International Communications in Heat and Mass Transfer 85 (2017) 53–61

Contents lists available at ScienceDirect

International Communications in Heat and Mass Transfer

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

Relative permeability of two immiscible fluids flowing through porous MARK

media determined by lattice Boltzmann method
Huawei Zhaoa,b,⁎, Zhengfu Ninga,b, Qinjun Kangc, Li Chend, Tianyi Zhaoa,b
a
State Key Laboratory of Petroleum Resources and Engineering in China University of Petroleum, Beijing 102249, PR China
b
Ministry of Education Key Laboratory of Petroleum Engineering in China University of Petroleum, Beijing 102249, PR China
c
Computational Earth Science Group (EES-16), Los Alamos National Laboratory, Los Alamos, NM 87545, USA
d
Key Laboratory of Thermo-fluid Science and Engineering of MOE, School of Energy and Power Engineering, Xi'an Jiaoting University, Xi'an, Shaanxi 710049, PR China

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

Keywords: This article applied the multiple relaxation time multicomponent/multiphase pseudopotential lattice Boltzmann
Lattice Boltzmann method model to simulate two immiscible fluids flow in 2D porous media, and analyzed the effects of capillary number
Two phase flow (Ca), viscosity ratio (M) and wettability on the relative permeability curves. Simulation results indicate that the
Relative permeability nonwetting phase (NWP) relative permeability increases with increasing Ca; while the effect of Ca on the wetting
phase (WP) relative permeability depends on the wettability. When M > 1, the NWP relative permeability
increase with increasing M in a strong wetting condition because of the lubricating effect. The amplitude of the
NWP relative permeability may even exceed the single phase permeability. However, the exact value of the
amplitude and where it occurs depends on M and the structure of the porous media. The WP relative
permeability is insensitive to M. When the porous media converts from strong wetting condition to neutral
wetting condition, the NWP relative permeability decreases while the WP relative permeability increases.

1. Introduction
Two immiscible fluids flowing through porous media is a common
environmental phenomenon and is of significant importance for many
industrial problems, such as enhanced oil recovery, geological carbon
dioxide sequestration, and fuel cells [1–3]. The relative permeability is
the key descriptor of the flows of the two phase fluids. During the early
years, the flows of the two phase fluids were believed to be uncoupled,
and a simple extension of the single phase Darcy's law was made to
obtain the relative permeability curves. However, researchers were
soon afterwards aware that viscous coupling effect, which was a result
of the monument transfer between the two fluids, played a vital role
during the two immiscible fluids flow [4–6]. And it is now widely
accepted that the relative permeability curves, rather than a simple
result of saturation, are functions of many parameters including
capillary number, viscosity ratio, and wettability.
Historically, laboratory experimental methods were first and widely
used to determine the relative permeability curves of the two phase
flow. For example, Dullien and Dong measured the relative perme-
ability curves of sand packs by two sets of co-current steady state
experiments [7]. External force was applied to one phase each time, and
the relative permeability curves were calculated based on the recorded
oil and water velocity. Bentsen and Manai estimated the permeability
coefficients by a combination of co-current and countercurrent flow
experiments [8]. Until now, experimental methods are still general
choices for the study of two phase flow [9,10]. However, the relative
permeability experiment is a time consuming process; the experiment
sometimes fails to reflect the real subsurface conditions; and the
heterogeneity between experimental porous media may affect the
analysis of the flow controlling factors. In recent years, benefited by
the evolution of the computational capacities, some numerical simula-
tion methods emerged, among which the pore network modeling has
great potential applications [11]. Zhao et al. used pore network
modeling to study the oil recovery by water flooding in sandstones
and carbonates, and discussed the effect of initial water saturation,
contact angle distribution and oil wet fraction [12]. Gharbi and Blunt
used the pore network modeling method to study the relative perme-
ability curves of carbonate samples, and discussed the impact of
wettability and connectivity on the relative permeability curves [13].
However, the pore network method is based on displacement and
transport equations, which fail to unveil the underlying microscopic
dynamics. Besides, the accuracy of the modeling is highly related to the
extracted network.
The lattice Boltzmann (LB) method emerged in the late 1980s, and

Abbreviations: LB, lattice Boltzmann;; MRT, multiple relaxation time; MCMP, multicomponent multiphase; WP, wetting phase; NWP, nonwetting phase
⁎
Corresponding author at: School of Petroleum Engineering, China University of Petroleum, Beijing 102249, PR China.
E-mail address: h.w.zhao2013@gmail.com (H. Zhao).

http://dx.doi.org/10.1016/j.icheatmasstransfer.2017.04.020

0735-1933/ © 2017 Elsevier Ltd. All rights reserved.

H. Zhao et al. International Communications in Heat and Mass Transfer 85 (2017) 53–61

Nomenclature Ffσ fluid-fluid interaction force

Fadsσ fluid-solid interaction force
fσ density distribution function Fbσ external body forces
fσ(eq) equilibrium density distribution function gσσ fluid-fluid interaction strength
e discrete velocities gσw fluid-solid interaction strength
ueq equilibrium velocity Kr , nw nonwetting phase relative permeability
ρσ density Kr , w wetting phase relative permeability
M transformation matrix Ca capillary number
S diagonal relaxation matrix M viscosity ratio
υσ viscosity θ contact angle
Fσ total force

has become an effective numerical tool to simulate and investigate a ⎛ ⎞

e = ⎜ 0 1 0 ‐1 0 1 ‐1 ‐1 1 ⎟
broad class of flows, including the two phase flow [14–16]. By now, ⎝ 0 0 1 0 ‐1 1 1 ‐1 ‐1⎠ (2)
there are four types of LB multiphase models, which are the color model
And the equilibrium distribution function fiσ(eq)(x, t) is written as
proposed by Rothman and Keller (also called R-K model) [17,18], the
pesudopotential model proposed by Shan and Chen (Shan-Chen model) ⎡ 3 9 3 ⎤
fiσ (eq) = ρσ ωi ⎢1 + 2 (ei⋅ueq
σ ) + (ei⋅ueq 2
σ ) − ueq
σ ⎥
2
[19,20], the free energy model introduced by Swift et al. [21], and the ⎣ c 2c 4 2c 2 ⎦ (3)
kinetic theory based model proposed by He et al. (He-Shan-Doolen
In which, ωi is the weight factor with ω0 = 4/9, ω1 − 4 = 1/9 and
model) [22]. There have been some studies investigating the multi-
ω5 − 8 = 1/36.
phase flow using different multiphase models based on different pore
In the MCMP model, the equilibrium velocity is calculated as
structures. Langaas and Papatzacos simulated concurrent and counter-
current flows in a uniform pore space geometry using the free energy ueq = ∑ sρσ ρσ u σ ∑ sρσ ρσ
model and studied the effects of capillary pressure, viscosity ratio under σ σ (4)
different wettability [23]. Kang et al. simulated displacement of a two- Where ρσ and uσ are the density and velocity of component σ, and
dimensional (2D) immiscible droplet in a channel using the pseudopo- are calculated as
tential model, and discussed fingering controlling factors [24,25]. Li
δt σ
et al. used a multiple relaxation time (MRT) pseudopotential model to ρσ = ∑ fiσ , ρσ u σ = ∑ ei fiσ + F
simulate two phase flow in a 3D porous medium and discussed the i i
2 (5)
impacts of capillary number, viscosity ratio, wettability and fluid-fluid M is a transformation matrix, and for the detailed value of M one is
interfacial area on the relative permeability curves [26]. Huang and Lu referred to [30]. S is a diagonal relaxation matrix, and is expressed as
simulated co-current and counter-current flow in a simplified 2D porous
medium using one component/two phase pseudopotential model [27]. Sσ = diag[sρσ , seσ , sεσ , sjσ , sqσ , sjσ , sqσ , sυσ , sυσ ] (6)
Dou and Zhou discussed the relative permeability affecting factors in σ
Where sρ corresponds to conserved mass, and sj corresponds to σ
both homogenous and heterogeneous 2D porous media [28]. All these conserved moments. They are both taken 1. seσ, sεσ and sqσ correspond to
studied have deepened the understanding of viscous coupling in two non-conserved moments and can be adjusted independently to improve
phase immiscible flows. the accuracy and stability of the MRT model. In this study, following
In this study, lattice Boltzmann simulation of two immiscible fluids [31], the three parameters are taken as: seσ = 0.6, sεσ = 1.54, and
flowing through 2D porous medium has been conducted. The applied sqσ = 1.2. sυσ is the dimensionless relaxation time, and is related to
lattice Boltzmann is the multicomponent multiphase (MCMP) pseudo- viscosity υσ as
potential model with a MRT collision operator. The 2D porous medium
structure was extracted from a computer tomography image of a tight ⎛1 1⎞
υσ = cs2 ⎜ σ − ⎟ δt
sandstone sample. After validating the model by three benchmarks (i.e., ⎝ sυ 2⎠ (7)
Laplace law, contact angle, and flow in 2D channel), the two immiscible
fluids flowing through the porous media were simulated. And the
effects of the capillary number (Ca), viscosity ratio (M), and wettability
Where cs2 is the lattice sound speed (
cs2 =
c2
3 ).
The force term F iσ in Eq. (1) follows Guo's force scheme, and is
(θ) on the relative permeability curves were discussed. defined as [30,32]
F σ ⋅(e − ueq) σ (eq)
F iσ = fi
2. Methodology ρσ cs2 (8)
σ
The MCMP pseudopotential model was used for the simulations in Where F is the total force acting on component σ, and is composed
this study. And a MRT collision operator was added to increase the of three parts
stability of the simulations [29]. The standard LB equation using the F σ = F σf + Fads
σ
+ F σb (9)
MRT collision operator with a force term is expressed as σ σ
Where Ff is the fluid-fluid interaction force, Fads is the fluid-solid
f σ (x + c eδt , t + δt ) − f σ (x, t ) = −(M−1Sσ M) f σ ( x, t )
− f σ (eq) (x, t ) interaction force, and Fbσ are other possible external forces such as
( )
⎡ −1⎛ gravitational force. For the MCMP model, the fluid-fluid interaction
Sσ ⎞ ⎤ σ
+ ⎢M ⎜I − ⎟ M⎥ F (x, t ) force between components σ and σ is defined as
⎣ ⎝ 2⎠ ⎦ (1)
N
σ
Where f (x, t) is the density distribution function of the component σ F σf (x) = −gσσ ψσ (x) ∑ w ( e i 2 ) ψ σ (x + e i ) e i
i =1 (10)
at position x and time t. c = δx/δtis the lattice speed with δx and δt being
the lattice spacing and time step respectively. e are the discrete Where gσσ represents the interaction strength between different
velocities, and fσ(eq)(x, t) is the equilibrium density distribution func- components. In this study, we set gσσ and gσσ as zero, and take
tion. The discrete velocities for a D2Q9 model are given by gσσ = gσσ = 0.65 to separate the two phase and maintain a moderate

54
H. Zhao et al. International Communications in Heat and Mass Transfer 85 (2017) 53–61

transition zone (for detailed process to determine g, one is referred to The simulations were conducted in a 101 × 101 lu2 system with the
[33]). ψ(x) is an effect mass depending on the local density, and is taken upper and lower solid walls using half-way bounce back boundaries and
as ψ(x) = 1 − exp(−ρ(x)). w(|ei |2) are the weight coefficients used for the inlet and outlet using periodic boundaries. A semicircular droplet of
calculation of the isotropic interaction force. For the 4th order isotropic one component was put in the middle of the lower wall inside another
tensor with nearest and next-nearest ones, the weight coefficients are component and no external force was added to the system. When the
taken as w(0) = 0, w(1) = 1 and w(2) = 1/4. steady state was reached, the base length b and the height h of the
Similar to the fluid-fluid interaction force, the fluid-solid interaction droplet were measured. The radius r of the droplet can be calculated as
force is defined as [34]
N
σ
4h2 + b 2
Fads (x) = −gσw ψσ (x) ∑ w (|ei |2 ) ψ (ρw ) s (x + ei ) ei r=
8h (15)
i =1 (11)
Where gσw is the interaction strength between the fluid component σ And the static contact angle can be calculated by
and the solid phase. For a two-component system, usually gσw = −gσw . ⎧ arcsin b ,
⎪ ( 2r ) θ ≤ 90°
ρw is a fictitious density of the solid phase, and gσw and ρw can be θ=⎨
⎪ π − arcsin ( b ) , θ > 90°
adjusted separately or jointly to obtain different contact angles. A ⎩ 2r (16)
negative or positive value of gσw leads to < 90°(wetting) or >
90°(nonwetting) contact angle, and gσw = 0 results in a contact angle We simulated five scenarios with the gσw being − 0.35, − 0.2, 0,
of 90°.s is an indicator function with 1 representing solid phase and 0 0.2,and 0.35. The corresponding contact angles (θ) are 13.26°, 54.10°,
representing fluid phase. 90.0°, 126.87° and 168.57°, respectively (Fig. 2). In this way, it is
Finally, the external body forces can be simply applied as demonstrated that different contact angles can be obtained by changing
gσw.
F σb = ρσ g (12)
Where g is the body force per unit mass. 3.1.3. Relative permeability in a 2D channel
The last benchmark is the two phase flow in an infinite 2D
3. Results and discussion horizontal channel. The wetting phase flows along the walls
(a ≤ |y | ≤ b) and nonwetting phase (0 ≤ | y | ≤ a) flows between the
3.1. Model validation wetting phase. Assuming a Poiseuille type flow in the x direction, the
analytical solution for the relative permeability as a function of wetting
Three tests including Laplace law, contact angle, and 2D channel phase saturation Sw and viscosity ratio M is expressed as [35]
flow were conducted to validate the effectiveness of the lattice ⎡3 ⎛ 3 ⎞⎤ 1 2
Boltzmann model. Kr,nw = (1 − Sw ) ⎢ M + (1 − Sw )2 ⎜1 − M ⎟ ⎥ , Kr,w = Sw (3 − Sw )
⎣2 ⎝ 2 ⎠⎦ 2
3.1.1. Laplace law (17)
The Laplace law states that the pressure difference pi − po inside and Where Kr , nw is the nonwetting phase relative permeability, and Kr , w
outside a static droplet is inversely proportional to the radius r of the is the wetting phase permeability; Sw is the wetting phase saturation,
droplet, and the proportion is surface tension coefficient σ. The which is defined as Sw = a/b.
relationship is expressed as We simulated the two phase flow in a 11 × 201 lu2 system with the
σ upper and lower solid walls using half-way bounce back boundaries and
pi − po =
r (13) the inlet and outlet using periodic boundaries. Following the method
proposed in [36], we applied the same body force (Fb = 2 × 10− 5) to
The simulations were conducted in a 101 × 101 lattice unites (lu2)
the two phase simultaneously. The mathematical formulas of the
periodic boundary system. A circular static droplet of one component
relative permeabilities are written as
was initially embedded in the center of the other component, and no
external force was added to the system. The two fluids had the same
density, and the sυσ of the fluids were changed to simulate different
viscosity ratios M = υnw/υw. The pressure was calculated as
cs2
p= ∑ ρσ cs2 + ∑ gσσ ψσ ψσ (14)
2
The maximal iteration step was 50,000, which was large enough for
the system to arrive the steady state. The pressure difference was
calculated following the method in [2]. The radius of the droplet was
counted with the two-component boundary taken where ρ = (ρ1 + ρ2)/
2. A series of simulations were conducted by changing the initial radius
of the droplet, and several cases with different M were also simulated.
Fig. 1 plots pi − po as a function of 1/r with different M. The dash line is
the least-square linear fit of the simulation data when M = 1. It is clear
that all points fit the straight line through the origin quite well,
meaning the Laplace law is excellently satisfied. The slope of the
straight line (i.e., surface tension coefficient) is 0.1816. It is also
observed in Fig. 1 that the simulation data of other viscosity ratios
almost overlap with that of M = 1, indicating an identical surface
tension.

3.1.2. Contact angle

We set ρw = 1 and adjusted gσw to achieve different contact angles. Fig. 1. Relationship between pressure difference and the reciprocal of the bubble radius.

55
H. Zhao et al. International Communications in Heat and Mass Transfer 85 (2017) 53–61

Fig. 2. Simulation results of different contact angles. (a) θ = 13.26°. (b) θ = 54.10°. (c) θ = 90.0°. (d) θ = 126.87°. (e) θ = 168.57°.

Fig. 3. Relative permeability curves of two phase flow in an infinite 2D channel. (a) M = 1; (b) M = 10; (c) M = 0.1.

56
H. Zhao et al.
a L
∫ y =0 u nw dy ∫y = a u w dy
Kr,nw (Sw ) = L
, Kr,w (Sw ) = L
∫y =0 u nw dy ∫y =0 u w dy
Three cases with different viscosity ratios are investigated: (a)
M = 1; (b) M = 10; and (c) M = 0.1. The steady state was considered
to be arrived when the following criterion was achieved
∑y [u y (y, t ) − u y (y, t − 500)]2
< 10−5
∑y u y (y, t )2
The comparison between the simulation results and analytical
solution are plotted in Fig. 3. The simulation results are in good
agreement with the analytical solutions. It is noticed in Fig. 3b that the
relative permeability of the nonwetting phase is larger than 1. This is
because of the so called lubricating effect [26], which will be discussed
in details in Section 3.2.2.
3.2. Relative permeability of 2D porous media
After the model validation, we simulated two phase flow in a 2D
porous medium. The 2D porous medium was generated from a slice of
computer tomography images of a tight sandstone sample. After the
noise removing and segmentation processes, the original gray-scale
image was converted to binary image. We added/broadened some
throats artificially to improve the connectivity of the porous medium.
Besides, we added ten columns of void nodes to both inlet and outlet to
make sure a periodic boundary of the porous medium. And the negative
relative permeability of the wetting phase reported by Landry et al. was
successfully avoided [37]. The size of the porous media is
3600 × 1920 lu2, and the porosity is 0.330. The solid nodes are marked
in gray, the wetting phase (WP) is in blue, and the nonwetting phase
(NWP) is in red (Fig. 5).
Similar to the model validation in Section 3.1.3, inlet and outlet are
periodic boundaries, upper and lower are solid walls using half-way
bounce back boundaries, and external force is applied to drive the
fluids. To realize a specific saturation, a random number was generated
for each void node. The node was defined as red phase if the random
number was greater than the specific saturation, otherwise the node
was defined as blue phase. The convergence criterion was similar to Eq.
(19). The only difference was that the volumetric flow rate rather than
the flow rate in y-direction were used for calculation. We simulated
cases with different capillary number (Ca), viscosity ratio (M), and
wettability (θ). And the effects of Ca, M, and θ on the relative
permeability curves are discussed in the following sections.
Fig. 4. Effect of capillary number on the relative permeability curves. (a) Neutral wetting, θ = 90∘; (b) strong wetting, θ = 168.57∘.
International Communications in Heat and Mass Transfer 85 (2017) 53–61
3.2.1. Effect of capillary number
The capillary number in this study was defined as Ca = Fb/σ. The
(18) interfacial tension was obtained from Section 3.1.1, and was indepen-
dent from the viscosity ratio because of the MRT collision operator. To
show the dependence of the relative permeability on the capillary
number, two scenarios with different external forces (i.e.,
Fb = 2 × 10− 5 and Fb = 4 × 10− 5) were simulated, and the correspond-
ing capillary numbers were Ca = 1.1 × 10‐4 and Ca = 2.2 × 10‐4, re-
spectively. The viscosity ratio was set as 1. For wettability, two cases
with neutral wetting (θ = 90∘) and strong wet (θ = 168.57°) conditions
(19) were considered, respectively.
The relative permeability curves of different capillary numbers for
the neutral wetting case are plotted in Fig. 4a. As expected, the NWP
and WP relative permeability curves are symmetrical, and the equal
permeability occurs where the wetting phase saturation is 0.5. These
are good indications that the system is strictly neutral wetting. Because
of the neutral wetting porous media, the two phase distributions are
quite arbitrary (Fig. 5, left column). When the Ca increases from
1.1 × 10− 4 to 2.2 × 10− 4, both the NWP and WP permeability
coefficients increase for all the saturation range. Still, the NWP and
WP relative permeability curves are symmetrical. Our simulation
results agree with Li et al. that both the NWP and WP relative
permeability were increasing function of the capillary number in the
neutral porous media [26]. However, by comparing our Fig. 4a with
Fig. 4 in Li et al., one may find that the relative permeability coefficient
of the WP in out simulation is more sensitive to the capillary number
than that of the WP was in Li's study. It is worth mentioning here that
the contact angle in the simulation of Li et al. is ~ 65°, which in fact
indicated a weak wetting condition. We continue to discuss the effect of
capillary number in a wetting porous media in the following paragraph.
The relative permeability curves in a strong porous media are
plotted in Fig. 4b. The relative permeability of the NWP increases
significantly with the increasing capillary number, while that of the WP
changes little with the increasing capillary number. This result is in
consistent with that of Huang and Lu [27]. In Fig. 5 (right column),
When the WP saturation is 0.3, it is clear that the NWP occupies the
middle of main stream path, while the WP flows along the surface of the
solid walls. When the saturation is 0.5 or 0.7, the NWP still takes up the
main stream path, only part of the NWP remains in dead-end pores.
Meanwhile, the WP occupies some of the flow path. By comparing two
cases with different wettability, a conclusion can be drawn that the
NWP relative permeability increases with increasing capillary number;
the effect of capillary number on WP relative permeability depends on
the wettability. The WP permeability increases with increasing capil-
lary pressure for the neutral wetting condition. But the increment
becomes small with the increasing wettability, until the relative
57
H. Zhao et al. International Communications in Heat and Mass Transfer 85 (2017) 53–61

Fig. 5. Two phase distribution at different saturations (Sw = 0.3 , 0.5 , 0.7, Ca = 1.1 × 10− 4). Left column is the neutral wetting cases (θ = 90∘); right column is strong wetting cases
(θ = 168.57∘).

Fig. 6. Effect of viscosity ratio on the relative permeability curves (Ca = 1.1 × 10− 4, Fig. 7. Effect of wettability on the relative permeability curves (Ca = 1.1 × 10− 4, M = 1).
θ = 168.57∘).

58
H. Zhao et al.
Fig. 8. Effect of wettability on the relative permeability curves in heterogeneous
wettability porous medias (Ca = 1.1 × 10− 4, M = 1).
permeability is insensitive to the capillary number.
Fig. 9. Two phase distributions of heterogeneous porous medias at different saturations (Sw = 0.3 , 0.5 , 0.7, Ca = 1.1 × 10− 4, M = 1). The gray is strong wetting matrix (θ = 168.57∘) and
the black is neutral wetting matrix (θ = 90∘).
59
International Communications in Heat and Mass Transfer 85 (2017) 53–61
3.2.2. Effect of viscosity ratio
The capillary number was taken as Ca = 1.1 × 10‐4, which was high
enough to make sure the flow was controlled by the viscous force. The
contact angle was taken as 168.57°, indicating a strong wetting porous
medium. Two scenarios with M = 1 and M = 10 were simulated. As
shown in Fig. 6, the relative permeability of the NWP increases
significantly when the viscosity ratio increases from 1 to 10. And in
certain range of saturation (~ 0.3 in this study), the NWP relative
permeability is larger than 1. This phenomenon has been observed by
both experimental and numerical studies, and is called the lubricating
effect [38]. Meanwhile, the WP relative permeability is less affected by
the change in viscosity ratio. There is a slight decrease in the WP
relative permeability. This may because the wetting phase is more
attached to the surface of the porous media when the viscosity ratio is
large.
The maximal value of the NWP relative permeability for the M = 10
case in our simulation is > 1, while that by some other researchers
is < 1 [26]. We attribute this high value to the large viscosity ratio we
took in our simulation, as it is believed that the greater the viscosity
ratio is, the greater the lubricating effect becomes [27]. The lubricating
effect is considered to be most significant at the intermediate saturation
range [27,35,39]. In our simulation, it occurs where the saturation is
~0.3. The lubricating effect is strongest when the solid surface is well
covered by the wetting phase, and the nonwetting phase has the best
connectivity. Thus, there must have certain saturation when both
criteria are satisfied by the most extent, and the critical saturation
should be related to the structure of the porous media. The maximal
H. Zhao et al.
value of the NWP relative permeability is ~1.07, much smaller than
that in the 2D channel flow (~ 5.47). This may be due to the tortuosity
of the porous media and uneven distribution of the two phases.
There are also studies about the effect of the viscosity ratio when the
WP has a larger viscosity [35,39]. It is concluded that the relative
permeability values of both the NWP and WP are smaller than 1. The
results are similar to the curves shown in Fig. 3c, and we will not make
much interpretation about this.
3.2.3. Effect of wettability
Finally, we want to discuss the effect of the wettability on the
relative permeability curves. The capillary number is Ca = 1.1 × 10‐4,
and the viscosity ratio is M = 1. The compared two cases are the neutral
wetting (θ = 90∘) and strong wetting (θ = 168.57°) conditions, and the
results are shown in Fig. 7. When the porous media convert from strong
wetting to neutral wetting, there is an evident decrease in the relative
permeability of the NWP; while there is a slight increase in the relative
permeability of the WP. According to Li et al., the wettability works on
the relative permeability through two mechanisms [26]. One is
resistance to flow because of the interaction between the fluid and
solid walls; the other is the connectivity of the fluids by wetting the
solid walls. And the net effect of wettability is a combination of the
above two mechanisms. In our simulation, the former seems to be the
dominate factor.
To further clarify the effect of the wettability on the relative
permeability curves, we run several simulations in relatively simple
porous media. The size of the porous media is 500 × 400 lu2, and the
porosity of the porous medium is 0.422. The gray represents the strong
wetting matrix (θ = 168.57°), while the black represents the neutral
wetting matrix (θ = 90∘). In Fig. 9, the left column (S1) is the base case
of porous media with homogeneous wettability, while the middle (S2)
and the right (S3) are porous media with heterogeneous wettability. By
keeping the capillary number (Ca = 1.1 × 10‐4) and viscosity ratio
(M = 1) as constants, we run simulations using these three structures
respectively. The results are plotted in Fig. 8. As is shown in the picture,
when the WP saturation is smaller than 0.3, the relative permeability of
the WP is zero. And the two phase flow is equivalent to the NWP single
phase flow, resulting in almost equal NWP relative permeability for
different porous media. With the increase of the WP saturation, the
relative permeability of the WP increases gradually, and that of the
NWP decreases accordingly. And the relative permeability curves of
different porous media show differences because of the variance in
wetting conditions. When the WP saturation is 0.9, the relative
permeability of the WP reaches the maximal value, while that of the
NWP is the smallest. And the specific values of the relative permeability
are in accordance with the order of wetting matrix percentage. It is
clear in Fig. 9 that the nonwetting phase is more attached to the solid
walls with increasing neutral wetting matrix. This results in the
decrease of the NWP relative permeability. On the contrary, the
decrease in the solid-fluid force results in the increase of the WP
relative permeability.
4. Conclusions
In this study, we used the MCMP pseudopotential model to simulate
two immiscible fluids flowing through 2D porous media, and analyzed
the effects of capillary number (Ca), viscosity ratio (M), and wettability
on the relative permeability curves. The main conclusions of the study
are as follows.
(1) The NWP relative permeability increases with increasing Ca; the
effect of Ca on WP relative permeability depends on the wettability.
For the neutral wetting condition, the WP permeability increases
with increasing capillary pressure; for the strong wetting condition,
the WP permeability is insensitive to Ca.
(2) When M > 1, the NWP relative permeability increases with
60
International Communications in Heat and Mass Transfer 85 (2017) 53–61
increasing M in the strong wetting condition because of the
lubricating effect. The amplitude of the NWP relative permeability
may even exceed 1. However, the exact value of the amplitude and
where it occurs depend on M and the structure of the porous media.
Meanwhile, the WP relative permeability is insensitive to the
viscosity ratio.
(3) When the porous media convert from strong wetting to neutral
wetting, the NWP relative permeability decreases while the WP
relative permeability increases. This trend is further certified by the
simulations in simplified porous media with heterogeneous wett-
ability. The results of this study provide some new insights into the
characteristics of two immiscible fluids flowing through porous
media.
Acknowledgements
The financial support of National Natural Science Foundation of
China (Grant No. 51474222 and 51504265) and Outstanding PhD
Thesis Fund and Innovation Fund of China University of Petroleum-
Beijing (2462016YXBS01) are acknowledged.
References
[1] Q. Kang, P.C. Lichtner, H.S. Viswanathan, A.I. Abdel-Fattah, Pore scale modeling of
reactive transport involved in geologic CO2 sequestration, Transport Porous. Med.
82 (1) (2010) 197–213.
[2] L. Chen, Q. Kang, B.A. Robinson, Y.-L. He, W.-Q. Tao, Pore-scale modeling of
multiphase reactive transport with phase transitions and dissolution-precipitation
processes in closed systems, Phys. Rev. E 87 (4) (2013) 043306.
[3] Z. Dai, R. Middleton, H. Viswanathan, J. Fessenden-Rahn, J. Bauman, R. Pawar, S.-
Y. Lee, B. McPherson, An integrated framework for optimizing CO2 sequestration
and enhanced oil recovery, Environ. Sci. Technol. Lett. 1 (1) (2013) 49–54.
[4] D.G. Avraam, A.C. Payatakes, Flow mechanisms, relative permeabilities, and
coupling effects in steady-state two-phase flow through porous media. The case of
strong wettability, Am. Chem. Soc. 38 (3) (1999) 778–786.
[5] R.G. Bentsen, Effect of momentum transfer between fluid phases on effective
mobility, J. Pet. Sci. Eng. 21 (1) (1998) 27–42.
[6] M. Ayub, R.G. Bentsen, Interfacial viscous coupling: a myth or reality? J. Pet. Sci.
Eng. 23 (1) (1999) 13–26.
[7] F.A.L. Dullien, M. Dong, Experimental determination of the flow transport
coefficients in the coupled equations of two-phase flow in porous media, Transport
Porous. Med. 25 (1) (1996) 97–120.
[8] R.G. Bentsen, A.A. Manai, On the use of conventional cocurrent and countercurrent
effective permeabilities to estimate the four generalized permeability coefficients
which arise in coupled, two-phase flow, Transport Porous. Med. 11 (3) (1993)
243–262.
[9] M. Akbarabadi, M. Piri, Relative permeability hysteresis and capillary trapping
characteristics of supercritical CO2/brine systems: an experimental study at
reservoir conditions, Adv. Water Resour. 52 (2013) 190–206.
[10] S. Krevor, R. Pini, L. Zuo, S.M. Benson, Relative permeability and trapping of CO2
and water in sandstone rocks at reservoir conditions, Water Resour. Res. 48 (2)
(2012).
[11] M.J. Blunt, B. Bijeljic, H. Dong, O. Gharbi, S. Iglauer, P. Mostaghimi, A. Paluszny,
C. Pentland, Pore-scale imaging and modelling, Adv. Water Resour. 51 (2013)
197–216.
[12] X. Zhao, M.J. Blunt, J. Yao, Pore-scale modeling: effects of wettability on water-
flood oil recovery, J. Pet. Sci. Eng. 71 (3) (2010) 169–178.
[13] O. Gharbi, M.J. Blunt, The impact of wettability and connectivity on relative
permeability in carbonates: a pore network modeling analysis, Water Resour. Res.
48 (12) (2012).
[14] S. Chen, G.D. Doolen, Lattice Boltzmann method for fluid flows, Annu. Rev. 30 (1)
(1998) 329–364.
[15] J. Wang, C. Li, Q. Kang, S.S. Rahman, The lattice Boltzmann method for isothermal
micro-gaseous flow and its application in shale gas flow: a review, Int. J. Heat Mass
Transf. 95 (2016) 94–108.
[16] J. Yi, H. Xing, Pore-scale simulation of effects of coal wettability on bubble-water
flow in coal cleats using lattice Boltzmann method, Chem. Eng. Sci. 161 (2017)
57–66.
[17] A.K. Gunstensen, D.H. Rothman, S. Zaleski, G. Zanetti, Lattice Boltzmann model of
immiscible fluids, Phys. Rev. E 43 (8) (1991) 4320.
[18] D.H. Rothman, J.M. Keller, Immiscible cellular-automaton fluids, J. Stat. Phys. 52
(3-4) (1988) 1119–1127.
[19] X. Shan, H. Chen, Lattice Boltzmann model for simulating flows with multiple
phases and components, Phys. Rev. E 47 (3) (1993) 1815.
[20] X. Shan, H. Chen, Simulation of nonideal gases and liquid-gas phase transitions by
the lattice Boltzmann equation, Phys. Rev. E 49 (4) (1994) 2941.
[21] M.R. Swift, E. Orlandini, W.R. Osborn, J.M. Yeomans, Lattice Boltzmann simula-
tions of liquid-gas and binary fluid systems, Phys. Rev. E 54 (5) (1996) 5041.
[22] X. He, X. Shan, G.D. Doolen, Discrete Boltzmann equation model for nonideal gases,
H. Zhao et al.
Phys. Rev. E 57 (1) (1998) R13.
[23] K. Langaas, P. Papatzacos, Numerical investigations of the steady state relative
permeability of a simplified porous medium, Transport Porous. Med. 45 (2) (2001)
241–266.
[24] Q. Kang, D. Zhang, S. Chen, Displacement of a two-dimensional immiscible droplet
in a channel, Phys. Fluids 14 (9) (2002) 3203–3214.
[25] Q. Kang, D. Zhang, S. Chen, Immiscible displacement in a channel: simulations of
fingering in two dimensions, Adv. Water Resour. 27 (1) (2004) 13–22.
[26] H. Li, C. Pan, C.T. Miller, Pore-scale investigation of viscous coupling effects for
two-phase flow in porous media, Phys. Rev. E 72 (2) (2005) 026705.
[27] H. Huang, X.-y. Lu, Relative permeabilities and coupling effects in steady-state gas-
liquid flow in porous media: a lattice Boltzmann study, Phys. Fluids 21 (9) (2009)
092104.
[28] Z. Dou, Z.-F. Zhou, Numerical study of non-uniqueness of the factors influencing
relative permeability in heterogeneous porous media by lattice Boltzmann method,
Int. J. Heat Fluid. Flow 42 (2013) 23–32.
[29] P. Lallemand, L.-S. Luo, Theory of the lattice Boltzmann method: dispersion,
dissipation, isotropy, Galilean invariance, and stability, Phys. Rev. E 61 (6) (2000)
6546.
[30] Z. Guo, C. Zheng, B. Shi, Lattice Boltzmann equation with multiple effective
relaxation times for gaseous microscale flow, Phys. Rev. E 77 (3) (2008) 036707.
[31] Z. Yu, L.-S. Fan, Multirelaxation-time interaction-potential-based lattice Boltzmann
model for two-phase flow, Phys. Rev. E 82 (4) (2010) 046708.
61
International Communications in Heat and Mass Transfer 85 (2017) 53–61
[32] Z. Guo, C. Zheng, B. Shi, Discrete lattice effects on the forcing term in the lattice
Boltzmann method, Phys. Rev. E 65 (4) (2002) 046308.
[33] L. Chen, H.B. Luan, Y.L. He, W.Q. Tao, Pore-scale flow and mass transport in gas
diffusion layer of proton exchange membrane fuel cell with interdigitated flow
fields, Int. J. Therm. Sci. 51 (4) (2012) 132–144.
[34] L. Chen, Q.J. Kang, Y.T. Mu, Y.L. He, W.Q. Tao, A critical review of the
pseudopotential multiphase lattice Boltzmann model: methods and applications,
Int. J. Heat Mass Transf. 76 (2014) 210–236.
[35] A.G. Yiotis, J. Psihogios, M.E. Kainourgiakis, A. Papaioannou, A.K. Stubos, A lattice
Boltzmann study of viscous coupling effects in immiscible two-phase flow in porous
media, Colloids Surf. A 300 (1) (2007) 35–49.
[36] M.L. Porter, E.T. Coon, Q. Kang, J.D. Moulton, J.W. Carey, Multicomponent
interparticle-potential lattice Boltzmann model for fluids with large viscosity ratios,
Phys. Rev. E 86 (3) (2012) 036701.
[37] C.J. Landry, Z.T. Karpyn, O. Ayala, Relative permeability of homogenous-wet and
mixed-wet porous media as determined by pore-scale lattice Boltzmann modeling,
Water Resour. Res. 50 (5) (2014) 3672–3689.
[38] A.S. Odeh, Effect of viscosity ratio on relative permeability, Trans. AIME 216 (1959)
346–353.
[39] A. Ghassemi, A. Pak, Numerical study of factors influencing relative permeabilities
of two immiscible fluids flowing through porous media using lattice Boltzmann
method, J. Pet. Sci. Eng. 77 (1) (2011) 135–145.