Soil Dynamics and Earthquake Engineering 31 (2011) 491–501

Contents lists available at ScienceDirect

Soil Dynamics and Earthquake Engineering

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

Transient dynamic response of fluid-saturated soil under a moving

cyclic loading
Yu-Jia Hu a, Yuan-Yuan Zhu b, Chang-Jun Cheng c,n
a
Department of Mechanical Engineering, University of Shanghai for Science and Technology, Shanghai 200092, PR China
b
Department of Computer Science and Technology, Shanghai Normal University, Shanghai 200234, PR China
c
Shanghai Institute of Applied Mathematics and Mechanics, Department of Mechanics, Shanghai University, Shanghai 200072, PR China

a r t i c l e in f o abstract

Article history: Based on the Theory of Porous Media (TPM), a mathematical model of a two-dimensional incompressible
Received 16 March 2010 fluid-saturated elastic soil is established, and the periodic boundary conditions are presented to analyze
Received in revised form the transient dynamic response of this soil under a moving cyclic loading. The differential quadrature
2 November 2010
method (DQM) and the second-order backward difference scheme are applied to discretize the governing
Accepted 3 November 2010
equations on the spatial and temporal domains, respectively. As application, a typical two-dimensional
wave-induced transient problem with a seabed of finite thickness is analyzed, and the numerical results
Keywords: are compared with the analytical results presented in the present work. In addition, a transient dynamic
Theory of Porous Media (TPM) response of fluid-saturated soil under limit moving vehicle loadings is studied. The effects of the velocity
Fluid-saturated soil
of vehicle and the volume fraction on the settlement and the pore water pressure are studied.
Cyclic loading
Crown Copyright & 2010 Published by Elsevier Ltd. All rights reserved.
Differential quadrature method
Transient dynamic response

1. Introduction
The investigation of porous media plays an important role in its
classical application fields, such as soil mechanics, hydrology, etc.;
furthermore, it is essential to the development of emerging
sciences and technologies as well, such as, mechanical character-
istic of skin, soft tissue and articular cartilage in biology, etc. Hence,
it is important to provide the relevant theoretical analyses and
numerical simulation methods. In addition, the transient response
of a saturated soil is also essential to understand the deformations
as well as the pore water pressures generated by ground motion.
This response is a key factor to the analysis of foundation of
buildings, offshore structures, and wave propagation in geological
medium due to blast or earthquake.
There are some theories describing the characteristics of porous
media, for example, Biot Theory [1,2], Porous Media Theory [3–5]
and Hybrid Mixture Theory [6], etc. Detailed comments on these
theories can be found in Ref. [7] presented by Schanz. In 2005, de
Boer [5] established a complete porous elastic theory based on the
continuous media mixture laws and the concept of volume fraction,
in which several micro-mechanical characteristics can be directly
applied to describe the macro-mechanical characteristics. Thus, the
Porous Media Theory may be an alternative method to study the
mechanical properties of porous media.
n
Corresponding author.
E-mail address: chjcheng@mail.shu.edu.cn (C.-J. Cheng).
Most of the transient response studies for porous media are
solved by numerical methods, such as finite element method (FEM)
[8–11], boundary integral–boundary element method (BEM)
[12–15],etc. Ghaboussi and Wilson [8] seem to be the first to have
formulated a FEM for dynamical processes in poroelastic media
based on the Biot’s theory. Zienkiewicz and Shiomi [9] published
the most complete FEM presentation for poroelastodynamics. Chen
[12,13] introduced the Green’s functions of saturated porous media
and applied BEM to study the dynamic response of porous media.
Khalili and Yazdchi [16] used the finite element–infinite element
method to study the wave propagation in saturated soil. Karim et al.
[17] applied the element-free Galerkin method (EFGM) to study the
transient response and obtained the stable numerical results.
Korsawe et al. [18] used the Galerkin finite element method
(GFEM) and the least-squares mixed finite element method
(LS-MFEM) to analyze the consolidation in porous media, and
thoroughly compared the effect of the two different numerical
approaches on the numerical results. Breuer and Jagering [19] used
a standard FEM to investigate elastic and plastic dynamical
problems of saturated porous media based on TPM. Hu et al.
[20] applied the DQM to analyze the dynamic response of fluid-
saturated visco-elastic porous media.
In this investigation, based on TPM, a mathematical model of a
two-dimensional incompressible saturated elastic soil is estab-
lished, and the periodic boundary conditions are presented for
analyzing the transient dynamic response of the saturated soil
under a moving cyclic loading. The differential quadrature method
(DQM) and the second-order backward difference scheme are
applied to discretize the governing equations on the spatial and

0267-7261/$ - see front matter Crown Copyright & 2010 Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.soildyn.2010.11.004
492 Y.-J. Hu et al. / Soil Dynamics and Earthquake Engineering 31 (2011) 491–501

temporal domains, respectively. As application, a typical two- skeleton may be expressed as

dimensional wave-induced transient problem with a seabed of @ux @uz @ux @uz
finite thickness is analyzed. In addition, the transient dynamic eSx ¼ , eSz ¼ , gSxz ¼ þ ð2Þ
@x @z @z @x
response of a fluid-saturated soil under limit moving vehicle
loading is analyzed. The effects of the velocity of vehicle and the Assume that the material of solid skeleton is linear elastic, the
volume fraction on the settlement and the pore water pressure are constitutive relations are
8
studied. > SE S S S S S
< sx ¼ ðl þ 2m Þex þ l ez
>
S S S
>
sSE S S
z ¼ ðl þ 2m Þez þ l ex ð3Þ
>
: tSE ¼ mS gS
2. Formulation of the problem xz rz

in which, e e g and sSE

S S S SE SE
x , sz , trz are the strain and effective stress
x , z , rz
Consider a two-dimensional half-plane incompressible fluid-satu-
components of the solid skeleton.
rated soil with depth H. The physical model is shown in Fig. 1, and the
The total stress components of a fluid-saturated porous media
domain occupied by the fluid-saturated soil is O ¼{(x,z): NrxrN,
are given as
0rzrH}. Assume that a moving cyclic loading qðx,H,tÞ or a pore water
pressure pðx,H,tÞ is applied on the top of the fluid-saturated soil. An sSx ¼ sSE S
x n p, sSz ¼ sSE S
z n p, tSxz ¼ tSE
xz ð4Þ
arbitrary periodic domain Oi is chosen for analyzing (shown in Fig.1), in
For a two-dimensional incompressible saturated elastic media
which l is the wavelength of periodic loading. Set Gl and Gr are the
in the case of the small deformation, the kinematics equations,
periodic boundaries of the domain Oi at the left and right vertical
from (1) to (4), are given as follows:
borders, respectively, which can be expressed as Gl ¼ {(x,z):x¼0,
8 S
0rzrH}, Gr ¼{(x,z):x¼l, 0rzrH}. >
> ðl þ mS Þ @x
@e
þ mS Dux nS @p S€
@x r u x þ Sv w x ¼ 0
~
>
>
>
> S S @e S S @p S€
>
> ðl þ m Þ @z þ m Duz n @z r u z þSv w z ¼ 0 ~
<
2.1. Governing equations of incompressible saturated elastic soils F _~
nF @p
@x þ r ðw x u x Þ þ Sv w x ¼ 0
€ ~ ð5Þ
>
>
> F @p
> F ~_
Based on TPM [6], the governing equations of an incompressible >
> n þ r ðw z u z Þ þ Sv w z ¼ 0
€ ~
> @z
>
: @ ðu_ þ nF w ~ x Þ þ @z ðu z þ nF w
@ _ ~ zÞ ¼ 0
saturated elastic soil in the case of small deformation can be @x x
expressed as follows:
in which, D ¼ ð@2 =@x2 Þ þ ð@2 =@z2 Þ,e ¼ div u ¼ ð@ux =@xÞ þ ð@uz =@zÞ; w
~x¼
S
ðl þ mS Þgrad di v uS þ mS di v grad uS nS grad prS u€ S þ Sv ðw
_ S u_ S Þ ¼ 0 w_ x u_ x , w
~ z ¼w
_ z u_ z are the relative velocities in the x- and z-
ð1aÞ directions, respectively.

nF grad p þ rF w
€ S þ Sv ðw
_ S u_ S Þ ¼ 0 ð1bÞ 2.2. Boundary conditions

divðnS u_ S þ nF w
_ SÞ ¼ 0 ð1cÞ Usually, the boundary conditions of the solid skeleton and the
in which, uS ,wS are the displacement vectors of the solid skeleton fluid phase are prescribed in a saturated soil, respectively. For the
and the fluid phase, respectively, in which uS ¼(ux,uy,uz), physical model shown in Fig. 1, assume that a time-related vertical
wS ¼(wx,wy,wz) in the Cartesian coordinate system. p is the pore loading qðx,H,tÞ or a pore water pressure pðx,H,tÞ is applied on the
S
water pressure of the fluid phase. mS , l are the Lame constants of top of the soil; the bottom of soil is rigid and impermeable and the
F 2 FR F
the solid skeleton. Sv ¼ ðn Þ g =k is the coupled interaction outer vertical boundaries meet the periodic boundary conditions.
coefficient between the solid skeleton and the fluid phase, in which Thus, the boundary conditions can be given as follows.
gFR is the effective specific weight of the fluid; kF is the Boundary conditions on the top of the soil (0rx rl,z¼H) are
Darcy permeability coefficient. rS , rF are the partial densities of sSE txz ¼ 0, p ¼ pðx,H,tÞ
z ¼ qðx,H,tÞ, ð6Þ
the solid skeleton and fluid phase, respectively, in which ri ¼ ni riR ,
riR(i¼S,F) are the real densities of the solid skeleton and the Boundary conditions on the bottom of the soil (0rx rl,z¼0)
fluid phase, ni(i¼S,F) are the corresponding volume fractions. are
The volume fraction of the solid skeleton nS can be approximately @p
ux ¼ 0, uz ¼ 0, ¼0 ð7Þ
denoted as nS ¼ nS0S ðdetðxi,K ÞÞ1  nS0S in the case of small @z
deformation, and nS0S is the solid volume fraction in the initial On the outer vertical boundaries Gl and Gr, the following periodic
configuration. boundary conditions should be satisfied, namely, (1) the displace-
For a two-dimensional incompressible saturated elastic media ment components are equal; (2) the corresponding total stress
in the case of small deformation, the geometric relations of the solid components are equal; (3) the pore pressures are equal;
and (4) the inflow quality is equal to the outflow quality. Hence,
we have
p (x,H,t) or q (x,H,t)      
uS Gr ¼ uS Gl , t S Gl ¼ t S Gr , pGl ¼ pGr ,
 
h/2 mF G ¼ mF G , ðX,tÞ A C  ½0, þ 1Þ
l r
ð8Þ
Load
R
in which, uS, tS, p and mF ¼ Gi ri Uw _ dsði ¼ l,rÞ are the displacement
Γl Saturated soil Γr H and stress vectors, pore water pressure and the inflow quality
(or the outflow quality). For the physical model ðrl ¼ rr Þ shown in
z Ωi
x Fig. 1, the periodic boundary conditions can be further simplified
from Eq. (8).
l l l l The periodic boundary conditions of the solid skeleton are
       
Fig. 1. Physical model of saturated soil under the cyclic loading. ux Gl ¼ ux Gr , uz Gl ¼ uz Gr , sSx Gl ¼ sSx Gr , tSxz Gl ¼ tSxz Gr ð9aÞ
 Y.-J. Hu et al. / Soil Dynamics and Earthquake Engineering 31 (2011) 491–501 493

The periodic boundary conditions of the fluid phase are Nx

X
nF Að1Þ F _~
zk ðpÞkZ þ r ððw x ÞzZ ðu x ÞzZ Þ þ Sv ðw x ÞzZ ¼ 0,
    € ~
pGl ¼ pGr , ðw
~ x þ u_ x ÞGl ¼ ðw
~ x þ u_ x ÞGr ð9bÞ k¼1
z ¼ 1,2,. . .,Nx , Z ¼ 1,2,. . .,Nz ð13cÞ
2.3. Initial conditions

Nz
X  
In general, the initial conditions can be given as _~
nF Bð1Þ F
Zl ðpÞzl þ r ðw z ÞzZ ðu z ÞzZ þ Sv ðw z ÞzZ ¼ 0,
€ ~
uS ¼ 0, ~ S ¼ 0,
w u_ S ¼ 0, _~ S ¼ 0,
w ð10Þ
l¼1
z ¼ 1,2,. . .,Nx , Z ¼ 1,2,. . .,Nz ð13dÞ
Thus, Eqs. (2)–(10) compose the dynamic governing equations of a
two dimensional incompressible saturated elastic soil under a
moving cyclic loading. Nx
X   XNz  
F ~ F ~
Að1Þ _
zk ðu x ÞkZ þ n ðw x ÞzZ þ Bð1Þ _
Zl ðu z ÞkZ þ n ðw z ÞzZ ¼ 0,
k¼1 l¼1

3. Solution methods z ¼ 2,3,. . .,Nx 1, Z ¼ 2,3,. . .,Nz 1 ð13eÞ

The DQM [21,22] is used to discretize the dynamic governing

equations in the spatial domain. The fundamental idea of the DQM is to
3.2. DQ discretization of boundary conditions
approximate an unknown function and its derivative at any discrete
point as the linear weighted sums of its values at all of the discrete
~ x,w
~ z ,p and The DQ discretization of the boundary conditions can be yielded
points chosen in the solution domain. The functions ux ,uz , w
from Eqs. (6)–(9).
their kth order derivatives with respect to x can be approximated by
On the top of soil (z ¼ 1,2,. . .,Nx , Z ¼ Nz ), we have the DQ
  X
Nx n o discretization forms of the boundary conditions (6) as
~ x,w
ux ,uz , w ~ z ,p ¼ ~ x Þmj ,ðw
lm ðxÞ ðux Þmj ,ðuz Þmj ,ðw ~ z Þmj ,ðpÞmj
m¼1 ðsSE SE
z ÞzNz ¼ qzNz ,ðtz ÞzNz ¼ 0, ðpÞzNz ¼ pzNz ð14aÞ
ð11Þ
in which, qzNz , pzNz are the values of the loading qðr,H,tÞ and the pore
@k  water pressure pðr,H,tÞ at grid points.
~ x,w
ðux ,uz , w ~ z ,pÞx ¼ xi
@xk On the bottom of soil (z ¼ 1,2,. . .,Nx , Z ¼ 1), we have the DQ
X Nx   discretization forms of the boundary conditions (7) as
¼ AðkÞ
im
~ x Þmj ,ðw
ðxÞ ðux Þmj ,ðuz Þmj ,ðw ~ z Þmj ,ðpÞmj ð12Þ
m¼1
Nz
X
ðux Þz1 ¼ 0, ðuz Þz1 ¼ 0, Bð1Þ
1l
ðpÞzl ¼ 0 ð14bÞ
l¼1
where,fðux Þmj ,ðuz Þmj ,ðw ~ x Þmj ,ðw
~ z Þmj ,ðpÞmj g ¼ fux ðxm ,zj ,tÞ,uz ðxm ,zj ,tÞ,
On the periodic boundaries Gl(zl ¼ 1, Zl ¼ 2,3,. . .,Nz 1) and Gr
~ x ðxm ,zj ,tÞ, w
w ~ z ðxm ,zj ,tÞ,pðxm ,zj ,tÞg, where Nx is the total number of grid
(zr ¼Nx, Zr ¼ 2,3,. . .,Nz 1), the DQ discretization forms of the
points distributed along the x-axis, lm(x) is the Lagrange interpolation boundary conditions (9) are given as
polynomial, and AðkÞim
is the weighting coefficient, the corresponding
ðux Þ1Zl ¼ ðux Þ1Zr , ðuz Þ1Zl ¼ ðuz Þ1Zr
recursive formula can be found in Refs. [21,22]. The k-order partial
S S
derivative of the function with respect to z has the similar expressions, ðs x Þ1Zl ¼ ðs x Þ1Zr , ðtSxz Þ1Zl ¼ ðtSxz Þ1Zr ð14cÞ
and let BZðkÞl be the corresponding weighting coefficients of the kth order
derivative. ðpÞ1Zl ¼ ðpÞ1Zr , ~ x Þ1Z þðu_ x Þ1Z ¼ ðw
ðw ~ x Þ1Z þ ðu_ x Þ1Z ð14dÞ
l l r r
The accuracy of DQM relates to the distribution of grid points. In
this study, the zero of the Chebyshev–Lobatto polynomial is Hence, Eqs. (13) and (14) and the initial conditions form the DQ
adopted as the coordinates of the grid point, and assume that discretization forms of the initial-boundary-value problem of
Nx  Nz grid points are collocated in the x-, z-directions of the spatial saturated soil in the domain Oi.
domains on the xz-plane, respectively.
3.3. Discretization of derivative terms with respect to time
3.1. DQ discretization of governing equations
Set U~ ¼ ½u~ r , u~ z , w
~ r ,w ~ T is an unknown vector collecting all the
~ z , p
Substituting the relationships (12) into Eqs. (5), one can obtain values of unknown variables at nodes, and u~ r , u~ z , w ~ r ,w
~ z , p~ are the
the following DQ of Eq. (5) discretization equations: sub-vectors collecting the values of the corresponding unknown
X
Nx Nz
X X
Nx Nz
X variables at all nodes. The second-order backward difference
S S
ðl þ 2mS Þ Að2Þ
zk ðux ÞkZ þ m
S
Bð2Þ S
Zl ðux Þzl þ ðl þ m Þ Að1Þ
ik
Bð1Þ
jl
ðuz Þkl scheme is applied to deal with the derivative terms with respect
k¼1 l¼1 k¼1 k¼1
to time, namely
X
Nx

nS Að1Þ i S €
zk ðp ÞkZ r ðu x ÞzZ þ Sv ðw x ÞzZ ¼ 0,
~ dU~ 3 ~t t1 1 t2
k¼1 ¼ U 2U~ þ U~ Þ=Dt ð15aÞ
dt 2 2
z ¼ 2,3,. . .,Nx 1, Z ¼ 2,3,. . .,Nz 1 ð13aÞ
!
1 t2
2
d U~ 9 ~t t1 3 t2 _ t1
! ! ¼ U 3U~ þ U~ 2U~ Dt þ U_~ Dt =Dt 2 ð15bÞ
X
Nx Nz
X Nz
X Nz
X dt 2 4 4 2
S S S
ðl þ 2m Þ Að1Þ
ik
Bð1Þ
jl
ðux Þkl þ Bð2Þ
Zl ðuz Þzl þm Bð2Þ
Zl ðuz Þzl
k¼1 k¼1 l¼1 l¼1 in which, t denotes the current time. By using step by step solution
t
Nz
X process, the unknown vector U~ can be obtained when the relevant
nS Bð1Þ S €
Zl ðpÞzl r ðu z ÞzZ þSv ðw z ÞzZ ¼ 0,
~ initial conditions are given. Unless otherwise specified, let the time
l¼1 step Dt ¼0.02, and 17  17 grid points are collected on the spatial
z ¼ 2,3,. . .,Nx 1, Z ¼ 2,3,. . .,Nz 1 ð13bÞ domain Oi in the computing procedure.
494 Y.-J. Hu et al. / Soil Dynamics and Earthquake Engineering 31 (2011) 491–501

4. Transient dynamic response of two-dimensional saturated
soils
4.1. Comparison of numerical results with analytical results
Before proceeding to analyze the transient dynamic response of
a two-dimensional saturated soil under a moving cyclic loading,
the convergence of the proposed solution method is evaluated by
comparison of numerical results with analytical solutions. de Boer
et al [23] presented an analytical solution for the one-dimensional
transient wave propagation in a fluid-saturated incompressible
porous media by the Laplace transform. In order to simulate this
problem by DQM, set the depth of soil as 10 m, the width of soil as
1 m and assume that the upper surface of soil is ideal drained and
subjected to a time-related vertical loading qðx,H,tÞ, and its bottom
surface is rigid and undrained, the outer vertical borders are rigid
along the x-direction and ideal undrained (shown in Fig. 2).
Consider two kinds of vertical loadings subjected to the top of
soil: (1) the cyclic loading qðx,H,tÞ ¼ q0 ½1cosðotÞ, 0 rxr1 m; (2)
the step loading qðx,H,tÞ ¼ q0 hðtÞ, 0 rx r1 m, in which, q0 ¼3 kN/
m2, o ¼75 s  1, and h(t) is the Heaviside function. The material
parameters of a fluid-saturated soil are shown in Table 1. Tables 2
and 3 list the comparison of the DQ numerical results with the
analytical results at different times under the two kinds of dynamic
loadings. It can be seen that the DQM is an effective numerical
method to solve the transient dynamic response of saturated soils.

1m
q 4.2. Wave-induced transient problem with a seabed of finite thickness

The physical model of a seabed with finite thickness is shown in

Fig. 1, the material parameters of a fluid-saturated soil are shown in
xz = 0
Table 1. The pore wave pressure pðx,tÞ ¼ p0 cosðaxotÞ, in which,
a¼2p/l is the wave number; l is the wavelength, o ¼ 2p/T is the wave
ux = 0 ux = 0 frequency, T is the wave period and x is the horizontal coordinate. The
amplitude factor p0 is related to the wave pressure. Based on the first
order linear wave theory [24], p0 ¼ gw h=2 cosh adw , in which, h is the
xz = 0 xz = 0 wave height and dw is the still water depth. In computation, set T¼1 s,
p0 ¼1000 Pa.
Based on TPM, by introducing the potential functions and using
the variable separation method, the steady-state analytical solu-
10 m tions of the problem can be obtained as
hh
ux ¼ Re iaðc1 ezl1 þ c2 ezl1 þ c3 ezl2 þ c4 ezl2 Þ
∂p ∂p
=0 = 0 pffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi i i
∂x ∂x þc5 b2 sinð b2 zÞc6 b2 cosð b2 zÞ eiðaxotÞ ð16aÞ

hh
uz ¼ Re ðc1 ezl1 l1 c2 ezl1 l1 þ c3 ezl2 l2 c4 ezl2 l2 Þ
pffiffiffiffiffi pffiffiffiffiffi i i
þiaðc5 cosð b2 zÞ þ c6 sinð b2 zÞÞ eiðaxotÞ ð16bÞ

∂p ""
u x = uz = =0 2 2
∂z ezl1 ðb1 þ l1 Þ ezl1 ðb1 þ l1 Þ
p ¼ Re c1 þ c2
a2 a2
Fig. 2. Physical model and grid points. # #
zl2 2 zl2 2
e ðb1 þ l2 Þ e ðb1 þ l2 Þ iðaxotÞ
þ c3 þ c4 e ð16cÞ
a2 a2
Table 1
Material parameters [23].
""
2
iaezl1 ðnf ðb1 þ l1 Þa2 rf o2 Þ
s 3
u ¼ 5:583  10 kN=m 2 FR
r ¼ 1000 kg=m 3 ~ x ¼ Re
w c1
a2 ðSv irf oÞ
ls ¼ 8:375  103 kN=m2 r FR ¼ 10  1000 N=m3
2 2
rSR ¼ 2000 kg=m3 kF ¼ 0.01 m/s iaezl1 ðnf ðb1 þ l1 Þa2 rf o2 Þ iaezl2 ðnf ðb1 þ l2 Þa2 rf o2 Þ
nSOS ¼ 0:67 nFOF ¼ 0:33
þ c2 þ c3
a2 ðSv ir oÞ
f a2 ðSv irf oÞ

Table 2
Settlement uz(0,10,t) of the soil under the step loading.

Time Nx ¼Nz ¼ 7 Nx ¼ Nz ¼8 Nx ¼Nz ¼ 9 Analytical result [23] (mm)

Dt¼ 0.01 Dt¼ 0.02 Dt ¼0.01 Dt ¼0.02 Dt ¼ 0.01 Dt¼ 0.02

t ¼ 0.2 0.33638 0.33190 0.33628 0.33178 0.33636 0.33188 0.3413

t ¼ 0.4 0.48009 0.47699 0.48006 0.47696 0.48004 0.47694 0.4835
t ¼ 0.6 0.58971 0.58720 0.58969 0.58718 0.58968 0.58717 0.5925
t ¼ 0.8 0.68182 0.67966 0.68182 0.67965 0.68181 0.67965 0.6843
t ¼ 1.0 0.76248 0.76056 0.76249 0.76057 0.76249 0.76057 0.7652
 Y.-J. Hu et al. / Soil Dynamics and Earthquake Engineering 31 (2011) 491–501 495

Table 3
Settlement uz(0,10,t) of the soil under the cyclic loading.

Time Nx ¼Nz ¼7 Nx ¼ Nz ¼ 8 Nx ¼Nz ¼ 9 Analytical result [23] (mm)

Dt¼ 0.01 Dt ¼0.02 Dt ¼0.01 Dt ¼ 0.02 Dt¼ 0.01 Dt¼ 0.02

t ¼ 0.2 0.33336 0.34424 0.34861 0.36003 0.33735 0.34844 0.3413

t ¼ 0.4 0.53791 0.51754 0.51899 0.50564 0.53112 0.51530 0.4835
t ¼ 0.6 0.51696 0.53684 0.53029 0.53899 0.52316 0.53592 0.5925
t ¼ 0.8 0.74377 0.73385 0.74237 0.74242 0.74103 0.73738 0.6843
t ¼ 1.0 0.74892 0.74405 0.73770 0.72886 0.74683 0.73955 0.7433

2
iaezl2 ðnf ðb1 þ l2 Þa2 rf o2 Þ 0.0003
þ c4
a2 ðSv irf oÞ
pffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi # #
f 2
b2 r o sinð b2 zÞ b2 rf o2 cosð b2 zÞ iðaxotÞ 0.0002
c5 þ c6 e
ðSv irf oÞ ðSv irf oÞ
0.0001

Displacement /m
ð16dÞ
0.0000
""
zl1 f 2 f 2 zl1 f 2 f 2
~ z ¼ Re
iae ðn ðb1 þ l1 Þa2 ro Þ iae ðn ðb1 þ l1 Þa2 ro Þ
w c1 þ c2 -0.0001
ro
a2 ðSv i f Þ a2 ðSv i f Þ ro
2 2
iaezl2 ðnf ðb1 þ l2 Þa2 rf o2 Þ iaezl2 ðnf ðb1 þ l2 Þa2 rf o2 Þ -0.0002
þc3 þ c4
a2 ðSv irf oÞ a2 ðSv irf oÞ
# 7*7 grid pointds, Δt=0.10s
pffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi pffiffiffiffiffi
f 2
b2 r o sinð b2 zÞ f 2
b2 r o cosð b2 zÞ iðaxotÞ -0.0003 11*11 grid pointds, Δt=0.02s
þc5 c6 e ð16eÞ 15*15 grid pointds, Δt=0.02s
ðSv irf oÞ ðSv irf oÞ
Analysical results
-0.0004
-2 0 2 4 6 8 10 12 14 16
in which
Time /s
rS o2 ðirF o þSv ÞrF o2 Sv nS ðirF o þ Sv Þ þ nF Sv
a1 ¼ S
, a2 ¼  S Fig. 3. Settlement-history curve at point (0,15).
ðl þ 2mS ÞðirF o þ Sv Þ ðl þ 2mS ÞðirF o þ Sv Þ
rS o2 Sv rF o2 ðnF Þ2
a3 ¼  S , a4 ¼ F 2
m S F
m ðir o þ Sv Þ r o þ ioSv þ nF rF o2
0.0003
b1 ¼ a1 ðaÞ2 , b2 ¼ a3 ðaÞ2 , b3 ¼ ða1 a2 =a4 2a2 Þ,
b4 ¼ a4 a2 ða1 a2 =a4 Þ 0.0002
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Displcaement /m

b þ b3 4b4 b  b3 4b4 0.0001

l1 2 ¼ 3 , l2 2 ¼ 3
2 2
Above constants c1–c6 are the solutions of the following six 0.0000
linear equations:
2 32
3 2 3
-0.0001
k1 k1 k1 k1 0 k2 c1 0
6 k k3 k4 0 0 76 7 6 7
6 3 k4 76 c2 7 6 0 7 -0.0002
6 76 7 6 7
6 l1 l1 l2 l2 k1 0 76 c3 7 6 0 7
6 pffiffiffiffiffi pffiffiffiffiffi 76 7 6 7 7*7 grid pointds, Δt=0.10s
6 k k5 =e 2Hl1
k6 k6 =e 2Hl2
k7 sinð b2 HÞ k7 cosð b2 HÞ 7 6 7¼6 7
6 5 76 c 7 6 0 7 -0.0003 11*11 grid pointds, Δt=0.02s
6 pffiffiffiffiffi pffiffiffiffiffi 76 4 7 6 7
6 k k8 =e2Hl1 k9 =e2Hl1 k10 sinð b2 HÞ 7 6 7 6 7 15*15 grid pointds, Δt=0.02s
4 8 k9 k10 cosð b2 HÞ 54 c5 5 4 0 5
Analytical results
k11 k11 =e2Hl1 k12 k12 =e2Hl1 0 0 c6 p0
-0.0004
ð17Þ -2 0 2 4 6 8 10 12 14 16
Time /s
where the parameters are defined as
pffiffiffiffiffi 3 Fig. 4. Settlement-history curve at point (7.5,15).
k1 ¼ ia, k2 ¼  b2 , k3 ¼ ðl1 þ b1 l1 Þ=a2
3 2 S S
k4 ¼ ðl2 þ b1 l2 Þ=a2 , k5 ¼ eHl1 ðl1 ðl þ 2 S Þa2 l Þm
 
2 S 2 S
k6 ¼ eHl2 l2 ðl þ2mS Þa l to the steady-state. Figs. 5–7 show the contours of the displace-
pffiffiffiffiffi ments and the pore water pressures and the streamline charts
k7 ¼ 2ia b2 mS , k8 ¼ 2iaeHl1 l1 mS , k9 ¼ 2iae H l2
l2 m S

2 2
k10 ¼ ðb2 a2 ÞmS , k11 ¼ eHl1 ðb1 þ l1 Þ=a2 ,
Figs. 3 and 4 show the displacement-history curves at the
positions (0,15) and (7.5,15). Three kinds of nodal distributions,
namely 7  7, 11  11 and 15  15, are used to study the effect of the
nodal distributions on the numerical results. At the same time, two
time steps Dt ¼0.02 s, 0.1 s are used. It can be found that the effect
of the nodal distribution (for example, 11  11 and 15  15) on the
numerical result is small when the time step Dt ¼0.02 s. In addition,
from Figs. 3 and 4, one can see that the transient numerical results
are consistent with the analytical results when the motion reaches
k12 ¼ eHl2 ðb1 þ l2 Þ=a2
of the porous seabed at times t ¼10.0 s, 10.5 s (after 10 cycles).
It can be seen that there are no oscillations in these numerical
results. This characteristic is not readily achievable by standard
FEM [17].
4.3. Transient dynamic response of a saturated soil under limit
moving vehicle loadings
Consider a two-dimensional half-plane incompressible satu-
rated soil with the depth H. Assume that vehicles which have the
496 Y.-J. Hu et al. / Soil Dynamics and Earthquake Engineering 31 (2011) 491–501

15 15

10 10

z/m
z/m

5 5

0 0
0 5 10 15 0 5 10 15
x/m x/m

Fig. 5. Contours of the displacements at different times: (a) t ¼ 10.0 s and (b) t ¼10.5 s.

15 15

10 10
z/m

z/m

5 5

0 0
0 5 10 15 0 5 10 15
x/m x/m

Fig. 6. Contours of the pore water pressures at different times: (a) t¼ 10.0 s and (b) t ¼10.5 s.

15 15

10 10
z/m
z/m

5 5

0 0
0 5 10 15 0 5 10 15
x/m x/m

Fig. 7. Streamline charts of water in porous seabed at different times: (a) t¼ 10.0 s and (b) t ¼10.5 s.
 Y.-J. Hu et al. / Soil Dynamics and Earthquake Engineering 31 (2011) 491–501 497

c c c c c 0.0002

lv q (x, t) lv q (x, t) lv q (x, t) lv q (x, t) lv q (x, t) 0.0001 Volocity=20m/s

Γl Ωi Γr H Volocity=30m/s

Displacement /m
Soil 0.0000
z x

-0.0001
l l l l

Fig. 8. Physical model under the limit moving vehicle loadings. -0.0002

-0.0003

q0 -0.0004
i i+1 0 10 20 30 40 50
Time /s

Fig. 12. The effect of velocity of vehicle on the settlement-history curves at

l1 l2
point (0,20).

Fig. 9. Physical model of the concentrated loading.

0.0002

0.0001
0.0002 Velocity=20m/s
0.0000 Velocity=30m/s
15*15
Displacement /m

0.0001 17*17 -0.0001

-0.0002
Displacement /m

0.0000 -0.0003

-0.0004
-0.0001
-0.0005

-0.0006
-0.0002
-0.0007
0 10 20 30 40 50
-0.0003 Time /s
0 1 2 3 4 5 6 7
Fig. 13. The effect of velocity of vehicle on the settlement-history curves at point
Time /s
(10,20).
Fig. 10. The effect of grid points on the settlement-history curves at point (0,20).

0.1

0.0002 0.0
15*15
0.0001
17*17 -0.1
Settlement /mm

0.0000 Velocity=20m/s,maximum of settlement

-0.2 Velocity=30m/s,maximum of settlement
Displacement /m

-0.0001 Velocity=20m/s,minimum of settlement

-0.3 Velocity=30m/s,minimum of settlement
-0.0002

-0.0003 -0.4

-0.0004 -0.5
-0.0005
-0.6
-0.0006 0.1 0.2 0.3 0.4 0.5
0 1 2 3 4 5 6 7 volume fraction nF
Time /s
Fig. 14. The effect of volume fraction and velocity of vehicle on the settlement at
Fig. 11. The effect of grid points on the settlement-history curves at point (10,20). point (0,20).
498 Y.-J. Hu et al. / Soil Dynamics and Earthquake Engineering 31 (2011) 491–501

same mass move on the top of the soil with a constant velocity and loading, c is the velocity of vehicles, lv is the wheel gap, and l is the
spacing interval, namely the saturated soil is subjected to limit spacing interval of vehicle.
moving vehicle loadings (shown in Fig. 8). Set q0 is the vehicle An arbitrary periodic sub-domain Oi is used to study the transient
dynamic response of saturated soil under the limit moving vehicle
loadings. It is not difficult to obtain the moving vehicle loading which is
given as: qðx,tÞ ¼ q0 ðdðxct þ f1 ðx,tÞÞ þ dðxctlv þ f2 ðx,tÞÞÞ=2, in
0.0
which f1 ðx,tÞ ¼ IntððctÞ=lÞl, f2 ðx,tÞ ¼ Intððct þ lv Þ=lÞl; Int, truncates a
value to a whole number, which can be defined as IntðxÞ ¼
-0.1
n, n r x on þ1, in which n is an integer. The concentrated loadings
in the DQ simulation can be disposed as the following method.
Settlement /mm

-0.2
Velocity=20m/s,maximum of settlement Assume that a concentrated loading of vehicle is applied
Velocity=30m/s,maximum of settlement between the ith point and the (i+ 1)th point, and the distances
-0.3
Velocity=20m/s,minimum of settlement between the concentrated loading and the ith point and the (i+ 1)th
Velocity=30m/s,minimum of settlement point are l1 and l2, respectively (shown in Fig. 9). The loadings
-0.4
applied on the ith point and the (i+1)th point can be approximately
-0.5 expressed as

l2 l1
-0.6 qi ¼ q0 , qi þ 1 ¼ q0 ð18Þ
l1 þ l2 l1 þ l2

-0.7 In computation, the material parameters of the fluid-saturated

0.1 0.2 0.3 0.4 0.5 soil are shown in Table 1. Set l ¼20 m, H¼20 m, q0 ¼3000 N/m2,
c¼20 m/s, 30 m/s and lv ¼ 2 m. Figs. 10 and 11 show the settle-
volume fraction nF
ment-history curves at the points (0,20) and (10,20). Two kinds of
Fig. 15. The effect of volume fraction and velocity of vehicle on the settlement at grid points, namely 15  15 and 17  17, are used in the simula-
point. tions. From these figures, it can be noted that the numerical results

20 20

15 15
z/m

z/m

10 10

5 5

0 0
0 5 10 15 20 0 5 10 15 20
x/m x/m

20 20

15 15
z/m

z/m

10 10

5 5

0 0
0 5 10 15 20 0 5 10 15 20
x/m x/m

Fig. 16. Contours of the settlements at different times: (a) t ¼40.0 s, (b) t¼ 40.2 s, (c) t ¼40.5 s and (d) t¼ 40.8 s.
 Y.-J. Hu et al. / Soil Dynamics and Earthquake Engineering 31 (2011) 491–501 499

20 20

15 15
z/m

z/m
10 10

5 5

0 0
0 5 10 15 20 0 5 10 15 20
x/m x/m

20 20

15 15
z/m

z/m

10 10

5 5

0 0
0 5 10 15 20 0 5 10 15 20
x/m x/m

Fig. 17. Contours of the pore pressures at different times: (a) t ¼ 40.0 s, (b) t¼ 40.2 s, (c) t ¼40.5 s and (d) t ¼40.8 s.

with these two kinds of grid points are consistent, which denotes
that the simulation method on the concentrated loading is
effective. Figs. 12 and 13 show the effect of the velocity of vehicle
on the displacement of the saturated soil, and set c ¼20 m/s, 30 m/s,
17  17grid points and Dt¼0.02 s in the analysis. It can be seen that
the amplitudes are 2.40  10  4 and 3.17  10  4 at the point (0,20)
when the motion arrives at the steady state and the velocities of
vehicle are 20 and 30 m/s, respectively. In addition, the amplitudes
are 4.16  10  4 and 5.54  10  4 at the point (10,20), and the
velocities of vehicle are 20 and 30 m/s, respectively. Therefore, the
amplitude of the settlement increases with the increase of velocity
when the volume fraction nFOF ¼ 0:33. Figs. 14 and 15 systematically
study the effect of the volume fraction nF and the velocity of vehicle
c on the settlement of the saturated soil when the motion arrives at
the steady state. From these simulation results, one can find that
the velocity of vehicle has great effects on the settlement of the
saturated soil when the volume fraction nF ¼0.2. However, the
effect is relatively small when the volume fraction nF is small or
large. Figs. 16–18 show the contours of the settlement and the pore
water pressure and the streamline chart of water in porous soils at
t ¼40.0, 40.2, 40.5, 40.8 s (after 40 cycles) when the velocity is
20 m/s. From Fig. 17, one can see that there exists a zone of negative
pressure and a zone of positive pressure at the rear and the front of
the vehicle loadings, respectively. And these two zones will move
with the motion of vehicle. From Fig. 18, it is noticed that the
drainage zone at the top of the soil is only under the bottom of the
vehicle loadings.
5. Conclusion
In order to describe the general periodic problem of saturated
porous elastic soils, four kinds of periodic boundary conditions of a
saturated porous elastic soil under a moving cyclic loading are
presented in the present work, namely, (1) the displacement compo-
nents are equal; (2) the corresponding total stress components are
equal; (3) the pore water pressures are equal; and (4) the inflow
quality is equal to the outflow quality. A new numerical procedure is
presented based on TPM and DQM to analyze the transient response
of saturated porous elastic soil under a moving cyclic loading. The
present analysis is validated through the direct comparison between
the present numerical results and the analytical results. As applica-
tion, the transient dynamic response of a typical two-dimensional
wave-induced transient problem with a seabed of finite thickness is
analyzed. Finally, the transient dynamic response of a fluid-saturated
soil under limit moving vehicle loadings is studied. The effects of the
velocity of vehicle and the volume fraction on the settlement of the
fluid-saturated soil are obtained.
500 Y.-J. Hu et al. / Soil Dynamics and Earthquake Engineering 31 (2011) 491–501

20 20

15 15
z/m

z/m
10 10

5 5

0 0
0 5 10 15 20 0 5 10 15 20
x/m x/m

20 20

15 15
z/m

z/m

10 10

5 5

0 0
0 5 10 15 20 0 5 10 15 20
x/m x/m

Fig. 18. Streamline charts of water in porous soil at different times: (a) t¼ 40.0 s, (b) t ¼40.2 s, (c) t ¼ 40.5 s and (d) t ¼ 40.8 s.

Acknowledgements
Sponsored by the National Science Foundation for Post-doctoral
Scientists of China (Grant no. 20080440613); the Shanghai Post-
doctoral Sustentation Fund, China (Grant no. 09R21412700);
National Science Foundation for Distinguished Young Scholars of
China (Grant no. 11002084); and Shanghai Pujiang Program
(07pj14073).
References
[1] Biot MA. General theory of three-dimensional consolidation. J Appl Phys
1941;12:155–64.
[2] Biot MA. Theory of propagation of elastic waves in a fluid-saturated porous
media, I, low frequency range. J Acoust Soc Am 1956;28(2):168–78.
[3] Bowen RM. Compressible porous media models by use of the theory of
mixtures. Int J Eng Sci 1982;20:693–735.
[4] de Boer R. Theory of porous media: highlights in the historical development
and current state. Berlin, Heidelberg: Springer-Verlag; 2000.
[5] Schrefler BA. Mechanics and thermodynamics of saturated/unsaturated porous
material and quantitative solution. Appl Mech Rev 2002;5:351–80.
[6] de Boer R. Theoretical poroelasticity—a new approach. Chaos Solitons Fractals
2005;25:861–78.
[7] Schanz M. Poroelastodynamics: linear models, analytical solutions, and
numerical methods. Appl Mech Rev 2009;62:030803-1.
[8] Ghaboussi J, Wilson EL. Variational formulation of dynamics of fluid-saturated
porous elastic solids. J Eng Mech Div ASCE EM4 1972;98(4):947–63.
[9] Zienkiewicz OC, Shiomi T. Dynamic behavior of saturated porous media;
the generalised Biot formulation and its numerical solution. Int J Numer Anal
Methods Geomech 1984;8:71–96.
[10] Chen Z, Steeb H, Diebels S. A time-discontinuous Galerkin method for the
dynamical analysis of porous media. Int J Numer Anal Methods Geomech
2006;30:1113–34.
[11] Manzari MT. On finite deformation dynamic analysis of saturated soils. Arch
Mech 1996;48:281–310.
[12] Chen J. Time domain function solution to Biot’s complete equations of dynamic
poroelasticity part I: three dimensional solution. Int J Solid Struct 1994;31(2):
169–202.
[13] Chen J. Time domain function solution to Biot’s complete equations of dynamic
poroelasticity, part II: two dimensional solution. Int J Solid Struct 1994;31(10):
1447–90.
[14] Chopra M. Poro-elasticity using BEM. In: Kassab AJ, Aliabadi MH, editors.
Coupled field problems—advances in boundary elements, vol. 11_. WIT; 2001.
p. 211–40.
[15] Lu JF, Jeng DS, Williams S. A 2.5-D dynamic model for a saturated porous
medium: part II. Boundary element method. Int J Solids Struct 2008;45:
359–77.
[16] Khalili N, Yazdchi M, Valliappan S. Wave propagation analysis of two-phase
saturated porous media using coupled finite-infinite element method. Soil Dyn
Earthquake Eng 1999;18(2):533–53.
[17] Karim MR, Nogami T, Wang JG. Analysis of transient response of saturated
porous elastic soil under cyclic loading using element-free Galerkin method.
Int J Solid Struct 2002;39:6011–33.
[18] Korsawe J, Starke G, Wang W, Kolditz O. Finite element analysis of poro-elastic
consolidation in porous media: standard and mixed approaches. Comput
Methods Appl Mech Eng 2006;195:1096–115.
[19] Breuer S, Jagering S. Numerical calculation of the elastic and plastic behavior of
saturated porous media. Int J Solid Struct 1999;36:4821–40.
 Y.-J. Hu et al. / Soil Dynamics and Earthquake Engineering 31 (2011) 491–501 501

[20] Hu-jia Hu, Yuan-yuan Zhu, Chang-jun Cheng. Dynamic analysis of fluid-
saturated porous visoelastic materials. Int J Solid Struct 2009;46:1667–75.
[21] Bellmam RE, Kashef BG, Casti J. Differential quadrature: a technique for the
rapid solution of nonlinear partial differential equations. J Comput Phys
1972;10:40–52.
[22] Bert CW, Malik M. Differential quadrature method in computational
mechanics: a review. Appl Mech Rev 1996;49:1–28.
[23] de Boer R, Ehlers W, Liu Z. One-dimensional transient wave propagation in
fluid-saturated incompressible porous media. Arch Appl Mech 1993;63:
59–72.
[24] Madsen OS. Wave climate of the continental margin: elements of its
mathematical description. In: Stanley DF, Swift DJP, editors. Marine
sediment transport and environmental management. New York: John Wiley;
1976. p. 65–90.