oxford, England: Pergamon, Elsevier Science Ltd. [can be down-
loaded from 〈http://www.nicee.org/wcee/〉]; 1996.
[43] Trifunac MD, Todorovska MI. A note on energy of strong ground motion during
North ridge, California, earthquake of January17, 1994. Soil Dyn Earthq Eng
2013;47:175–84.
[44] Das BM, Luo Z. Principles of soil dynamics, 3rd edition. CL Engineering; 2016.
[45] Tang YQ, Zhou J, Liu S, Yang P, Wang JX. Test on cyclic creep behavior of mucky
clay in Shanghai under step cyclic loading. Environ Earth Sci 2011;63(2):321–7.
[46] Tang YQ, Cui ZD, Zhang X, Zhao SK. Dynamic response and pore pressure model of
the saturated soft clay around the tunnel under vibration loading of Shanghai
subway. Eng Geol 2008;98:16–132.
[47] Ren XW, Tang YQ, Xu YQ, Wang YD, Zhang X, Liu S. Study on dynamic response of
saturated soft clay under the subway vibration loading I: instantaneous dynamic
response. Environ Earth Sci 2011;64:1875–83.
[48] Mitchell RJ, King RD. Cyclic loading of an Ottawa area Champlain sea clay. Can
Geotech J 1977;14:52–63.
[49] Tang YQ, Huang Y, Ye WM, Wang YL. Critical dynamic stress ratio and dynamic
strain and analysis of soils around the tunnel under subway train loading. Chn J
Rock Mech Eng 2003;22(9):1566–70.
[50] Ren XW. Dynamic response and deformation of soft soils under high speed train
loads [PhD dissertation]. Shanghai, China: Tongji Univ.; 2014.