Available online at www.sciencedirect.

com

ScienceDirect
Procedia Engineering 126 (2015) 471 475

7th International Conference on Fluid Mechanics, ICFM7

Modelling of migration of CO2 in porous media under conditions of

saline aquifers using lattice Boltzmann method
Y.Q.Zu a,*
a
Department of Mechanics and Engineering Science, Fudan University, Shanghai, 200433,China

Abstract

In this paper, migration of supercritical CO2 in porous media under conditions of saline aquifers is numerically simulated
using lattice Boltzmann method (LBM). A new LBM model has been proposed to minimize the unphysical interfacial current.
Using the newly formulated LBM model, buoyancy-driven two phase flow past through body centered sphere beds, as the
simplified geometries of porous media, and also real rock geometry obtained by Micro-CT scanner is numerically simulated.
CO2/saline-water properties under conditions of CO2 storage in saline aquifers are used in the simulation. Through the simulation,
the effects of different parameters including bond number, contact angle, viscosity ratioand porosity of porous media are
analyzed and discussed.

2015
2015The
TheAuthors. Published
Authors. by by
Published Elsevier Ltd.Ltd.
Elsevier This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of The Chinese Society of Theoretical and Applied Mechanics (CSTAM).
Peer-review under responsibility of The Chinese Society of Theoretical and Applied Mechanics (CSTAM)
Keywords:CO2 Geological storage; Saline Aquifers; Porous Media; Multiphase Flow; Lattice Boltzmann Method;

1. Introduction

Geological storage of CO2 in saline aquifers is presently one of the most promising technologies for reducing the
emissions of CO2 into the atmosphere [1].So far, numerical investigations of CO2 transport in porous media have
been conducted using two different types of models, i.e. continuum field-scale models [2-4] and pore-scale models
[5-7]. The continuum field-scale models have been widely used, especially in industry, due to the higher
computational efficiency. However, these models can only provide information under average conditions and rely

* Corresponding author. Tel.: +86 (0) 21 5566 4965; Fax: +86 (0) 21 6564 2742.
E-mail address: yzu@fudan.edu.cn

1877-7058 2015 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of The Chinese Society of Theoretical and Applied Mechanics (CSTAM)
doi:10.1016/j.proeng.2015.11.285
472 Y.Q. Zu / Procedia Engineering 126 (2015) 471 475

heavily on the knowledge of the effective properties of the rock such as relative permeability and capillary pressure
functions, which depend critically on the geometry and topology of the pore space and the interaction between the
fluids and the pore boundaries. Thus, it is necessary to conduct the pore-scale modeling such as network models [5,6]
to identify key parameters or physical processes that control macroscopic phenomena but which are not captured by
continuum field-scale models. Additionally, the conduction of the pore-scale modeling can provide values for
constitutive parameters needed in the continuum models through up-scaling of the pore-scale results [7,8]. Network
models use a network of discrete pores connected by uniform throats to approximate a porous structure. Recently,
LBM [7,9,10] has been used to carry out the simulation of the detailed flow physics in the pore. Such models are
advantageous over the network models being capable of incorporating the exact pore geometry and topology.
In this paper, a new LBM methodis developed to combine the LBE of Zheng et al. [11,12] for interface capturing
with the LBE of He et al. [13] for velocity and pressure fields. Using the newly formulated LBM model, buoyancy-
driven two-phase flow past through porous media is numerically simulated. CO2 and saline-water properties under
conditions of CO2 storage in saline aquifers are used in the simulation.

2. Methodology

2.1. LBE for Interface Capturing

In the diffuse interface modeling with phase-field theory, an interface-capturing equation is formulated to mimics
the convective Cahn-Hilliard equation [14],

w tI  (I u) M2 PI (1)
where u and t are the velocity and time respectively; and M is a diffusion parameter named as mobility, PI is a
chemical potential, I acts as an order parameter to distinguish the different fluids.
A LBE for interface capturing of Zheng et al. [11] is introduced below to recover the Cahn-Hilliard equation,
fD (x  eG t , t  G t )  fD (x, t ) (1  q)[ fD (x  eG t , t )  fD (x, t )]  [ fD (x, t )  fD(eq ) (x, t )] / W f (2)
where, xis the space coordinate; eD is discrete velocity vector in the D th direction; Gt is the time step during
which the particles travel the lattice spacing; W f is a dimensionless relaxation time; fD (x, t ) is a distribution function
for the order parameter I , while f D (x, t ) is its corresponding equilibrium state, q 1/(W f  0.5) is a constant
( eq)

Eq. (2) was initially proposed to track the interface of two phase flow in 2D [11]. In this study, it isextended to
( eq)
solve 3D problem by developing new equilibrium distribution function f D (x, t ) . On the basis of three-dimensional
fifteen-velocity (D3Q15) lattice structure[15], f D(eq) (x, t ) is developed as

M  (1  Z0 )[PM / cs2 ,
D 0
fD( eq ) (3)
ZD [[PM  (eD u)M ] / cs , D 1  14
2

Where c s is the sound speed for an ideal fluid, ZD is the weighting coefficient in the D th direction, [ is a
constant controlling the strength of mobility, M q(qW f  1/ 2)[G t . I is computed by I
D
fD .

2.2. LBE for Velocity and Pressure

A LBE for velocity and pressure fields of He et al. [13] is introduced to solve the velocity and pressure field,
gD (x  eG t , t  G t )  gD (x, t ) [ gD (x, t )  gD( eq ) (x, t )] / W g 
(4)
[(2W g  1) / (2W g )](eD  u)[*D (u)(F  G)  (*D (u)  *D (0))( U cs2 )]G t
 Y.Q. Zu / Procedia Engineering 126 (2015) 471 475 473

where, F is the force associated with surface tension, W g is a relaxation time related to fluid kinetic viscosity, and Gis
the body force. In this study, a "potential form" of the surface tension F IPI rather than the "pressure form" in
the original model of He et al. [13] is employed in the interfacial force modeling. By this means, parasitic currents
(eq)
across the interface of the fluids can be reduced [16]. gD is the distribution function for pressure and velocity, gD
is its equilibrium states given by
e u (e u) 2 u u
gD(eq) (x, t ) *D (0) p  [*D (u)  *D (0)]Ucs2 , D (u) ZD [1  D 2 (eD u)  D 4  2 ] (5)
cs 2cs 2cs
Hydrodynamic pressure p and the macroscopic velocity u are computedby,
1 1
p  u Ucs2Gt , u [ e g  (F  G)cs2Gt ] / Ucs2
g (6)
D2 2
In order to obtain no-slip velocity conditions at the solid surface of the porous media, a particle distribution
function bounce-back scheme [17] is used for fD and gD at the solid walls. To study the capillary effect of the
porous media, awetting boundary treatment proposed by Yan & Zu [15]is applied in the present simulation.

3. Numerical Simulation

3.1. Multiphase Flow in Porous Media with Simplified Geometry

Fig. 1 shows the computational domain and initial distribution of the two fluids. Saline (fluid A) is located at the
top half of the computational domain while super critical CO2 (fluid B), which is of relatively lower density, is at the
bottom, so that the flow is driven by buoyancy. The computational domain is divided into 60 u 60 u 60 lattice cells.
All the boundaries of the computational domain are periodic.

CS-I

x y

Fig. 1. Computational domain and Fig. 2. Three phase contact line: Fig. 3. Variation with time of volumetric
initial distribution of two fluids: (green: effects of contact angle (Bo=162, flow rate: effects of viscosity ratio. (Bo=162,
solid; blue: CO2; clear: saline water.) =0.39, P A / P B 10 , t*=26.53) =0.39, =30o)

Non-dimensional parameters, including porosity M , bond number Bo ( U A  UB ) gL2 / V AB , viscosity ratio

P A / P B , non-dimensional time t* t g / L and non-dimensional volumetric flow rate Q* Q /(L2 gL ) are defined
for the current study. Here,Lis the width of the computational domain. To be consistent with the CO2/saline-water
properties [18] under conditions of CO2 storage, the ranges of parameters considered in the present study are
Bo [16.2, 810] , T [0 , 30 ] , P A / P B [4, 10] and M [0.13, 0.82] respectively.
D D

The effects of contact angle from 0o to 30o on the two phase flow are studied with Bo=162, =0.39,
P A / P B 10 . To check the effects of contact angle, three phase contact lines are plotted at a specific region as
shown in Fig. 2. For contact angles of 0oand 10o, the interface of the two fluids have detached from the solid wall
due to the relative lower contact angles. While, for =20 o and =30o, there are still small contact regions between
CO2 and the solid surface, when the obtained contact angles agree well with those of the initial setting.
474 Y.Q. Zu / Procedia Engineering 126 (2015) 471 475

The effects of viscosity ratio on the fluid behavior are studied by performing a series of simulations with various
viscosity ratios ( P A / P B = 4, 6, 8, 10, respectively) when the Bond number, porosity and contact angle are fixed at
Bo=162, =0.39, =30o. During the simulation, the viscosity of saline is fixed and that of CO2 is adjusted to achieve
the suitable viscosity ratio. It is found that CO2 in higher viscosity ratio case travels faster as shown Fig. 3, where
variations with time of volumetric flow rates for different viscosity ratio are plotted. This is due to the lower
viscosity drag force acting on the fluids in the higher viscosity ratio case.

(a) =0.87; (b) =0.69; (c) =0.39; (d) =0.18;

Fig. 4. Distribution of saline water and CO2: effects of porosity. (Bo=162, =30o, Fig. 5. Variation with time of volumetric
P A / P B 10 , t*=12.25) flow rate. (=0.39, =30o, )

Figs. 4(a-d) shows the distribution of saline-water and CO2 at t*=12.25 for porosities of 0.87, 0.69, 0.39 and 0.18
respectively when Bond number, contact angle and viscosity ratio are fixed at Bo=162, =30 o, P A / P B 10 .
Obviously, the fluids move faster in porous media with higher porosity. As shown in Fig.5, an interesting
phenomenon observed is that the volumetric flow rate in porous media of =0.18 is very small at the early stage of
the evolution and decreases with time, approaching zero after a long time. In other words, CO 2is confined in the
porous media.

(a) t=400s; (b)t=500s

Fig. 6. Digital image of a rock Fig. 7. Evolution of phase distribution with brine saturation of 0.79 (green: solid; blue:
reproduced by the CMT. supercritical CO2; clear: saline water).

3.2. Towards the modelling of CO2 migration in real rock geometry

The sandstone rocks from Bramhope[19]are studied. Samples of the rocksare scanned using the X-ray computed
micromotography (CMT) and digital imageswith a size of 1mm1mm1mm are produced, an example of which is
shown in Fig. 6. It is assumed that the fluid flow through rocks is laminar and fluid parameters used in the LBM
simulationsare: brine/CO2 =2, brine/CO2 =8 and contact angle is set as 30o.The fluidsflow is driven by the pressure
gradient and gravity along the vertical (upward) direction and the periodicboundary condition is imposed on the top
and bottom faces. Two layers of voids are added on top and bottom faces to enable the boundary conditions to be
applied. Theside surfaces are made symmetric boundary conditions. Fig. 7 shows the preliminary results of time
dependent phase distributions when the situation of brine=0.79.
 Y.Q. Zu / Procedia Engineering 126 (2015) 471 475 475

Conclusions

In this paper, migration of supercritical CO2 in porous media under conditions of saline aquifers is numerically
investigated using LBM. A new LBM method has been developed. This model is capable of being consistent with
the Cahn-Hilliard equation, reducing parasitic currents and meanwhile simulating multiphase flows with moderate
density ratios. Using the newly formulated LBM model, buoyancy-driven two phase flow past through a porous bed
formed by spherical beads is numerically simulated. CO 2 and saline-water properties under conditions of CO2
storage in saline aquifers are used in the simulation. Through the simulation, the effects of different parameters
including bond number, contact angle, viscosity ratio and porosity of porous media have been analyzed and
discussed. Moreover, the "residual trapping" phenomena have been observed under suitable conditions.
It should be pointed out that, in the present study, two phase flows in some simplified homogeneous porous
media geometries have been simulated. And it is also found that the flow phenomena in the real rock geometry are
more complicated. The LBM simulation for the CO2 migration in the rock pore structure has been initially simulated
and more detailed simulation and analysis will be carried out in the near future.

Acknowledgements

The digital rock geometries/images are provided by Dr. Benoit Lamy-Chappuis and Prof. Bruce W. D. Yardleyat
the University of Leeds, UK. This work is partly supported by Shanghai Pujiang Program under grant 14PJ1401600,
the Scientific Research Foundation for the Returned Overseas Chinese Scholars and Fudan University Initiative
Scientific Research Program.

References
[1] IPCC Special Report on Carbon Dioxide Capture and Storage, Cambridge University Press, Cambridge, UK, IPCC, 2005.
[2] N. Spycher, and K. Pruess, CO2-H2O mixtures in the geological sequestration of CO2 center dot. II. Partitioning in chloride brines at 12-100
degrees C and up to 600 bar. Geochim Cosmochim Ac 69 (2005) 3309-3320.
[3] B.A. Robinson, H.S. Viswanathan, and A.J. Valocchi, Efficient numerical techniques for modeling multicomponent ground-water transport
based upon simultaneous solution of strongly coupled subsets of chemical components. Adv Water Resour 23 (2000) 307-324.
[4] X. Jiang, A review of physical modelling and numerical simulation of long-term geological storage of CO2. Appl Energ 88 (2011) 3557-3566.
[5] M.J. Blunt, M.D. Jackson, M. Piri, and P.H. Valvatne, Detailed physics, predictive capabilities and macroscopic consequences for pore-
network models of multiphase flow. Adv Water Resour 25 (2002) 1069-1089.
[6] J.S. Ellis, and A. Bazylak, Dynamic pore network model of surface heterogeneity in brine-filled porous media for carbon sequestration. Phys
Chem Chem Phys 14 (2012) 8382-8390.
[7] L. Hao, and P. Cheng, Pore-scale simulations on relative permeabilities of porous media by lattice Boltzmann method. Int J Heat Mass Tran
53 (2010) 1908-1913.
[8] R.S. Middleton, G.N. Keating, P.H. Stauffer, A.B. Jordan, H.S. Viswanathan, Q.J.J. Kang, J.W. Carey, M.L. Mulkey, E.J. Sullivan, S.P.P. Chu,
R. Esposito, and T.A. Meckel, The cross-scale science of CO2 capture and storage: from pore scale to regional scale. Energ Environ Sci 5
(2012) 7328-7345.
[9] C.L. Lin, A.R. Videla, and J.D. Miller, Advanced three-dimensional multiphase flow simulation in porous media reconstructed from X-ray
Microtomography using the He-Chen-Zhang Lattice Boltzmann Model. Flow Meas Instrum 21 (2010) 255-261.
[10] Y.Q. Zu, and S. He, Lattice Boltzmann modelling of migration of CO2 in porous media under conditions of saline aquifers, Int. Sym. Heat
Transfer, ISHT-8, Beijing China, 2012.
[11] H.W. Zheng, C. Shu, and Y.T. Chew, Lattice Boltzmann interface capturing method for incompressible flows. Phys. Rev. E 72 (2005) 056705.
[12] H.W. Zheng, C. Shu, and Y.T. Chew, A lattice Boltzmann model for multiphase flows with large density ratio. J Comput Phys 218 (2006)
353-371.
[13] X.Y. He, S.Y. Chen, and R.Y. Zhang, A lattice Boltzmann scheme for incompressible multiphase flow and its application in simulation of
Rayleigh-Taylor instability. J. Comput. Phys. 152 (1999) 642-663.
[14] Y.Q. Zu, and S. He, Phase-field-based lattice Boltzmann model for incompressible binary fluid systems with density and viscosity contrasts.
Physical Review E 87 (2013).
[15] Y.Y. Yan, and Y.Q. Zu, A lattice Boltzmann method for incompressible two-phase flows on partial wetting surface with large density ratio. J.
Comput. Phys. 227 (2007) 763-775.
[16] T. Lee, and P.F. Fischer, Eliminating parasitic currents in the lattice Boltzmann equation method for nonideal gases. Phys. Rev. E 74 (2006)
046709.
[17] S. Chen, and G.D. Doolen, Lattice Boltzmann method for fluid flows. Annu Rev Fluid Mech 30 (1998) 329-364.
[18] P. Chiquet, J.L. Daridon, D. Broseta, and S. Thibeau, CO2/water interfacial tensions under pressure and temperature conditions of CO2
geological storage. Energ Convers Manage 48 (2007) 736-744.
[19] B. Lamy-Chappuis, D. Angus, Q. Fisher, C. Granttoni, and B.W.D. Yardley, Rapid Porosity and Permeability Changes of Calcareous
Sandstone Due to CO2-enriched Brine Injection. Geophy Res Lett 41 (2013) 399-406.