IRJET- Numerical Analysis for Nonlinear Consolidation of Saturated Soil using Lattice Boltzmann Method

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 06 Issue: 04 | Apr 2019 www.irjet.net p-ISSN: 2395-0072
© 2019, IRJET | Impact Factor value: 7.211 | ISO 9001:2008 Certified Journal | Page 3611
Numerical Analysis for Nonlinear Consolidation of Saturated Soil using
Lattice Boltzmann Method
Pyol Kim1, Yong-Gun Kim 2, Hak-Bom Myong 1, Chung-Hyok Paek 1, Jun Ma 1
1 Faculty of Geology, Kim Il Sung University, Pyongyang 999093, Democratic People’s Republic of Korea
2 Faculty of Global Environmental Science, Kim Il Sung University, Pyongyang 999093, Democratic People’s
Republic of Korea
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - In this paper, a new numerical method for one-
dimensional (1D) nonlinear consolidation analysis of
saturated soil is proposed on the basis of the lattice
Boltzmann method. At first, the lattice Bhatnagar-Gross-
Krook (LBGK) model is used for 1D nonlinear consolidation
problem of saturated soil subjected to time-dependent
loading under different types of boundary conditions. In
addition, the multiscale Chapman-Enskog expansion is
applied to recover mesoscopic lattice Boltzmann equation to
macroscopic nonlinear consolidation equation. As a result of
the numerical simulation for verification, the numerical
results are proved to be in good agreement with the
analytical solutions available in previous literature. Finally,
the numerical simulation is performed to investigate the
consolidation behavior of saturated soil subjected to two
different types of time-dependent loading.
Key Words: Nonlinear consolidation, Saturated soil,
Lattice Boltzmann method, Time-dependent loading
1.INTRODUCTION
It is very important in predicting settlement of ground
composed of soft soil to analyze one-dimensional (1D)
consolidation by taking nonlinear behavior of the ground
into account. Since the study on 1D nonlinear consolidation
theory was started in the 1960s, many researchers have
suggested different kinds of 1D nonlinear consolidation
theory. Davis and Raymond [1] developed a nonlinear
consolidation theroy and derived an analytical solution for
a constant loading case, assuming that the decrease in
permeability is proportional to the decrease in
compressibility during the consolidation process and that
the distribution of initial effective stress is constant with
depth. Based on the relationship between the void ratio
and the logarithm of effective stress and permeability (i.e.
e-log and e-log wk ), many scholars have solved the
similar problem using finite difference method [2-4].
Gibson et al. [5, 6] proposed the general theories of 1D
finite nonlinear consolidation of thin and thick
homogeneous layers for a constant loading condition. Xie et
al. [7] developed analytical solution for 1D consolidation of
soft soil subjected to time-dependent loading on the basis
of the nonlinear consolidation theory proposed by Davis
and Raymond. Chen et al. [8] and Zheng et al. [9] carried
out numerical analysis for 1D nonlinear consolidation of
saturated soil by differential quadrature method. Cheng et
al. [10] developed the finite analytic method to simulate 1D
nonlinear consolidation under different time-dependent
loading and initial conditions.
It is worth to note that the nonlinear consolidation of
soil is governed by partial differential equation which is
difficult to obtain analytical solution, except for specifec
conditions, and thus numerical methods are still the most
important means for analyzing the nonlinear consolidation
problem. Recently, unlike conventional numerical methods
based on macroscopic equation, the lattice Boltzmann
method (LBM) which is based on mesoscopic equation has
emerged as an alternative powerful method for solving
fluid dynamics problems and achieved much success in
studying nonlinear equations of complex systems [11,12].
Compared to traditional numerical methods, due to the
advantages such as the simplicity of programming and the
numerical efficiency, the LBM has been widely applied not
only to fluid dynamics but also to many other areas, such as
advection-diffusion problem [13], soil dynamics [14] and
so on. More recently, Kim et al. [15] employed the LBM to
analyze 1D linear consolidation of saturated clay. Previous
studies show that LBM can be used in various engineering
disciplines. Nevertheless, the LBM has hardly ever been
used for nonlinear consolidation analysis of soil. Thus, the
goal of the present study is to extend the LBM into 1D
nonlinear consolidation analysis of saturated soil.
In this paper, a new numerical method for 1D nonlinear
consolidation analysis of saturated soil is proposed on the
basis of the lattice Boltzmann method. The lattice
Bhatnagar-Gross-Krook (LBGK) model is employed for 1D
nonlinear consolidation of saturated soil subjected to time-
dependent loading under various boundary conditions. In
order to recover mesoscopic lattice Boltzmann equation to
macroscopic nonlinear consolidation equation, the
multiscale Chapman-Enskog expansion is applied. As a
result of the numerical simulation for verification, the
numerical results are proved to be in good agreement with
the analytical solutions available in previous literature. The
numerical simulation is performed to investigate the
consolidation behavior of saturated soil subjected to two
different types of time-dependent loading.
This paper is organized as follows. In Section 2, 1D
nonlinear consolidation equation of saturated soil
subjected to time-dependent loading is presented and in
Section 3, the lattice Boltzmann method for 1D nonlinear
consolidation of saturated soil is proposed. In Section 4, the
numerical results are compared with the analytical

solutions available in the literature in order to verify the
proposed method, and also the numerical simulation is
performed to investigate the consolidation behavior of
saturated soil subjected to two different time-dependent
loading. Finally, Section 5 gives conclusions.
2. MATHEMATICAL MODEL
A saturated soil layer of thickness tH ( 2tH H for
double drainage condition; tH H for single drainage
condition) subjected to time-dependent loading is
considered as shown in Fig. 1. Assuming the validity of the
nonlinear consolidation theory proposed by Davis and
Raymond [1], except for the assumption of a constant
loading, the governing equation of 1D nonlinear
consolidation of saturated soil subjected to time-
dependent loading is as follows:
22
2
1 d
d
v
u u u q
c
t z tz 
    
         
(1)
where vu, q, c and σ are excess pore water pressure,
time-dependent loading, the coefficient of consolidatioin
and the effective stress, respectively; t and z are the
variables of time and space respectively.
According to the assumption that the coefficient of
consolidation is constant while the decrease in
permeability is propotional to the decrease in
compressibility,
0
0
=const.w
v
w v
k
c
m
 (2)
where 0wk is the initial coefficient of permeability; w is
the unit weight of water; 0vm is the initial coeffiecient of
compressibility defined as 0 0 00.434 / (1 )v cm C e   in
which cC is the compression index, 0e is the initial void
ratio and 0 is the initial effective stress.
According to Terzaghi’s principle of effective stress,
σ can be expressed as:
0q u     (3)
By defining a new parameter  ,
0
ln
q





 
(4)
Eq. (1) can be simplified to the following form:
2
2
( )vc R t
t z
  
 
 
(5)
where
0
1 d
( )
d
q
R t
q t
 
 
(6)
Fig. 1: A saturated soil layer subjected to time-dependent
loading.
Initial and boundary conditions are considered as
follows:
0
0t
 
 (7)
0
0z
 
 , 2
0z H
 
 (for double drainage condition) (8)
0
0, 0z
z Hz

 


 

(for single drainage condition) (9)
3. LATTICE BOLTZMANN METHOD
3.1 Lattice Boltzmann Equation
In the present study, 1D consolidation analysis is
carried out using lattice Bhatnagar-Gross-Krook (LBGK)
D1Q3 model which is the most popular one in the 1D
lattice Boltzmann models. The distribution function
( , )if z t is first defined on the basis of the general
principles of lattice Boltzmann model as follows:
( , ) ( , ) 0, 1, 2eq
i i
i i
f z t f z t , i     (10)
where if and eq
if are the distribution function and the
equilibrium distribution function along direction i .
Based on the LBGK model, the lattice Boltzmann
equation for Eq. (10) is given by He and Luo [16]:
1
( , ) ( , ) ( , ) ( , )eq
i i i i i if z e t t t f z t f z t f z t tF

         
 
(11)
where ie and t are the discrete velocity and the discrete
time step respectively;  is the dimensionless relaxation
time; iF is the force term calculated by:
i iF w S
where iw is the weight factor with 0 2/ 3w  and
1 2 1/ 6w w  ; S is the rate of loading.
The LBKG model given by Eq. (11) is performed by two
procedures, a collision and a streaming expressed as:
Collision:
1
( , ) ( , ) ( , ) ( , )eq
i i i i if z t t f z t f z t f z t tw S

       
 
(12)

Streaming:
( , ) ( , )i if z z t t f z t t       (13)
In the D1Q3 model, the discrete velocities ( 0,1,2)ie i 
are defined as follows:
0, 0
, 1
, 2
i
i
e c i
c i


 
 
(14)
z
c
t



where c is the lattice speed.
The equilibrium distribution function is given by:
( 0,1,2)eq
i if w , i  (15)
Boundary conditions for Eqs. (8) and (9) is given by:
1( , ) ( , )0, 0i i Nf z t f z t  (for double drainage) (16)
1 1( , ) ( , ) ( ,0, )i i N i Nf z t f z f z tt  (for single drainage) (17)
where 1( , )if z t , ( , )i Nf z t and 1( , )i Nf z t are the
distribution functions for the first lattice node 1z , the last
lattice node Nz and the N-1th lattice node 1Nz  ,
respectively.
3.2 Recovery of 1D nonlinear consolidation
equation
The multiscale Chapman-Enskog expansion is used to
recover the macroscopic 1D nonlinear consolidation
equation of saturated soil. The 1D nonlinear consolidation
equation can be scaled spatially as, z is set to 1 /z  , t is set
to 2
1 /t  where  is a small parameter.
For D1Q3 model,  can be expressed by Eqs. (10) and
(15) as follows:
0 1 2( , ) ( , ) ( , ) ( , )i
i
f z t f z t f z t f z t     (18)
0 1 2( , ) ( , ) ( , ) ( , )eq
i
i
f z t w z t w z t w z t       (19)
The distribution function and the source term can be
expressed in terms of a small parameter  as:
(1) 2 (2)eq
i i i if f f f    (20)
2 (2)
i iF F (21)
where ( )k
if and (2)
iF are the non-equilibrium distribution
functions and non-equilibrium force term defined by:
(2) (2)
i
i
F F (22)
Since
(1) 2 (2)eq
i i i i
i i
f f f f      
   (23)
and eq
i
i
f   , hence other expanded term in the above
equation should be zero, i.e.,
( )
0, ( 1)k
i
i
f k  (24)
Applying the Taylor series, the updated distribution
function is expanded:
 
2 2 22
3
2 2
( , ) ( , )
2
2
i i
i i i i
i i i
i i i
f f
f z e t t t f z t e t t
z t
f f ft
e e e t
z tz t
 
        
 
   
          
(25)
Introducing scaling for the above equation, i.e., 1 z  is
replaced by 1z   , 1 t  is replaced by 2
1/ t  :
 
2
1 1
2 2 3 2 4 22
3
2 2
1 11 1
( , ) ( , )
2
2
i i
i i i i
i i i
i i i
f f
f z e t t t f z t e t t
z t
f f ft
+ e e e t
z tz t
 
  
 
        
 
   
         
(26)
Substituting Eqs. (20) and (26) into Eq. (11) and
retaining terms up to order of 2
 , yields:
   
(1)
(1) 2 (2) 2 (2) 2
1 1
2 2
2 3 2
2
1 1
1
2
eq
i i
i i i i i
eq eq
i i
i i
f f
f f F e e
t z z
f ft
e e t
t z
    


 
 
     
   
 
      
 
(27)
Comparing the two sides of Eq. (27) and treating terms
in order of  and 2
 , yields:
Terms order of  :
(1)
1
1 eq
i
i i
f
f e
t z

 
 
(28)
Terms order of 2
 :
(1) 2
(2) (2)
2
1 1 1
1
2
eq eq
i i i
i i i i i
f f ft
f F e e e
t z t z
  
    
   
(29)
Applying Eq. (28) to the right side of Eq. (29), yields:
2
(2) (2)
2
1 1
1
( )
2
eq eq
i i
i i i i
f ft
f F t e e
t t z


 
     
  
(30)
For recovering the 1D nonlinear consolidation
equation Eq. (5), Eq. (30) is summed over all states, i.e.,
(2) (2)
1
2
2
1
1
( )
2
eq
i
i i
i i i
eq
i
i i
i
f
f F
t t
ft
t e e
z


 
     
 
      
  

(31)
According to Eqs. (22) and (24), (2) (2)
i
i
F F and
(2)1
0i
i
f
t
 
   
 . Then, both terms of the right side of
the above equation are written as, respectively:
2
1 1 1
1
eq
ieq
i i
i
f
f
t t t t
 


  
  
   

 (32)

 
2 2 2 2
2 2 2 2
1 1
eq
eqi
i i i i i
i i
f c
e e f e e
z z z


   
      
  (33)
Meanwhile, the source term F is taken as:
( )i i
i i
F F w S R t    (34)
Hence, Eq. (31) can be simplified and the 1D nonlinear
consolidation equation can be recovered as:
2
2
2
1
( ) ( )
2
c t R t
t z
 

 
   
 
(35)
By comparing Eq. (5) with Eq. (35), the relationship
between the coefficient of consolidation vc and the
dimensionless relaxation time  can be obtained.
2 1
( )
2
vc c t    (36)
4. RESULTS AND DISCUSSION
4.1 Verification
In order to verify the proposed LBGK model for 1D
nonlinear consolidation analysis, the numerical simulation
is performed for two cases in terms of instantaneous and
ramp loading. And the numerical results are compared
with the analytical solutions available in previous
literature. For the numerical simulation, the domain is
discretized into 100 nodes; the dimensionless relaxation
time, the discrete lattice spacing and the time step are set
as 1.
4.1.1 Verification under instantaneous loading
In order to verify the proposed LBGK model for 1D
nonlinear consolidation analysis under instantaneous
loading, analytical solutions derived by Davis and
Raymond [1] is used for reference. Figs. 2 and 3 show the
excess pore water pressure isochrones and the average
degree of consolidation with time factor defined by
2
/v vT c t H for double drainage condition. It can be seen
that the numerical results are in good agreement with
analytical solution.
4.1.2 Verification under time-dependent loading
For time-dependent loading, analytical solutions for
ramp loading given by Xie et al. [7] is selected for
reference. The ramp loading as shown in Fig. 4 can be
expressed as follows:
,
( )
,
u
c
c
u c
q
t t t
tq t
q t t

 
 
 
(37)
where uq is the ultimate loading; ct is the time of
application of any load (i.e. construction time).
Fig. 2: The excess pore water pressure isochrones.
Fig. 3: The average degree of consolidation with time
factor.
Fig. 4: Ramp loading
Fig. 5 presents the normalized excess pore water
pressure with time factor vT at half the depth for double
drainage condition and different construction time factors.
The construction time factor is defined by 2
/vc v cT c t H .

It is found that two solutions are in excellent agreement.
Fig. 6 shows the average degree of consolidation under the
different construction time factors. It is also shown that
the LBM results are agree well with the analytical
solutions at all construction time factors.
Fig. 5: Variation of the excess pore water pressure under
different construction time factors.
Fig. 6: The average degree of consolidation under different
construction time factors
4.2 Example and discussion
Two different types of time-dependent loadings shown
in Fig. 7, i.e., exponential and haversine cyclic loadings are
considered and the loading functions are expressed as
follows.
The exponential loading is expressed as
( ) (1 )bt
uq t q e
  (38)
where b is the loading parameter controlling the rate of
exponential loading.
The haversine cyclic loading is expressed as
2
0
( ) sinu
t
q t q
t

 (39)
where 0t is the loading parameter which is period of
haversine cyclic loading.
(a): Exponential loading
(b): Haversine cyclic loading
Fig. 7: Example loading type.
Furthermore, based on the proposed method, the
numerical simulation is carried out to investigate the
influence of the ratio of ultimate loading intensity to initial
effective stress and the loading parameters on 1D
nonlinear consolidation of saturated soil under double
drainage condition.
4.2.1 Consolidation under different values of the ratio
of ultimate loading intensity to initial effective stress
Fig. 8 shows the variation of average degree of
consolidation with time factor under different ratio of
ultimate loading intensity to initial effective stress for the
exponential loading with the dimensionless loading
parameter 1.04 1b E , which is defined by 2
/ vb H b c .
Figs. 8(a) and 8(b) present the changes of average degree
of consolidation defined by settlement and effective stress
with time factor, respectively. It can be seen that the
average degree of consolidation Us defined by settlement
increases with the increase of the ratio 0/uq  , but the
average degree of consolidation Up defined by effective
stress decreases with the increase of the ratio 0/uq  .
Fig. 9 shows the results at the different ratio 0/uq 
under the haversine cyclic loading with the dimensionless
loading parameter 0 0.05T  , which is defined by
2
0 0 /vT c t H . Similar to the results from the exponential

loading, it can be found that the average degree of
consolidation Us increases with the increase of the ratio
0/uq  , but the average degree of consolidation Up
decreases.
(a): Us-Tv curves
(b): Up-Tv curves
Fig. 8: Variation of average degree of consolidation with
time factor under different ratio 0/uq  for exponential
loading.
4.2.2. Consolidation under different loading
parameters
Fig. 10 shows the variation of average degree of
consolidation with time factor at different dimensionless
loading parameter b under exponential loading with the
ratio 0/ 1.5uq = . It can be found that the dimensionless
loading parameter b has significant effects on the rate of
settlement and the rate of dissipation of excess pore water
pressure, and a bigger dimensionless loading parameter
b leads to a faster rate of settlement and dissipation of
excess pore water pressure. Moreover, both the settlement
and the dissipation of excess pore water pressure tend to
proceed more quickly at the early stage of consolidation as
the demensionless loading parameter b increases.
Fig. 11 shows the results at different dimensionless
loading parameter 0T under the harversine cyclic loading
with the ratio 0/ 1.5uq = . It can be seen that a smaller
value of the parameter 0T induces more cycles of
oscillation, while a bigger value of the parameter 0T
results in a larger oscillation in the settlement and the
dissipation of excess pore water pressure.
(a): Us-Tv curves
(b): Up-Tv curves
Fig. 9. Variation of average degree of consolidation with
time factor under different ratio 0/uq  for haversine
cyclic loading.
5. CONCLUSION
In this paper, a new numerical method for 1D nonlinear
consolidation analysis of saturated soil is proposed on the

basis of the lattice Boltzmann method. The lattice
Bhatnagar-Gross-Krook (LBGK) model is employed for 1D
nonlinear consolidation analysis of saturated soil
subjected to time-dependent loading under different types
of boundary conditions. In addition, the multiscale
Chapman-Enskog expansion is applied to recover
mesoscopic lattice Boltzmann equation to macroscopic
consolidation equation. As a result of the numerical
simulation for verification, the numerical results are
proved to be in good agreement with the analytical
solutions available in previous literature.
The consolidation behavior of saturated soil subjected
to exponential loading and haversine cyclic loading is
investigated through the numerical simulation. The
following conclusions can be drawn:
(a): Us-Tv curves
(b): Up-Tv curves
time factor under different loading parameter for
exponential loading.
(1) As the ratio 0/uq  increases, the rate of settlement
increases but the rate of dissipation of excess pore
water pressure decreases.
(2) A bigger dimensionless loading parameter b induces
a faster rate of the settlement and the dissipation of
excess pore water pressure. Moreover, both the
settlement and the dissipation of excess pore water
pressure tend to proceed more quickly at the early
stage of consolidation as the demensionless loading
parameter b increases.
(3) A smaller value of the dimensionless loading
parameter 0T induces more cycles of oscillation, while
a bigger value of the parameter 0T results in a larger
oscillation in the settlement and the dissipation of
excess pore water pressure.
(a): Us-Tv curves
(b): Up-Tv curves
time factor under different loading parameter for
haversine cyclic loading.

ACKNOWLEDGEMENT
I would like to take the opportunity to express my
hearted gratitude to all those who make a contribution
to the completion of my paper.
REFERENCES
[1] E.H. Davis and G.P. Raymond, A non-linear theory of
consolidation, Geotechnique 15(2), 1965, pp. 161-173.
[2] L. Barden and P.L. Berry, Consolidation of normally
consolidated clay, Journal of the Soil Mechanics and
Foundation Division, ASCE 91 (SM5), 1965, pp. 5–35.
[3] T.I. Poskitt, The numerical solution of non-linear
consolidation problems, International Journal for
Numerical Methods in Engineering, Vol. 3, 1971, pp. 5-
11.
[4] G. Mesri and A. Rokhsar, Theory of consolidation for
clays. ASCE 100(GT8), 1974, pp. 889–903.
[5] R.E. Gibson, G.L. England and M.J.L. Hussey, The theory
of one-dimensional consolidation of saturated clays. I.
Finite non-linear consolidation of thin homogeneous
layers [J]. Geotechnique 17(2), 1967, pp. 261-273.
[6] R.E. Gibson, R.L. Schiffman and K.W. Cargill, The
theory of onedimensional soil consolidation of
saturated clays: II. Finite nonlinear consolidation of
thick homogeneous layers, Can. Geotech J. 18, 1981,
pp. 280–293.
[7] K.H. Xie, B.H. Li and Q.L. Li, A nonlinear theory of
consolidation under time-dependent loading, In:
Proceedings of the Second International Conference
on Soft Soil Engineering, Nanjing, China, 27–30 May
1996, pp. 193–196.
[8] R.P. Chen et al., One-dimensional nonlinear
consolidation of multi-layered soil by differential
quadrature method, Computers and Geotechnics 32
(2005), pp. 358–369.
[9] G.Y. Zheng et al., Analysis of non-linear consolidation
of soft clay by differential quadrature method, Applied
Clay Science 79(2013), pp. 2-7.
[10] D. Cheng et al., Finite Analytic Method for One-
Dimensional Nonlinear Consolidation under Time-
Dependent Loading. Shock and Vibration, Vol. 2017,
2017, pp. 12, Article ID 4071268.
[11] R. Benzi, S. Succi, M. Vergassola, The Lattice
Boltzmann Equation: Theory and Application, Physics
Reports, Vol. 222, No. 3, 1992, pp. 145-197.
[12] B.C. Shi, Z.L. Guo, Lattice Boltzmann Model for
Nonlinear Convection-Diffusion Equations. Physics
Review E, Vol. 79, 2009, 016701, 1-13,
DOI:https://doi.org/10.1103/PhysRevE.79.016701.
[13] Z.L. Guo, B.C. Shi, N.C. Wang, Fully Lagrangian and
Lattice Boltzmann Method for the Advection-Diffusion
Equation. Journal of Scientific Computing Vol. 14, No.
3, 1999, pp. 291-300.
[14] U.E. Shamy, Y. Abdelhamid, Modeling granular soils
liquefaction using coupled lattice Boltzmann method
and discrete element method. Soil Dynamics and
Earthquake Engineering 67 (2014), 2014, pp. 119–
132.
[15] P. Kim et al., Lattice Boltzmann method for
consolidation analysis of saturated clay, Journal of
Ocean Engineering and Science, 2019, DOI:
10.1016/j.joes.2019.04.004.
[16] X. He and L.S. Luo, Lattice Boltzmann model for the
incompressible Navier– Stokes equation, J. Stat. Phys.
88, 1997, pp. 927–944.

IRJET- Numerical Analysis for Nonlinear Consolidation of Saturated Soil using Lattice Boltzmann Method

More Related Content

What's hot (18)

Similar to IRJET- Numerical Analysis for Nonlinear Consolidation of Saturated Soil using Lattice Boltzmann Method (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Numerical Analysis for Nonlinear Consolidation of Saturated Soil using Lattice Boltzmann Method