Sie sind auf Seite 1von 205

Journal of Chemistry

Transport Phenomena in Porous


Media and Fractal Geometry
Guest Editors: Peng Xu, Jianchao Cai, Agus Pulung Sasmito,
Sachin Vinayak Jangam, and Boming Yu
Transport Phenomena in Porous Media and
Fractal Geometry
Journal of Chemistry

Transport Phenomena in Porous Media and


Fractal Geometry

Guest Editors: Peng Xu, Jianchao Cai, Agus Pulung Sasmito,


Sachin Vinayak Jangam, and Boming Yu
Copyright © 2015 Hindawi Publishing Corporation. All rights reserved.

This is a special issue published in “Journal of Chemistry.” All articles are open access articles distributed under the Creative Commons
Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is prop-
erly cited.
Contents
Transport Phenomena in Porous Media and Fractal Geometry, Peng Xu, Jianchao Cai,
Agus Pulung Sasmito, Sachin Vinayak Jangam, and Boming Yu
Volume 2015, Article ID 486501, 2 pages

Simulation of Gas Transport in Tight/Shale Gas Reservoirs by a Multicomponent Model Based on PEBI
Grid, Longjun Zhang, Daolun Li, Lei Wang, and Detang Lu
Volume 2015, Article ID 572434, 9 pages

Experimental Investigation of Flow Resistance in a Coal Mine Ventilation Air Methane Preheated
Catalytic Oxidation Reactor, Bin Zheng, Yongqi Liu, Ruixiang Liu, Jian Meng, and Mingming Mao
Volume 2015, Article ID 375789, 7 pages

Effect of Matrix-Wellbore Flow and Porosity on Pressure Transient Response in Shale Formation
Modeling by Dual Porosity and Dual Permeability System, Daolun Li, Wenshu Zha, Dewen Zheng,
Longjun Zhang, and Detang Lu
Volume 2015, Article ID 426860, 9 pages

Laboratory Measurement and Interpretation of the Changes of Physical Properties after Heat Treatment
in Tight Porous Media, Yili Kang, Mingjun Chen, Lijun You, and Xiangchen Li
Volume 2015, Article ID 341616, 10 pages

Stability of Fluorosurfactant Adsorption on Mineral Surface for Water Removal in Tight Gas
Reservoirs, Lijun You, Wanchun Zhang, Yili Kang, Zhangxin Chen, and Xuefen Liu
Volume 2015, Article ID 980439, 8 pages

Review on Fractal Analysis of Porous Metal Materials, J. Z. Wang, J. Ma, Q. B. Ao, H. Zhi, and H. P. Tang
Volume 2015, Article ID 427297, 6 pages

Finite Element Method Using a Characteristic-Based Split for Numerical Simulation of a Carbonate
Fracture-Cave Reservoir, Liehui Zhang, Yuhui Zhou, Lei Zhao, and Deliang Zhang
Volume 2015, Article ID 815051, 13 pages

Compositional Modeling for Optimum Design of Water-Alternating CO2 -LPG EOR under Complicated
Wettability Conditions, Jinhyung Cho, Sung Soo Park, Moon Sik Jeong, and Kun Sang Lee
Volume 2015, Article ID 604103, 9 pages

Heat Transport and Water Permeability during Cracking of the Landfill Compacted Clay Cover,
Haijun Lu, Jianjun Liu, Yuan Li, and Yiqie Dong
Volume 2015, Article ID 685871, 6 pages

Modeling Wettability Variation during Long-Term Water Flooding, Renyi Cao, Changwei Sun,
and Y. Zee Ma
Volume 2015, Article ID 592951, 11 pages

Numerical Study of the Influence of Cavity on Immiscible Liquid Transport in Varied-Wettability


Fractures, Zhi Dou, Zhifang Zhou, Yefei Tan, and Yanzhang Zhou
Volume 2015, Article ID 961256, 10 pages
Study on Skin Factor and Productivity of Horizontal Well after Acidizing with Nonuniform Damage,
Lin Yuan, Xiao-Ping Li, Xiao-Hua Tan, and Lie-Hui Zhang
Volume 2015, Article ID 706312, 10 pages

Intermittent Plurisink Model and the Emergence of Complex Heterogeneity Patterns: A Simple
Paradigm for Explaining Complexity in Soil Chemical Distributions, Miguel Ángel Martín,
Miguel Reyes, and F. Javier Taguas
Volume 2015, Article ID 138202, 5 pages

The Applicability of Different Fluid Media to Measure Effective Stress Coefficient for Rock Permeability,
Ying Wang, Guozhi Li, Min Li, and Jing Zhang
Volume 2015, Article ID 391851, 11 pages

Numerical Simulation of Driven Convective Heat Transfer Based Lattice Boltzmann Method in a
Porous Cavity, You-Sheng Xu, Rui-Min Wang, Guo-Neng Li, and You-Qu Zheng
Volume 2015, Article ID 710341, 8 pages

Numerical Simulation of Unsteady-State Flow in Dual Porous Coalbed Methane Horizontal Wells with
Complex Boundary Conditions, Cheng-yong Li, Jun Zhou, Xiang-yi Yi, Yi Luo, and Ping-zhi Gong
Volume 2015, Article ID 173975, 11 pages

Pressure Transient Analysis and Flux Distribution for Multistage Fractured Horizontal Wells in
Triple-Porosity Reservoir Media with Consideration of Stress-Sensitivity Effect, Jingjing Guo,
Haitao Wang, Liehui Zhang, and Chengyong Li
Volume 2015, Article ID 212901, 16 pages

Numerical Well Test Analysis for Polymer Flooding considering the Non-Newtonian Behavior,
Jia Zhichun, Li Daolun, Yang Jinghai, Xue Zhenggang, and Lu Detang
Volume 2015, Article ID 107625, 10 pages

A Mathematical Model for the Analysis of the Pressure Transient Response of Fluid Flow in Fractal
Reservoir, Jin-Zhou Zhao, Cui-Cui Sheng, Yong-Ming Li, and Shun-Chu Li
Volume 2015, Article ID 596597, 8 pages

Production Decline Analysis for Two-Phase Flow in Multifractured Horizontal Well in Shale Gas
Reservoirs, Wei-Yang Xie, Xiao-Ping Li, Lie-Hui Zhang, Xiao-Hua Tan, Jun-Chao Wang, and Hai-Tao Wang
Volume 2015, Article ID 212103, 10 pages

Nonlinear Seepage Model of Gas Transport in Multiscale Shale Gas Reservoirs and Productivity
Analysis of Fractured Well, Ting Huang, Xiao Guo, and Kun Wang
Volume 2015, Article ID 349507, 10 pages
Hindawi Publishing Corporation
Journal of Chemistry
Volume 2015, Article ID 486501, 2 pages
http://dx.doi.org/10.1155/2015/486501

Editorial
Transport Phenomena in Porous Media and Fractal Geometry

Peng Xu,1 Jianchao Cai,2 Agus Pulung Sasmito,3 Sachin Vinayak Jangam,4 and Boming Yu5
1
College of Science, China Jiliang University, Hangzhou 310018, China
2
Hubei Subsurface Multi-Scale Imaging Key Laboratory, Institute of Geophysics and Geomatics,
China University of Geosciences, Wuhan 430074, China
3
Department of Mining and Materials Engineering, McGill University, Montreal, QC, Canada H3A 2A7
4
Department of Chemical and Biomolecular Engineering, National University of Singapore, Singapore 117585
5
School of Physics, Huazhong University of Science and Technology, Wuhan 430074, China

Correspondence should be addressed to Peng Xu; pengxuhust@gmail.com

Received 1 July 2015; Accepted 2 July 2015

Copyright © 2015 Peng Xu et al. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The concept of porous media is used in many areas of the transient flow in fractured/shale reservoirs. L. Yuan et
applied science and engineering [1–3]. Transport phenomena al. proposed a model for local skin factor and productivity
in porous media, including single phase and multiphase fluid of horizontal well. L. You et al. investigated the adsorption
flow through porous media, heat transfer in porous media, stability of surfactant on the rock mineral surface. R. Cao
and electrical and acoustical transport in porous media, are et al. theoretically modeled the wettability variation during
a subject of most common interest and have emerged as long-term water flooding. B. Zheng et al. reported the results
a separate field of study [4–6]. Transport phenomena in of experimental investigation of flow resistance in a coal
porous media, from microscopic to upward scales, cover mine ventilation air methane preheated catalytic oxidation
general theories behind flow and transport in porous media reactor. Y. Kang et al. measured the physical properties of
and form the basis for deterministic and stochastic mod- samples from representative tight gas reservoirs before and
els that describe them. However, the microstructures and after high temperature treatment. L. Zhang et al. developed a
transport mechanisms in porous media are very complex in new compositional model to characterize non-Darcy flow in
nature. Fractal geometry has been shown to have potential tight/shale gas reservoirs. L. Zhang et al. established a discrete
in the analysis of flow and transport properties of porous network model for fracture-cave reservoirs to study fluid flow
media (consult recent special issue [7–11]). This special issue characteristics and pressure distributions.
includes a number of recent works on transport phenomenon Z. Dou et al. numerically simulated the immiscible
and applications of fractal geometry theory in porous media. liquid transport in cavity-fractures by Lattice Boltzmann
The first group of papers focuses on the transport (LB) method. Y. Xu et al. presented a LB model of the
phenomena in porous media. L. Wang et al. presented the uniform velocity and driven convective thermal conductivity
applicability of different fluid media to measure effective in a porous cavity. H. Lu et al. investigated the hydraulic
stress coefficient for rock permeability. T. Huang et al. conductivity of soil under different wetting-drying cycles.
established a multiscale comprehensive mathematical model The second group of papers focuses on the applications
to simulate different flow regimes. C. Li et al. constructed of fractal geometry in porous media. M. A. Martı́n et al. pro-
a mathematical model to describe gas from porous coalbed vided an explanation to the complexity of spatial variability of
methane reservoirs with complex boundary conditions. A chemicals or pollutants in soil via an extremely simple model.
water alternating CO2 -LPG enhanced oil recovery simulation Z. Wang et al. presented a review on the major development
model was developed by J. Cho et al. on fractal analysis of porous metal materials. J. Zhao et
For transport process in complex reservoir media, J. Guo al. used similar construction method of solution to solve
et al., D. Li et al., Z. Jia et al., and W. Xie et al. studied mathematical models of spherical flow in a fractal reservoir.
2 Journal of Chemistry

We hope that this special issue will help to further advance


this multidisciplinary endeavor of fractal geometry theory
and porous media.

Acknowledgments
The guest editors are thankful to all authors for their excellent
contributions and to the anonymous referees for their efforts
in providing valuable comments.
Peng Xu
Jianchao Cai
Agus Pulung Sasmito
Sachin Vinayak Jangam
Boming Yu

References
[1] M. Sahimi, “Flow phenomena in rocks: from continuum mod-
els to fractals, percolation, cellular automata, and simulated
annealing,” Reviews of Modern Physics, vol. 65, no. 4, pp. 1393–
1534, 1993.
[2] B. M. Yu, “Analysis of flow in fractal porous media,” Applied
Mechanics Reviews, vol. 61, no. 5, Article ID 050801, 2008.
[3] P. Xu and B. Yu, “Developing a new form of permeability and
Kozeny-Carman constant for homogeneous porous media by
means of fractal geometry,” Advances in Water Resources, vol.
31, no. 1, pp. 74–81, 2008.
[4] J. Cai, B. Yu, M. Zou, and L. Luo, “Fractal characterization of
spontaneous co-current imbibition in porous media,” Energy
and Fuels, vol. 24, no. 3, pp. 1860–1867, 2010.
[5] A. Mirzaei-Paiaman and M. Masihi, “Scaling of recovery by
cocurrent spontaneous imbibition in fractured petroleum reser-
voirs,” Energy Technology, vol. 2, no. 2, pp. 166–175, 2014.
[6] M. Xu and H. Dehghanpour, “Advances in understanding
wettability of gas shales,” Energy Fuels, vol. 28, no. 7, pp. 4362–
4375, 2014.
[7] E. Perfect, Y. Pachepsky, and M. A. Martin, “Fractal and
multifractal models applied to porous media,” Vadose Zone
Journal, vol. 8, no. 1, pp. 174–176, 2009.
[8] J. Cai, F. San José Martiı́nez, M. A. Martı́n, and E. Perfect, “An
introduction to modeling of flow and transport in fractal porous
media: part I,” Fractals, vol. 22, no. 3, Article ID 1402001, 2014.
[9] B. Ghanbarian, A. G. Hunt, R. P. Ewing, and T. E. Skinner,
“Universal scaling of the formation factor in porous media
derived by combining percolation and effective medium theo-
ries,” Geophysical Research Letters, vol. 41, no. 11, pp. 3884–3890,
2014.
[10] J. Cai, F. San José Martı́nez, M. A. Martı́nez, and X. Hu, “An
introduction to flow and transport in fractal models of porous
media: part II,” Fractals, vol. 23, no. 1, Article ID 1502001, 2015.
[11] J. Cai, L. Luo, R. Ye, X. Zeng, and X. Hu, “Recent advances
on fractal modeling of permeability for fibrous porous media,”
Fractals, vol. 23, no. 1, Article ID 1540006, 2015.
Hindawi Publishing Corporation
Journal of Chemistry
Volume 2015, Article ID 572434, 9 pages
http://dx.doi.org/10.1155/2015/572434

Research Article
Simulation of Gas Transport in Tight/Shale Gas Reservoirs
by a Multicomponent Model Based on PEBI Grid

Longjun Zhang,1 Daolun Li,1,2 Lei Wang,3 and Detang Lu1


1
Department of Modern Mechanics, University of Science and Technology of China, Hefei 230027, China
2
Hefei University of Technology, Hefei 230026, China
3
Institute of Nuclear Energy Safety Technology, Chinese Academy of Sciences, Hefei 230031, China

Correspondence should be addressed to Daolun Li; ldaol@mail.ustc.edu.cn

Received 4 August 2014; Accepted 11 September 2014

Academic Editor: Peng Xu

Copyright © 2015 Longjun Zhang et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The ultra-low permeability and nanosize pores of tight/shale gas reservoir would lead to non-Darcy flow including slip flow,
transition flow, and free molecular flow, which cannot be described by traditional Darcy’s law. The organic content often adsorbs
some gas content, while the adsorbed amount for different gas species is different. Based on these facts, we develop a new
compositional model based on unstructured PEBI (perpendicular bisection) grid, which is able to characterize non-Darcy flow
including slip flow, transition flow, and free molecular flow and the multicomponent adsorption in tight/shale gas reservoirs. With
the proposed model, we study the effect of non-Darcy flow, length of the hydraulic fracture, and initial gas composition on gas
production. The results show both non-Darcy flow and fracture length have significant influence on gas production. Ignoring non-
Darcy flow would underestimate 67% cumulative gas production in lower permeable gas reservoirs. Gas production increases with
fracture length. In lower permeable reservoirs, gas production increases almost linearly with the hydraulic fracture length. However,
in higher permeable reservoirs, the increment of the former gradually decreases with the increase in the latter. The results also show
that the presence of CO2 in the formation would lower down gas production.

1. Introduction proposed a formulation of apparent permeability in terms


of Knudsen number. Civan [6] improved the function of
Gas production from unconventional gas reservoirs, such as the dimensionless rarefaction coefficient proposed by Beskok
tight gas/shale gas reservoir, has grown great interest in recent and Karniadakis [4] and established a mathematical model
years. Because of the ultra-low permeability (usually under for gas flow in tight gas formation [7]. Zheng et al. [8–10]
0.1 mD) and small pore diameter (usually under 50 nm) proposed a predictive model for gas slippage factor and gas
[1], gas flow in such tight formations reveals multiflow diffusivity in microporous media based on fractal theory.
mechanisms that cannot be described by traditional Darcy’s Freeman et al. [11] incorporated the dusty-gas model into
law, such as slip flow and Knudsen diffusion [2, 3]. TOUGH+ family code to study gas flow behavior in tight
Some modeling work has been conducted to study flow gas/shale gas reservoirs. Freeman et al. [12] also incorporate
mechanisms in such reservoirs. Javadpour [2] combined extended Langmuir isotherm into a compositional model
convective flow and Knudsen diffusion into gas mass bal- to represent gas desorption in shale. However, they did not
ance equation and found that the apparent permeability consider multiflow mechanisms this time, such as slip flow
derived from the new mass balance equation can lead and transition flow. Clarkson et al. [13] modeled transport
to one to two orders of magnitude difference from the in tight gas/shale using dynamic slippage concept which
intrinsic permeability in origin Darcy’s law. Beskok and developed by Ertekin et al. [14]. Swami et al. [15] and Li et
Karniadakis [4] derived a unified Hagen-Poiseuille-type al. [16] both separately incorporated multiflow mechanisms
equation for volumetric gas flow through a single pipe. into a numerical model to simulate gas behavior in shale. Yao
Based on Beskok and Karniadakis [4], Florence et al. [5] et al. [17] compared gas production predicted by Civan [6],
2 Journal of Chemistry

Javadpour [2], and Dusty-gas model and studied effect of The Knudsen number 𝐾𝑛 expresses the mean free path
fracture parameters on gas production in shale. of molecules as a fraction of a representative path (mean
However, most models above are single component hydraulic radius, e.g.) [24]:
model and are based on structured grid [2, 7, 13, 15–
17]. Some models [2, 11, 12] did not combine multiflow 𝜆𝑔
𝐾𝑛 = . (1)
mechanisms and gas sorption together. In this paper, we first 𝑅ℎ
developed a compositional model which incorporates mul-
tiflow mechanisms and multicomponent adsorption based Here, 𝜆 𝑔 is the mean free path for gas and is defined by the
on PEBI (perpendicular bisection) grid. We also studied following equation:
effect of apparent permeability, initial gas composition, and
fracture length on gas production under various intrinsic 𝜇𝑔 𝜋𝑅𝑇
permeability conditions. 𝜆𝑔 = √ , (2)
𝑃 2𝑀𝑔

2. Mathematical Model where 𝜇𝑔 is the gas viscosity in Pa⋅s, 𝑃 is the absolute gas pres-
sure in Pa, 𝑅 is the universal gas constant (8,314 J/(kmol⋅K)),
2.1. PEBI Grid. PEBI grids are also known as Voronoi 𝑇 is the absolute temperature in Kelvin, and 𝑀𝑔 is the
cells and are defined as the region in which all points are molecular weight of the gas in kg/kmol.
closer to the corresponding seed than any other seeds [18– The mean hydraulic radius of flow tubes in porous media
20]. The boundary of each Voronoi cell is normal to the 𝑅ℎ is defined as
line connecting the seeds on the two sides. Compared to
structured Cartesian and Corner gird, the unstructured PEBI 𝑘0
grids have the following advantages. 𝑅ℎ = 2√2𝜏√ , (3)
𝜙
(1) Flexibility: it can represent the characteristic of com- where 𝜏 is the tortuosity, 𝜙 is the porosity, and 𝑘0 is the
plex boundary, pinch-out, and faults in the formation intrinsic permeability of the reservoir in m2 .
precisely. Based on the unified model for gas flow in microtubes
(2) Easiness of local refinement: because of the flexibility derived by Beskok and Karniadakis [4] and Florence et al. [5]
of PEBI grid and the arbitrariness of arranging PEBI proposed a formulation of apparent permeability in terms of
grid point, it is easier to refine grid in the local place, Knudsen number to characterize the non-Darcy flow in the
such as domain around wells. porous media. Consider
(3) Less grid orientation effect: the unstructured hexag- 4𝐾𝑛
onal PEBI grid makes the grid orientation effect less 𝑘 = 𝑘0 (1 + 𝛼𝐾𝑛 ) (1 + ), (4)
1 + 𝐾𝑛
significant than structured grid.
(4) Easiness of discretizing and solving equations: the where 𝛼 is the dimensionless rarefaction coefficient and is
local orthogonality of PEBI grid makes it easier to defined by Beskok and Karniadakis as
discretize and solve equations by using finite volume 128
method. 𝛼= tan−1 (4𝐾𝑛0.4 ) . (5)
15𝜋2
The PEBI grid used in this paper is based on previous Substituting (5) into (4) yields
researches [21, 22]. The grid points are arranged following
streamline and based on well type, location, and reservoir 128 4𝐾𝑛
geometry. The generated grids are denser near wells and 𝑘 = 𝑘0 (1 + 2
tan−1 (4𝐾𝑛0.4 ) 𝐾𝑛 ) (1 + ). (6)
15𝜋 1 + 𝐾𝑛
looser far away from wells, as shown in Figure 1. This
arrangement of PEBI grids can keep computation accuracy This equation is valid for all flow regimes for gas flow in
and save computation time. porous media [4, 5].

2.2. Apparent Permeability. Gas flow in low-permeability 2.3. Multicomponent Langmuir Isotherm. To simulate and
tight and shale gas reservoirs occurs following various mech- distinguish sorption capacity for different components, the
anisms, such as slip flow, transition flow, and free molecular extended Langmuir isotherm which is widely accepted by
flow. The matrix permeability in such reservoirs needs to be petroleum industry is used.
modified to enable traditional Darcy’s law describe such non- For component 𝑖, the sorption volume is as follows:
Darcy flow.
Note that the non-Darcy flow in the paper includes slip 𝑦𝑖 𝑃
𝑉ads,𝑖 = 𝑉𝐿,𝑖 𝑛ℎ
, 𝑖 = 1, . . . , 𝑛ℎ , (7)
flow, transition flow, and free molecular flow which is defined 𝑃𝐿,𝑖 (1 + ∑𝑗=1 𝑦𝑗 (𝑃/𝑃𝐿,𝑗 ))
as below.
Chambre and Schaaf [23] have classified four flow where 𝑉ads,𝑖 is the standard volume of sorbed component
regimes based on Knudsen number (𝐾𝑛 ), as shown in Table 1. 𝑖, 𝑦𝑖 is the mole fraction of the component 𝑖, and 𝑛ℎ is
Journal of Chemistry 3

Table 1: Flow regimes classified by Chambre and Schaaf.

𝐾𝑛 >10.0 (0.1, 10) (10−3 , 0.1) <10−3


Flow regimes Free molecular flow Transition flow Slip flow Continuum flow

(a) One vertical well in the middle (b) One vertical well with a hydrau-
lic fracture in the middle

Figure 1: Schematic of PEBI grid.

the total number of components. The Langmuir volume 𝑉𝐿,𝑖 right term denotes the advection term. The second right term
and Langmuir pressure 𝑃𝐿,𝑖 are measured values for the pure denotes source or sink in the well, which stands for the rate
component 𝑖. The total sorption is given by of gas mass produced from or injected into a well.
For vertical well, the well production rate 𝑞𝑔,std is
𝑛ℎ 𝑛ℎ
𝑦𝑖 𝑃 expressed by Peaceman model:
𝑉ads = ∑𝑉ads,𝑖 = ∑𝑉𝐿,𝑖 𝑛ℎ
. (8)
𝑖=1 𝑖=1 𝑃𝐿,𝑖 (1 + ∑𝑗=1 𝑦𝑗 (𝑃/𝑃𝐿,𝑗 )) 𝜃𝑘ℎ
𝑞𝑔,std = (𝑝 − 𝑝𝑤𝑓 ) , (10)
𝜇𝑔 ln (𝑟𝑒 /𝑟𝑤 ) 𝑗
2.4. Mass Conservation Equations. Based on finite volume
method, the governing mass balance equation for the com- where ℎ is the effective height of well perforation in m, 𝑟𝑤 is
ponent 𝑖 considering gas sorption is given by wellbore radius in m, 𝑝𝑗 is well grid pressure in Pa, 𝑝𝑤𝑓 is
bottom-hole flowing pressure in Pa, and 𝑟𝑒 is the equivalent
𝜕 radius in the well grid in m. For structured Cartesian grid, 𝑟𝑒
(𝑉𝜙𝜌𝑔 𝑦𝑖 + 𝑉𝜌𝑠 𝑉ads,𝑖 𝜌𝑔,std )
𝜕𝑡 can be expressed as
0.5 0.5 0.5
1 (9) [(𝑘𝑦 /𝑘𝑥 ) Δ𝑥2 + (𝑘𝑥 /𝑘𝑦 ) Δ𝑦2 ]
= ∑(𝑇𝑟 𝜌 𝑦 ΔΦ ) − 𝜌𝑔,std 𝑦𝑖 𝑞𝑔,std ,
𝑙
𝜇𝑔 𝑔 𝑖 𝑔 𝑙 𝑟𝑒 = 0.28 0.25 0.25
. (11)
(𝑘𝑦 /𝑘𝑥 ) + (𝑘𝑥 /𝑘𝑦 )
𝑖 = 1, . . . , 𝑛ℎ ,
For unstructured PEBI grid, we derive 𝑟𝑒 as follows.
where 𝑉 is the gas volume in m3 , 𝜌𝑠 is the rock density in Assume pressure at equivalent radius 𝑟𝑒 as 𝑝𝑒 which is
kg/m3 , 𝜌𝑔,std is the gas density under standard conditions equivalent to well grid pressure 𝑝𝑗 . Arranging (10) yields
(1 atm and 15∘ C) in mol/m3 , 𝜌𝑔 is the gas density under 𝜇𝑔 ln (𝑟𝑒 /𝑟𝑤 ) 𝑞𝑔,std
formation condition in mol/m3 , and 𝑞𝑔,std is the gas pro- (𝑝𝑒 − 𝑝𝑤𝑓 ) = . (12)
𝜃𝑘ℎ
duction rate under standard condition in m3 /s. Values of 𝜌𝑔 The flux from adjacent grids to well grid 𝑄 is the same as 𝑞𝑔,std .
and 𝜌𝑔,std were calculated using the Peng-Robinson equation 𝑄 can be expressed as
of state (PR EOS) [25]. The symbol 𝑙 represents connection
between adjacent grids, and ΔΦ𝑔 is the difference in gas 𝑇𝑟 𝑇
𝑄 = 𝑞𝑔,std = ∑ (𝑝 − 𝑝𝑗 ) = ∑ 𝑟 (𝑝𝑙 − 𝑝𝑒 ) . (13)
potentials between adjacent grids 𝑙1 and 𝑙2 and is given 𝑙
𝜇𝑔 𝑙 𝑙
𝜇𝑔
by ΔΦ𝑔 = Φ𝑙1 − Φ𝑙2 = 𝑝𝑙1 − 𝑝𝑙2 − (𝜌𝑙1 𝑍𝑙1 − 𝜌𝑙2 𝑍𝑙2 )𝑔,
where 𝑍 is the depth of the formation. The viscosity 𝜇𝑔 was Arranging (10) and (12) and substituting into (13) yield
calculated using the Lohrenz-Bray-Clark (LBC) correlation ∑𝑙 𝑇𝑟 ln Δ𝐿 − 𝜃𝑘ℎ
[26]. The transmissibility parameter 𝑇𝑟 = 𝑘𝐴/Δ𝐿, where ln 𝑟𝑒 = . (14)
∑𝑙 𝑇𝑟
𝑘 is the apparent permeability represented by (6), 𝐴 is the
cross-sectional area between the adjacent grids, and Δ𝐿 is the Consequently, 𝑟𝑒 can be expressed as follows:
distance between the adjacent grids.
∑𝑙 𝑇𝑟 ln Δ𝐿 − 𝜃𝑘ℎ
The left term of (9) is the mass accumulation term and 𝑟𝑒 = exp ( ). (15)
includes the mass of both free gas and sorbed gas. The first ∑𝑙 𝑇𝑟
4 Journal of Chemistry

Table 2: Properties and parameters of a tight gas reservoir in


31
Xinjiang, China.

Name Value Units 30

Pressure (MPa)
Production
Reservoir dimensions 3500 × 2000 × 18 Meter Well shut-in
29
Fracture half-length 230 Meter
Permeability 0.056 mD 28
Porosity 0.0376 N/A
Reservoir temperature 355.0 Kelvin 27
Initial pressure 31.5 MPa
26
Bulk density 2500.0 kg/m3 0 10 20 30 40 50
Wellbore diameter 0.084 meter Time (day)
Langmuir volume (𝑉𝐿 ) 0.0012 m3 /kg
Model output
Langmuir pressure (𝑃𝐿 ) 7.3 MPa Build-up pressure data

Figure 2: The bottom-hole pressure history of a production well in


For vertical well with a hydraulic fracture, we use the a tight gas reservoir in Xinjiang, China. The production rate from
infinite conductivity model [27] to represent the conductivity day 0 to 30 was 8.36 × 104 m3 /day, where pressure data was not
of the fracture in the reservoir. Considering the wellbore available. The predicted pressure build-up (black line) during the
storage, the well production rate 𝑞𝑔,std can be expressed as shut-in period compares well with data (red symbol).

∑𝑙 ((𝑇𝑟 /𝜇𝑔 ) 𝜌𝑔 ΔΦ𝑔 )𝑙 𝐶 𝑡 𝑡𝑛


𝑞𝑔,std = − (𝑝 𝑛+1 − 𝑝𝑤𝑓 ), (16) Specific reservoir parameters and simulation conditions
𝜌𝑔,std Δ𝑡 𝑤𝑓
are listed in Table 3, unless noted otherwise. The sorption
parameters for gas species are listed in Table 4. The hydrauli-
where 𝐶 is the wellbore storage in m3 /MPa and 𝑡𝑛+1 and 𝑡𝑛 are
cally fractured vertical well is located in the middle of
(𝑛 + 1)th and 𝑛th time steps.
simulated gas reservoir of which the boundary is sealed.
The unknown variables include pressure 𝑝, bottom-hole
flowing pressure 𝑝𝑤𝑓 , and the mole fraction 𝑦𝑖 in mass
conservation equations (9). The nonlinear equations are 4.1. The Effect of Non-Darcy Flow on Gas Production. Here,
solved using Newton iteration method and the matrix solver the gas production was calculated with and without con-
GMRES (generalized minimal residual method) [28]. sidering non-Darcy flow under three different intrinsic per-
meability conditions which are 1.01 × 10−2 mD (10−17 m2 ),
3. Model Validation 1.01 × 10−4 mD (10−19 m2 ), and 1.01 × 10−6 mD (10−21 m2 ),
respectively.
The model was validated by reproducing the pressure build- The non-Darcy flow is incorporated into apparent per-
up data from a hydraulic fractured vertical well in a tight meability which is represented by (6). The non-Darcy flow
gas reservoir in Xinjiang, China. The pilot test operation in tight formations includes slip flow, transition flow, and
comprised two stages: gas production at the rate of 8.36 × free molecular flow which will enhance the gas flow ability,
104 m3 /day for 30 days, followed by a shut-in period of 25 especially under lower pressure condition. Thus, as gas
days. produces, the reservoir pressure continues to decrease and
The model was set up using parameters and properties the influence of non-Darcy flow on gas production will
listed in Table 2. The initial reservoir gas contains 94.7% become more significant.
CH4 , 1.2% CO2 , 2.7% C2 H6 , and 1.4% C3 H8 . The average gas In the higher permeable reservoir (𝑘0 = 1.01 × 10−2 mD),
sorption parameters (𝑉𝐿 and 𝑃𝐿 ) in Table 2 were provided the influence of non-Darcy flow on gas production is very
without distinguishing different type of gas components. limited, as depicted by Figure 3(a). The flow regime of gas
Figure 2 shows the predicted pressure dropdown during mainly stays in slip flow regime. After 4.38 × 104 days (120
the production stage and the pressure build-up during the years) production, the difference of predicted produced gas
shut-in period. The modeling output was compared to the volume with and without considering non-Darcy flow is only
monitored high frequency/high-resolution BHP data during about 2%, as shown in Figure 4.
the shut-in operation [29]. The pressure data during the While in lower permeable gas reservoirs (𝑘0 = 1.01 ×
production was not available. The simulation output shows 10−4 mD, 𝑘0 = 1.01 × 10−6 mD) which are more common for
reasonable reproduction of the pressure build-up data. tight/shale gas reservoirs, the non-Darcy flow (mainly under
transition flow regime) reveals more significant influence
4. Results and Discussion on produced gas volume, as depicted in Figures 3(b) and
3(c). The predicted produced gas volume without considering
The developed model was used to understand the effect of non-Darcy flow is 24% and 67% smaller than that with
non-Darcy flow, length of the hydraulic fracture, and initial considering non-Darcy flow for reservoir 𝑘0 = 1.01×10−4 mD
gas composition on gas production. and reservoir 𝑘0 = 1.01 × 10−6 mD, respectively.
Journal of Chemistry 5

Table 3: Input parameters for simulated hydraulically fractured vertical well and the gas reservoir.

Input parameter Values Unit


Dimension 2000 × 1000 × 25 m
Fracture half-length 200 m
Wellbore storage 1 m3 /MPa
Flowing bottom-hole pressure 10 MPa
Porosity 0.05

Temperature 60 C
Initial reservoir pressure 35 MPa
Initial gas composition 2% CO2 , 90% CH4 , 8% C2 H6
Bulk density 2500 Kg/m3

×108 ×107
×106
Cumulative production (std m3 )

4
Cumulative production (std m3 )

Cumulative production (std m3 )


3.5 3 6
3 2.5 5
2.5 2 4
2
1.5
1.5 k0 = 1.01 × 10−2 mD
3 k0 = 1.01 × 10−6 mD
1 2
1
0.5 0.5 k0 = 1.01 × 10−4 mD 1

1 2 3 4 1 2 3 4 1 2 3 4
Production time (day) ×104 ×104 ×104
Production time (day) Production time (day)

Non-Darcy flow Non-Darcy flow Non-Darcy flow


Darcy flow Darcy flow Darcy flow
(a) (b) (c)

Figure 3: The gas production predicted with and without considering non-Darcy flow under three different intrinsic permeability conditions.

70 Table 4: Langmuir constants for gas species in gas reservoir.

60 CO2 CH4 C2 H6
Gas production difference (%)

𝑉𝐿 (m3 /t) 4.1 1.6 2.6


50 𝑃𝐿 (MPa) 5.76 10.77 5.59
40

30 on slip flow, early transition flow, and late transition flow,


respectively.
20 Based on the results above, non-Darcy flow is critical
for accurate predicting gas flow behavior in low permeable
10
reservoirs, such as tight gas or shale gas reservoirs. Ignoring
0 it would significantly underestimate gas production capacity.
1E − 16 1E − 17 1E − 18 1E − 19 1E − 20 1E − 21 1E − 22
Intrinsic permeability (m2 )
4.2. The Effect of Length of Hydraulic Fracture on Gas Produc-
Figure 4: Gas production difference predicted with considering tion. Here, we aim to understand the role of hydraulic frac-
non-Darcy flow and without considering it after 120 years of ture length in determining gas production. Gas production
production. rate and cumulative gas production were predicted when the
half-length of hydraulic fracture equals 0, 50, 100, 200, and
400 meters in the reservoir of which intrinsic permeability
Lower permeability often indicates smaller pore diameter equals 1.01 × 10−2 mD, 1.01 × 10−4 mD, and 1.01 × 10−6 mD.
which means that more chance of non-Darcy flow would Take the medium permeable reservoir (𝑘0 = 1.01 ×
happen during gas flow. From Figures 3(a) to 3(c), as the 10−4 mD), for example. The gas cumulative production and
permeability goes smaller, the gas flow regimes mainly stay production rate are depicted in Figure 5.
6 Journal of Chemistry

×107 ×106
2.2
6
2
Cumulative production (std m3 )

1.8

Production rate (std m3 )


5
1.6
4 1.4
1.2
3 1
0.8
2
0.6
0.4
1
0.2

0.5 1 1.5 2 2.5 3 3.5 4 10−4 10−2 100 102 104


Production time (day) ×104 Production time (day)

No fracture Half L = 200 No Fracture Half L = 200


Half L = 50 Half L = 400 Half L = 50 Half L = 400
Half L = 100 Half L = 100
(a) Cumulative production (b) Production rate

Figure 5: Effect of fracture half-length on predicted gas cumulative production and production rate.

The results show that gas production rate and cumulative increment relationship with fracture half-length. And in the
production increase with half-length of hydraulic fracture. lower permeable reservoir (𝑘0 = 1.01 × 10−6 mD), the
The vertical well with no fracture (represented by red line) relationship between the two can be identified as linear.
indicates very limited production potential. The averaged Thus, we conclude that the lower the permeability is,
production rate is only 79 m3 /day and the cumulative pro- the more significant the increasing fracture length influences
duction after 120 years is only 3.4 million cubic meters, as gas production. But for some relatively higher permeable
shown in Figures 5(a) and 5(b), while for the vertical well reservoirs, an appropriate length of fracture should be found
with hydraulic fracture of which the half-length is only 50 considering the balance between economical cost of fractur-
meters, the cumulative production increases to 15 million ing a well and gas production.
cubic meters, about five times as that of the well with no
fracture. As the half-length of fracture increases to 100 m, 4.3. The Effect of Initial Gas Composition on Gas Production.
200 m, and 400 m, the cumulative production increases to Here, gas cumulative productions are calculated when initial
22, 36, and 65 million cubic meters, respectively. With the CO2 composition equals to 0%, 10%, 20%, and 30%, respec-
increase of half-length of fracture, the extent of cumulative tively. The half-length of hydraulic fracture is confined to
production increased is remarkable, indicating the significant 200 meters. The intrinsic permeability is confined to 1.01 ×
importance of length of hydraulic fracture on gas production. 10−4 mD. The initial gas species in the reservoir is assumed
Gas production rate in such low permeable reservoir to be CO2 and CH4 . All other parameters are the same as in
decreases very quickly, especially in the early production Tables 3 and 4. The results are shown in Figure 7.
time, as depicted by Figure 5(b). For the well with a fracture The results show initial gas composition has some influ-
half-length of 400 meters, in the first 2 days of production, ence on cumulative production. Higher percentage of CO2
the rate stays above 100,000 m3 /day, while after 200 days of in the initial gas leads to lower gas production which can be
production, the rate decreases rapidly to 10,000 m3 /day. In seen clearly on the enlarged figure. This is mainly because the
the late production time, the rate decreases much slower. presence of CO2 increases gas viscosity, which would slow
After 100 years of production, the rate remains fairly at down flow speed in the formation and eventually reduce gas
1,000 m3 /day. production.
For all simulated reservoirs, the increase of half-length The decrement of cumulative gas production slows down
of hydraulic fracture shows a positive influence on gas with the increase of CO2 percentage in initial gas content,
production, as depicted in Figure 6. In the higher permeable as depicted in Figure 8(a), while that of cumulative methane
reservoir (𝑘0 = 1.01 × 10−2 mD), the cumulative production production does not, as depicted in Figure 8(b). The cumula-
shows nonlinear relationship with fracture half-length. With tive methane production shows a linear relationship with the
increase of fracture half-length, the increment of cumulative percentage of CO2 in initial gas content. When initial CO2
production decreases gradually and may finally become percentage equals 0%, the produced cumulative gas volume is
negligible. In the medium permeable reservoir (𝑘0 = 1.01 × 4.4% (1.5 million m3 ) higher than that of the case when initial
10−4 mD), the cumulative production shows almost linear CO2 percentage equals 30%, while for produced cumulative
Journal of Chemistry 7

×108 ×107 ×106


Cumulative production (std m3 )

Cumulative production (std m3 )

Cumulative production (std m3 )


5 14
6
12
4 5 10
3 4 8
3 6
2
2 4
1
1 2
0 0 0
0 100 200 300 400 0 100 200 300 400 0 100 200 300 400
Fracture half length (m) Fracture half length (m) Fracture half length (m)

(a) 𝑘0 = 1.01 × 10−2 mD (b) 𝑘0 = 1.01 × 10−4 mD (c) 𝑘0 = 1.01 × 10−6 mD

Figure 6: Effect of fracture half-length on cumulative production after 120 years of production in three different permeable gas reservoirs.

×107 ×107
3.5
3.5
Cumulative production (std m3 )

Cumulative production (std m3 )


3 3.45

2.5 3.4

3.35
2
3.3
1.5
3.25
1 3.2

0.5 3.15

3.1
1 2 3 4 3.8 3.9 4 4.1 4.2
×104 ×104
Production time (day) Production time (day)

30% CO2 10% CO2 30% CO2 10% CO2


20% CO2 0% CO2 20% CO2 0% CO2

Figure 7: Effect of initial gas composition on cumulative production.

×107 ×107
3.60
3.6
Cumulative production (std m3 )

Cumulative production (std m3 )

3.4
3.55
3.2

3.0
3.50
2.8

3.45 2.6

2.4

0 5 10 15 20 25 30 0 5 10 15 20 25 30
CO 2 composition (%) CO 2 composition (%)
(a) (b)

Figure 8: Effect of initial gas composition on produced (a) cumulative mixed gas volume and (b) methane volume after 120 years of
production.
8 Journal of Chemistry

methane volume, the former is 49.8% (11.9 million m3 ) higher [5] F. Florence, J. Rushing, K. E. Newsham, and T. A. Blasingame,
than the latter. “Improved permeability prediction relations for low permeabil-
ity sands,” in Proceedings of the Rocky Mountain Oil & Gas
Technology Symposium, 2007.
5. Conclusions
[6] F. Civan, “Effective correlation of apparent gas permeability in
In this paper, we proposed a compositional model for tight porous media,” Transport in Porous Media, vol. 82, no. 2,
tight/shale gas reservoirs based on unstructured PEBI grid. pp. 375–384, 2010.
The non-Darcy flow, including slip flow, transition flow, [7] F. Civan, C. S. Rai, and C. H. Sondergeld, “Shale-gas per-
and free molecular flow, is considered in terms of apparent meability and diffusivity inferred by improved formulation of
permeability in the model. Multicomponent adsorption is relevant retention and transport mechanisms,” Transport in
Porous Media, vol. 86, no. 3, pp. 925–944, 2011.
also considered in terms of extended Langmuir isotherm.
With the proposed model, we studied the effect of non- [8] Q. Zheng and B. Yu, “A fractal permeability model for gas flow
Darcy flow, length of the hydraulic fracture, and initial gas through dual-porosity media,” Journal of Applied Physics, vol.
111, no. 2, Article ID 024316, 2012.
composition on gas production. The results showed the
following: (1) the non-Darcy flow shows significant influence [9] Q. Zheng, B. Yu, S. Wang, and L. Luo, “A diffusivity model
on gas production, especially in low permeable reservoirs. for gas diffusion through fractal porous media,” Chemical
Engineering Science, vol. 68, no. 1, pp. 650–655, 2012.
Ignoring this effect would lead to 67% underestimation on
produced cumulative gas volume according to the simu- [10] Q. Zheng, B. Yu, Y. Duan, and Q. Fang, “A fractal model for
lated case. (2) Gas production increases with half-length of gas slippage factor in porous media in the slip flow regime,”
Chemical Engineering Science, vol. 87, pp. 209–215, 2013.
hydraulic fracture. However, in higher permeable reservoirs,
the increment of gas production decreases with increase in [11] C. M. Freeman, G. J. Moridis, and T. A. Blasingame, “ A
fracture length, while in lower permeable reservoirs, gas numerical study of microscale flow behavior in tight gas and
shale gas reservoir systems,” Transport in Porous Media, vol. 90,
production almost increases linearly with fracture length. Gas
no. 1, pp. 253–268, 2011.
production would impossibly reach to economic rate without
fracturing the well. (3) Higher initial CO2 percentage would [12] C. M. Freeman, G. J. Moridis, and T. A. Blasingame, “Modeling
lead to lower gas production, for its ability to increase gas and performance interpretation of flowing gas composition
changes in shale gas wells with complex fractures,” in Pro-
viscosity in the formation.
ceedings of the International Petroleum Technology Conference:
With considering non-Darcy flow and multicomponent Challenging Technology and Economic Limits to Meet the Global
gas adsorption in tight/shale gas reservoirs, the proposed Energy Demand (IPTC ’13), pp. 4868–4883, Beijing, China,
compositional model can be a powerful tool to predict gas March 2013.
behavior in such unconventional reservoirs and be used to [13] C. R. Clarkson, M. Nobakht, D. Kaviani, and T. Ertekin,
offer valuable insights into reservoir engineers to make better “Production analysis of tight-gas and shale-gas reservoirs using
exploitation schemes. the dynamic-slippage concept,” SPE Journal, vol. 17, no. 1, pp.
230–242, 2012.
Conflict of Interests [14] T. Ertekin, G. R. King, and F. C. Schwerer, “Dynamic gas
slippage: a unique dual-mechanism approach to the flow of gas
The authors declare that there is no conflict of interests in tight formations,” SPE Formation Evaluation, vol. 1, no. 1, pp.
regarding the publication of this paper. 43–52, 1986.
[15] V. Swami, F. Javadpour, and A. Settari, “A numerical model for
Acknowledgments multi-mechanism flow in shale gas reservoirs with application
to laboratory scale testing,” in Proceedings of the 75th EAGE
The authors are grateful for funding from Major State Basic Conference & Exhibition Incorporating SPE EUROPEC, 2013.
Research Development Program of China (973 Program) (no. [16] J. Li, C. Wang, D. Ding, Y. S. Wu, and Y. Di, “A generalized
2011CB707305) and National Key Science and Technology framework model for simulation of gas production in uncon-
Project (2011ZX05009-006). ventional gas reservoirs,” in Proceedings of the SPE Reservoir
Simulation Symposium, The Woodlands, Tex, USA, February
2013.
References [17] J. Yao, H. Sun, D.-Y. Fan, C.-C. Wang, and Z.-X. Sun, “Numer-
[1] P. H. Nelson, “Pore-throat sizes in sandstones, tight sandstones, ical simulation of gas transport mechanisms in tight shale gas
and shales,” AAPG Bulletin, vol. 93, no. 3, pp. 329–340, 2009. reservoirs,” Petroleum Science, vol. 10, no. 4, pp. 528–537, 2013.
[2] F. Javadpour, “Nanopores and apparent permeability of gas [18] Z. E. Helnemann, C. W. Brand, M. Munka, and Y. M. Chen,
flow in mudrocks (shales and siltstone),” Journal of Canadian “Modeling reservoir geometry with irregular grids,” SPE Reser-
Petroleum Technology, vol. 48, no. 8, pp. 16–21, 2009. voir Engineering, vol. 6, no. 2, pp. 225–232, 1991.
[3] F. Javadpour, D. Fisher, and M. Unsworth, “Nanoscale gas [19] Z. E. Heinemann and C. W. Brand, “Gridding techniques in
flow in shale gas sediments,” Journal of Canadian Petroleum reservoir simulation,” in Proceedings of the 1st International
Technology, vol. 46, no. 10, pp. 55–61, 2007. Forum on Reservoir Simulation, Alpbach, Austria, 1988.
[4] A. Beskok and G. E. Karniadakis, “A model for flows in [20] C. L. Palagi and K. Aziz, “Use of Voronoi grid in reservoir
channels, pipes, and ducts at micro and nano scales,” Microscale simulation,” SPE Advanced Technology Series, vol. 2, no. 2, pp.
Thermophysical Engineering, vol. 3, no. 1, pp. 43–77, 1999. 69–77, 1994.
Journal of Chemistry 9

[21] L. Zhang, D. Li, W. Zha et al., “Generation and application


of adaptive PEBI grid for numerical well testing(NWT),” in
Proceedings of the International Conference on Mechatronic
Sciences, Electric Engineering and Computer (MEC ’13), pp.
3002–3006, Shenyang, China, 2013.
[22] W. Zha, Numerical Reservoir Calculation on PEBI Grid and
Implementation, University of Science and Technology of
China, 2009.
[23] P. A. Chambre and S. A. Schaaf, Flow of Rarefied Gases, 1961.
[24] L. B. Loeb, The Kinetic Theory of Gases, Courier Dover, 2004.
[25] S. M. Walas, Phase Equilibria in Chemical Engineering, vol. 4,
Butterworth, Boston, Mass, USA, 1985.
[26] J. Lohrenz, B. Bray, and C. Clark, “Calculating viscosities of
reservoir fluids from their compositions,” Journal of Petroleum
Technology, vol. 16, no. 10, pp. 1171–1176, 1964.
[27] B. Horsfield and H. M. Schulz, “Shale gas exploration and
exploitation,” Marine and Petroleum Geology, vol. 31, no. 1, pp.
1–2, 2012.
[28] Y. Saad and M. H. Schultz, “GMRES: a generalized minimal
residual algorithm for solving nonsymmetric linear systems,”
SIAM Journal on Scientific and Statistical Computing, vol. 7, no.
3, pp. 856–869, 1986.
[29] W. Ding, C. Li, C. Li et al., “Fracture development in shale and
its relationship to gas accumulation,” Geoscience Frontiers, vol.
3, no. 1, pp. 97–105, 2012.
Hindawi Publishing Corporation
Journal of Chemistry
Volume 2015, Article ID 375789, 7 pages
http://dx.doi.org/10.1155/2015/375789

Research Article
Experimental Investigation of Flow Resistance in a Coal Mine
Ventilation Air Methane Preheated Catalytic Oxidation Reactor

Bin Zheng, Yongqi Liu, Ruixiang Liu, Jian Meng, and Mingming Mao
School of Transportation and Vehicle Engineering, Shandong University of Technology, Zhangzhou Road 12, Zibo 255049, China

Correspondence should be addressed to Yongqi Liu; sdutliu@163.com

Received 26 September 2014; Revised 6 January 2015; Accepted 6 January 2015

Academic Editor: Agus Sasmito

Copyright © 2015 Bin Zheng et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This paper reports the results of experimental investigation of flow resistance in a coal mine ventilation air methane preheated
catalytic oxidation reactor. The experimental system was installed at the Energy Research Institute of Shandong University of
Technology. The system has been used to investigate the effects of flow rate (200 Nm3 /h to 1000 Nm3 /h) and catalytic oxidation bed
average temperature (20∘ C to 560∘ C) within the preheated catalytic oxidation reactor. The pressure drop and resistance proportion
of catalytic oxidation bed, the heat exchanger preheating section, and the heat exchanger flue gas section were measured. In addition,
based on a large number of experimental data, the empirical equations of flow resistance are obtained by the least square method.
It can also be used in deriving much needed data for preheated catalytic oxidation designs when employed in industry.

1. Introduction solution seems to be autothermal combustion in reverse-


flow reactors. So far, most studies have focused on catalytic
Coal mine methane emissions make up approximately 8% combustion in catalytic flow-reversal reactors (CFRRs) and
of the world’s anthropogenic methane emissions, and the thermal flow-reversal reactors (TFRRs). The earliest attempt
quantity of methane emissions from coal mining alone is at ventilation air methane (VAM) combustion in CFRR
over 25 million tons every year. Roughly 70% of the methane was supposedly carried out at the Boreskov Institute of
emissions come from coal mine ventilation air methane Catalysis. At present, Salomons et al. [2] and Litto et al. [3]
(VAM), which is not only a greenhouse gas but also a considered that CANMET in Canada is the most advanced
wasted energy resource if not utilized. As a greenhouse gas, in terms of the development of the technology of lean air
CH4 , which (17%) is the second largest contributor to global methane mixtures combustion. Marı́n et al. [4] systematically
warming after CO2 (55%), is over 21 times more effective in compared the performance of particle beds and monolithic
trapping heat in the atmosphere than carbon dioxide over a beds in catalytic reverse-flow reactors. Shi et al. [5] classified
100-year period. As the air volume is large and the methane the existing technologies for coal mine methane mitigation
resource is diluted and varies in concentration and flow rate, and utilization. Tischer et al. [6] developed transient two-
ventilation air methane is the most difficult source of CH4 and three-dimensional simulations of catalytic combustion
to be used as an energy source. Karakurt et al. [1] reported monoliths. Shahamiri and Wierzba [7] developed a one-
that methane concentration in coal mine ventilation air is low dimensional model to investigate the effects of operational
(0.1∼1.5 vol.%), and its effective utilization is considerably low. conditions on CFRR. Shahamiri and Wierzba [8] studied
The combustion of lean methane-air mixtures combined with the combustion of preheated lean mixtures of hydrogen with
the recovery of the heat of reaction is an essential problem for methane in a catalytic packed-bed reactor. Gosiewski et al.
the mining industry. [9, 10] carried out a comparative assessment of TFRR and
A number of studies have been conducted to determine CFRR. It has been shown that both solutions have advantages
the most efficient way to utilize this lean methane via com- and drawbacks. Li et al. [11, 12] built a flow-reversal strategy
bustion as well as to recover the energy thus produced. Due of CFRR based on the thermal front propagation velocity
to the pretty low CH4 concentrations, the most promising and a logic-based controller. Gosiewski et al. [13], Qi et al.
2 Journal of Chemistry

Flow rate Methane


meter sensor
Ventilation
air methane
Methane
sensor
Catalytic
Heat exchanger Outlet
reactor

Figure 1: Preheated catalytic oxidation reactor operating principle


and experimental schema.

Figure 2: General view of the research and demonstration of


[14], Gao et al. [15], and Zheng et al. [16] considered that preheated catalytic oxidation reactor apparatus.
TFRRs are regarded as an attractive alternative. Salvador et
al. [17] reported that such reactors have long been employed,
for example, for the homogeneous (thermal) combustion of and flows into the catalytic reactor, where VAM is catalytically
volatile organic compounds (VOCs). combusted. The high temperature flue gas from the catalytic
The common characteristic of CFRRs and TFRRs is reactor flows into the heat exchanger to preheat VAM, and
reverse-flow in reactors. The drawbacks of their reactors then the cooled flue gas is emitted into the atmosphere.
are as follows. Firstly, the device occupies a large area
because the reactor is packed with a large number of ceramic 3. Experimental Reactor System
monolith blocks. Secondly, after a long time operation, a high
maintenance cost is needed due to the damage of ceramic Tests were carried out in a preheated catalytic oxidation
monolith blocks. Thirdly, the pressure drop is high and more reactor built at the Energy Research Institute of Shandong
energy is required to supply ventilation air methane. Finally, University of Technology. A general view of the experimental
the control machine of the time-varying airflow direction is apparatus is shown in Figure 2, while a simplified flowsheet
quite large and complex, and as a result, the control system is of the apparatus is given in Figure 3.
rather complex. The experimental system is divided into six parts: the
Zheng et al. [18] investigated a preheated catalytic oxida- gas supply system, the parameter acquisition system, the
tion reactor to combustion coal mine ventilation air methane. startup system, the differential pressure analysis system, the
This reactor employs a new working principle and different heat exchanger, and the catalytic oxidation bed. A schematic
CFRRs and TFRRs, which is not reverse-flow. The equipment diagram of the experimental apparatus is shown in Figure 3.
used to recover exhaust flue gas heat is a wall-type heat Mine ventilation air used in the experiment was a mixture
exchanger. In this regard, Tahseen et al. [19], Golman and air and natural gas, where the CH4 purity was 99.9%.
Julklang [20], Gao and Sun [21], and Zheng [22] studied waste Compressed air from the air compressor through the gas
heat recovery by a wall-type heat exchanger. The performance regulating valve and pressure-relief valve mixes with the
of the heat exchanger is reliable. The flow resistance of reactor methane from methane cylinder in the mixer, and the mix-
is an important parameter. Wu and Yu [23] developed a ture is then fed into the heat exchanger. The catalytic reactor
fractal model for resistance of flow through porous media. consists of the startup system and the catalytic oxidation bed,
Kushwaha et al. [24], Amelio and Morrone [25], and Liu et which was surrounded with a layer of insulation ceramic
al. [26] studied the pressure drop of a flow-reversal reactor. fiber blanket which is 350 mm thick to prevent heat loss
The use of a preheated catalytic oxidation reactor for the flow from the reactor to the surroundings. The startup system
resistance therefore requires detailed studies to determine consists of some electrical heaters. To initiate the reaction,
the reasonable operating conditions. The objective of the the catalytic oxidation bed was preheated from an ambient
present study is to investigate the effects of key operational temperature to about 350∘ C by electrical heaters in the inlet
parameters on flow resistance processes within the preheated position of the reactor. Once the bed was preheated, the
catalytic reactor. It can also be used in deriving much needed electrical heater was extinguished and the lean methane
data for preheated catalytic oxidation designs when employed mixture was supplied to the bed. The heating temperature and
in industry. heating power in the startup process were controlled by the
temperature control instrument. The catalytic oxidation bed,
2. Operating Principle of which the length was 300 mm and the section was 600 ×
600 mm squared, was packed with catalytic ceramic monolith
A schematic diagram of the preheated catalytic reaction blocks (Figure 4) and a large number of straight and parallel
reactor being considered in the experiment is shown in channels. The block size was 100 mm ∗ 100 mm ∗ 100 mm,
Figure 1. VAM flows into the heat exchanger, where it is and the density of the hole, which was square, was 300
heated by the flue gas of high temperature from the catalytic meshes. The catalyst was Pd, which was loaded 0.28 wt%. The
reactor until it reaches the temperature of methane catalytic type of support was 𝛾-Al2 O3 . The heat exchanger was dividing
oxidation. Then preheated VAM flows of the heat exchanger wall-type, which consists of heat exchangers 1 and 2. Heat
Journal of Chemistry 3

Computer

Acquisition card Flow rate


meter Methane
Pressure
Mixer
sensor 2 Air
Catalytic reactor Pressure sensor 1
Electrical Catalytic
heater oxidation bed Pressure
Heat exchanger Outlet
difference
transmitter
Pressure Pressure sensor 4
sensor 3
Thermocouple

Figure 3: General overview of the research and demonstration of preheated catalytic oxidation reactor apparatus.

(2) Turn on the computer that is connected with the data


acquisition card; open the data acquisition operating
system on the computer; set the required acquisition
channels and parameters, and the values of each
parameter are displayed online.
(3) Open the methane cylinder and adjust the pressure-
relief valve.
(4) Open both the valves for air and methane and set the
startup flow rate values.
(5) Turn on the startup system. When the catalytic
oxidation bed is preheated to about 350∘ C, turn off the
Figure 4: Catalytic ceramic monolith blocks. startup system and set the desired flow rate values and
the experimental system starts up successfully.
(6) Adjust the operating parameters according to the
exchanger 1 was 700 mm wide, 1100 mm high, and 2000 mm experimental conditions, and collect and record the
long. Heat exchanger 2 was 700 mm wide, 3000 mm high, data. Repeat the above process for the next set of
and 700 mm long. There were 5 thermocouple measuring conditions.
points arranged along the axis of the catalytic oxidation bed, (7) When the experiment is finished, first turn off the
and the output signal of the thermocouple connected to methane gas. Then turn off the methane pressure-
the computer can monitor the transient temperature field in relief valve, stop data acquisition, and store the exper-
the bed. Meanwhile, the signal of VAM flow rate was also imental data and then shut down the computer.
connected to the computer. There were 4 pressure sensors (8) Close the flowmeter and then turn off the air source.
installed in the intake and exhaust of the heat exchanger and Turn off the power control supply.
the catalytic reactor; the output signal of pressure connected
(9) For the next experiment, repeat the above operations.
to pressure difference transmitter can monitor the pressure
drop field in the catalytic reactor and the heat exchanger. The To investigate the flow resistance performance of the
pressure sensor was stagnation pressure tube, whose type was reactor, some pressure drop parameters were used. ΔP was the
TPS-08-500-T. The experimental uncertainty for the pressure loss of total pressure in the reactor, which was the difference
drop was ±1%. between pressure sensor 1 and pressure sensor 4. ΔP1 was
After the experimental system is well connected and the pressure drop of VAM preheating section in the heat
the experimental instrument, equipment, and airtightness exchanger, which was the difference between pressure sensor
are checked, the experiment is started. The experimental 1 and pressure sensor 2. ΔP2 was the pressure drop of catalytic
procedures are as follows. oxidation bed (the pipe resistance is ignored), which was the
difference between pressure sensor 2 and pressure sensor 3.
(1) Turn on the power switch of the device; the system ΔP3 was the pressure drop of flue gas section in the heat
is then connected to the electricity supply and the air exchanger, which was the difference between pressure sensor
compressor is turned on. 3 and pressure sensor 4.
4 Journal of Chemistry

1200
60

900
Pressure drop (Pa)

Proportion (%)
40

600

20
300

0
0
200 400 600 800 1000 200 400 600 800 1000
Flow rate (Nm3 /h) Flow rate (Nm3 /h)

ΔP ΔP2 ΔP1
ΔP1 ΔP3 ΔP2
ΔP3
Figure 5: Pressure drop distribution in preheated catalytic oxidation
reactor with different flow rates. Figure 6: Resistance proportion distribution with different flow
rates.

4. Results and Discussions 2000

4.1. Effect of Flow Rate. Figure 5 shows the pressure drop


results of flow rate from 200 Nm3 /h to 1000 Nm3 /h in normal
condition. As can be seen from Figure 5, with the increase of 1500
flow rate, the pressure drop of different sections all increases,
Pressure drop (Pa)

but the trends of development are different. The pressure


drops of the preheating section and the flue gas section in
1000
the heat exchanger increase quadratically, but the pressure
drop of the catalytic oxidation bed increases linearly. The
reason for this is that the heat exchanger is dividing wall-
type with a large number of fins arranged in the internal heat 500
exchanger in order to enhance the heat transfer effect, so the
heat exchanger can be seen as a porous medium. According
to Darcy’s law, the pressure drops of the preheating section 0
and the flue gas section increase quadratically because the 0 100 200 300 400 500 600
Reynolds number in the heat exchanger is between 17773 Temperature (∘ C)
and 88640, which is far higher than the critical Reynolds
number (2320), but the pressure drop of catalytic oxidation ΔP ΔP2
ΔP1 ΔP3
bed increases linearly because the Reynolds number in the
catalytic oxidation bed is between 398 and 1987, which is Figure 7: Pressure drop distribution with different catalytic oxida-
under critical Reynolds number (2320). Figure 6 shows the tion bed average temperatures.
resistance proportion results of flow rate from 200 Nm3 /h
to 1000 Nm3 /h in cold test condition. Figure 6 shows that
if the other conditions remain the same, when the flow rate remain the same, as the average temperature increases, the
ranges from 200 Nm3 /h to 1000 Nm3 /h, that is, as the flow pressure drop of different sections all increases, but the
rate increases, the resistance proportion of the preheating trends of development are different. The pressure drops of
section and the flue gas section increases, but the resistance the preheating section and the flue gas section in the heat
proportion of catalytic oxidation bed decreases. The pressure exchanger increase exponentially, but the pressure drop of
drop of the heat exchanger becomes a chief component in the the catalytic oxidation bed increases linearly. The reason for
preheated catalytic oxidation reactor. this is that the higher the catalytic oxidation bed average
temperature is, the higher the VAM velocity and viscosity
4.2. Effect of Average Temperature in Catalytic Oxidation Bed. are and the higher the pressure drop is. Figure 8 shows the
Figure 7 shows the pressure drop results of catalytic oxidation resistance proportion results of the catalytic oxidation bed
bed average temperature from 20∘ C to 560∘ C and a flow rate average temperature from 20∘ C to 560∘ C and a flow rate of
of 500 Nm3 /h. Figure 7 shows that if the other conditions 500 Nm3 /h. As can be seen from Figure 8, with the increase
Journal of Chemistry 5

80 1000

60 800

Pressure drop (Pa)


Proportion (%)

40
600

20
400

0
200
0 100 200 300 400 500 600 0 50 100 150 200 250
Temperature (∘ C) Temperature (∘ C)

ΔP1 Experimental results


ΔP2 Calculated results
ΔP3
Figure 9: Contrast of the experimental results and the calculated
Figure 8: Resistance proportion distribution with different catalytic results in the heat exchanger preheating section.
oxidation bed average temperatures.

of the bed average temperature, the resistance proportion from 293 K to 553 K and in the V3 region from 3.7 m/s to
of the catalytic oxidation bed increases, while the resistance 6.2 m/s:
proportion of the preheating section and the flue gas section
decreases. 𝑇2 1.5 403.4 2𝜌02 V32 𝑇3
Δ𝑃3 = 16.99 × ( ) × × , (2)
293 𝑇3 + 110.4 1.293 × 273
4.3. Empirical Equation of Flow Resistance in Preheated
Catalytic Oxidation Reactor. The widely employed resistance
model for flow through porous media was proposed by Ergun where 𝑇3 is the average temperature of intake and exhaust
[27] in 1952. This model is called Ergun equation. Based on in the flue gas section, K; 𝜌0 is the VAM density in normal
a large number of experimental data and Ergun equation, condition, kg/m3 ; V3 is the VAM flow velocity in the flue gas
the empirical equations of flow resistance are obtained by the section in normal condition, m/s.
least square method. Equation (1) is the empirical equation Equation (3) is the empirical equation of the catalytic
of the preheating section in the heat exchanger, and Figure 9 oxidation bed, and Figure 11 shows the contrast of the
shows the contrast of the experimental results and the experimental results and the calculated results. Figure 11
calculated results. As can be seen from Figure 9, the goodness shows that the goodness of fit is high and the maximum
of fit is high, the correlation coefficient is 0.991, and the relative error is less than ±4.6%. The following equation is
maximum relative error is less than ±4.7%. The following quite suitable in the 𝑇2 region from 293 K to 833 K and in the
equation is quite suitable in the 𝑇1 region of 293 K to 503 K V3 region of 0.15 m/s to 1.74 m/s:
and in the V1 region of 3.7 m/s to 6.2 m/s:
(1 − 𝜀)2 𝜇 1−𝜀
𝑇 1.5 403.4 2𝜌02 V12 𝑇1 Δ𝑃2 = [56.75 V + 0.0067 3 𝜌0 V2 ]
Δ𝑃1 = 43.07 ( 1 ) × × , (1) 𝜀3 𝑑𝑝2 𝜀 𝑑𝑝
293 𝑇1 + 110.4 1.293 × 273 (3)
𝑇 0.18
where 𝑇1 is the average temperature of intake and exhaust ⋅ ( 2 ) ⋅ 𝐿,
in the preheating section, K; 𝜌0 is the VAM density in 293
normal condition, kg/m3 ; V1 is the VAM flow velocity in the
preheating section in normal condition, m/s. where 𝜀 is the void ratio of catalytic ceramic monolith block;
Equation (2) is the empirical equation of the flue gas 𝜇 is the coefficient of kinetic viscosity, kg/(m⋅s); 𝑑𝑝 is the
section in the heat exchanger, and Figure 10 shows the equivalent diameter of hole in catalytic ceramic monolith
contrast of the experimental results and the calculated results. block, m; v is the VAM velocity in the catalytic oxidation
Figure 9 shows that the goodness of fit is high, the correlation bed intake, m/s; 𝑇2 is the average temperature of the catalytic
coefficient is 0.996, and the maximum relative error is less oxidation bed, K; L is the axial length of the catalytic oxidation
than ±6.5%. Equation (1) is quite suitable in the 𝑇3 region bed, m; 𝜌0 is the VAM density in normal condition, kg/m3 .
6 Journal of Chemistry

linearly. As the flow rate increases, the resistance


proportion of the preheating section and the flue gas
400
section increases, but the resistance proportion of the
catalytic oxidation bed decreases. The pressure drop
of the heat exchanger becomes a chief component in
Pressure drop (Pa)

the preheated catalytic oxidation reactor.


300
(2) If the other conditions remain the same, as the
average temperature increases, the pressure drop of
different sections all increases, but the trends of
200 development are different. The pressure drops of the
preheating section and the flue gas section in the heat
exchanger increase exponentially, but the pressure
drop of the catalytic oxidation bed increases linearly.
100
With the increase of the bed average temperature,
0 50 100 150 200 250 300 the resistance proportion of the catalytic oxidation
Temperature (∘ C) bed increases, but the resistance proportion of the
preheating section and the flue gas section decreases.
Experimental results
(3) Based on a large number of experimental data, the
Calculated results
empirical equations of flow resistance are obtained by
Figure 10: Contrast of the experimental results and the calculated the least square method; the goodness of fit is high.
results in the heat exchanger flue gas section. It can also be used in deriving much needed data for
preheated catalytic oxidation designs when employed
in industry.
1000

Conflict of Interests
800 The authors declare that there is no conflict of interests
regarding the publication of this paper.
Pressure drop (Pa)

600 Acknowledgments
This work was supported by the China National 863
High Technology Fund Project (2009AA063202), Shan-
400 dong Provincial Science and Technology Development Pro-
gram, China (2012GGX10417), and Shandong Provincial
Natural Science Foundation, China (ZR2013EEQ005 and
ZR2013EEQ008).
200
300 400 500 600
Temperature (∘ C) References
3
900 Nm /h experimental results [1] I. Karakurt, G. Aydin, and K. Aydiner, “Mine ventilation
900 Nm3 /h calculated results air methane as a sustainable energy source,” Renewable and
700 Nm3 /h experimental results Sustainable Energy Reviews, vol. 15, no. 2, pp. 1042–1049, 2011.
700 Nm3 /h calculated results
[2] S. Salomons, R. E. Hayes, M. Poirier, and H. Sapoundjiev, “Flow
Figure 11: Contrast of the experimental results and the calculated reversal reactor for the catalytic combustion of lean methane
results in the catalytic oxidation bed. mixtures,” Catalysis Today, vol. 83, no. 1–4, pp. 59–69, 2003.
[3] R. Litto, R. E. Hayes, H. Sapoundjiev et al., “Optimization of
a flow reversal reactor for the catalytic combustion of lean
methane mixtures,” Catalysis Today, vol. 117, no. 4, pp. 536–542,
5. Conclusions 2006.
[4] P. Marı́n, M. A. G. Hevia, S. Ordóñez, and F. V. Dı́ez, “Com-
The main conclusions of the present study are as follows. bustion of methane lean mixtures in reverse flow reactors:
comparison between packed and structured catalyst beds,”
(1) With the increase of flow rate, the pressure drop Catalysis Today, vol. 105, no. 3-4, pp. 701–708, 2005.
of different sections all increases, but the trends of [5] S. Shi, B. Andrew, G. Hua, and M. Cliff, “An assessment of mine
development are different. The pressure drops of methane mitigation and utilization technologies,” Progress in
the preheating section and the flue gas section in Energy and Combustion Science, vol. 31, no. 2, pp. 123–170, 2005.
the heat exchanger increase quadratically, but the [6] S. Tischer, C. Correa, and O. Deutschmann, “Transient three-
pressure drop of catalytic oxidation bed increases dimensional simulations of a catalytic combustion monolith
Journal of Chemistry 7

using detailed models for heterogeneous and homogeneous petroleum coke fin-and-tube waste heat exchanger,” The Open
reactions and transport phenomena,” Catalysis Today, vol. 69, Fuels & Energy Science Journal, vol. 7, no. 1, pp. 20–25, 2014.
no. 1–4, pp. 57–62, 2001. [23] J. Wu and B. Yu, “A fractal resistance model for flow through
[7] S. A. Shahamiri and I. Wierzba, “Modeling catalytic oxidation of porous media,” International Journal of Heat and Mass Transfer,
lean mixtures of methane-air in a packed-bed reactor,” Chemical vol. 50, no. 19-20, pp. 3925–3932, 2007.
Engineering Journal, vol. 149, no. 1–3, pp. 102–109, 2009. [24] A. Kushwaha, M. Poirier, H. Sapoundjiev, and R. E. Hayes,
[8] S. A. Shahamiri and I. Wierzba, “Simulation of catalytic oxida- “Effect of reactor internal properties on the performance of
tion of lean hydrogen-methane mixtures,” International Journal a flow reversal catalytic reactor for methane combustion,”
of Hydrogen Energy, vol. 34, no. 14, pp. 5785–5794, 2009. Chemical Engineering Science, vol. 59, no. 19, pp. 4081–4093,
[9] K. Gosiewski, Y. S. Matros, K. Warmuzinski, M. Jaschik, 2004.
and M. Tanczyk, “Homogeneous vs. catalytic combustion of [25] M. Amelio and P. Morrone, “Numerical evaluation of the ener-
lean methane-air mixtures in reverse-flow reactors,” Chemical getic performances of structured and random packed beds in
Engineering Science, vol. 63, no. 20, pp. 5010–5019, 2008. regenerative thermal oxidizers,” Applied Thermal Engineering,
[10] K. Gosiewski and A. Pawlaczyk, “Catalytic or thermal reversed vol. 27, no. 4, pp. 762–770, 2007.
flow combustion of coal mine ventilation air methane: what [26] Y.-Q. Liu, Z.-X. Zhang, Z.-Q. Gao, R.-X. Liu, and B. Zheng,
is better choice and when?” Chemical Engineering Journal, vol. “Numerical simulation on resistance of the regenerative oxida-
238, pp. 78–85, 2014. tion bed for ventilation air methane,” Journal of the China Coal
[11] Z. Li, Z. Qin, Y. Zhang et al., “A control strategy of flow reversal Society, vol. 35, no. 6, pp. 946–950, 2010.
with hot gas withdrawal for heat recovery and its application in [27] S. Ergun, “Fluid flow through packed columns,” Chemical
mitigation and utilization of ventilation air methane in a reverse Engineering Progress, vol. 48, pp. 89–94, 1952.
flow reactor,” Chemical Engineering Journal, vol. 228, no. 15, pp.
243–255, 2013.
[12] Z. Li, Z. Qin, Z. Wu et al., “Fuzzy logic control of a reverse
flow reactor for catalytic oxidation of ventilation air methane,”
Control Engineering Practice, vol. 25, no. 1, pp. 112–122, 2014.
[13] K. Gosiewski, A. Pawlaczyk, and M. Jaschik, “Thermal com-
bustion of lean methane-air mixtures: flow reversal research
and demonstration reactor model and its validation,” Chemical
Engineering Journal, vol. 207-208, pp. 76–84, 2012.
[14] X. Qi, Y. Liu, H. Xu, Z. Liu, and R. Liu, “Modeling thermal
oxidation of coal mine methane in a non-catalytic reverse-flow
reactor,” Strojniški vestnik—Journal of Mechanical Engineering,
vol. 60, no. 7-8, pp. 495–505, 2014.
[15] Z. Gao, Y. Liu, and Z. Gao, “Heat extraction characteristic of
embedded heat exchanger in honeycomb ceramic packed bed,”
International Communications in Heat and Mass Transfer, vol.
39, no. 10, pp. 1526–1534, 2012.
[16] B. Zheng, Y.-Q. Liu, R.-X. Liu, Z.-Q. Gao, and J. Meng,
“Oxidation of coal mine ventilation air methane in thermal
reverse-flow reactor,” Journal of the China Coal Society, vol. 34,
no. 11, pp. 1475–1478, 2009.
[17] S. Salvador, J. M. Commandré, and Y. Kara, “Thermal recuper-
ative incineration of VOCs: CFD modelling and experimental
validation,” Applied Thermal Engineering, vol. 26, no. 17-18, pp.
2355–2366, 2006.
[18] B. Zheng, Y. Liu, R. Liu, S. Chen, M. Mao, and J. Meng, “Starting
characteristics of ventilation air methane preheating catalytic
oxidation reactor,” Journal of China Coal Society, vol. 39, no. 6,
pp. 1084–1088, 2014.
[19] T. A. Tahseen, M. Ishak, and M. M. Rahman, “An overview on
thermal and fluid flow characteristics in a plain plate finned
and un-finned tube banks heat exchanger,” Renewable and
Sustainable Energy Reviews, vol. 43, pp. 363–380, 2015.
[20] B. Golman and W. Julklang, “Simulation of exhaust gas heat
recovery from a spray dryer,” Applied Thermal Engineering, vol.
73, no. 15, pp. 899–913, 2014.
[21] W. Gao and J. Sun, “Heat transfer characteristics of flowing
granular waved moving bed under vibration condition,” China
Powder Science and Technology, vol. 9, no. 3, pp. 1–4, 2003.
[22] B. Zheng, Y. Liu, R. Liu, Z. Wang, R. Yu, and X. Qi, “Experi-
mental investigation of heat transfer characteristics of calcined
Hindawi Publishing Corporation
Journal of Chemistry
Volume 2015, Article ID 426860, 9 pages
http://dx.doi.org/10.1155/2015/426860

Research Article
Effect of Matrix-Wellbore Flow and Porosity on
Pressure Transient Response in Shale Formation Modeling by
Dual Porosity and Dual Permeability System

Daolun Li,1,2 Wenshu Zha,1,2 Dewen Zheng,3 Longjun Zhang,2 and Detang Lu2
1
Hefei University of Technology, Hefei, China
2
University of Science and Technology of China, Hefei, China
3
Research Institute of Petroleum Exploration & Development, Langfang, China

Correspondence should be addressed to Daolun Li; ldaol@ustc.edu.cn

Received 11 October 2014; Revised 10 December 2014; Accepted 11 December 2014

Academic Editor: Agus Sasmito

Copyright © 2015 Daolun Li et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

A mathematical dual porosity and dual permeability numerical model based on perpendicular bisection (PEBI) grid is developed
to describe gas flow behaviors in shale-gas reservoirs by incorporating slippage corrected permeability and adsorbed gas effect.
Parametric studies are conducted for a horizontal well with multiple infinite conductivity hydraulic fractures in shale-gas reservoir
to investigate effect of matrix-wellbore flow, natural fracture porosity, and matrix porosity. We find that the ratio of fracture
permeability to matrix permeability approximately decides the bottom hole pressure (BHP) error caused by omitting the flow
between matrix and wellbore and that the effect of matrix porosity on BHP is related to adsorption gas content. When adsorbed
gas accounts for large proportion of the total gas storage in shale formation, matrix porosity only has a very small effect on BHP.
Otherwise, it has obvious influence. This paper can help us understand the complex pressure transient response due to existence of
the adsorbed gas and help petroleum engineers to interpret the field data better.

1. Introduction the pressure transient behaviors or production data analysis


for fractured horizontal wells [9–13]. Clarkson et al. [14] and
Shale formation has both free gas and adsorbed gas [1], Yan et al. [15] use numerical simulation for flow behaviors
which usually is described by Langmuir isotherm equation in single porosity or three porosities model to study flow in
and affects pressure behaviors and production performance shale gas. In order to decrease capillary pressure to improve
significantly. Another characteristic of shale formation is gas flow, wettability alteration by chemical treatments is
the ultralow permeability. Because of that, flow in shale-gas proposed [16].
reservoir leads to several transport mechanisms, such as slip, In the above studies, the dual porosity and dual perme-
transition, and molecular diffusion, which depends on the ability (DPDP) model, in which there are flow in matrix
value of Knudsen number [2–7]. blocks and fracture blocks and flow between them, is seldom
Due to the incomplete development of the natural frac- mentioned.
ture network in shale, natural fracture network is not well In this paper, we first establish a mathematical dual
interconnected and cannot form effective flow channels. In porosity and dual permeability shale-gas flow model incor-
most cases, economic production is possible only if a very porating the gas retention and slip flow mechanism. Then the
complex hydraulic fracture network is created that effectively nonlinear pressure equation is solved by numerical way based
connects a huge reservoir surface area to the wellbore [8]. on PBEI grid. Lastly, we use the numerical model to study the
Therefore, the shale formation consists of matrix, natural effect of flow between matrix and fracture, effect of natural
fracture, and hydraulic fracture. Many researchers use analyt- fracture porosity, and effect of matrix porosity on pressure
ical or semianalytical solution of dual porosity model to study transient response for wells in shale-gas reservoir.
2 Journal of Chemistry

2. Flow Model Scheme of Shale Gas According to the definition of mean free path of gas mole-
cules, Knudsen number is expressed as
2.1. Flow Model. Taking the adsorbed gas as accumulation
term and using apparent permeability to describe the slip 𝜇 𝜋𝑅𝑇
𝐾𝑛 = √ , (6)
flow, we can obtain the following equation for matrix system 𝑑𝑝 2𝑀
and fracture system, respectively:
where 𝜇 is the gas viscosity in Pa⋅s, 𝑅 = 8314 J/kmol/K is
𝑘𝑚 𝑝𝑚 − 𝑝𝑓 the universal gas constant, 𝑇 is the absolute temperature in
∇⋅[ (∇𝑝𝑚 − 𝛾𝑔 ∇𝑍)] − 𝛼𝑘𝑚 K, and 𝑀 is the molecular mass of gas in kg/kmol. Civan [4]
𝜇𝑔 𝐵𝑔 𝜇𝑔 𝐵𝑔
estimates the equivalent hydraulic radius:
(1)
𝜕 𝜙𝑚 𝜕
= ( ) + (𝜌𝑠 𝑉ads ) + 𝑞𝑔𝑠𝑐,𝑚 , 𝑘
𝜕𝑡 𝐵𝑔 𝜕𝑡 𝑑 = 2√2𝜏√ , (7)
𝜙
𝑘𝑓 𝑝𝑚 − 𝑝𝑓 where 𝜏 is the tortuosity and 𝜙 is the porosity of porous media.
∇⋅[ (∇𝑝𝑓 − 𝛾𝑔 ∇𝑍)] + 𝛼𝑘𝑚
𝜇𝑔 𝐵𝑔 𝜇𝑔 𝐵𝑔 According to Schaaf and Chambre [19], free molecular
(2) flow happens if 𝐾𝑛 > 10, when collision between the gas
𝜕 𝜙𝑓 molecules and the pore wall dominates. For 0.1 ≤ 𝐾𝑛 ≤ 10,
= ( ) + 𝑞𝑔𝑠𝑐,𝑓 ,
𝜕𝑡 𝐵𝑔 it is considered as transition regime. For 10−3 < 𝐾𝑛 < 0.1,
the flow at the wall cannot be neglected, and slip flow exists.
where subscript 𝑚 denotes matrix system, subscript 𝑓 For 𝐾𝑛 < 10−3 , collision between molecules and the pore wall
denotes natural fracture system, 𝜌𝑠 is shale density in Kg/m3 , can be neglected, and the flow behaviors can be described by
𝐵𝑔 is formation volume factor of gas in fraction, 𝜙 is porosity Darcy’s law. In this paper, the flow regime is limited to slip
in fraction, 𝑉ads is adsorption content in m3 /Kg, and 𝑞𝑔𝑠𝑐,𝑚 flow.
and 𝑞𝑔𝑠𝑐,𝑓 are volume flow rate of the well divided by the
volume of the well grid in 1/s. 2.3. Isotherm Sorption of Gas. The amount of adsorption in
Many attempts have been made in the past to define the (1) can be modeled by the Langmuir isotherm equation:
shape factor. The classical formula is given by Kazemi [17] as 𝑉𝐿 𝑝
follows: 𝑉ads = , (8)
𝑝𝐿 + 𝑝
1 1 1 where 𝑉ads is the standard volume of gas adsorbed per unit
𝛼 = 4( 2
+ 2 + 2 ), (3)
𝐿𝑥 𝐿𝑦 𝐿𝑧 shale mass in m3 /Kg under the pressure 𝑝, 𝑉𝐿 denotes the
ultimate adsorption amount in m3 /Kg, 𝑝 is the gas pressure
where 𝐿 𝑥 , 𝐿 𝑦 , and 𝐿 𝑧 are the natural fracture spacing in 𝑥, 𝑦, in Pa, and 𝑝𝐿 is the pressure when adsorption reaches half of
and 𝑧 directions in meter, respectively, and 𝛼 is shape factor the ultimate adsorption amount in Pa.
to describe the interporosity flow between matrix and natural
fracture. In this paper, for simplicity, we suppose that 𝐿 𝑥 , 𝐿 𝑦 ,
and 𝐿 𝑧 are equal and just use 𝐿 𝑥 to express the natural fracture 3. Numerical Calculation Scheme
spacing. PEBI grids can overcome structure grid disadvantages, such
Let 𝑄 in m3 /s be the production of the well at surface con- as inflexibility in description of geologic features, inaccurate
dition, and considering wellbore storage, the flow equation in representation of multiwell locations, and grid orientation,
the wellbore is given by and are especially widely used in numerical well test [20–23].
PEBI grid is actually Voronoi block, which is defined as the
2𝜋𝑘ℎ 𝐶 𝑛+1 𝑛
(𝑝 − 𝑝wf ) − (𝑝 − 𝑝wf ) = 𝑄, region of space that is closer to its grid point than to any other
𝜇𝑔 𝐵𝑔 [ln (𝑟𝑒 /𝑟𝑤 ) + 𝑆] 𝑖 Δ𝑡 wf grid points, and its boundaries are normal to lines joining the
(4) nodes on the two sides of each boundary.
Figure 1(a) gives the grids of horizontal well with 5
where 𝑝wf is BHP in Pa, 𝑟𝑒 is equivalent well radius in m and hydraulic fractures. Figure 1(b) gives the fracture grids and
𝑟𝑤 is actual well radius in m, 𝐶 is wellbore storage in m3 /Pa, horizontal well grids in detail. In order to capture the
and 𝑛 is time step. characteristics of the transient flow, the grid points are
distributed by radial growth near the well and fractures. The
2.2. Apparent Permeability. In order to describe the flow well and fractures are modeled as grids and treated as infinite
mechanism caused by the ultralow permeability, the intrinsic conductivity [24].
permeability in (1) and (2) is replaced by apparent permeabil- Equations (1), (2), and (4) form an equation system. There
ity, which can be described with Knudsen number for the slip are a total of 2 variables of 𝑝𝑓 , 𝑝𝑚 for a grid and one variable
flow [18] as follows: for one well producing under constant flow rate.
Nonlinear pressure equations (7) and (8) are implicitly
4𝐾𝑛 solved by using the matrix solver GMRES (generalized
𝑘 = 𝑘0 (1 + ). (5)
1 + 𝐾𝑛 minimal residual method) [25].
Journal of Chemistry 3

Hydraulic fractures

Fractured horizontal well

Horizontal well

(a) Grid (b) Part of the horizontal well and fractures

Figure 1: Hydraulically fractured horizontal well grid without cell points in 2D form.

Table 1: Data used in validation. 20.00

Name Value Unit


Reservoir size 810 × 810 m 19.99
Boundary condition Closed
Initial pressure 20 MPa Pressure (MPa)
19.98
Reservoir thickness 30 m
Rock compressibility 0.0000145 1/MPa
∘ 19.97
Temperature 122 F
Gas gravity 0.7
Half of hydraulic fracture 75 m 19.96
Natural fracture space 10 m
Fracture 0 50 100 150 200
Permeability 700 mD Time (day)
Porosity 0.05 STARS, kf = 40000
Matrix Our code, infinite conductivity
Permeability 0.1 mD
Figure 2: Comparison of bottom hole flowing pressure.
Porosity 0.25
Horizontal well
Skin factor 0
Well storage 0 m3 /MPa The horizontal wellbore only connects the fractures in
Constant rate (STC) 12000 m3 /day this paper; thus, horizontal well with one fracture is equiva-
Production time 200 Day lent to fractured vertical well. In our numerical code, the well
Dranchuk-Abou- and hydraulic fracture are treated as inner boundary.
Deviation factor 𝑍 Kassem The grid number is 81 × 81 in IMEX model. The size of
equation grid is 10 m. The blocks including the hydraulic fracture are
Gas viscosity Lee equation subdivided into 11×3 grids. The main parameters used in the
validation are shown in Table 1.
The well is in the middle of reservoir and produces gas
4. Simulation and Analysis at a constant flow rate of 12000 m3 /day in a closed boundary
reservoir. The BHP is compared between our code and STARS
4.1. Validation. We use the IMEX model of STARS software as shown in Figure 2. The almost overlap of the two output
to validate our core codes. STARS is a commercial software curves validates our code.
developed by CMG and is widely accepted in petroleum
industry. The code we used here to compare with STARS
(2012 version) is only the conventional mass balance equation 4.2. Reservoir Parameters and Grid. Table 2 gives the main
containing gas phase flowing in DPDP and without adsorp- parameters used for the simulation.
tion and apparent permeability correction. The conventional Table 3 gives parameters about the DPDP and hydraulic
balance equation is the core part of our model. Our code will fractures. The total simulation time is 20000 days (about
be convincing if the core part of our model is validated. 54.8 years). Gas viscosity and volume formation factor are
4 Journal of Chemistry

Table 2: Simulation parameters (if not specified additionally). 1.157 MPa under dual well model. If well-matrix flow is
omitted, the BHP error is 9.03% (Table 4).
Name Value Unit
When the fracture permeability is 100 times of the matrix
Reservoir size 80000 × 80000 m permeability (Figure 3(b)), the BHP difference at 10000th day
Initial pressure 20 MPa is 0.0146 MPa. If matrix-wellbore flow is omitted, the BHP
Reservoir thickness 30 m error is 1.004% (Table 4).
Rock compressibility 0.00015 1/MPa By comparison of the BHP error with the ratio of 𝑘𝑚 /𝑘𝑓 ,
Langmuir adsorption 𝑉𝐿 0.02 Std m3 /Kg we know that the ratio of the matrix permeability to the
Langmuir pressure 𝑝𝐿 7.5 MPa fracture permeability approximately decides the BHP error
Horizontal well
caused by omitting the matrix-wellbore flow.
The corresponding pressure change and pressure deriva-
Skin factor 0
tive comparison is shown in Figures 3(c) and 3(d).
Well storage 10 m3 /MPa In Figure 4, except the natural fracture spacing of 𝐿 𝑥 =
Constant rate 10000 m3 /day 10 m, other data are the same as data in Figure 3, from which
Production time 10000 Day we can see the influence of interporosity flow ability on the
Shut time 10000 Day BHP error. The BHP error for 𝑘𝑓 = 10𝑘𝑚 in Figure 4(a) is
Deviation factor 𝑍 Dranchuk-Abou-Kassem equation 9.26% and BHP error for 𝑘𝑓 = 100𝑘𝑚 in Figure 4(b) is 1.028%,
Gas viscosity Lee equation which is shown in Table 4. The BHP errors in Figure 4 are
correspondingly larger than that in Figure 3, which is caused
by smaller interporosity ability.
Table 3: DPDP data (if not specified additionally). When interporosity ability is small, due to the large
permeability of fracture system, some gas needs to flow
Natural from matrix system to fracture system. If the interporosity
Matrix data Hydraulic fracture data
fracture data ability is not large enough, larger pressure difference between
Three hydraulic fractures with matrix and fracture is required, which causes larger error
𝑘𝑚 = 0.001 mD, 𝑘𝑓 = 0.1 mD,
40 meters long and fracture if the matrix-wellbore flow is omitted. Therefore, the low
𝜙𝑚 = 0.1 𝜙𝑓 = 0.001
spacing is 280 meters interporosity flow ability can enlarge the error when we omit
the matrix-wellbore flow.
In Figure 5, the constant flow rate increases from
Table 4: Error percentage of BHP without flow between wellbore
and matrix.
10000 m3 /day to 50000 m3 /day compared to the parameters
in Figure 4. In Figure 5(a), the fracture permeability is 10
Figure 𝑄 Error times of the matrix permeability, and the BHP difference
Parameter
percentage at 10000th day is 0.77247 MPa. The BHP error is 12.6% if
𝐿 𝑥 = 1 m, matrix-wellbore flow is neglected. In Figure 5(b), the fracture
Figure 3(a) 10000 m3 /day 9.03%
𝑘𝑓 = 10𝑘𝑚 permeability is 100 times of the matrix permeability, the BHP
𝐿 𝑥 = 10 m, difference at 10000th day is about 0.09457 MPa, and the BHP
Figure 4(a) 10000 m3 /day 9.26%
𝑘𝑓 = 10𝑘𝑚 error is 1.1924% if the flow between wellbore and matrix is
𝐿 𝑥 = 10 m, omitted.
Figure 5(a) 50000 m3 /day 12.6%
𝑘𝑓 = 10𝑘𝑚 Compared to the value of BHP error in Figure 4 (see
𝐿 𝑥 = 1 m, Table 4), the BHP error under 𝑄 = 50000 m3 /day increases,
Figure 3(b) 10000 m3 /day 1.004%
𝑘𝑓 = 100𝑘𝑚 which shows that large flow rate leads to the large BHP error.
𝐿 𝑥 = 1 m, We also find that the increase of BHP error under 𝑘𝑓 = 10𝑘𝑚
Figure 4(b) 10000 m3 /day 1.028%
𝑘𝑓 = 100𝑘𝑚 is larger than the increase of BHP error under 𝑘𝑓 = 100𝑘𝑚 ,
𝐿 𝑥 = 10 m, which shows that large ratio of 𝑘𝑓 /𝑘𝑚 decreases BHP error if
Figure 5(b) 50000 m3 /day 1.19% matrix-wellbore flow is omitted.
𝑘𝑓 = 100𝑘𝑚
Table 4 gives the BHP error comparison for Figures 3, 4,
and 5. When natural fracture spacing increases from 1 m to
10 m under 𝑄 = 10000 m3 /day, the BHP error increases from
calculated with PVT [26, 27] and are changing with pressure 9.03% to 9.26% under 𝑘𝑓 = 10𝑘𝑚 , which means that BHP
and obtained through iterative process. error increases with decrease of interporosity flow ability.
When constant flow rate increases 5 times, the BHP error
4.3. Effect of Flow between Matrix and Wellbore with increases from 9.26% to 12.6% under 𝑘𝑓 = 10𝑘𝑚 , while
Adsorption. In Figure 3, ultimate adsorption capacity (𝑉𝐿 ) is the BHP error only increases from 1.028% to 1.19% under
0.02 m3 /Kg, and constant flow rate (𝑄) is 10000 m3 /day. The 𝑘𝑓 = 100𝑘𝑚 , which has a smaller increase rate than that
other parameters are shown in the figures. under 𝑘𝑓 = 10𝑘𝑚 . Therefore, when ratio of 𝑘𝑓 /𝑘𝑚 is large,
Figure 3(a) shows that when fracture permeability is 10 neglecting of the matrix-wellbore flow only causes very small
times of the matrix permeability, the BHP difference at BHP error. However, if ratio of 𝑘𝑓 /𝑘𝑚 is relatively large, the
10000th day is about 0.11 MPa. The pressure change is about matrix-wellbore flow should not be omitted.
Journal of Chemistry 5

20.2
20.2
20.0
20.0
19.8 19.8
Pressure (MPa)

Pressure (MPa)
19.6 19.6

19.4 19.4
Poro f = 0.0001 19.2 Poro f = 0.0001
19.2
Poro m = 0.05 Poro m = 0.05
19.0
19.0 kf = 0.1 mD kf = 0.1 mD
18.8
Lx = 1 m
18.8 18.6
18.6 18.4
0 5000 10000 15000 20000 0 5000 10000 15000 20000
Time (day) Time (day)
L x = 1 m, kf = 10km , dual well L x = 1 m, kf = 100km , dual well
L x = 1 m, kf = 10km , not dual well L x = 1 m, kf = 100km , not dual well

(a) BHP comparison for 𝑘𝑓 = 10𝑘𝑚 (b) BHP comparison for 𝑘𝑓 = 100𝑘𝑚

10
1
1
0.1
dp, dp󳰀 (MPa)
dp, dp󳰀 (MPa)

0.1

0.01
0.01

1E − 3
1E − 3

1E − 4
1E − 4
0.01

0.1

10

100
1E − 7

1E − 6

1E − 5

1E − 4

1E − 3

1000

10000
10

100

1000

10000
0.01

0.1

1
1E − 7

1E − 6

1E − 5

1E − 4

1E − 3

Time (day) Time (day)


kf = 10km kf = 10km , not dual well kf = 100km kf = 100km , not dual well
kf = 10km kf = 10km , not dual well kf = 100km kf = 100km , not dual well

(c) Derivative comparison for 𝑘𝑓 = 10𝑘𝑚 (d) Derivative comparison for 𝑘𝑓 = 100𝑘𝑚

Figure 3: Effect of flow during matrix and wellbore on BHP under 𝐿 𝑥 = 1 m and 𝑄 = 10000 m3 /day.

4.4. Effect of Porosity on Transient Pressure Response with 0.0001. This is because the natural fracture system is the main
Adsorption Gas flow channel. Larger fracture porosity can make the linear
flow in nature fracture last longer and linear flow behavior
4.4.1. Effect of Fracture Porosity on Transient Pressure Response. becomes obvious.
Figure 6 shows the effect of natural fracture porosity on
pressure transient responses for a horizontal well with 3 4.4.2. Effect of Matrix Porosity on Transient Pressure Response.
hydraulic fractures with ultimate adsorption capacity of 𝑉𝐿 = Figure 7 shows the effect of matrix porosity on pressure
0.1 m3 /Kg and natural fracture spacing of 𝐿 𝑥 = 1 m. transient responses of a horizontal well with 3 hydraulic
The effect of fracture porosity on BHP is so small that fractures for ultimate adsorption capacity of 𝑉𝐿 = 1 m3 /Kg,
it is not easy to identify the difference in Figure 6(a). From natural fracture spacing of 𝐿 𝑥 = 1 m.
pressure derivative curves in Figure 6(b), we can see the Figure 7 shows that the matrix porosity has minimal
difference of the early linear flow regime. However, the time effect on BHP when the gas ultimate adsorption capacity is
of the difference only lasts 0.01 days. When fracture porosity 1 m3 /Kg. Matrix system is the storage space and its porosity
is 0.001, the early linear flow regime is obvious, while the should affect behaviors of BHP. However, in shale gas, the
early linear flow regime is not clear under fracture porosity of situation is different. When pressure drops, the adsorbed
6 Journal of Chemistry

20.0 20.0

Pressure (MPa)
Pressure (MPa)

19.5
19.5

Poro f = 0.0001 Poro f = 0.0001


19.0
Poro m = 0.05 Poro m = 0.05
19.0
kf = 0.1 mD kf = 0.1 mD

18.5

0 5000 10000 15000 20000 0 5000 10000 15000 20000


Time (day) Time (day)
L x = 10 m, dual well, kf = 10km L x = 10 m, dual well, kf = 100km
L x = 10 m, kf = 10km L x = 10 m, kf = 100km

(a) BHP comparison for 𝑘𝑓 = 10𝑘𝑚 (b) BHP comparison for 𝑘𝑓 = 100𝑘𝑚

Figure 4: Effect of flow during matrix and wellbore on BHP under 𝐿 𝑥 = 10 m and 𝑄 = 10000 m3 /day.

20 20

19
18
18
Pressure (MPa)

Pressure (MPa)

17
16
16

15 14

14
12
13

0 5000 10000 15000 20000 0 5000 10000 15000 20000


Time (day) Time (day)
L x = 10 m, dual well, kf = 10km L x = 10 m, dual well, kf = 100km
L x = 10 m, kf = 10km L x = 10 m, kf = 100km

(a) BHP comparison for 𝑘𝑓 = 10𝑘𝑚 (b) BHP comparison for 𝑘𝑓 = 100𝑘𝑚

Figure 5: Effect of flow during matrix and wellbore under 𝐿 𝑥 = 10 m and 𝑄 = 50000 m3 /day.

gas desorbed. If the ultimate adsorption capacity is big, the Figure 9 gives the case of no adsorption gas. The BHP
magnitude of the free gas obtained from the adsorption gas is difference in Figure 9(a) is 0.46 MPa, which is much larger
much bigger than the free gas from the matrix porosity, which than the BHP difference under 𝑉𝐿 = 0.05 m3 /Kg in
reduces the effect of the matrix porosity. Figure 8(a). The pressure change and pressure derivative in
When the ultimate adsorption capacity is small, the Figures 7, 8, and 9 together clearly show the effect of matrix
desorbed gas will have less effect on the BHP, and the effect porosity under different ultimate adsorption capacities.
of the matrix porosity on BHP will appear. Figure 8 gives
effect of matrix porosity on pressure response with ultimate 5. Conclusion
adsorption capacity of 0.05 m3 /Kg. With small adsorption
content, the BHP difference in Figure 8(a) is 0.28 MPa at the In this paper, we first developed a dual porosity and dual
end of the production period due to the porosity difference permeability model for a multistage fractured horizontal
between matrix systems. When adsorption content decreases, well in shale-gas reservoirs incorporating the slippage cor-
the free gas will account for more proportion of the flow rate. rected gas permeability and gas adsorption. Then we devel-
Therefore, matrix porosity has more influence on the pressure oped fully implicit numerical simulation model using PEBI
transient response. grid and finally conducted parametric studies to quantify
Journal of Chemistry 7

10
20.0 L x = 1 m, VL = 0.1 m3 /kg
L x = 1 m, VL = 0.1 m3 /kg

19.8 1

log(dp), log(dp)󳰀 (MPa)


19.6
0.1
pwf (MPa)

19.4
0.01
19.2

19.0 1E − 3

18.8 Early linear flow


1E − 4
0

2000

4000

6000

8000

10000

12000

14000

16000

18000

1E − 7

1E − 6

1E − 5

1E − 4

1E − 3

0.01

0.1

10

100

1000
Time (day) Time (day)
Poro f = 0.001 Poro f = 0.001
Poro f = 0.0001 Poro f = 0.0001

(a) BHP for different fracture porosities (b) Pressure derivative of the pressure drawdown period

Figure 6: Effect of porosities of fracture system on pressure response for 𝑉𝐿 = 0.1 m3 /Kg, 𝜑𝑚 = 0.05, and 𝐿 𝑥 = 1 m.

10
20.1 Poro f = 0.0001, L x = 1 m, VL = 1 m3 /kg
3
Poro f = 0.0001, L x = 1 m, VL = 1 m /kg
20.0 1
19.9
0.1
19.8
dp, dp󳰀 (MPa)
pwf (MPa)

19.7 0.01
19.6
1E − 3
19.5

19.4 1E − 4
19.3
1E − 5
19.2
1E − 7

1E − 6

1E − 5

1E − 4

1E − 3

0.01

0.1

10

100

1000
0

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

Time (hour) Time (day)


Poro m = 0.05 Poro m = 0.01
Poro m = 0.01 Poro m = 0.05

(a) BHP for different matrix porosities (b) Pressure derivative for different matrix porosities

Figure 7: Effect of matrix porosities on pressure response for 𝑉𝐿 = 1 m3 /Kg, 𝐿 𝑥 = 1 m, and 𝜑𝑓 = 0.0001.

the transient pressure behaviors for flow in shale-gas forma- content. The desorbed free gas in matrix can offset
tion. Our conclusions could be summarized as below. the effect of the matrix porosity. When adsorbed gas
(1) The ratio of fracture permeability to matrix perme- accounts for a big part of total gas storage in shale
ability approximately decides the BHP error caused formation, the change of the matrix porosity only has
by omitting the flow between the matrix and the well- a very small effect of the total gas, which leads to a
bore. The BHP error is also related to the interporosity very small effect on BHP.
flow ability. The smaller interporosity flow ability is,
the larger BHP error is. (3) Although effect of natural fracture system porosity
(2) In shale-gas formation with adsorption gas, the effect on BHP is not discernable, it influences the flow
of matrix porosity on BHP is related to adsorption gas behaviors during the early stage significantly.
8 Journal of Chemistry

20.0
Poro f = 0.0001, L x = 1 m, VL = 0.05 m3 /kg 20.0
19.8
19.6
19.5
19.4
19.2
pwf (MPa)

pwf (MPa)
19.0
19.0
18.8
18.5
18.6
18.4
18.0
18.2
18.0
17.5
0 5000 10000 15000 20000

16000

18000
0

2000

4000

6000

8000

14000
10000

12000

Time (day)
Time (day) Poro m = 0.01, Poro f = 0.0001, L x = 1, VL = 0
Poro m = 0.05, Poro f = 0.0001, L x = 1, VL = 0
Poro m = 0.05
Poro m = 0.01 (a) BHP for different matrix porosities
(a) BHP for different matrix porosities
1
0.5 Poro f = 0.0001, L x = 1 m, VL = 0.05 m3 /kg

0.0
0.1
dp, dp󳰀 (MPa)
log(dp), log(dp)󳰀 (MPa)

−0.5

−1.0 0.01
−1.5

−2.0 1E − 3

−2.5
1E − 4
−3.0
1E − 7

1E − 6

1E − 5

1E − 4

1E − 3

0.01

0.1

10

100

1000
−3.5
−5 0 5 Time (day)
Time (day)
Poro m = 0.01, Poro f = 0.0001
Poro m = 0.01 Poro m = 0.01, Poro f = 0.0001
Poro m = 0.05 Poro m = 0.05, Poro f = 0.0001
Poro m = 0.05, Poro f = 0.0001
(b) Pressure derivative for different matrix porosities
(b) Pressure derivative for different matrix porosities
Figure 8: Effect of matrix porosities on pressure response with 𝑉𝐿 =
0.05 m3 /Kg, 𝐿 𝑥 = 1 m, and 𝜑𝑓 = 0.0001. Figure 9: Effect of matrix porosities on pressure response with 𝑉𝐿 =
0 m3 /Kg.

This paper can help us understand the complex pressure Project (2011ZX05009-006), and CAS Strategic Priority
transient response due to existence of adsorbed gas and help Research Program (XDB10030402).
petroleum engineers to interpret the field data better.

References
Conflict of Interests
[1] D. J. K. Ross and R. M. Bustin, “Impact of mass balance cal-
The authors declare that there is no conflict of interests culations on adsorption capacities in microporous shale gas
regarding the publication of this paper. reservoirs,” Fuel, vol. 86, no. 17-18, pp. 2696–2706, 2007.
[2] F. Javadpour, “Nanopores and apparent permeability of gas
Acknowledgments flow in mudrocks (shales and siltstone),” Journal of Canadian
Petroleum Technology, vol. 48, no. 8, pp. 16–21, 2009.
This work was sponsored by Major State Basic Research [3] A. Beskok and G. E. Karniadakis, “Report: a model for flows in
Development Program of China (973 Program) (no. channels, pipes, and ducts at micro and nano scales,” Microscale
2011CB707305), National Key Science and Technology Thermophysical Engineering, vol. 3, no. 1, pp. 43–77, 1999.
Journal of Chemistry 9

[4] F. Civan, “Effective correlation of apparent gas permeability in [21] C. L. Pinzon, H.-Y. Chen, and L. W. Teufel, “New MexicoTech
tight porous media,” Transport in Porous Media, vol. 82, no. 2, numerical well test analysis of stress-sensitive reservoirs,” in
pp. 375–384, 2010. Proceedings of the SPE Rocky Mountain Petroleum Technology
[5] A. Sakhaee-Pour and S. L. Bryant, “Gas permeability of shale,” Conference, vol. 71034, Keystone, Colo, USA, May 2001.
SPE 146944, 2011. [22] O. R. Alcalde and L. W. Teufel, “Diagnosis of formation damage
[6] C. M. Freeman, G. J. Moridis, and T. A. Blasingame, “A numer- by rock deformation/compaction through numerical well-test
ical study of microscale flow behavior in tight gas and shale gas simulations,” in Proceedings of the SPE International Sympo-
reservoir systems,” Transport in Porous Media, vol. 90, no. 1, pp. sium and Exhibitionon Formation Damage Control, SPE 98053,
253–268, 2011. Lafayette, La, USA, February 2006.
[7] C. Niu, Y.-Z. Hao, D. Li, and D. Lu, “Second-order gas- [23] W. S. Zha, D. L. Li, D. T. Lu, and X. H. Kong, “PEBI grid division
permeability correlation of shale during slip flow,” SPE Journal, in inter-well interference area,” Acta Petrolei Sinica, vol. 29, no.
vol. 19, no. 5, pp. 786–792, 2014. 5, pp. 742–746, 2008 (Chinese).
[24] H. Cinco-Ley and F. Samaniego, “ Transient pressure analysis
[8] C. L. Cipolla, E. P. Lolon, J. C. Erdle, and B. Rubin, “Reservoir
for fractured wells,” Journal of Petroleum Technology, vol. 33, no.
modeling in shale-gas reservoirs,” SPE Reservoir Evaluation and
9, pp. 1749–1766, 1981.
Engineering, vol. 13, no. 4, pp. 638–653, 2010.
[25] Y. Saad and M. H. Schultz, “GMRES: a generalized minimal
[9] J. E. Warren and P. J. Root, “The behavior of naturally fractured
residual algorithm for solving nonsymmetric linear systems,”
reservoirs,” Society of Petroleum Engineers Journal, vol. 3, pp.
SIAM Journal on Scientific and Statistical Computing, vol. 7, no.
245–255, 1963.
3, pp. 856–869, 1986.
[10] F. Medeiros Jr., E. Ozkan, and H. Kazemi, “Productivity and
[26] P. M. Dranchuk and J. H. Abou-Kassem, “Calculation of Z
drainage area of fractured horizontal wells in tight gas reser-
factors for natural gases using equations of state,” Journal of
voirs,” SPE Reservoir Evaluation & Engineering, vol. 11, no. 5, pp.
Canadian Petroleum Technology, vol. 14, no. 3, pp. 34–36, 1975.
902–911, 2008.
[27] A. L. Lee, M. H. Gonzalez, and B. E. Eakin, “The viscosity of
[11] F. Medeiros, B. Kurtoglu, E. Ozkan, and H. Kazemi, “Analysis of natural gases,” Journal of Petroleum Technology, vol. 18, no. 8,
production data from hydraulically fractured horizontal wells pp. 997–1000, 2013.
in shale reservoirs,” SPE Reservoir Evaluation and Engineering,
vol. 13, no. 3, pp. 559–568, 2010.
[12] E. Ozkan, M. Brown, R. Raghavan, and H. Kazemi, “Compari-
son of fractured-horizontal-well performance in tight sand and
shale reservoirs,” SPE Reservoir Evaluation and Engineering, vol.
14, no. 2, pp. 248–259, 2011.
[13] A. A. Hasan, M. A. Anas, and R. A. Wattenbarger, “Application
of linear flow analysis to shale gas wells-field cases,” in Pro-
ceedings of the SPE Unconventional Gas Conference, Paper SPE
130370, Pittsburgh, Pa, USA, February 2010.
[14] C. R. Clarkson, M. Nobakht, D. Kaviani, and T. Ertekin,
“Production analysis of tight-gas and shale-gas reservoirs using
the dynamic-slippage concept,” SPE Journal, vol. 17, no. 1, pp.
230–242, 2012.
[15] B. Yan, Y. Wang, and J. E. Killough, “Beyond dual-porosity
modeling for the simulation of complex flow mechanisms in
shale reservoirs,” in Proceedings of the SPE Reservoir Simulation
Symposium, The Woodlands, Tex, USA, February 2013.
[16] Y.-L. Wang, L. Ma, B.-J. Bai, G.-C. Jiang, J.-F. Jin, and Z.-
B. Wang, “Wettability alteration of sandstone by chemical
treatments,” Journal of Chemistry, vol. 2013, Article ID 845031,
8 pages, 2013.
[17] H. Kazemi, “Pressure transient analysis of naturally fractured
reservoirs with uniform fracture distribution,” in Proceedings of
the 43rd Annual Fall Meeting, Houston, Tex, USA, September-
October 1968.
[18] F. Civan, C. S. Rai, and C. H. Sondergeld, “Shale-gas per-
meability and diffusivity inferred by improved formulation of
relevant retention and transport mechanisms,” Transport in
Porous Media, vol. 86, no. 3, pp. 925–944, 2011.
[19] S. A. Schaaf and P. A. Chambre, Flow of Rarefied Gases, Prin-
ceton University Press, 1961.
[20] C. L. Palagi and K. Aziz, “Use of Voronoi grid in reservoir
simulation,” in Proceedings of the Annual Technical Conference
and Exhibition, Paper SPE 22889, Dallas, Tex, USA, October
1991.
Hindawi Publishing Corporation
Journal of Chemistry
Volume 2015, Article ID 341616, 10 pages
http://dx.doi.org/10.1155/2015/341616

Research Article
Laboratory Measurement and Interpretation of the Changes of
Physical Properties after Heat Treatment in Tight Porous Media

Yili Kang, Mingjun Chen, Lijun You, and Xiangchen Li


State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum University, Chengdu 610500, China

Correspondence should be addressed to Mingjun Chen; chenmj1026@163.com

Received 25 September 2014; Accepted 30 December 2014

Academic Editor: Sachin Jangam

Copyright © 2015 Yili Kang et al. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Prevention of water blocking and optimization of multiscale flow channels will increase gas production of tight reservoirs. Physical
properties of samples from representative tight gas reservoirs were measured before and after high temperature treatment. Results
show that, with the increase of treatment temperature, mass decreases, acoustic transit time increases, and permeability and porosity
increase. Permeability begins to increase dramatically if treatment temperature exceeds the threshold value of thermal fracturing,
which is 600∼700∘ C, 500∼600∘ C, 300∼500∘ C, and 300∼400∘ C for shale, mudstone, tight sandstone, and tight carbonate rock,
respectively. Comprehensive analyses indicate that the mechanisms of heat treatment on tight porous media include evaporation
and dehydration of water, change of mineral structure, generation of microfracture, and network connectivity. Meanwhile, field
implementation is reviewed and prospected. Interpretations indicate that, according to the characteristics of multiscale mass
transfer in tight gas formation, combining heat treatment with conventional stimulation methods can achieve the best stimulation
result.

1. Introduction Therefore, effectiveness of conventional stimulation method


is still not so good for tight gas reservoirs. A certain kind of
Tight gas reservoirs, including shale gas, tight sandstone gas, stimulation method is urgent to be developed, and formation
and tight carbonate gas reservoirs, are playing an increasingly heat treatment, based on thermal property of rock, is devel-
important role in the growth of natural gas reserves and oped as an innovative well stimulation technique, which is
energy supply. Gas production from tight gas reservoir is a focused on prevention of water blocking and generation or
process of multiscale mass transfer. In order to maximize propagation of small-scale fractures.
the ability of gas production, it is necessary to optimize all Formation heat treatment, which is recognized as a state-
the scales of mass transfer [1–4]. Single hydraulic fracturing of-the-art technology for near-wellbore formation [8], might
mainly generates great-scale fractures based on preexisting play a significant role in well stimulation. On the one hand,
fractures, but for tight gas reservoir, many of the reserves the water in pores would be removed perfectly, due to the
occur in small-scale pores. Meanwhile, in order to maximize evaporation or dehydration of water at high temperature.
fracture propagation, much more fracturing fluid is required Therefore, formation damage like water blocking could be
to be pumped into formation. If the fluid system contains prevented, and permeability of rock is enhanced by remov-
fresh water, the potential of clay swelling and migration ing water in gas flow channel [9–11]. On the other hand,
would be extremely high. Considering that huge volume of induced fracture generates, preexisting fracture propagates,
liquid into formation is very difficult to flow back from the and finally various kinds of fractures connect to be network
pores which have the ultrasmall volume and poor connec- under the action of thermal stress at high temperature [12–
tivity, various types of formation damage would be easily 14]. As a result, the ability of mass transfer can be enhanced
induced and the production cannot be satisfactory [5–7]. dramatically through formation heat treatment. Compared
2 Journal of Chemistry

with conventional stimulation methods, like hydraulic frac- by LGPM700 with transient pulse decay method for perme-
turing and acid treatment, the advantages of formation heat ability less than 0.01 × 10−3 𝜇m2 . Initial physical properties
treatment mainly consist in the following. of these samples under conventional conditions (435 psi and
20∘ C) are presented in Table 1.
(1) It prevents water blocking by evaporation of blocked
water and dehydration of clay structure [15]. 2.2. Experimental Temperature Setting. X-ray diffraction
(2) It enhances permeability and porosity in microscale (XRD) analyses of these samples show that clay mineral and
uniformly. Meanwhile, it makes anisotropy of rock in quartz are the main minerals (Table 2), except that tight
mesoscale and macroscale under control. carbonate rock is mainly composed of dolomite. The clay
minerals of samples mainly consist of illite, chlorite, kaolinite,
(3) It accelerates desorption and diffusion of methane in
and mixed-layer mineral of illite/smectite, except that shale
matrix, especially for rocks rich in organic [16].
does not contain kaolinite. Within a certain temperature
(4) Formation heat treatment does not need water source range, structure of clay minerals would be destroyed. Besides
and it also does not contaminate groundwater. It is clay minerals, thermal reactions can occur in other mineral
totally eco-friendly. constituents such as quartz and carbonate. Particularly the
quartz, which has a volume expansion of 2.7% in 2∼5 seconds
However, systematic studies on heat treatment are still when temperature elevates to 573∘ C due to the 𝛼 → 𝛽
rare and the application on stimulation of tight gas formation inversion of quartz, has a dominant effect on the magnitude of
is urgent to understand. The previous studies on formation thermal expansion [19, 22]. Heats of reaction for several main
heat treatment are based on sandstone, and samples’ perme- minerals of samples are concluded in Table 3. According to
ability is relatively high [8, 9, 15, 17]. Also, thermal cracking the temperature range of heat reaction for minerals shown in
in rock, which has been studied a lot in nuclear waste storage, Table 3, various temperature values, that is, 100, 200, 300, 400,
mining technique, HDR geothermal extraction, and stability 500, 600, 700, and 800∘ C, were specified as the temperature
analysis of constructions, is mostly based on granite that set-points to evaluate the effect of heat treatment on physical
is not the natural gas reservoir [13, 18–21]. Therefore, the properties of tight rocks.
effect of high temperature on tight gas reservoirs, which is
important in the development of petroleum industry, needs 2.3. Experimental Equipment and Methods. In order to simu-
to be evaluated. Furthermore, the effect of heat treatment late the in situ anoxic condition, samples with corresponding
on different kinds of tight rocks also still needs to be distin- irreducible water saturation were mounted in a tube furnace
guished. In this work, samples from the representative tight filled with argon gas. Mass of samples was measured before
gas reservoirs were treated under high temperature in argon and after heat treatment by precision electronic balance in
gas environment to simulate the in situ anoxic condition. a controlled humidity oven at 20∘ C and 0% RH. Acoustic
Several lab experimental methods were comprehensively compressional wave (P-wave) and shear wave (S-wave) transit
utilized to investigate the effect of heat treatment on physical time were measured before and after heat treatment under
properties of tight rocks. Then the essence of the changes conventional conditions (435 psi and 20∘ C) through SCMS-
of physical properties for tight rocks after heat treatment J Acoustics-Resistivity Measurement Equipment, which is
is analyzed comprehensively. Lastly, field implementation of researched and developed independently by the State Key
heat treatment process is discussed. Laboratory of Oil and Gas Reservoir Geology and Exploita-
tion of SWPU, and the frequency of ultrasonic wave is
2. Experimental Sample and Procedure 170 kHz. Porosity and permeability were measured before and
after heat treatment through CMS-300 for permeability more
2.1. Samples’ Description. Tight rock samples in this study are than 0.01×10−3 𝜇m2 and through LGPM700 for permeability
taken from three kinds of representative tight gas reservoirs. less than 0.01 × 10−3 𝜇m2 under conventional conditions
Shale samples are from the Longmaxi formation of Lower (435 psi and 20∘ C).
Silurian in Sichuan Basin, which is recognized as the most As is shown in Figure 1, the testing procedures are as
producible and profitable shale gas reservoir in China. Tight follows.
sandstone samples are from Upper Palaeozoic in Permian
(1) Measure one sample’s mass, acoustic transit time,
in Ordos Basin, which is the most giant tight sandstone gas
and porosity and permeability under the conditions
reservoir in China. And the tight carbonate rock sample is
shown above.
from Feixianguan formation of Lower Triassic in northeast
Sichuan Basin, which is the representative tight carbonate (2) Mount the sample in the tube furnace. Fully evacuate
gas reservoir in China. Meanwhile, in order to investigate the sample at 62∘ C to ensure that air, including
the influence of organic matter in thermal process, mudstone adsorbed gas, is excluded from the rock. Then break
samples from Sichuan Basin are selected to be compared with vacuum with argon gas until ambient pressure of
shale samples. These reservoirs are typically characterized sample returns to atmospheric pressure.
by various kinds of pore types, complicated pore struc- (3) Heat the sample to 100∘ C at a rate of 5∘ C/min,
ture, and strong heterogeneity. Porosity and permeability starting with atmospheric temperature (20∘ C). When
were measured by CMS-300 Core-Automatic Determination temperature in tube furnace is up to 100∘ C, the testing
Instrument for permeability more than 0.01 × 10−3 𝜇m2 and temperature would be maintained for 2 h and then
Journal of Chemistry 3

Table 1: Physical properties of samples without any heat treatment.

Permeability
Mass Length Diameter Porosity P-wave transit S-wave transit
Sample number 𝐾0 , Lithology
𝑀0 , g 𝐿 0 , cm 𝐷0 , cm Φ0 , % time Δ𝑇𝑝 , 𝜇s/m time Δ𝑇𝑠 , 𝜇s/m
10−3 um2
Shale1 63.3463 5.400 2.504 3.5 0.00780 221.2963 376.6667 Shale
Shale2 67.0235 7.396 2.482 4.6 0.0256 245.9438 408.8697 Shale
Shale3 42.9417 4.966 2.500 4.7 0.0947 247.8856 432.5413 Shale
Shale4 64.2955 6.040 2.500 3.8 0.0860 232.6159 390.0662 Shale
Sand1 41.0425 3.252 2.502 7.3 0.0167 287.3846 428.9231 Sandstone
Sand2 66.0089 5.268 2.478 4.6 0.0181 209.0288 383.9150 Sandstone
Sand3 59.5467 4.650 2.506 5.8 0.0465 274.0964 456.9707 Sandstone
Sand4 60.1367 4.862 2.504 8.3 0.0571 328.9420 523.6723 Sandstone
Sand5 63.8542 5.308 2.488 8.8 0.0539 269.3178 421.0328 Sandstone
Carbonate 55.8123 3.958 2.496 4.3 0.0135 160.2629 311.9312 Carbonatite
Mud1 44.9856 4.210 2.482 5.3 0.0323 358.7732 482.4061 Mudstone
Mud2 46.8326 4.196 2.484 3.1 0.00711 322.8422 539.8188 Mudstone
Mud3 37.6457 3.392 2.488 2.3 0.00559 415.6342 688.4956 Mudstone

Table 2: Quantitative analyses of minerals by XRD.

Content of minerals, %
Lithology
Clay mineral Quartz K-feldspar Anorthose Calcite Dolomite Siderite Pyrite
Shale 39.96 41.15 3.12 5.12 2.91 4.11 0.00 3.63
Sandstone 24.90 72.89 0.01 0.13 2.06 0.01 0.00 0.00
Carbonatite 7.90 7.70 4.67 2.93 0.00 76.80 0.00 0.00
Mudstone 56.52 11.14 0.00 7.04 25.30 0.00 0.00 0.00

Table 3: Heats of reaction for several minerals [34, 35].

Minerals Temperature range, ∘ C Reaction after heat treatment


Illite and clay mica 125∼250 Loss of hydroscopic water
Mg-chlorite 650 14 Å spacing is intensified
Fe-chlorite 500 14 Å spacing less intense, becoming broad and diffuse
Mixed-layer clays <600 Varies with amounts and types of minerals present
Kaolinite, well crystallized 575∼625 Replacement by amorphous metakaolin
Quartz 573 𝛼 → 𝛽 inversion
Ca-carbonate 700–830 Decomposition

Precision electronic
balance
Controlled humidity
Control oven
valve
Pressure Sample
gauge Acoustic Sample Acoustic
transmitter probe receiver probe

Acoustic signal (SCMS-J)


Argon gas
Vacuum Tube furnace source
source pump
Oscilloscope

CMS-300 LGPM700

Figure 1: Schematic diagram of the experimental apparatus.


4 Journal of Chemistry

decreases to the atmospheric temperature with a rate 1.00


of 5∘ C/min.
0.99
(4) Measure the mass, acoustic transit time, and porosity
and permeability through the same methods and
0.98
conditions of the first procedure.

Mi /M0
(5) Repeat the above steps with heating temperatures of 0.97
200, 300, 400, 500, 600, 700, and 800∘ C, respectively.
During the whole process of the heat treatment, the 0.96
sample is always in argon gas atmosphere.
0.95
3. Experimental Results
0.94
3.1. Effect of Heat Treatment on Mass. Change of mass after 0 200 400 600 800
heat treatment mainly reflects the loss of free water, adsorbed T (∘ C)
water, interlayer water, and constitution water. As is shown Shale1 Shale2 Shale3
in Figure 2, mass of samples tends to decrease as treatment Shale4 Sand1 Sand2
temperature increases. Meanwhile, the evident abrupt change Sand3 Sand4 Sand5
of mass occurs at temperature lower than 200∼300∘ C with Carbonate Mud1 Mud2
Mud3
temperature increasing. The mass is substantially unchanging
if temperature is higher than that range. Figure 2: Normalized mass as a function of treatment temperature
Change of mass varies in different lithologies. Mudstone for all the samples.
is the most affected. After heat treatment under 800∘ C, its
mass decreases as much as 2.38% on average compared with
that before any heat treatment. Magnitude of mass decrease is 2.5
1.85% for shale, 0.55% for tight sandstone, and 0.10% for tight
carbonate rock. 2
ΔT800∘ C /ΔT20∘ C

3.2. Effect of Heat Treatment on Acoustic Transit Time. The- 1.5


oretically, acoustic transit time could increase if the porous
media become less tight. Since evaporation or dehydration 1
of water phase and thermal-induced fracturing are the most
0.5
important mechanisms of heat treatment, change of the
acoustic transit time in this work mainly reflects the change
0
of pore structure. Experimental results show that transit time
Shale1
Shale2
Shale3
Shale4
Sand1
Sand2
Sand3
Sand4
Sand5
Carbonate
Mud1
Mud2
Mud3
for both compressional wave (P-wave, Δ𝑇𝑝 ) and shear wave
(S-wave, Δ𝑇𝑠 ) tends to increase as temperature increases
(Figure 3), but the change is not very remarkable, as well ΔTp
as the change of porosity presented below. Comparing the ΔTs
acoustic transit time after heat treatment at 800∘ C with that
before heat treatment, Δ𝑇𝑝 and Δ𝑇𝑠 increase as much as 1.19 Figure 3: Ratio of acoustic transit time at 800∘ C to that at 20∘ C for
times and 1.14 times for shale, 1.62 times and 1.55 times for all the samples.
tight sandstone, 1.10 times and 1.17 times for tight carbonate
rock, and 1.13 times and 1.18 times for mudstone.
Since the responses of Δ𝑇𝑝 and Δ𝑇𝑠 to temperature are that engineers are most concerned about. The permeability
different, ratio of Δ𝑇𝑠 to Δ𝑇𝑝 (Δ𝑇𝑠 /Δ𝑇𝑝 ) is necessary to be change caused by heat treatment is presented in Figure 5. Per-
concerned with in order to comprehensively analyze the meability tends to increase with the treatment temperature
effect of heat treatment on physical properties, such as pore getting higher. What is more, permeability has an evident
size and fracture propagation [23]. Outcome of the acoustic abrupt change within a certain temperature range. In detail,
transit time measurement shows that change of Δ𝑇𝑠 /Δ𝑇𝑝 does if temperature is lower than the threshold value, increase of
not have obvious regularity as the treatment temperature the permeability is not so obvious, and if the temperature is
increases (Figure 4). Compared with Δ𝑇𝑠 /Δ𝑇𝑝 without any higher than the threshold value, significant increase of the
heat treatment, the value after heat treatment at 800∘ C has permeability occurs. Chen et al. (1999) recognized that per-
a tendency of increase for shale and most of tight sandstone colation model could describe that change behavior perfectly
samples, but for tight carbonate rock and most of mudstone [18].
samples, it has a tendency of decrease. Meanwhile, the variation of permeability by heat treat-
ment is not the same for different lithologies. For the thresh-
3.3. Effect of Heat Treatment on Permeability and Porosity. old temperature, shale is 600∼700∘ C, tight sandstone is 300∼
In general, permeability is one of the physical properties 500∘ C, tight carbonate rock is 300∼400∘ C, and mudstone is
Journal of Chemistry 5

2.5 4.5

4.0
2
3.5
1.5
ΔTs /ΔTp

3.0

Φi /Φ0
1 2.5

0.5 2.0

1.5
0
Shale1
Shale2

Sand1

Sand4
Sand5
Shale3
Shale4

Sand2
Sand3

Carbonate
Mud1
Mud2
Mud3
1.0

0.5
0 100 200 300 400 500 600 700 800
ΔTs /ΔTp at 20∘ C
T (∘ C)
ΔTs /ΔTp at 800∘ C
Shale1 Shale2 Shale3
Figure 4: Ratio of Δ𝑇𝑠 to Δ𝑇𝑝 at 20∘ C and 800∘ C for all the samples. Shale4 Sand1 Sand2
Sand3 Sand4 Sand5
Carbonate Mud1 Mud2
Mud3
90
Figure 6: Normalized porosity as a function of treatment tempera-
80 ture for all the samples.
70
60
50
abrupt change within a range of threshold temperature.
Ki /K0

Meanwhile, the variation of porosity is not the same for


40 different lithologies. Compared with porosity without any
30 heat treatment, the porosity of samples after heat treatment
at 800∘ C increases as much as 1.85∼2.14 times (1.95 times on
20
average) for shale, 1.61∼2.87 times (2.03 times on average)
10 for tight sandstone, 2.15 times for tight carbonate rock, and
0 1.58∼4.05 times (2.82 times on average) for mudstone.
0 100 200 300 400 500 600 700 800
T (∘ C)
4. Discussions
Shale1 Shale2 Shale3
Shale4 Sand1 Sand2
Tight gas reservoir has characteristics such as relatively small
Sand3 Sand4 Sand5
Carbonate Mud1 Mud2 pore, richness in clay minerals or fragile minerals, complex
Mud3 flow paths, and serious anisotropy [24, 25]. Meanwhile,
potential water blocking is quite easy to occur and difficult
Figure 5: Normalized permeability as a function of treatment tem- to prevent from the initial drilling and completion of wellbore
perature for all the samples. to the depletion of reservoir during production. According to
the experimental results shown above, physical properties of
samples change remarkably after high temperature treatment.
500∼600∘ C. Compared with permeability without any heat It is necessary to investigate mechanisms of the changes of
treatment, the permeability of samples after 800∘ C treatment physical properties.
increases as much as 10.53∼36.84 times (24.18 times on
average) for shale, 14.63∼33.55 times (21.92 times on average) 4.1. Evaporation and Dehydration of Water Phase. Although
for tight sandstone, 11.34 times for tight carbonate rock, and the free water phase in pore can be easily excluded at about
12.11∼80.36 times (53.47 times on average) for mudstone. 100∘ C, other water phases, that is, adsorbed water, interlayer
Effect of heat treatment on porosity is similar to that on water, and constitution water that exist within minerals, are
permeability (Figure 6). It still has an abrupt increase within not easily excluded [26]. Generally speaking, if temperature
a specified temperature range, which is nearly the same as increases to 100∼200∘ C, adsorbed water and interlayer water
the change of permeability. In detail, as the treatment tem- can be excluded and constitution water in lattice can be
perature elevates, increase of porosity is relatively slow if the excluded if the temperature is increased to 400∼800∘ C. The
temperature is lower than that range, but when temperature main reason for mass decrease in the experiment is the
exceeds that range, the porosity has a relatively large increase. exclusion of adsorbed water and interlayer water. Evaporation
However, the increase of porosity is not as remarkable as that and dehydration of water phase in pores expand the gas flow
of permeability, and it also does not have a very evidently channel, and then permeability increases.
6 Journal of Chemistry

Meanwhile, the dehydration of constitution water in clay fractures were initiated or propagated after 800∘ C treatment.
minerals can make the newly generated minerals tighter Meanwhile, SEM images show that thermal-induced frac-
and strengthen fragility of rock, which could be recognized tures are generated in different scales depending on thermal
as the function of consolidation and would accelerate the expansion of different minerals. These fractures increase
effectiveness of hydraulic fracturing [27, 28]. Besides, if the permeability remarkably.
heating rate is high enough, such as microwave heating [1],
instant evaporation of interlayer water would occur to make 4.4. Comprehensive Mechanisms of High Temperature Treat-
the mineral crystal fracture in the middle and separate from ment. Comprehensive analyses of the above mechanisms
the edge of particle, which would generate microfracture and indicate that essence of the changes of physical properties
enhance the permeability evidently. for tight rock after high temperature treatment is a set of
multiscale processes involving evaporation and dehydration
4.2. Mineral Phase Change and Decomposition at High Tem- of water phase, change of mineral structure, and generation
perature. As is shown in Table 3, some physical or chemical of fracture network. As is presented in Figure 8, the red arrow
reactions occur when the minerals absorb a certain amount represents the process of thermal fracturing which produces
of heat. The most consistent reaction is the inversion of microfractures from initiation and propagation to network
quartz from the 𝛼 → 𝛽 inversion at 573∘ C. The amount connectivity.
of heat needed to complete this inversion is known to be According to the experimental results of permeability
4.825 cal/gm [29]. The phase is fully reversible, so upon change, permeability has an evident abrupt change within a
cooling, an equivalent amount of heat is liberated. The reason certain temperature range, which can be recognized as the
of highlighting the phase change of quartz is that quartz has a range of threshold value based on percolation model. If the
quick (2∼5 seconds) volume expansion of 2.7% when quartz temperature is lower than the threshold value, effect of heat
is heated to 573∘ C, which can easily cause the strong thermal- treatment is mainly reflected in evaporation or dehydration
induced stress. Therefore, when sample’s temperature reaches of water phase and minor generation of microfractures. Its
573∘ C, it is prone to some degree of break. main mechanism is to prevent water blocking in pores, but it
cannot enhance permeability of tight rock essentially. There-
4.3. Thermal-Induced Fracturing. Although thermal fractur- fore, the main mechanism of permeability enhancement
ing of rock has minor influence on bulk volume or density, it through heat treatment is the fracture network developed
has a significant effect on pore structure, mainly reflected as from initiation of microfracture and fracture propagation
generation and propagation of fracture. As rock is made up of under the action of thermal stress when the treatment
different kinds of minerals, differences in thermal expansion temperature is higher than the threshold value. Furthermore,
of different minerals and differences in thermal expansion kerogen in shale can strengthen the action of thermal stress
along different crystallographic axes of the same mineral can compared to others without any organic materials. In detail,
result in heterogeneity and anisotropy of thermal expansion, the kerogen would generate large amount of gas and oily
which generate thermal-induced stress [22]. Besides, if tem- product at certain high temperature, and when it is heated,
perature gradient exists in rock, thermal expansion must be these products would expand seriously, resulting in extending
different in every part of rock even though thermal expansion pressure [32]. If the value of extending pressure exceeds a
coefficient of each mineral is the same, which could also certain critical value, the rock would develop more dramatic
generate thermal stress. fracturing.
If thermal stress exceeds the ultimate tensile strength
(tensile strength or compressive strength) at somewhere of 5. Prospects on Field Testing
the rock, microfracture would occur. Also, different heating
rate and interval can cause different degree of thermal fractur- The concept of formation heat treatment was first proposed
ing. Generally speaking, thermal fracturing tends to occur in by Jamaluddin et al. (1995) to solve formation damage
the short axis direction of mineral particles [30]. Therefore, induced by water blocking, which mainly aimed at relatively
when temperature is relatively low, intercrystalline fracture high permeability samples compared with the samples in this
is the main result of thermal fracturing, and as temperature work [8]. For the application in industry, the earliest report
increases, intracrystal fracture and transcrystalline fracture was that of Albaugh (1954), on an oil well in California, which
begin occurring [31]. had an increase of 76% in production compared with that of
According to the permeability measurement results, after pretreatment [33]. The other typical application in industry
800∘ C treatment, permeability of samples increases as much was that of Jamaluddin et al. (1999), on a field test that was
as 24.18 times on average for shale, 21.92 times on average carried out in a disused gas well, which made the permeability
for tight sandstone, 11.34 times for tight carbonate rock, and increase from 0.66 × 10−3 𝜇m2 to 20 × 10−3 𝜇m2 [17].
53.47 times on average for mudstone. In order to detect Since the high temperature has the risk of destroying
the mechanism of permeability enhancement caused by the casing/cement integrity, it needs to be considered in field
generation of fracture, SEM (scanning electron microscopy) application. Many excellent ideas have been presented, such
imaging was conducted before and after 800∘ C treatment, as that of Jamaluddin et al. (1999) who designed and
respectively. Meanwhile, to image the microstructure of shale constructed an electrical down-hole heater by using high-
more clearly, argon-ion milling was utilized to produce a pressure nitrogen gas as the heat carrier [17]. What is more,
much flatter surface. As is shown in Figure 7, various kinds of several other ways to transport heat to objective formation
Journal of Chemistry 7

Before heat treatment After heat treatment Before heat treatment After heat treatment
(a) Shale (b) Tight sandstone

Before heat treatment After heat treatment Before heat treatment After heat treatment
(c) Tight carbonate rock (d) Mudstone

Figure 7: SEM of samples showing thermal-induced fractures after 800∘ C treatment.


8 Journal of Chemistry

Table 4: Contribution of various stimulation methods on different scales of mass transfer.


Scale Formation heat treatment Hydraulic fracturing Gas-based fracturing Acid treatment Acid fracturing
𝑆𝑤 decrease 󳀅󳀅󳀅󳀅 e f e e
Mineral reaction 󳀅󳀅󳀅󳀅 f f 󳀅󳀅󳀅 󳀅󳀅󳀅
Microfracture generation 󳀅󳀅󳀅󳀅 󳀅󳀅 󳀅󳀅 f 󳀅󳀅
Meso/macrofracture generation 󳀅󳀅 󳀅󳀅󳀅󳀅 󳀅󳀅󳀅󳀅 f 󳀅󳀅󳀅󳀅
󳀅󳀅󳀅󳀅: predominant, 󳀅󳀅󳀅: secondary, 󳀅󳀅: weak, f: ineffective, and e: harmful.

Rock Thermal-induced
Fracture network fracturing
Dehydration

Scale
Mineral Phase change
Decomposition
Evaporation Water phase Evaporation
Temperature
100∘ C

Ambient
573∘ C temperature
Quartz 200∘ C
phase change

Dehydration
400∘ C
Minerals
decomposition

Figure 8: Mechanisms of heat treatment for tight rocks.

point, such as microwave [1] and high-power laser technology heat treatment method mainly works in the development of
[20], are still very effective. These heating methods are all microfracture and prevention of water blocking. Therefore,
designed and constructed to enhance permeability most combining heat treatment stimulation and other nonthermal
remarkably and avoid destroying the casing/cement integrity. stimulations can perfectly match all the scales of mass trans-
Meanwhile, in order to heat the target depth accurately, port processes, resulting in the most effective stimulation.
coiled tubing system would be a very good choice. Also, the
technology of coiled tubing is applicable in horizontal well, 6. Conclusions
which can meet the requirement of staged stimulation in
horizontal well. In this work, physical properties after heat treatment for
Gas production of tight reservoirs is a typical multiscale different lithologies are studied experimentally. Meanwhile,
mass transfer process, which is related to decrease of water SEM imaging was implemented to detect microfracture
saturation (𝑆𝑤 ), change of mineral structure, and generation development and other structural changes. Since formation
and propagation of microfracture and meso/macrofracture. heat treatment is a state-of-the-art technology for tight gas
Table 4 summarizes the contribution of different stimula- formation and the systematic studies are still in the infancy,
tion methods, that is, formation heat treatment, hydraulic results of this work are significant to deeply understand the
fracturing, gas-based fracturing, acid treatment, and acid advantage of heat treatment on gas production enhancement.
fracturing, on the above four scales. Compared with other Conclusions from this work are summarized as follows.
stimulation methods, the advantages of heat treatment are
mainly reflected in the scale of reducing 𝑆𝑤 and genera- (1) Physical properties of tight rocks change signifi-
tion of microfracture. The development of large amount of cantly after specified temperature treatment. Gen-
microfracture plays an important role in reducing fracturing erally speaking, shale and mudstone change more
pressure and generating fracture network for tight rock. remarkably than tight sandstone and tight carbonate
Generally speaking, gas production of tight reser- rock.
voirs contains processes of desorption, diffusion, and slip (2) The decrease of mass mainly occurs lower than 200∼
flow. Therefore, only if the matrix pore, microfracture, 300∘ C. Acoustic transit time increases as temperature
and meso/macrofracture were suitably matched, could the increases, except that the change of Δ𝑇𝑠 /Δ𝑇𝑝 does
highest production be achieved. Conventional stimulation not have obvious regularity. As temperature increases,
method, such as hydraulic fracturing, mainly plays an impor- permeability of shale, mudstone, tight sandstone,
tant role in the propagation of meso/macrofracture and and tight carbonate rock increases remarkably at
Journal of Chemistry 9

600∼700∘ C, 500∼600∘ C, 300∼500∘ C, and 300∼400∘ C, Proceedings of the 8th European Formation Damage Conference,
respectively, which is the threshold temperature range Paper SPE 122034, Scheveningen, The Netherlands, May 2009.
of thermal fracturing for each lithology. [7] J. Cai and S. Sun, “Fractal analysis of fracture increasing
(3) Essence of the changes of physical properties after spontaneous imbibition in porous media with gas-saturated,”
International Journal of Modern Physics C, vol. 24, no. 8, Article
heat treatment for tight rock is a set of multiscale
ID 1350056, 2013.
processes involving evaporation and dehydration of
[8] A. K. M. Jamaluddin, M. Vandamme, and B. K. Mann, “For-
water phase, change of mineral structure, and gener-
mation heat treatment (FHT): a state-of-the-art technology for
ation of fracture network.
near-wellbore formation damage treatment,” Paper SPE 95-67,
(4) Typical field applications are reviewed to confirm presented at the Annual Technical Meeting, Calgary, Alberta,
the feasibility of heat treatment in industry. Heat- Canada, June 1995.
ing methods, such as high-pressure nitrogen and [9] A. K. M. Jamaluddin, D. B. Bennion, F. B. Thomas, and T. Y. Ma,
microwave, are presented to be effective in enhancing “Application of heat treatment to enhance permeability in tight
permeability remarkably and avoiding destroying the gas reservoirs,” Journal of Canadian Petroleum Technology, vol.
casing/cement integrity. 39, no. 11, pp. 19–24, 2000.
(5) Heat treatment can dramatically enhance permeabil- [10] J. C. Cai and B. M. Yu, “A discussion of the effect of tortuosity on
the capillary imbibition in porous media,” Transport in Porous
ity in the scale of matrix pore and microfracture.
Media, vol. 89, no. 2, pp. 251–263, 2011.
However, for the scale of meso/macrofracture, it is
necessary to use conventional stimulation method. [11] F. Meng, P. Zhong, Z. Li, X. Cui, and H. Zheng, “Surface
structure and catalytic performance of Ni-Fe catalyst for low-
Therefore, integrating heat treatment with conven-
temperature CO hydrogenation,” Journal of Chemistry, vol.
tional stimulation might be the best choice. 2014, Article ID 534842, 7 pages, 2014.
[12] C. Yong and C.-Y. Wang, “Thermally induced acoustic emission
Conflict of Interests in westerly granite,” Geophysical Research Letters, vol. 7, no. 12,
pp. 1089–1092, 1980.
The authors declare that there is no conflict of interests [13] Y. Zhao, Z. Wan, Y. Zhang et al., “Experimental study of
regarding the publication of this paper. related laws of rock thermal cracking and permeability,” Chinese
Journal of Rock Mechanics and Engineering, vol. 29, no. 10, pp.
Acknowledgments 1970–1976, 2010.
[14] H. Tian, T. Kempka, N. X. Xu, and M. Ziegler, “Physical
This work was supported by the National Basic Research properties of sandstones after high temperature treatment,”
Program of China (2010CB226705), China Scholarship Fund, Rock Mechanics and Rock Engineering, vol. 45, no. 6, pp. 1113–
and Open Fund (PLN1117) of the State Key Laboratory of 1117, 2012.
Oil and Gas Reservoir Geology and Exploitation (Southwest [15] A. K. M. Jamaluddin, L. M. Vandamme, T. W. Nazarko, and
Petroleum University). D. B. Bennion, “Heat treatment for clay-related near wellbore
formation damage,” Journal of Canadian Petroleum Technology,
vol. 37, no. 1, pp. 56–63, 1998.
References [16] C. R. Hartman, J. R. Ambrose, I. Y. Akkutlu, and C. R. Clarkson,
[1] G. Li, Y. F. Meng, and H. M. Tang, “Clean up water blocking “Shale gas-in-place calculations part II—multicomponent gas
in gas reservoirs by microwave heating: laboratory studies,” in adsorption effects,” in Proceedings of the North American Uncon-
Proceedings of the International Oil & Gas Conference and Exhi- ventional Gas Conference and Exhibition, Paper SPE 144097, The
bition in China, Paper SPE 101072, Beijing, China, December Woodlands, Tex, USA, June 2011.
2006. [17] A. K. M. Jamaluddin, M. Hamelin, K. Harke, and H. McCaskill,
[2] F. Javadpour, D. Fisher, and M. Unsworth, “Nanoscale gas “Field testing of the formation heat treatment process,” Journal
flow in shale gas sediments,” Journal of Canadian Petroleum of Canadian Petroleum Technology, vol. 38, no. 3, pp. 38–45,
Technology, vol. 46, no. 10, pp. 55–61, 2007. 1999.
[3] S. A. Najeeb, A. K. Mohammed, K. Hossein, and M. G. Ramona, [18] Y. Chen, X. D. Wu, and F. Q. Zhang, “Experimental study of
“Physics and modeling of gas flow in shale reservoirs,” in Pro- rock thermal cracking,” Chinese Science Bulletin, vol. 44, no. 8,
ceedings of the Abu Dhabi International Petroleum Conference pp. 880–883, 1999.
and Exhibition, Paper SPE 161893, pp. 2822–2836, Abu Dhabi, [19] M. G. Keaney, C. Jones, P. Meredith, and S. Murrell, “Thermal
UAE, November 2012. damage and the evolution of crack connectivity and permeabil-
[4] J. Cai, E. Perfect, C.-L. Cheng, and X. Hu, “Generalized ity in ultra-low permeability rocks,” in Proceedings of the 6th
modeling of spontaneous imbibition based on hagen-poiseuille North America Rock Mechanics Symposium, Paper ARMA 04-
flow in tortuous capillaries with variably shaped apertures,” 537, Houston, Tex, USA, June 2004.
Langmuir, vol. 30, no. 18, pp. 5142–5151, 2014. [20] R. M. Graves and E. T. Bailo, “Analysis of thermally altered rock
[5] D. B. Bennion, F. B. Thomas, R. F. Bietz, and D. W. Bennion, properties using high-power laser technology,” in Proceedings
“Water and hydrocarbon phase trapping in porous media— of the Canadian International Petroleum Conference, Calgary,
diagnosis, prevention and treatment,” Journal of Canadian Canada, June 2005.
Petroleum Technology, vol. 35, no. 10, pp. 29–36, 1996. [21] T. Mehmannavaz, M. Ismail, S. Radin Sumadi, M. A. Rafique
[6] L. J. You and Y. L. Kang, “Integrated evaluation of water Bhutta, M. Samadi, and S. M. Sajjadi, “Binary effect of fly ash
phase trapping damage potential in tight gas reservoirs,” in and palm oil fuel ash on heat of hydration aerated concrete,” The
10 Journal of Chemistry

Scientific World Journal, vol. 2014, Article ID 461241, 6 pages,


2014.
[22] W. H. Somerton, Thermal Properties and Temperature-Related
Behavior of Rock/Fluid Systems, Elsevier Science Publisher B.V.,
Amsterdam, The Netherlands, 1992.
[23] C. I. Killian, “The Vp/Vs ratio after 40 years: uses and abuses,”
in Proceedings of the SEG Annual Meeting, Paper SEG 2006-1183,
New Orleans, La, USA, October 2006.
[24] W. C. James and P. B. Alan, “Petrophysics of low-permeability
medina sandstone, Northwestern Pennsylvania, Appalachian
Basin,” The Log Analyst, vol. 39, no. 4, pp. 36–46, 1998.
[25] Y.-L. Kang and P.-Y. Luo, “Current status and prospect of key
techniques for exploration and production of tight sandstone
gas reservoirs in China,” Petroleum Exploration and Develop-
ment, vol. 34, no. 2, pp. 239–245, 2007.
[26] J. Mahadevan, M. M. Sharma, and Y. C. Yortsos, “Evaporative
cleanup of water blocks in gas wells,” SPE Journal, vol. 12, no. 2,
pp. 209–216, 2007.
[27] R. H. Friedman, B. W. Suries, and D. E. Kleke, “High-
temperature sand consolidation,” SPE Production Engineering,
vol. 3, no. 2, pp. 167–168, 1988.
[28] M. C. Ross, E. Rangel, M. L. Castanier, S. P. Hara, and R. A.
Kovscek, “A laboratory investigation of temperature-induced
sand consolidation,” SPE Journal, vol. 11, no. 2, pp. 206–215,
2006.
[29] K. K. Kelley and E. G. King, Contributions to the Data on The-
oretical Metallurgy, U.S. Bureau of Mines Bulletin, Washington,
DC, USA, 1960.
[30] J.-P. Zuo, H.-P. Xie, H.-W. Zhou, and S.-P. Peng, “Experimental
research on thermal cracking of sandstone under different
temperature,” Chinese Journal of Geophysics (Acta Geophysica
Sinica), vol. 50, no. 4, pp. 1150–1155, 2007.
[31] Y. Zhang, X. Zhang, and Y.-S. Zhao, “Process of sandstone
thermal cracking,” Chinese Journal of Geophysics, vol. 48, no. 3,
pp. 656–659, 2005.
[32] Q.-R. Meng, Z.-Q. Kang, Y.-S. Zhao, and D. Yang, “Experiment
of thermal cracking and crack initiation mechanism of oil
shale,” Journal of China University of Petroleum (Edition of
Natural Science), vol. 34, no. 4, pp. 89–98, 2010.
[33] F. W. Albaugh, “Oil well production process,” U.S. Patent
2,685,930, 1954.
[34] I. Barshad, “Temperature and heat of reaction calibration of the
differential thermal apparatus,” American Mineralogist, vol. 37,
no. 8, pp. 667–694, 1952.
[35] D. Carroll, “Clay minerals: a guide to their x-ray identification,”
Special Paper of the Geological Society of America, vol. 126, pp.
1–80, 1970.
Hindawi Publishing Corporation
Journal of Chemistry
Volume 2015, Article ID 980439, 8 pages
http://dx.doi.org/10.1155/2015/980439

Research Article
Stability of Fluorosurfactant Adsorption on Mineral Surface for
Water Removal in Tight Gas Reservoirs

Lijun You,1 Wanchun Zhang,1 Yili Kang,1 Zhangxin Chen,2 and Xuefen Liu1
1
State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum University, Chengdu 610500, China
2
Department of Chemical & Petroleum Engineering, Schulich School of Engineering, University of Calgary,
Calgary, AB, Canada T2N 1N4

Correspondence should be addressed to Lijun You; youlj0379@126.com

Received 10 October 2014; Revised 18 January 2015; Accepted 24 February 2015

Academic Editor: Sachin Jangam

Copyright © 2015 Lijun You et al. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Long-term effectiveness of rock wettability alteration for water removal during gas production from tight reservoir depends on
the surfactant adsorption on the pore surface of a reservoir. This paper selected typical cationic fluorosurfactant FW-134 as an
example and took advantage of Fourier transform infrared spectroscopy (FTIR), X-ray photoelectron spectroscopy (XPS), and
atomic force microscope (AFM) to investigate its adsorption stability on the rock mineral surface under the oscillation condition
at high temperature for a long time. The experimental results indicate that the F element content on the sample surface increases
obviously, the surface structure of fluorine-carbonization also undergoes a significant change, and the fluorine surfactant exhibits a
good interfacial modification and wettability alteration ability due to its adsorption on the pore surface transforming the chemical
structure of the original surface. The adsorption increases indistinctly with the concentration of over 0.05% due to a single layer
adsorption structure and is mainly electrostatic adsorption because the chemical bonding between the fluorosurfactant and the
rock mineral surface, the hydrogen bonding, is weak and inconspicuous.

1. Introduction mechanical, textile, papermaking, and oil industries. In the


oil and gas industry, fluorine surfactant can be used as an oil-
Tight gas reservoirs, one of main development objectives in displacing agent, a wettability alternation agent, and a cleanup
the future dozens of years, are characterized by a high water additive by improving the ability of fluid flow in porous
phase trapping damage potential owing to the characteristic media. Li and Firoozabadi [6] put forward the term “gas-
of water-wet and high capillary pressure. The water phase wetness.” Their motive was to change rock wettability from
trapping damage can induce a fluid sensitivity damage, an water or oil wetness to gas-wetness for reduction or removal
aggravate stress sensitivity damage, and other damage [1], of liquid blocking or phase trapping damage. Since then,
seriously affects an economic development of the tight gas many papers presented research results about improving
reservoirs [2–4], and is difficult to be removed due to high gas production by wettability alteration [7–15] and tested
capillary pressure of tight sandstone. different surfactants at different reservoir temperatures. Li’s
Surfactant interface modification is increasingly attract- group [6–8, 10, 13] conducted experiments to change the
ing more attention of reservoir engineers because it can wettability of tight sandstone with permeability of less than
reduce interface intension and alter wettability of rocks. 0.1 mD and they also reported a pilot field test in a tight
Compared with ordinary hydrocarbon surfactant, fluorosur- gas-condensate reservoir to improve the gas production by
factant possesses the outstanding characteristic of a high wettability alteration. The production of the test well was
surface activity, high thermal stability, and high chemical sta- increased to about 30,000 m3 /day, but the gas production
bility [5], and its fluorine hydrocarbon exhibits a hydrophobic declined very fast and it was only about 4,000 m3 /day after
and oleophobic property. Owing to these properties, it has a four days [16]. However, there have been few reports about the
very wide range of applications in the chemical, biological, reason that the period of validity of improving gas production
2 Journal of Chemistry

by wettability alteration was so short. Fluorine surfactant (4) Shake up the mixtures, cover tightly the conical flasks
can modify the interface of a gas-water-rock system by with a plug, and immerse the flasks into the water
adsorption on mineral surface to alter the wettability and bath oscillator to oscillate at 90∘ C for 40 hours. The
surfactant adsorption behavior on the mineral surface and purpose of oscillation is to simulate the reservoir
influences the long-term effect of wettability alteration. A detention water flow scouring to the pore surface dur-
key to maintaining the long-term effectiveness of wettability ing gas production. In order to ensure that the FW-
alteration is understanding the interaction of chemicals and 134 reaches saturation adsorption on solid surfaces,
rock [17, 18]. Seiedi et al. [19] presented AFM images for quan- we test the adsorptive capacity Γ changing with time
titative assessment of the surface roughness for a wettability on the mica surface. We can see from the results
alteration investigation. Wu and Firoozabadi [17, 18] analyzed (Figure 1) that the adsorption equilibrium time is
the long-term effectiveness of wettability alteration using core not more than 2 hours. We conduct the adsorption
laboratory experiments. experiments for 40 hours, which is much more than
In this study, we select a kind of fluorine surfactant FW- the adsorption equilibrium time.
134 which is able to successfully change water-wet tight sand- (5) Take out the conical flasks and leave them to rest for
stone into gas-wetness and improve the water flow rate by 30 minutes.
modifying the interface [12], and we select the mica to repre-
sent the main rock-forming minerals of tight sandstone. The (6) Filter the precipitation separation of solid particles
purpose is to investigate the adsorption behavior and the abil- and soak up the residual solution on the solid particle
ity of interface modification on the surface of rock mineral surface with a filter paper.
from microaspect under the conditions of simulated long- (7) Dry solid particles to a constant weight at 60∘ C,
term water flow back from a high temperature gas reservoir. number them sample A (distilled water), sample
B (0.05% FW-134), and sample C (0.10% FW-134)
2. Materials and Methods according to the concentration of FW-134 from low
to high, and seal them for WQF-520 Fourier infrared
2.1. Materials. Only a small part of clay minerals and carbon- (FTIR) spectrometer and Thermo Scientific Escalab
ate cementation play an important role in surfactant adsorp- 250 X-ray photoelectron spectroscopy (XPS) tests.
tion, and their main components are layered silicate minerals,
containing oxygen, silicon, aluminum, calcium, sodium, and (8) Select fresh natural dissociated mica (1 cm × 1 cm,
potassium. Mica is easily split into pieces to obtain a smooth thickness < 2 mm) and immerse them into 0.1% FW-
atomic hydrophilic surface [20], whose naked aluminum 134 solution according to a 1 : 50 ratio of solid to liquid
silicate lattice surface is similar to silica and many clay at 90∘ C constant temperature water bath.
surfaces. In reality, we selected several rock-forming minerals (9) Take out the mica plate after the oscillation for 40
like mica, kaolinite, and silica to conduct the adsorption, but hours, dry the surface with nitrogen flow to obtain the
this paper just selects fresh natural cleaved mica as a substrate mica surface after the adsorption of FW-134, and fix
to investigate fluorine surfactant adsorption on the high tem- the sample on a sample table of MMAFMLN 1728EX
perature solid-liquid interface. Fluorinated surfactant FW- atomic force microscope (AFM) to test surface mor-
134 used is a kind of typical cationic fluorinated surfactant phology. Nitrogen flow drying is to simulate gas flow
C8 F17 SO2 NH(CH2 )3 (CH3 )3 NI containing a small amount of scouring to the pore surface in the reservoir during
N, S, and O elements and its critical micelle concentration is gas production.
0.10%. The surfactant is 99.9% pure. Distilled water is used (10) Measure the surface contact angle.
to simulate the retention liquid like condensate water in the
reservoir in order to avoid the effect of water salinity on the
interaction of surfactants and rock mineral. 3. Experimental Results
3.1. Characterization of Adsorption Behavior. Sample A, sam-
2.2. Methods. Experimental procedures are as follows.
ple B, and sample C were analyzed by FTIR. These samples
(1) Wash with distilled water to remove impurities from were obtained from mica powder immersed into distilled
the mica surface and put them into a constant tem- water, 0.05% and 0.10% FW-134 solution under 90∘ C for 40
perature oven to dry to a constant weight at 60∘ C. hours’ oscillation, as mentioned above. The FTIR results can
(2) Crush the mica into powder, sieve the powder with be seen in Figure 1.
100∼200 meshes, and dry it to a constant weight at The peak at 3,625 cm−1 is the O-H stretching vibration
60∘ C. absorption peak of mica Al-OH and Mg-OH, 3,442 cm−1
(3) Make up 0.05% and 0.10% FW-134 solution with the O-H stretching vibration absorption peak of mica sur-
distilled water and FW-134 according to a 1 : 50 ratio of face adsorbing water, and 630∼1,640 cm−1 the O-H bending
solid to liquid into the conical flask, respectively [21]. vibration absorption peak; a high intensity peak at 1,024 cm−1
The 1 : 50 ratio of solid to liquid is from the results of is the Si-O absorption peak of stretching vibration, the
a series of adsorption experiments of different solid- absorption peak at 950∼600 cm−1 is O-H bending vibration,
liquid ratio to avoid the effect of the solid-liquid ratio and a vibration zone at 800∼500 cm−1 belongs to Si-O-Al
on the surfactant adsorption. spectral absorption peak. Using a comparative analysis of
Journal of Chemistry 3

2.0
(A)
0.10% FW-134 (B)
1.5

(C)

Transmittance
Γ (mg/g)

1.0

0.5

0.0
0 20 40 60 80 100 120 140 160 180 200
4400 4000 3600 3200 2800 2400 2000 1600 1200 800 400
t (min)
Wavenumber (cm−1 )
30∘ C (A) Distilled water (C) 0.10% FW-134
(B) 0.05% FW-134
Figure 1: FW-134 adsorption capability on mica versus time.
Figure 2: FT-IR spectra of samples adsorbing FW-134.

the FTIR diagram, methyl CH3 and methylene CH2 stretch-


ing vibration absorption peaks appeared obvious at 3,000∼
A, sample B, and sample C, and the results are seen in
2,800 cm−1 in sample B and sample C, and the absorption
Figure 3. The XPS spectra curve analysis and fitting were
peak of stretching vibration of fluoroalkyl groups CF3 and
conducted using the software XPS Peaks 4.1 developed by
CF2 also appeared at 1,350∼1,120 cm−1 . These indicate that University of Hong Kong. We analyze the binding energy
FW-134 produces a significant adsorption on mica surfaces of Al2p, Si2p, C1s, K2p, O1s, F1s, N1s, and S2p of different
under the conditions of high temperature and long-term samples. As shown in Figure 3, the binding energy of Al2p,
oscillations, but there is basically no obvious change of Si2p, K2p, and O1s and the curve shape do not change with
displacement, strength, or shape of the absorption peak the concentration of FW-134; the binding energy of F1s, N1s,
from the whole FTIR graph characteristics before and after and S2p of sample B and sample C increases significantly
adsorption, suggesting that the chemical bond between FW- compared with that of sample A and appears at 689 eV, 402 eV,
134 and the mica surface is weak and the adsorption is mainly and 169 eV, respectively.
physical adsorption. The peaks 1800∼1200 cm−1 in sample A The content and the binding energy of each element are
and sample B are the same, which hardly appear in sample calculated according to the peak area (Tables 1 and 2).
C. This shows that this peak is not related to the surfactant Sample A, with no FW-134 adsorption on the mica, only
adsorption and will not affect the experiment results. contains trace F, N, and S element, while the F, N, and
A mica surface has a strong capacity of bonding hydroxyl S element contents in sample B and sample C with the
groups, polar atoms like F, N, and O are easy to form a adsorption of FW-134 increase significantly, and the content
hydrogen bond with the surface hydroxyl groups, and mica of the F element increases from 0.598% to 17.545% and
surfaces in aqueous solution are negatively charged, so FW- 22.037%, the content of N element from 0.577% to 2.092% and
134, a kind of cationic fluorinated surfactant, is easy to be 2.616%, and the S content from 0.502% to 1.166% and 1.400%
adsorbed on mica surfaces due to the electrostatic force, (Table 1), respectively. These showed that a large amount of
dispersion force, hydrogen bond, and other physical and FW-134 is adsorbed on the mica surface under the conditions
chemical effects. However, it can be seen in Figure 2 that of 40 hours’ oscillations at 90∘ C.
O-H stretching vibration absorption displacement, strength, The element binding energy (B.E.) in these samples can
and shape of surface water on the samples with different be seen in Table 2. The binding energy of Al2p, Si2p, K2p,
concentration of FW-134 are the same, and the peak is at and O1s in samples B and C has little difference with that of
3,442 cm−1 , which shows that the bonding force between sample A (Table 2), which shows that, under the conditions of
FW-134 molecules and the mica surface hydroxyl bond is high temperature, the chemical bonding force between FW-
weak. This may be due to the fact that hydrogen bonds 134 and the mica surface is weak. This is also consistent with
are destroyed at high temperature, or the hydrogen bonding the result from the infrared spectrum analysis. In order to
adsorption effect is weak, and it is not reflected in the FTIR further determine the FW-134 adsorption pattern on the mica
diagram. These suggest that the main mechanism of FW-134 surface, F1s XPS peaks of each sample were analyzed by Gauss
adsorption under the condition of high temperature is not a fitting spectra, as shown in Figure 4.
hydrogen bond. The F1s peaks of all samples are a single peak (Figure 4), so
the chemical structure of their F element is relatively single.
3.2. Energy Spectrum Analysis of Surface Structure. We take The pulse signal of the F1s peak of sample A is strong, and
an advantage of XPS to test the energy spectrum of sample the peak intensity and peak area are small, which show that
4 Journal of Chemistry

Table 1: XPS element content data of sample systems adsorbing FW-134.

Content (%) Al2p Si2p C1s K2p O1s F1s N1s S2p
Sample A 13.245 17.383 11.470 4.950 51.768 0.598 0.577 0.502
Sample B 10.269 13.458 7.588 4.689 43.194 17.545 2.092 1.166
Sample C 9.3410 11.949 10.377 4.327 37.953 22.037 2.616 1.400

Table 2: XPS 𝐸𝐵 data of sample systems adsorbing FW-134.

Binding energy (eV) Al2p Si2p C1s K2p O1s F1s N1s S2p
Sample A 74.26 102.42 284.76 293.1 531.58 685.07 401.92 169.16
Sample B 74.62 102.80 286.55 292.97 531.66 689.19 402.77 169.23
Sample C 74.50 102.65 286.31 292.67 531.69 689.06 402.79 169.02

than hydrogen bonding is the main mechanism of FW-


134 adsorption under the condition of the 40 hours’ water
F1s oscillation.
The Gauss fitting spectra of the C1s XPS peaks of samples
are shown in Figure 5. As shown in Figure 5(a), the C1s peak
of sample A was split into four peaks; peak 1 and peak 2 are
Counts (s)

(C) classified as a type of peak; the binding energies from low to


N1s C1s S2p high are for C-H\C-C, C-O, and C=O peaks. As shown in
(B) Figures 5(b) and 5(c), the C1s peaks of sample B and sample
C were divided into three peaks, respectively. Their binding
(A) energies from low to high are for C-H\C-C, C-N\C-C, and
C-F peaks.
The analysis results of the peak binding energy and peak
area are shown in Table 3. The C-H\C-C peak area of sample
1200 1050 900 750 600 450 300 150 0 A reaches 88.80%, and a small piece of the C-O and C=O
Binding energy (eV)
peak area appears in the C1s peak (Table 3). This may be
caused by the chemical structure of carbon on the sample
(A) Distilled water (C) 0.10% FW-134 original surface or adsorption of CO2 in the air on the surface.
(B) 0.05% FW-134 Compared with sample A, sample B and sample C have a large
Figure 3: XPS spectra of sample systems adsorbing FW-134. area of the C-F and C-N\C-C peaks, whose total peak area
is, respectively, 85.88% and 81.63%, while the peak area of C-
H\C-C is greatly reduced. The reason is that a large amount
of FW-134 adsorbed covers on the mica surface to change the
there is less F content in sample A (Figures 4(a) and 4(d)). chemical structure of the original surface carbon under the
The curves of samples B and C in Figures 4(b) and 4(c) are high temperature water oscillation.
smooth and orderly, and the intensity and peak area of the
F1s peak are large, which suggests that a lot of the F element 3.3. Evaluation of Surface Conformation and Modification
is in the surface structure of samples B and C and a great deal Effect. The water adsorption on modified surface of minerals
of FW-134 adsorption is on the mica surface. depends on the surface structure [22]. In order to observe the
The F1s peak positions of samples B and C are different morphology to further understand the FW-134 adsorption
from those of sample A, and the F1s peak binding energy of structure and interfacial modification effect on the mica
sample B and sample C is basically the same (Figure 4(d) and surface, a mica surface treated by a 40 hours’ oscillation with
Table 2). That is, the F1s peak position and chemical structure 0.1% FW-134 solution and nitrogen flow drying under high
of F atom before and after adsorption on the surface are temperature was tested by atomic force microscope (AFM).
different. The F1s peak binding energy of sample A is 685.2 eV, Figures 6 and 7 are the AFM scanning plane height map and
the Ca-F structure of mica; the F1s peak binding energies of three-dimensional map of the treated mica surface before and
sample B and sample C are about 689.1 eV, the structure of after cleaning of distilled water.
chain C-F; there are no hydrogen bonding peaks in sample As intuitively seen in Figure 6, a lot of FW-134 molecules
B and sample C. These indicate that the chemical bonding adsorbed on the mica surface are unequal granular and
strength between FW-134 and the mica surface is weak under massive distribution phenomenon; the adsorption of FW-
the condition of high temperature, that covering is the main 134 covers up to 80% of the testing scope with 10 𝜇m ×
mechanism of different concentration of FW-134 adsorption 10 𝜇m; the adsorption layer is thin. The contact angles of
on the surface of mica, and that physical adsorption rather water on the adsorption surface treated by 0.05%, 0.1%, and
Journal of Chemistry 5

Table 3: XPS results of samples adsorbing FW-134 calculated by Gauss method.

Sample Type of structure Binding energy (eV) Peak area (%)


C-H C-C 285.0 88.80
A C-O 286.5 5.70
C=O 288.8 5.50
C-H C-C 285.0 14.22
B C-N C-C 285.7 46.07
C-F 286.8 39.81
C-H C-C 285.0 18.37
C C-N C-C 285.7 38.90
C-F 286.8 42.73

60000
10800

50000
10500
Fitting spectra
10200 40000
Original spectra
Counts

Original spectra
Counts

9900 30000

9600
20000
Fitting spectra
9300
10000
9000
692 690 688 686 684 682 680 694 692 690 688 686 684
B.E. (eV) B.E. (eV)

Distilled water 0.05% FW-134


(a) (b)
60000
50000
50000

40000
40000
(B)
30000
Counts
Counts

30000 (C)

20000 Original spectra 20000


Fitting spectra (A)
10000 10000

0 0
694 692 690 688 686 684 682 694 692 690 688 686 684 682 680
B.E. (eV) B.E. (eV)

0.10% FW-134 (A) Distilled water (C) 0.10% FW-134


(B) 0.05% FW-134
(c) (d)

Figure 4: XPS C1s spectra of sample systems adsorbing FW-134 and contrast diagram.
6 Journal of Chemistry

10500
7000
9000
Fitting spectra Original spectra
Original spectra 6000
7500
Split spectra 1
Split spectra 2 Split spectra 1 Split spectra 2
6000 5000

Counts
Counts

4500 Split spectra 3 4000 Fitting spectra


Split spectra 4 Split spectra 3
3000
3000

1500
2000
0
290 289 288 287 286 285 284 283 282 281 289 288 287 286 285 284 283 282
B.E. (eV) B.E. (eV)

Distilled water 0.05% FW-134


(a) (b)
7000

9000
6000
Split spectra 1 Fitting spectra 7500
5000
6000
Counts
Counts

(C)
4000 Original spectra
4500 (B)
Split spectra 3 (A)
3000
3000

2000 1500
Split spectra 2

1000 0
288 287 286 285 284 283 282 281 292 290 288 286 284 282 280
B.E. (eV) B.E. (eV)

0.10% FW-134 (A) Distilled water (C) 0.10% FW-134


(B) 0.05% FW-134
(c) (d)

Figure 5: XPS C1s spectra of samples adsorbing FW-134 and contrast diagram.

0.2% are 71.9∘ , 101.2∘ , and 92.3∘ , respectively, which mean a 4. Discussions
good hydrophobic performance and interfacial modification
ability of FW-134. After the mica surface is being cleaned Under the conditions of water oscillations, FW-134 on the
with distilled water for 48 hours, there is still a lot of mica surface is mainly physical adsorption, generally includ-
surfactant FW-134 adsorption on the mica surface (Figure 7). ing the electrostatic and dispersion force adsorption. Because
We conducted comparative experiments to verify whether the of characteristics of F atoms in fluorine surfactant molecules
AFM data are reproducible. The adsorption area differences (the most electronegative and the minimum atomic polar-
of FW-134 in 10 𝜇m × 10 𝜇m on mica surface in comparative izability) and the carbon chain of “shielding,” the C-F bond
experiments are less than ±6%, which verified that AFM data is not easy to be polarized, and dispersion forces of fluorine
reproducibility was good. As seen in Figure 7, compared surfactant molecules are small. FW-134 adsorption on the
with Figure 6, a part of FW-134 molecules adsorbed on the mica surface force is electrostatic adsorption.
mica surface are cleared away and the adsorption thickness With an increase of concentration, FW-134 reaches a
is reduced, but the adsorbed layer still shows a uniform saturated adsorption state on the surface of mica. Because of
distribution, which indicts that 0.1% FW-134 has reached a “hydrophobic effect” [9] of FW-134 molecules, micelle may
saturated adsorption state on the surface of mica. form on the mica surface, but the low dispersion force of
Journal of Chemistry 7

10.0 50.0

7.5

(𝜇m)
5.0

(nm)
25.0

2.5
8
6
4 )
2 (𝜇m
0 0.0
0 2.5 5.0 7.5 10.0
(𝜇m) x: 2.000 𝜇m/div
z: 50.000 nm/div

Figure 6: AFM plane height map and the three-dimensional topography (before cleaning).

10.0 30.0

7.5
(𝜇m)

(nm)

5.0 15.0

2.5 8
6
4 )
2 (𝜇m
0 0.0
0 2.5 5.0 7.5 10.0
(𝜇m) x: 2.000 𝜇m/div
z: 30.000 nm/div

Figure 7: AFM plane height map and the three-dimensional topography (after cleaning).

fluorine surfactant molecules and the thermodynamics effect alteration effect is good due to low surface energy and the
under high temperature longtime oscillation lead to greatly hydrophobic surface. Thus 0.05%∼0.1% FW-134 solution can
weaken surface micelle formation and multilayer adsorption, also have low surface tension and a good wettability reversal
so FW-134 adsorption on the mica surface has a single layer effect. Because of fluorinated surfactant being expensive, the
structure under the oscillation at 90∘ C when it reaches the compound and synergistic reaction of fluorine surfactant
saturated adsorption (Figure 8). and common surfactant considering the cost and the effect
Owing to an outstanding high surface activity of fluo- might be worth developing, but it should be emphasized that
rine surfactant, its 0.05%∼0.1% solution concentrations can counterions have great influence on the fluorine surfactant
decrease the surface tension of the aqueous solution to properties and adsorption behavior.
below 20 mN/m. Low surface energy of fluorine or a silicon
treatment agent is widely used in the surface modification to 5. Conclusions
enhance the surface hydrophobicity. Genzer and Efimenko
[21] and Peng et al. [23] used the fluorine surfactant treatment Fluorine surfactant adsorption on mica surfaces under the
agent on the material surface fluorination modification treat- condition of oscillation at 90∘ C causes a hydrophilic surface
ment and successfully obtained a super hydrophobic surface. to become hydrophobic structure with the strong ability of
By electrostatic adsorption on the mica surface, a cationic modifying interface and reversing wettability.
hydrophilic ionic bond of FW-134 molecules is inward and its Fluorine surfactant adsorption on the mica surface is
fluorocarbon hydrophobic chain is outward. The maximum mainly physical adsorption and has a single layer structure
C-F peak and the highest percentage of the area (Figure 5 when it reaches the saturated adsorption, which may be the
and Table 3) show that the micro- and nanostructure like a reason that the period of interface modification validity is
“lotus effect” is formed on a solid surface, so the wettability short.
8 Journal of Chemistry

+ + + + + ++ ++ +++ +++ + +++ +++++


Mica surface Mica surface Mica surface Mica surface

Figure 8: FW-134 adsorption schematic diagram on mica surface.

0.05%∼0.1% FW-134 can decrease the surface tension [10] M. Fahes and A. Firoozabadi, “Wettability alteration to inter-
of the aqueous solution to below 20 mN/m and show high mediate gas-wetting in gas-condensate reservoirs at high tem-
quality of wettability alteration effect. peratures,” SPE Journal, vol. 12, no. 4, pp. 397–407, 2007.
[11] A. Delavarmoghaddam, S. A. Mirhaj, and P. L. Zitha, “Gas
condensate productivity improvement by chemical wettability
Conflict of Interests alteration,” in Proceedings of the 8th European Formation Dam-
age Conference, Scheveningen, The Netherlands, 2009.
There is not any conflict of interests regarding the publication
of this paper. [12] X.-F. Liu, Y.-L. Kang, L.-J. You, and P. Wu, “Experimental study
on amphiphobic surface treating to prevent reservoir from
aqueous phase trapping damage,” Natural Gas Geoscience, vol.
Acknowledgments 20, no. 2, pp. 292–296, 2009.
[13] K. Li, Y. Liu, H. Zheng, G. Huang, and G. Li, “Enhanced gas-
The authors gratefully acknowledge financial support from condensate production by wettability alteration to gas wetness,”
The National Basic Research Program (973) of China Journal of Petroleum Science and Engineering, vol. 78, no. 2, pp.
(no. 2010CB226705), China Scholarship Council fund, and 505–509, 2011.
The National Municipal Science and Technology Project [14] S. Wu and A. Firoozabadi, “Simultaneous increase in gas and
(2011ZX05018). liquid relative permeabilities and reduction of high-velocity
coefficient from wettability alteration,” Journal of Canadian
References Petroleum Technology, vol. 50, no. 2, pp. 17–23, 2011.
[15] Y. Li, G. Jiang, L. Li et al., “The effect of a novel gas-wetting
[1] L. You, Y. Y. Kang, Y. Chen et al., “Stress sensitivity of reversal FC-1 on the condensate gas reservoir core,” Petroleum
fractured tight gas sands in consideration of fractures and water Science and Technology, vol. 32, no. 1, pp. 1–7, 2014.
saturation,” Journal of China University of Petroleum (Edition of [16] H. Milsch, G. Blöcher, and S. Engelmann, “The relationship
Natural Science), vol. 30, no. 2, pp. 59–63, 2006. between hydraulic and electrical transport properties in sand-
[2] D. B. Bennion, R. F. Bietz, F. B. Thomas, and M. P. Cimolat, stones: an experimental evaluation of several scaling models,”
“Reductions in the productivity of oil and low permeability gas Earth and Planetary Science Letters, vol. 275, no. 3-4, pp. 355–
reservoirs due to aqueous phase trapping,” Journal of Canadian 363, 2008.
Petroleum Technology, vol. 33, no. 9, pp. 45–54, 1994. [17] S. Wu and A. Firoozabadi, “Effect of salinity on wettability alter-
[3] D. B. Bennion, F. B. Thomas, R. F. Bietz, and D. W. Bennion, ation to intermediate gas-wetting,” SPE Reservoir Evaluation &
“Water and hydrocarbon phase trapping in porous media— Engineering, vol. 13, no. 2, pp. 228–245, 2010.
diagnosis, prevention and treatment,” Journal of Canadian [18] S. Wu and A. Firoozabadi, “Permanent alteration of porous
Petroleum Technology, vol. 35, no. 10, pp. 29–36, 1996. media wettability from liquid-wetting to intermediate gas-
[4] L. You and Y. Kang, “Integrated evaluation of water phase trap- wetting,” Transport in Porous Media, vol. 85, no. 1, pp. 189–213,
ping damage potential in tight gas reservoirs,” in Proceedings of 2010.
the 8th European Formation Damage Conference, Scheveningen, [19] O. Seiedi, M. Rahbar, M. Nabipour, M. A. Emadi, M. H.
The Netherlands, May 2009. Ghatee, and S. Ayatollahi, “Atomic force microscopy (AFM)
[5] T. Yoshimura, A. Ohno, and K. Esumi, “Equilibrium and investigation on the surfactant wettability alteration mechanism
dynamic surface tension properties of partially fluorinated of aged mica mineral surfaces,” Energy and Fuels, vol. 25, no. 1,
quaternary ammonium salt gemini surfactants,” Langmuir, vol. pp. 183–188, 2011.
22, no. 10, pp. 4643–4648, 2006. [20] A. Niecikowska, J. Zawala, and K. Malysa, “Influence of adsorp-
[6] K. Li and A. Firoozabadi, “Experimental study of wettability tion of n-alkyltrimethylammonium bromides (C8, C12, C16)
alteration to preferential gas-wetting in porous media and its and bubble motion on kinetics of bubble attachment to mica
effects,” SPE Reservoir Evaluation & Engineering, vol. 3, no. 2, surface,” Physicochemical Problems of Mineral Processing, vol. 47,
pp. 139–149, 2000. pp. 237–248, 2011.
[7] G.-Q. Tang and A. Firoozabadi, “Relative permeability modi- [21] J. Genzer and K. Efimenko, “Creating long-lived superhy-
fication in gas/liquid systems through wettability alteration to drophobic polymer surfaces through mechanically assembled
intermediate gas wetting,” SPE Reservoir Evaluation & Engineer- monolayers,” Science, vol. 290, no. 5499, pp. 2130–2133, 2000.
ing, vol. 5, no. 6, pp. 427–436, 2002.
[22] K. Cinku and B. Baysal, “Investigation of adsorption behavior of
[8] G.-Q. Tang and A. Firoozabadi, “Wettability alteration to inter- phosphonium salts onto Na-montmorillonite,” Physicochemical
mediate gas-wetting in porous media at elevated temperatures,” Problems of Mineral Processing, vol. 50, no. 2, pp. 417–432, 2014.
Transport in Porous Media, vol. 52, no. 2, pp. 185–211, 2003.
[23] M. Peng, H. Li, L. Wu, Q. Zheng, Y. Chen, and W. Gu, “Porous
[9] B. Adibhatla, K. K. Mohanty, P. Berger, and C. Lee, “Effect poly(vinylidene fluoride) membrane with highly hydrophobic
of surfactants on wettability of near-wellbore regions of gas surface,” Journal of Applied Polymer Science, vol. 98, no. 3, pp.
reservoirs,” Journal of Petroleum Science and Engineering, vol. 1358–1363, 2005.
52, no. 1–4, pp. 227–236, 2006.
Hindawi Publishing Corporation
Journal of Chemistry
Volume 2015, Article ID 427297, 6 pages
http://dx.doi.org/10.1155/2015/427297

Review Article
Review on Fractal Analysis of Porous Metal Materials

J. Z. Wang, J. Ma, Q. B. Ao, H. Zhi, and H. P. Tang


State Key Laboratory of Porous Metal Materials, Northwest Institute for Nonferrous Metal Research, Xi’an 710016, China

Correspondence should be addressed to H. P. Tang; hptang@c-nin.com

Received 24 September 2014; Revised 26 February 2015; Accepted 24 March 2015

Academic Editor: Gustavo Portalone

Copyright © 2015 J. Z. Wang et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Porous metal materials are multifunctional lightweight materials and have been used widely in industry. The structural and
functional characters of porous metal materials depend on the pore structure which can be described effectively by the fractal
theory. This paper reviews the major achievements on fractal analysis of pore structure of porous metal materials made by State
Key Laboratory of Porous Metal Materials, China, over the past few years. These include (i) designing and developing a set of novel
fractal analytical software of porous metal materials, (ii) the influence of material characterization and image processing method on
the fractal dimension, and (iii) the relationship between the material performance and the fractal dimension. Finally, the outlooks
of fractal theory applied in porous metal materials are discussed.

1. Introduction paper reviews the major achievements on fractal analysis of


pore structure of porous metal materials made by State Key
The fractal phenomenon is ubiquitous in a wide array Laboratory of Porous Metal Materials, China, over the past
of materials such as the growth of crystal, the fracture few years, including a set of novel fractal analytical software
or martensite morphology, the quasicrystal structure, the of porous metal materials developed, the influence of material
deposited film, and the porous materials [1, 2]. These mate- characterization and image processing method on the fractal
rials are a unique class of disordered materials and often dimension of pore structure of porous metal materials, and
display complex microstructures. Since the fractal theory
the relationship between the material performance and the
was presented [3], it had been widely used in many fields
fractal dimension. Finally, the outlooks of fractal theory
of modern science, such as studying permeability of porous
applied in porous metal materials are discussed.
media [4, 5] or dual-porosity medium [6], investigating the
mechanism of gas diffusion or fluid flow in porous media [7–
10], evaluating dislocation structure [11], discussing positive 2. Novel Fractal Analytical Software
pulsed streamer patterns in supercritical CO2 [12], analyzing
fracture surfaces or network [13, 14], simulating the failure In order to analyze precisely the pore structure of porous
of concrete [15], disclosing the temporal change of tropical metal materials, a set of novel fractal analytical software
cyclone Dan complex structure [16], and discussing the heat of porous metal materials was designed and developed by
flux of subcooled pool boiling [17] and thermal conductivity us based on the fractal theory and the computer image
performance [18]. processing technology. Figure 1 shows the processing page
For porous metal material, its performance is strongly layout view of the software. For the software, the shape
affected by the pore structure, so the accurate description factor and the sharpness factor of pore were introduced
of pore structure has become a research hotspot in recent as discussed elsewhere [20, 21]. In addition, the characters
years [19]. The fractal theory is an effective method for and the processing steps of the software were described
discussing the pore structure of porous metal materials. This systematically in our previous works [22–24].
2 Journal of Chemistry

Fractal analytical software of porous metal materials


File View Image processing Fractal dimension Help File Pretreat
Binarization
Open
Scale
Save
Filter
Exit
Segment

Extraction Statistics analysis


Analysis boundary Curvature distributed graph
Single pore Box method
Rectangular area
Whole image Sandbox method
Erasion boundary Parameter set

Number Perimeter (𝜇m) Area (𝜇m2 ) Maximum pore size (𝜇m) Average pore size (𝜇m)

Figure 1: Layout view of software.

3. Analysis of Affecting Factors of Table 1: Volume porosity of porous metal materials made by
Fractal Dimension irregular powder with the particle size of 44∼74 𝜇m.

The fractal dimension of pore structure of porous metal mate- Fractal 1.2304 1.2515 1.271 1.288 1.298
dimension
rials was affected by the porosity, the image magnification,
the powder particle size, the powder morphology, the image Volume porosity 19.3 21.4 24.7 29.7 34.3
tested, 𝜃/%
resolution, and the image region chosen [22–27].
Volume porosity 17.07 19.06 21.44 26.17 30.29
calculated, 𝜃/%
3.1. Effects of Porosity and Powder Particle Size on Fractal Δ𝜃/% 2.23 2.34 3.26 3.53 4.01
Dimension. At the same image magnification, the quantity
and the complexity of pore increase slowly as the porosity
of porous metal materials increases, so the fractal dimension Table 2: Volume porosity of porous metal materials made by
increases gradually with the porosity increasing [22, 24], irregular powder with the particle size of 74∼150 𝜇m.
shown in Figure 2.
In addition, the relationship between the fractal dimen- Fractal 1.1655 1.2072 1.2228 1.2441 1.2492
dimension
sion, the volume porosity, and the surface porosity can be
defined as [27] Volume porosity 20.1 22.6 26.1 29.6 33.5
tested, 𝜃/%
1 Volume porosity 18.43 21.23 24 26.59 31.85
𝜃 = 𝑀 (1 + ) 𝜑, (1) calculated, 𝜃/%
𝐷
Δ𝜃/% 1.67 1.37 2.1 3.01 1.65
where 𝜃 is the volume porosity, 𝐷 is the fractal dimension, 𝜑
is the surface porosity, and 𝑀 is the emendatory coefficient.
According to (1), the volume porosity of porous metal It also can be seen from Figure 2 that the fractal dimen-
materials made by irregular powder can be predicted and the sion of pore structure of porous metal materials manufac-
calculated results are in good agreement with the tested ones, tured by fine powder is higher than that by coarse powder
shown in Tables 1 and 2. at the identical porosity.
Journal of Chemistry 3

1.30 1.28

1.28 1.26

Fractal dimension, D
Fractal dimension, D

1.26 1.24

1.24 1.22
1.22
1.20
1.20
1.18
1.18
1.16
1.16
1.14
20 24 28 32 36 20 24 28 32 36 40
Porosity, 𝜃 (%) Porosity, 𝜃 (%)

44∼74 𝜇m 48∼106 𝜇m
74∼150 𝜇m 106∼180 𝜇m
(a) (b)

Figure 2: Fractal dimension versus porosity: (a) irregular powder and (b) spherical powder.

1.7 1.60

1.6
1.55
Fractal dimension, D

Fractal dimension, D

1.5
1.50
1.4
1.45
Irregular powder
1.3

1.40
1.2
Spherical powder

1.1 1.35
100 200 300 400 500 0 100 200 300 400 500
Magnification Magnification

Porosity of 30.8% Porosity of 85%


Experimental results Experimental results Experimental results
Fitting curves Fitting curves Fitting curve
(a) (b)

Figure 3: Fractal dimension versus image magnification for (a) porous metal powder materials and (b) porous metal fiber materials.

3.2. Effects of Image Magnification and Powder Morphology where 𝑥 is the image magnification and 𝑎0 , 𝑎1 , and 𝑎2 are the
on Fractal Dimension. At the same porosity, the quantity of correlative coefficients.
pore decreases slowly as the image magnification increases, In addition, it also can be seen from Figure 3(a) that the
so the fractal dimension decreases gradually with the image fractal dimension of pore structure of porous metal materials
magnification increasing, shown in Figure 3 [23, 24]. manufactured by irregular powder is higher than that by
Additionally, for porous metal powder materials spherical powder at the identical image magnification and the
(Figure 3(a)), the optimal image magnification ranges from same powder particle size.
200 to 500. Furthermore, the relationship between the fractal
dimension and the image magnification can be described by
[23, 24] 3.3. Effect of Image Resolution on Fractal Dimension. The
fractal dimension is also influenced by the image resolution,
𝑥 shown in Figure 4 [26, 27]. It can be seen from Figure 4
𝐷 = 𝑎0 exp (− ) + 𝑎2 , (2)
𝑎1 that the fractal dimension at the original image resolution is
4 Journal of Chemistry

1.80 Table 4: Air permeability (𝐾) of porous metal materials made by


irregular powder with the particle size of 44∼74 𝜇m.
1.78
Fractal 1.2304 1.2515 1.271 1.288 1.298
Fractal dimension, D

1.76 dimension
1.74 𝐾tested 14.20 15.30 19.50 30.01 47.90
(m3 /m2 ⋅kPa⋅h)
1.72 𝐾calculated 11.33 13.60 17.73 28.75 46.04
1.70 (m3 /m2 ⋅kPa⋅h)
Δ𝐾/% 2.87 1.70 1.77 1.26 1.86
1.68

1.66
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Pixel size (𝜇m) where 𝜎𝑐 is the compressive strength and 𝑎, 𝑘, and 𝑏 are the
correlative coefficients.
Figure 4: Fractal dimension versus image resolution for porous For porous metal fiber materials, the relationship between
metal fiber materials. the tensile strength and the fractal dimension can be regarded
as
Table 3: Fractal dimensions of pore structure for different regions
chosen. 𝜎𝑏 = 𝑎 − 𝑘 ⋅ 𝐷, (4)
Porosity/% 73.1% 88.2%
where 𝜎𝑏 is the tensile strength and 𝑎 and 𝑘 are the correlative
Region chosen 1 2 3 1 2 3
coefficients.
Fractal 1.7609 1.7623 1.7671 1.8241 1.8284 1.8251
dimension/𝐷
4.2. Relationship between the Sound Absorbing Performance
and the Fractal Dimension. The relationship between the
the lowest, about 1.665, while it increases gradually with the sound absorption coefficients and the fractal dimension is
image resolution increasing or decreasing. shown in Figure 6 [28]. The results show that the sound
absorption coefficients of porous metal fiber materials
decrease little by little as the fractal dimension increases.
3.4. Effect of Image Region Chosen on Fractal Dimension.
The effects of different image regions chosen on the fractal
dimensions are shown in Table 3. It can be seen from Table 3 4.3. Relationship between the Air Permeability and the Fractal
that the difference between the fractal dimensions is very Dimension. The relationship between the air permeability
small, lower than 1%, so the fractal dimension is not almost and the fractal dimension is shown in Figure 7 [28]. The
affected by the image region chosen. The results show that the results show that the air permeability of porous metal fiber
microstructure of porous metal materials is homogeneous. materials increases linearly as the fractal dimension increases.
In addition, the relationship between the air permeability and
the fractal dimension can be presented by
4. Relationship between the Material
Performance and the Fractal Dimension 2
𝐶 ⋅ (6.15𝑑𝑓 𝜃3.35 ) ⋅ 𝑀 (1 + 1/𝐷) 𝜑 (5)
The performance of porous metal materials is strongly 𝐾= ,
16
affected by the pore structure. In the paper, the relationships
between the mechanical properties, the acoustic properties, where 𝐾 is the air permeability, 𝐶 is the Kozeny Carman
the air permeability, and the microresistance of porous metal constant, 𝑑𝑓 is the diameter of metal fiber, 𝜃 is the volume
materials and the fractal dimension were reviewed. porosity, 𝜑 is the surface porosity, and 𝑀 is the emendatory
coefficient.
4.1. Relationship between the Mechanical Performance and the According to (5), the air permeability of porous metal
Fractal Dimension. The relationships between the compres- powder materials is calculated and the calculated results are
sive strength, the tensile strength, and the fractal dimension in good agreement with the tested ones, shown in Table 4.
are shown in Figure 5 [22, 28]. The results show that the So the air permeability of porous metal materials can be
compressive or the tensile strength decreases gradually with predicted using (5).
the fractal dimension increasing.
For porous metal powder materials, the relationship 4.4. Relationship between the Microresistance and the Fractal
between the compressive strength and the fractal dimension Dimension. The relationship between the microresistance of
can be regarded as porous metal materials and the fractal dimension is shown
in Figure 8 [29]. The results show that the microresistance
𝜎𝑐 = 𝑎 (1 − exp (𝑘 (𝐷 − 𝑏))) , (3) increases slowly as the fractal dimension increases.
Journal of Chemistry 5

80 14

70 12
Compressive strength, 𝜎c (MPa)

Tensile strength, 𝜎b (MPa)


10
60
8
50
6
40
4

30 2

20 0
1.18 1.20 1.22 1.24 1.40 1.44 1.48 1.52 1.56 1.60
Fractal dimension, D Fractal dimension, D

Experimental results
Fitted results
(a) (b)

Figure 5: Mechanical properties versus fractal dimension for (a) porous metal powder materials and (b) porous metal fiber materials.

0.9 16
Air permeability, K (×10 m /m ·KPa·h)

0.8 14
Sound absorption coefficient, 𝛼

0.7 12
3

0.6 10
3

0.5 8

0.4 6

0.3 4

0.2 2

0.1 0
1.40 1.44 1.48 1.52 1.56 1.60 1.40 1.44 1.48 1.52 1.56 1.60
Fractal dimension, D Fractal dimension, D

1000 Hz 2500 Hz Figure 7: Air permeability versus fractal dimension for porous
1600 Hz 3000 Hz metal fiber materials.
2000 Hz

Figure 6: Sound absorption coefficient versus fractal dimension for


27
porous metal fiber materials.
26
Microresistance R (𝜇Ω)

25
5. Summary and Conclusions
24
The fractal theory is a promising method for studying the
23
pore structure of porous metal materials. The paper attempts
a reasonably comprehensive review about our achievements 22
covering the affecting factors analysis of fractal dimension,
the relationships between the fractal dimension and the com- 21
pressive strength, the tensile strength, the sound absorption
1.45 1.50 1.55 1.60 1.65 1.70 1.75
coefficient, the air permeability, and the microresistance.
Fractal dimension, D
Most of the future research on fractal theory application in
porous metal materials will be driven by material structure Figure 8: Microresistance versus fractal dimension for porous metal
and properties. Furthermore, more accurate description of powder materials.
6 Journal of Chemistry

pore structure based on the fractal theory is indispensable for [12] T. Furusato, T. Ihara, S. Kameda et al., “Fractal analysis of pos-
optimizing the pore structure for obtaining excellent material itive pulsed streamer patterns in supercritical carbon dioxide,”
performance. IEEE Transactions on Plasma Science, vol. 40, no. 10, pp. 2425–
2430, 2012.
[13] W. J. Yuan, Q. Y. Zhu, C. J. Deng, and H. X. Zhu, “Fractal analysis
Conflict of Interests of fracture surfaces in refractories,” China’s Refractories, no. 1,
pp. 27–31, 2014.
The authors declare that they have no financial and personal
relationships with other people or organizations that can [14] M. Z. Gao, W. C. Jin, Z. X. Dai, and J. Xie, “Relevance between
abutment pressure and fractal dimension of crack network
inappropriately influence their work, and there is no profes-
induced by mining,” International Journal of Mining Science and
sional or other personal interest of any nature or kind in any Technology, vol. 23, no. 6, pp. 925–930, 2013.
product, service, and/or company that could be construed as
[15] Q. Xu, J. Y. Chen, J. Li, and M. M. Wang, “Multi-scale numerical
influencing the position presented in the paper. In addition, model for simulating concrete material based on fractal theory,”
they have no potential conflict of interests regarding each Acta Mechanica Solida Sinica, vol. 26, no. 4, pp. 344–352, 2013.
other.
[16] Z. X. Luo, Y. Wang, G. L. Ma et al., “Possible causes of the
variation in fractal dimension of the perimeter during the
Acknowledgments tropical cyclone Dan motion,” Science China: Earth Sciences, vol.
57, no. 6, pp. 1383–1392, 2014.
Authors gratefully acknowledge the financial support from [17] B. Q. Xiao, Y. Yang, and X. F. Xu, “Subcooled pool boiling heat
the National Natural Science Foundation of China (51301141, transfer in fractal nanofluids: a novel analytical model,” Chinese
51074130) and the Youth Scientific Star Project of Shaanxi Physics B, vol. 23, no. 2, Article ID 026601, 2014.
Province (2014KJXX-24) and the Opening Project of State [18] F. L. Zhu, S. Z. Cui, and B. H. Gu, “Fractal analysis for effective
Key Laboratory of Explosion Science and Technology (Bei- thermal conductivity of random fibrous porous materials,”
jing Institute of Technology) (KFJJ15-02M). Physics Letters A, vol. 374, no. 43, pp. 4411–4414, 2010.
[19] H. P. Tang, J. L. Zhu, Z. P. Xi, X. B. Di, J. Y. Wang, and Q. B. Ao,
“Impact factors of fractal analysis of porous structure,” Science
References China Technological Sciences, vol. 53, no. 2, pp. 348–351, 2010.
[1] A. S. Kulkarni, Nature of Branching in Disordered Materials, [20] N. Chawla and X. Deng, “Microstructure and mechanical
University of Cincinnati, 2007. behavior of porous sintered steels,” Materials Science and
[2] W. Y. Chu, Y. J. Su, K. W. Gao et al., Fractal in Materials Science, Engineering A, vol. 390, no. 1-2, pp. 98–112, 2005.
Chemical Industry Press, Beijing, China, 2004, (Chinese). [21] P. Beiss and M. Dalgic, “Structure property relationships in
[3] B. B. Mandelbrot, The Fractal Geometry of Nature, W. H. porous sintered steels,” Materials Chemistry and Physics, vol. 67,
Freeman and Company, New York, NY, USA, 1977. no. 1–3, pp. 37–42, 2001.
[4] B. M. Yu and P. Cheng, “A fractal permeability model for bi- [22] J. Z. Wang, H. P. Tang, J. L. Zhu, Q. B. Ao, H. Zhi, and J. Ma,
dispersed porous media,” International Journal of Heat and Mass “Relationship between compressive strength and fractal dimen-
Transfer, vol. 45, no. 14, pp. 2983–2993, 2002. sion of pore structure,” Rare Metal Materials and Engineering,
[5] X. H. Tan, X. P. Li, J. Y. Liu, G. D. Zhang, and L. H. Zhang, vol. 42, no. 12, pp. 2433–2436, 2013.
“Analysis of permeability for transient two-phase flow in fractal [23] J. Z. Wang, Z. P. Xi, H. P. Tang, W. D. Huang, J. L. Zhu, and Q.
porous media,” Journal of Applied Physics, vol. 115, no. 11, pp. B. Ao, “Fractal dimension for porous metal materials of FeCrAl
113502–113507, 2014. fiber,” Transactions of Nonferrous Metals Society of China, vol.
[6] Q. Zheng and B. Yu, “A fractal permeability model for gas flow 23, no. 4, pp. 1046–1051, 2013.
through dual-porosity media,” Journal of Applied Physics, vol. [24] H. P. Tang, J. Z. Wang, J. L. Zhu et al., “Fractal dimension of
111, no. 2, Article ID 024316, 2012. pore-structure of porous metal materials made by stainless steel
[7] L. Ma and R. He, “Steady-state gas diffusion in fractal porous powder,” Powder Technology, vol. 217, pp. 383–387, 2012.
media,” Journal of Tsinghua University (Science and Technology), [25] X. B. Di, Z. P. Xi, H. P. Tang, and J. L. Zhu, “Analysis of pore-
vol. 53, no. 10, pp. 1459–1463, 2013 (Chinese). structure by fractal in sintered porous stainless steel fibres,”
[8] M. Liang and H. Rong, “Unsteady state gas diffusion in fractal Rare Metal Materials and Engineering, vol. 36, supplement 3, pp.
porous media,” CIESC Journal, vol. 64, no. 9, pp. 3139–3144, 2013 S571–S573, 2007.
(Chinese). [26] H. P. Tang, J. L. Zhu, Z. P. Xi, X. B. Di, J. Y. Wang, and Q. B. Ao,
[9] J. C. Cai, “A fractal approach to low velocity non-Darcy flow in “Impact factors of fractal analysis of porous structure,” Science
a low permeability porous medium,” Chinese Physics B, vol. 23, China Technological Sciences, vol. 53, no. 2, pp. 348–351, 2010.
no. 4, Article ID 044701, 2014. [27] X. B. Di, Research on Pore-Structure Fractal Characterization in
[10] Y. Ju, Q. G. Zhang, Y. M. Yang, H. P. Xie, F. Gao, and H. J. Porous Metals, Northeastern University, Shenyang, China, 2008,
Wang, “An experimental investigation on the mechanism of (Chinese).
fluid flow through single rough fracture of rock,” Science China [28] H.-P. Tang, J.-Z. Wang, Q.-B. Ao, and H. Zhi, “Effect of pore
Technological Sciences, vol. 56, no. 8, pp. 2070–2080, 2013. structure on performance of porous metal fiber materials,” Rare
[11] A. Vinogradov, I. S. Yasnikov, and Y. Estrin, “Stochastic disloca- Metal Materials and Engineering. In press.
tion kinetics and fractal structures in deforming metals probed [29] C. Li, H. P. Tang, J. L. Zhu, Q. B. Ao, and H. Zhi, “Fractal
by acoustic emission and surface topography measurements,” characterization of ultra-thin porous Ti plate,” Heat Treatment
Journal of Applied Physics, vol. 115, no. 23, pp. 233506–233515, of Metals, vol. 36, no. 10, pp. 15–18, 2011.
2014.
Hindawi Publishing Corporation
Journal of Chemistry
Volume 2015, Article ID 815051, 13 pages
http://dx.doi.org/10.1155/2015/815051

Research Article
Finite Element Method Using a Characteristic-Based Split for
Numerical Simulation of a Carbonate Fracture-Cave Reservoir

Liehui Zhang, Yuhui Zhou, Lei Zhao, and Deliang Zhang


State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum University,
Chengdu, Sichuan 610500, China

Correspondence should be addressed to Liehui Zhang; zhangliehui@vip.163.com

Received 25 July 2014; Accepted 5 September 2014

Academic Editor: Jianchao Cai

Copyright © 2015 Liehui Zhang et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Fracture-cave carbonate reservoirs occur widely in source rocks and are prospects for exploitation worldwide. However, the
presence of massive caves and multiscale fracture systems results in extremely complex fluid flow patterns. Therefore, in this paper,
a discrete network model for fracture-cave reservoirs was established to study fluid flow characteristics and pressure distributions
in complex flow regimes. In this study, the cave system was treated as a free-flow region, and the fluid flow in fracture systems
followed the Navier-Stokes and Darcy equations, respectively. After discrete modeling, the Galerkin finite element method was
used for numerical calculation of the single-phase free flow; the method maintains a high-precision result with low grid orientations
during the simulation. In addition, because only one linear equation requires solving at each step, the solution is obtained quickly.
Moreover, based on the proposed discrete media network model of fracture-cave reservoirs and the finite element numerical
calculation method, a corresponding simulator was also developed. The finite element numerical simulation method based on
the characteristic-based split (CBS) algorithm has proven to be applicable to complex flow problems in fracture-cave reservoirs.

1. Introduction discrete medium models. Discrete medium models describe


the heterogeneity of local-scale reservoirs better than the
Fracture-cave reservoirs play a vital role in the petroleum other models do; therefore, these models describe actual
industry, as these reservoirs contain large volumes of oil reservoir characteristics more accurately. Three numerical
and are widely distributed in areas such as Saudi Arabia, methods are commonly used to solve flow models: the finite
Iran, Iraq, Mexico, Oman, and Syria [1]. Therefore, the difference method (FDM), the finite element method (FEM),
development of reliable analytical solutions that consider the and the boundary element method (BEM). The FEM has
geometrical properties, storage capacities, and flow prop- been widely used in the field of solid mechanics and is now
erties of porous media in fracture-cave reservoirs is of also widely used in the petroleum research, thus providing
utmost importance. Many of the world’s giant fields contain an important approach to numerical simulation of problems
naturally fractured and vuggy carbonate reservoirs, and the involving discrete fractured networks [2, 3].
pore systems in these reservoirs are complex as a result of Kim and Deo [4] proposed a finite element discrete
the characteristics of carbonate rocks, which are particularly fracture model to describe multiphase flow in porous media.
sensitive to postdepositional diagenesis that can involve Their study showed that oil recovery is determined by a com-
dissolution, dolomitization, and fracturing. plex interplay of absolute matrix permeability, permeability
A great deal of research and model development has been contrasts, and flow rates; fracture capillary pressure also plays
devoted to flow behavior in fractured vuggy reservoirs; the a significant role in determining water penetration into the
focus of such research, which has been conducted worldwide, matrix [5, 6].
can be categorized into efforts to develop three types of mod- Liu et al. [7] discussed the discretization of coupled
els: continuum models, equivalent continuum models, and finite element stress equations used in commercial reservoir
2 Journal of Chemistry

simulations and compared the stress predictions of these


equations with the results of a commonly used commercial
stress program. Sheng et al. [8] used an extended FEM to
investigate an integrated and improved shale-gas transport
model and characterized the main flow mechanisms and the
discrete fracture network. Wei et al. [9] presented a second- Cave Cave
order finite element algorithm that can be used to simulate
the frequency-dependent laterolog response of axially sym-
Fracture
metric, invaded, and anisotropic formations. Qin et al. [10]
developed a flow calculation method based on the Stokes- Matrix
Brinkman model and the discrete fracture network method,
which can be used for accurate and efficient upscaling of
Figure 1: Discrete fracture-cave network model and simplified cave
processes in naturally fractured carbonate karst reservoirs.
model.
Gulbransen et al. [11] developed a multiscale mixed finite
element (MsMFE) method that can be used to examine vuggy
and naturally fractured reservoirs; the MsMFE is a first step In conclusion, the FEM has been widely studied and
toward developing a uniform multiscale multiphysics frame- applied in petroleum engineering, especially for the analy-
work. In addition, a comparison of solutions of the MsMFE sis of flow in sandstone reservoirs and naturally fractured
method and the fine-scale Stokes-Brinkman model for two reservoirs, as well as in shale-gas reservoirs; however, the
different test cases, including both short- and long-range technique has rarely been applied to fracture-cave reservoirs.
fractures, was also presented; the results demonstrate how Because of the lack of a systematic theory and appropriate
fine-scale flow in fracture networks can be represented within modeling techniques, research on theories and techniques
a coarse-scale Darcy flow model using multiscale elements, of numerical simulation using discrete medium models to
with the solution computed using the Stokes-Brinkman model giant fracture-cave reservoirs is still in an exploratory
equations. The results demonstrate that the MsMFE method stage, with existing studies concentrating mostly on frac-
can simulate flow in vuggy and naturally fractured reservoirs tured media. The characterization of vuggy media with
using a highly detailed geocellular approach. extremely complex internal geometries using theoretical
Jouini and Vega [12] estimated the elastic properties of models remains difficult. The characteristic-based split (CBS)
core plug samples from a Middle East carbonate reservoir and algorithm has been used to solve problems in solid mechanics
computed tomography scanner images by X-ray. Using the and fluid mechanics, but until now the algorithm has not been
FEM, they simulated the elastic properties of the reservoir applied to a discrete media model of fracture-cave reservoirs.
rocks. Using Hertzian’s and Hookean’s contact models, the In this paper, we applied the FEM to the study of fracture-cave
discrete element method can be used to investigate and reservoirs based on a CBS discrete medium model.
predict the elastic properties of reservoir rocks.
Jackson et al. [13] developed approaches to simulate fluid 2. Discrete Fracture-Cave Network Model
flow using the pillar-grid concept. Results showed significant
improvements in the representation of multiscale geological
Concept and Mathematical Formulation
heterogeneities and predictions of flow through those hetero- As observed in numerous studies on the field of carbonate
geneities. In their research, which was based on more than formation, three types of porosity are typically present in
20 years of development of innovative numerical methods natural fracture-cave reservoirs; these are matrix porosity,
in ocean modeling, Jackson et al. refined and modified fracture porosity, and cave porosity [23–26]. Fractures and
their approach to address the unique challenges associated caves are irregularly distributed and vary in size from micro-
with reservoir simulation. Chen [14] presented a new finite scopic to macroscopic. Figure 1 presents the conceptualized
element in ABAQUS that incorporates the extended FEM discrete fracture-cave network model. The isolated caves are
(XFEM) in the solution of hydraulic fracture problems. connected via a discrete fracture network. Flow within the
Cinar et al. [15] developed a numerical model to analyze matrix and in macroscopic fracture systems follows the Darcy
transient pressure data from naturally fractured reservoirs; law and in cave systems follows a free-flow region.
their model calculated well pressures in a single well in a In this paper, we assume that fluid flow is isothermal,
naturally fractured reservoir. The characteristic properties of single phase, and slightly compressible and that the fluid
naturally fractured reservoirs, including fracture orientation, viscosity is constant. For a laminar viscous fluid flowing
fracture length, fracture density, and fracture distribution, through an open domain (e.g., a cave system), the flow is
were also simulated in this model. The simulation results governed by the continuity and momentum equations, which
showed that the model can predict the transient pressure can be expressed in vector form as [27, 28]
response in complex naturally fractured reservoirs; in addi-
𝜕𝜌
tion, it can be used to match pressure data and diagnose + ∇ ⋅ (𝜌𝑢) = 0,
characteristic properties of naturally fractured reservoirs. In 𝜕𝑡
(1)
addition to the studies reported above, a great deal of other 𝜕𝑢
research related to the FEM has been carried out [16–22]. 𝜌 + 𝜌 (𝑢 ⋅ ∇) 𝑢 = 𝜌𝑓 + ∇ ⋅ 𝜎,
𝜕𝑡
Journal of Chemistry 3

where 𝜌 is the fluid density (kg/m3 ), 𝑢 is the fluid velocity where 𝐵 is a matrix, 𝑁 is a deformation function, 𝑚 =
(m/s), ∇ is a gradient operator, 𝑓 is the body force per unit [1, 1, 0]𝑇 , and 𝐼0 = diag(2, 2, 1). After integration by parts,
mass (m/s2 ), and 𝜎 is the fluid stress tensor (Pa). 𝜎 is given by the expressions for 𝐾𝑢 and 𝑓𝑠 are
𝜎 = −𝑝𝑓 𝐼 + 2𝜇𝜀, (2) 1 𝑇
𝐾𝑢 = − ∫ (∇𝑇 (𝑢𝑁)) (∇𝑇 (𝑢𝑁)) 𝑑Ω,
where 𝑝𝑓 is the pressure in the fracture system (Pa), 𝐼 is a unit 2 Ω
tensor, 𝜇 is the fluid viscosity (Pa s), and 𝜀, the strain rate, is (10)
1 𝑇
1 𝑓𝑠 = − ∫ (∇𝑇 (𝑢𝑁)) 𝜌𝑔 𝑑Ω.
𝜀= (∇𝑢 + 𝑢∇) . (3) 2 Ω
2
For simplicity, in this study we treat the fluid as a Step 2. The value of Δ𝑝̃ can be easily determined using the
Newtonian fluid. The density of a slightly compressible fluid expression
can be assumed to be nearly constant (i.e., incompressible).
Therefore, the governing fluid motion equations can be (𝑀𝑃 + Δ𝑡2 𝜃1 𝜃2 𝐻) Δ𝑃̃
reduced to the Navier-Stokes equations, which are (11)
̃𝑛 + 𝜃1 𝐺𝑈
= Δ𝑡 [𝐺𝑈 ̃∗ − Δ𝑡𝜃1 𝐻𝑃̃𝑛 − 𝑓𝑝 ] .
∇ ⋅ 𝑢 = 0, (4)
𝜕𝑢 The matrices are
𝜌 + 𝜌 (𝑢 ⋅ ∇) 𝑢 = −∇𝑃𝑓 + 𝜇∇2 𝑢. (5)
𝜕𝑡
𝐻 = ∫ (∇𝑁)𝑇 ∇𝑁 𝑑Ω,
The boundary conditions for (2) and (3) can be classified Ω
into two types: Dirichlet (velocity) and Neumann (traction)
1 𝑛
[29, 30], which can be expressed, respectively, as 𝑀𝑝 = ∫ 𝑁𝑇( ) 𝑁 𝑑Ω,
Ω 𝑐2
𝑢 = 𝑢𝐷, on Γ𝐷, (6) (12)
𝐺 = ∫ (∇𝑁)𝑇 𝑁 𝑑Ω,
𝑛 ⋅ 𝜎 = 𝑡𝑁, on Γ𝑁, (7) Ω

where 𝑢𝐷 is the specific velocity at the Dirichlet boundary Γ𝐷, ̃𝑛 + 𝜃1 (Δ𝑈


̃∗ − Δ𝑡∇𝑝𝑛+𝜃2 )] 𝑑Γ.
𝑛 is an outward unit normal vector at the boundary, and 𝑡𝑁 is 𝑓𝑝 = Δ𝑡 ∫ 𝑁𝑇 𝑛𝑇 [𝑈
Γ
the specific traction.
̃ is determined from
Step 3. The value of Δ𝑈
3. Derivation of the Finite Element Method
Based on CBS ̃ = Δ𝑈
̃∗ − 𝑀−1 Δ𝑡 [𝐺𝑇 (𝑝̃𝑛 + 𝜃2 Δ𝑝) Δ𝑡 𝑛
Δ𝑈 ̃ + 𝑃𝑝̃ ] , (13)
2
3.1. CBS Procedure. The theory of the CBS algorithm is based
on the split- and characteristic-Galerkin procedures, which where
are extended in the FEM. The CBS algorithm involves three
indispensable steps: Step 1 is to compute the intermediate 𝑃 = ∫ (∇ (𝑢𝑁))𝑇 (∇𝑁) 𝑑Ω. (14)
momentum variable; Step 2 is to calculate the density or pres- Ω
sure field; and Step 3 involves combining the introduction of
the density or pressure fields to obtain the correct momentum 3.2. Derivation Process of the Free-Flow Model. From (5), we
variables. derive

̃∗ in matrix form is
Step 1. The solution for Δ𝑈 𝜕𝑢 𝜕𝑢 𝜕𝑢 1 𝜕𝑝 𝜇 𝜕2 𝑢 𝜕2 𝑢
= − (𝑢 +V )− + ( + ), (15)
𝜕𝑡 𝜕𝑥 𝜕𝑦 𝜌 𝜕𝑥 𝜌 𝜕𝑥2 𝜕𝑦2
̃∗ = −𝑀−1 Δ𝑡[(𝐶𝑈
Δ𝑈 ̃ + 𝐾̃ ̃ + 𝑓𝑠 )]𝑛 , (8)
𝑢 − 𝑓) − Δ𝑡 (𝐾𝑢 𝑈
𝜕V 𝜕V 𝜕V 1 𝜕𝑝 𝜇 𝜕2 V 𝜕2 V
where the quantities identified by ∼ symbols indicate nodal = − (𝑢 +V )− + ( + ). (16)
values, and 𝜕𝑡 𝜕𝑥 𝜕𝑦 𝜌 𝜕𝑦 𝜌 𝜕𝑥2 𝜕𝑦2

𝑀 = ∫ 𝑁𝑇 𝑁𝑑Ω, From Step 1, after the split-temporal discretization, (15)


Ω becomes (omitting the third-order derivative above)
𝐶 = ∫ 𝑁𝑇 (∇ (𝑢𝑁)) 𝑑Ω, Δ𝑢 = 𝑢𝑛+1 − 𝑢𝑛
Ω
(9)
2 𝜕𝑁𝑇 𝜕𝑁 𝜕𝑢 𝜕𝑢 1 𝜕𝑝 𝜇 𝜕2 𝑢 𝜕2 𝑢
𝐾 = ∫ 𝐵𝑇 𝜇 (𝐼0 − 𝑚𝑚𝑇 ) 𝐵𝑑Ω = ∫ 𝜇 𝑑Ω, = Δ𝑡 [− (𝑢 +V )− + ( + )]
Ω 3 Ω 𝜕𝑥𝑗 𝜕𝑥𝑗 𝜕𝑥 𝜕𝑦 𝜌 𝜕𝑥 𝜌 𝜕𝑥2 𝜕𝑦2 (17)

Δ𝑡2 𝜕 𝜕𝑢 𝜕𝑢 1 𝜕𝑝
𝑓 = ∫ 𝑁𝑇 (𝜌𝑔) 𝑑Ω + ∫ 𝑁𝑇 𝑡𝑑 𝑑Γ, + 𝑢 [(𝑢 +V )+ ].
Ω Γ 2 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜌 𝜕𝑥
4 Journal of Chemistry

From Step 2, after spatial discretization, we find that Δ𝑡2 𝜕𝑢 𝜕𝑢


+ ∫ ∇ (𝑁𝑢 𝑢) [− (𝑢 + V )] 𝑑Ω
2 Ω 𝜕𝑥 𝜕𝑦
Δ𝑢 = 𝑁𝑢 Δ̃
𝑢,

Δ𝑢∗ = 𝑁𝑢 Δ̃
𝑢∗ , 𝜇 𝜕𝑢 𝜕𝑢
+ Δ𝑡 ∫ 𝑁𝑢 ( + ) 𝑑Γ.
𝜌 Γ 𝜕𝑥 𝜕𝑦
𝑢 = 𝑁𝑢 𝑢̃, (19)
V = 𝑁V ̃V,
The left side of (19) can be expressed as
̃
𝑝 = 𝑁𝑝 𝑝,

𝑁 = [𝑁1 , 𝑁2 , 𝑁3 ] , ∫ 𝑁𝑢 Δ𝑢∗ 𝑑Ω
Ω

1 2 3 𝑇
Δ̃
𝑢 = [Δ𝑢 , Δ𝑢 , Δ𝑢 ] , (18) = ∫ 𝑁𝑢 (𝑁1 Δ𝑢1∗ + 𝑁2 Δ𝑢2∗ + 𝑁3 Δ𝑢3∗ ) 𝑑Ω
Ω
∗ 1∗ 2∗ 3∗ 𝑇
Δ̃
𝑢 = [Δ𝑢 , Δ𝑢 , Δ𝑢 ] ,
= [∫ 𝑁𝑢 𝑁1 𝑑Ω ∫ 𝑁𝑢 𝑁2 𝑑Ω ∫ 𝑁𝑢 𝑁3 𝑑Ω]
𝑇 Ω Ω Ω
𝑢̃ = [𝑢1 , 𝑢2 , 𝑢3 ] , 1∗ 2∗ 3∗
× [Δ𝑢 Δ𝑢 Δ𝑢 ]
1 2 3 𝑇
̃V = [V , V , V ] , 1 1 1 2 1 3 (20)
[∫Ω 𝑁 𝑁 𝑑Ω ∫Ω 𝑁 𝑁 𝑑Ω ∫Ω 𝑁 𝑁 𝑑Ω]
𝑇 [ ]
𝑝 = [𝑝1 , 𝑝2 , 𝑝3 ] , [ ]
=[ 2 1 2 2 2 3
[∫ 𝑁 𝑁 𝑑Ω ∫ 𝑁 𝑁 𝑑Ω ∫ 𝑁 𝑁 𝑑Ω]
]
𝑁𝑢 = 𝑁V = 𝑁𝑝 . [ Ω Ω Ω ]
[ ]
∫ 𝑁3 𝑁1 𝑑Ω ∫ 𝑁3 𝑁2 𝑑Ω ∫ 𝑁3 𝑁3 𝑑Ω
Then, [ Ω Ω Ω ]

𝜕𝑢 𝜕𝑢 Δ𝑢1∗

∫ 𝑁𝑢 Δ𝑢 𝑑Ω = Δ𝑡 [− ∫ 𝑁𝑢 (𝑢 + V ) 𝑑Ω [Δ𝑢2∗ ]
𝜕𝑥 𝜕𝑦 ×[ ].
Ω Ω
3∗
[Δ𝑢 ]
𝜇 𝜕𝑢 𝜕𝑢
− ∫ ∇ ⋅ 𝑁𝑢 ( + ) 𝑑Ω]
𝜌 Ω 𝜕𝑥 𝜕𝑦 The first term on the right side of (19) can be expressed as

𝜕𝑢 𝜕𝑢
∫ 𝑁𝑢 (𝑢 + V ) 𝑑Ω
Ω 𝜕𝑥 𝜕𝑦

𝜕𝑁1 1 𝜕𝑁2 2 𝜕𝑁3 3 𝜕𝑁1 1 𝜕𝑁2 2 𝜕𝑁3 3


= ∫ 𝑁𝑢 (𝑢 ( 𝑢 + 𝑢 + 𝑢 ) + V( 𝑢 + 𝑢 + 𝑢 ))
Ω 𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑦

𝜕𝑁1 𝜕𝑁1 𝜕𝑁2 𝜕𝑁2 𝜕𝑁3 𝜕𝑁3


= ∫ 𝑁𝑢 ((𝑢 +V ) 𝑢1 + (𝑢 +V ) 𝑢2 + (𝑢 +V ) 𝑢3 ) 𝑑Ω
Ω 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦

𝑢1
𝜕𝑁1 𝜕𝑁1 𝜕𝑁2 𝜕𝑁2 𝜕𝑁3 𝜕𝑁3
= [∫ 𝑁𝑢 (𝑢 +V ) 𝑑Ω ∫ 𝑁𝑢 (𝑢 +V ) 𝑑Ω ∫ 𝑁𝑢 (𝑢 +V ) 𝑑Ω] 𝑢2 ]
[
Ω 𝜕𝑥 𝜕𝑦 Ω 𝜕𝑥 𝜕𝑦 Ω 𝜕𝑥 𝜕𝑦 3
[𝑢 ]

1 𝜕𝑁1 𝜕𝑁1 1 𝜕𝑁2 𝜕𝑁2 1 𝜕𝑁3 𝜕𝑁3



[ Ω 𝑁 (𝑢 + V ) 𝑑Ω ∫ 𝑁 (𝑢 + V ) 𝑑Ω ∫ 𝑁 (𝑢 + V ) 𝑑Ω]
[ 𝜕𝑥 𝜕𝑦 Ω 𝜕𝑥 𝜕𝑦 Ω 𝜕𝑥 𝜕𝑦 ] 1
[ ] 𝑢
[ 𝜕𝑁 1
𝜕𝑁 1
𝜕𝑁 2
𝜕𝑁 2
𝜕𝑁 3
𝜕𝑁 3 ] 2
[ 2
= [∫ 𝑁 (𝑢 2 2
) 𝑑Ω] [ ]
𝜕𝑥
+V
𝜕𝑦
) 𝑑Ω ∫ 𝑁 (𝑢
𝜕𝑥
+V
𝜕𝑦
) 𝑑Ω ∫ 𝑁 (𝑢
𝜕𝑥
+V
𝜕𝑦 ] 𝑢3 ,
[ Ω Ω Ω ] 𝑢
[ ][ ]
[ 𝜕𝑁1 𝜕𝑁1 𝜕𝑁2 𝜕𝑁2 𝜕𝑁3 𝜕𝑁3 ]
∫ 𝑁3 (𝑢 +V ) 𝑑Ω ∫ 𝑁3 (𝑢 +V ) 𝑑Ω ∫ 𝑁3 (𝑢 +V ) 𝑑Ω
[ Ω 𝜕𝑥 𝜕𝑦 Ω 𝜕𝑥 𝜕𝑦 Ω 𝜕𝑥 𝜕𝑦 ]
(21)
Journal of Chemistry 5

1 𝑇
where 1 𝜕𝑁
[∫ 𝑁𝑢 𝑁 𝜕𝑥 𝑑Ω]
[ Ω ]
[ 1 ]
[∫ 𝑁 𝑁2 𝜕𝑁 𝑑Ω]
[ 𝑢
𝜕𝑥 ] 𝑢1
[ Ω ]
[ 1 ] [𝑢 2 ]
[∫ 𝑁 𝑁3 𝜕𝑁 𝑑Ω] [ 3]
𝜕𝑁1 𝜕𝑁1 [
[
𝑢
𝜕𝑥
]
] [𝑢 ]
∫ 𝑁𝑢 (𝑢 +V ) 𝑑Ω =[ Ω 1 ] [ 1] .
Ω 𝜕𝑥 𝜕𝑦 [∫ 𝑁 𝑁1 𝜕𝑁 𝑑Ω] [V ]
[ ] [ 2]
[ Ω
𝑢
𝜕𝑦 ] [V ]
[ 1 ]
𝜕𝑁1 [ 2 𝜕𝑁 ] 3
[V ]
= ∫ 𝑁𝑢 ( (𝑁1 𝑢1 + 𝑁2 𝑢2 + 𝑁3 𝑢3 ) [∫ 𝑁𝑢 𝑁 𝑑Ω]
𝜕𝑥 [ Ω 𝜕𝑦 ]
Ω [ 1 ]
[ 3 𝜕𝑁 ]
∫ 𝑁𝑢 𝑁 𝑑Ω
𝜕𝑁1 [ Ω 𝜕𝑦 ]
+ (𝑁1 V1 + 𝑁2 V2 + 𝑁3 V3 ) ) 𝑑Ω
𝜕𝑦 (22)

The second term on the right side of (19) can be expressed as

𝜕𝑁𝑢 𝜕𝑢 𝜕𝑢
∫ ( + ) 𝑑Ω
Ω 𝜕𝑥𝑖 𝜕𝑥 𝜕𝑦

𝜕𝑁𝑢 𝜕𝑁1 𝜕𝑁1 𝜕𝑁2 𝜕𝑁2 𝜕𝑁3 𝜕𝑁3


=∫ (( + ) 𝑢1 + ( + ) 𝑢2 + ( + ) 𝑢3 ) 𝑑Ω
Ω 𝜕𝑥𝑖 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 (23)

𝑢1
𝜕𝑁𝑢 𝜕𝑁1 𝜕𝑁1 𝜕𝑁𝑢 𝜕𝑁2 𝜕𝑁2 𝜕𝑁𝑢 𝜕𝑁3 𝜕𝑁3
=∫ [ ( + ) ( + ) ( + )] 𝑢2 ] 𝑑Ω.
[
Ω 𝜕𝑥𝑖 𝜕𝑥 𝜕𝑦 𝜕𝑥𝑖 𝜕𝑥 𝜕𝑦 𝜕𝑥𝑖 𝜕𝑥 𝜕𝑦 3
[𝑢 ]

The third term on the right side of (19) can be expressed as

𝜕𝑁𝑢 𝜕𝑢 𝜕𝑢
∫ (𝑁𝑢 𝑢) [− (𝑢 + V )] 𝑑Ω
Ω 𝜕𝑥𝑖 𝜕𝑥 𝜕𝑦

𝜕 𝜕𝑁1 𝜕𝑁1 𝜕𝑁2 𝜕𝑁2 𝜕𝑁3 𝜕𝑁3


= −∫ (𝑁𝑢 𝑢) [(𝑢 +V ) 𝑢1 + 𝑢 ( +V ) 𝑢2 + 𝑢 ( +V ) 𝑢3 ] 𝑑Ω
Ω 𝜕𝑥𝑖 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦

𝜕 1 𝜕𝑁1 𝜕𝑁1 𝜕 1 𝜕𝑁3 𝜕𝑁3 (24)


[ 𝜕𝑥 (𝑁 𝑢) (𝑢 + V ) ⋅ ⋅ ⋅ (𝑁 𝑢) (𝑢 + V )
[ 𝑖 𝜕𝑥 𝜕𝑦 𝜕𝑥𝑖 𝜕𝑥 𝜕𝑦 ] ] 𝑢1
[ 𝜕 𝜕𝑁1 𝜕𝑁1 𝜕 𝜕𝑁3 𝜕𝑁3 ]
[ ]
= −∫ [ 2
(𝑁 𝑢) (𝑢 +V ) ⋅⋅⋅ 2
(𝑁 𝑢) (𝑢 +V )] [𝑢2 ] 𝑑Ω,
Ω [ 𝜕𝑥𝑖 𝜕𝑥 𝜕𝑦 𝜕𝑥𝑖 𝜕𝑥 𝜕𝑦 ] 3
[ ] 𝑢
[ 𝜕 3 𝜕𝑁1 𝜕𝑁1 𝜕 3 𝜕𝑁3 𝜕𝑁3 ] [ ]
(𝑁 𝑢) (𝑢 +V ) ⋅⋅⋅ (𝑁 𝑢) (𝑢 +V )
[ 𝜕𝑥𝑖 𝜕𝑥 𝜕𝑦 𝜕𝑥𝑖 𝜕𝑥 𝜕𝑦 ]

where 𝜕𝑁1
+ (𝑁1 𝑢1 + 𝑁2 𝑢2 + 𝑁3 𝑢3 )
𝜕𝑥
𝜕
𝜕 (𝑁1 𝑢) 𝜕 (𝑁1 V) + 𝑁1 (𝑁1 V1 + 𝑁2 V2 + 𝑁3 V3 )
𝜕 𝜕𝑦
(𝑁1 𝑢𝑖 ) = +
𝜕𝑥𝑖 𝜕𝑥 𝜕𝑦
𝜕𝑁1
1 1 + (𝑁1 V1 + 𝑁2 V2 + 𝑁3 V3 )
𝜕𝑢 𝜕𝑁 𝜕V 𝜕𝑁 𝜕𝑦
= 𝑁1 +𝑢 + 𝑁1 + V
𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦
𝜕𝑁1 𝜕𝑁1
𝜕 = (𝑁1 + 𝑁1 ) 𝑢1
= 𝑁1 (𝑁1 𝑢1 + 𝑁2 𝑢2 + 𝑁3 𝑢3 ) 𝜕𝑥 𝜕𝑥
𝜕𝑥
6 Journal of Chemistry

𝜕𝑁2 𝜕𝑁1 then


+ (𝑁1 + 𝑁2 ) 𝑢2
𝜕𝑥 𝜕𝑥
𝜇 Δ𝑡
3 1
𝑢∗ = Δ𝑡 [−𝐶𝑢 𝑢̃ −
𝑀𝑢 Δ̃ 𝐾𝑙 𝑢̃ − 𝐾𝑢 𝑢̃] + 𝑓𝑢 . (27)
1 𝜕𝑁 3 𝜕𝑁 3 𝜌 2
+ (𝑁 +𝑁 )𝑢
𝜕𝑥 𝜕𝑥
Similarly, (16) can be expressed as
1 1
𝜕𝑁 𝜕𝑁
+ (𝑁1 + 𝑁1 ) V1 𝜇 Δ𝑡
𝜕𝑦 𝜕𝑦 𝑀𝑢 Δ̃V∗ = Δ𝑡 [−𝐶𝑢 ̃V − 𝐾 ̃V − 𝐾V ̃V] + 𝑓V . (28)
𝜌 𝑙 2
𝜕𝑁2 𝜕𝑁1 2
+ (𝑁1 + 𝑁2 )V Letting
𝜕𝑦 𝜕𝑦
𝑇
3 1 Δ𝑈∗ = [Δ𝑢1∗ ΔV1∗ Δ𝑢2∗ ΔV2∗ Δ𝑢3∗ ΔV3∗ ] , (29)
1 𝜕𝑁 3 𝜕𝑁 3
+ (𝑁 +𝑁 )V .
𝜕𝑦 𝜕𝑦 we can derive the result of Δ𝑈∗ , and (5) becomes
(25) 1
𝜕𝑢𝑛+𝜃 𝜕𝑢𝑛 𝜕Δ𝑢
Δ𝑡 = Δ𝑡 ( + 𝜃1 ) = 0. (30)
𝜕𝑥𝑖 𝜕𝑥𝑖 𝜕𝑥𝑖

According to CBS polymorphisms,


Letting 2
1 𝜕𝑝𝑛+𝜃
Δ𝑢 = 𝑢𝑛+1 − 𝑢𝑛 = Δ𝑢∗ − Δ𝑡
𝜌 𝜕𝑥𝑖
2
(31)
Δ𝑡2 1 𝜕 𝜕𝑝𝑛+𝜃
+ 𝑢 ( ),
2 𝜌 𝜕𝑥𝑖 𝜕𝑥𝑖
𝑀𝑢 = ∫ 𝑁𝑢𝑇 𝑁𝑢 𝑑Ω, and combining (31) and (30) gives
Ω

𝜕𝑢𝑖 𝜕𝑢 𝜕Δ𝑢∗ 1 𝜕2 𝑝 𝜕2 Δ𝑝
𝐶𝑢 = ∫ 𝑁𝑢𝑇 (𝑢𝑖 ) 𝑑Ω, Δ𝑡 ( + 𝜃1 − Δ𝑡𝜃1 ( 2 + 𝜃2 )) = 0. (32)
Ω 𝜕𝑥𝑖 𝜕𝑥𝑖 𝜕𝑥𝑖 𝜌 𝜕𝑥𝑖 𝜕𝑥𝑖2
𝜕𝑁𝑢 𝜕𝑢𝑖 The weak form of (32) is written using the characteristic-
𝐾𝑙 = ∫ 𝑑Ω,
Ω 𝜕𝑥𝑖 𝜕𝑥𝑖 Galerkin approximation as
𝑇 (26)
𝜕(𝑢𝑁𝑢 ) 𝜕𝑢 𝜕𝑁𝑝 1 𝜕𝑝 𝜕Δ𝑝
𝐾𝑢 = ∫ (𝑢𝑖 𝑖 ) 𝑑Ω, Δ𝑡 ∫ [𝑢 + 𝜃1 (Δ𝑢∗ − Δ𝑡 ( + 𝜃2 ))] 𝑑Ω
Ω 𝜕𝑥𝑖 𝜕𝑥𝑖 Ω 𝜕𝑥𝑖 𝜌 𝜕𝑥𝑖 𝜕𝑥𝑖

𝜕𝑢 𝜕𝑢 1 𝜕𝑝 𝜕Δ𝑝 (33)
𝑓𝑢 = Δ𝑡 ∫ 𝑁𝑢 ( + ) 𝑑Γ, − Δ𝑡 ∫ 𝑁𝑝 [𝑢 + 𝜃1 (Δ𝑢∗ − Δ𝑡 ( + 𝜃2 ))] 𝑑Γ
𝜕𝑥 𝜕𝑦 Γ 𝜌 𝜕𝑥𝑖 𝜕𝑥𝑖
Γ
= 0.
𝜕V 𝜕V
𝑓V = Δ𝑡 ∫ 𝑁𝑢 ( + ) 𝑑Γ,
Γ 𝜕𝑥 𝜕𝑦 The first term on the left side of (33) can be expressed as

𝜕𝑁𝑝 𝜕𝑁𝑝 𝜕𝑁𝑝


∫ 𝑢 𝑑Ω = ∫ ( 𝑢+ V) 𝑑Ω
Ω 𝜕𝑥𝑖 Ω 𝜕𝑥 𝜕𝑦
𝜕𝑁𝑝 𝜕𝑁𝑝
=∫ ( (𝑁1 𝑢1 + 𝑁2 𝑢2 + 𝑁3 𝑢3 ) + (𝑁1 V1 + 𝑁2 V2 + 𝑁3 V3 )) 𝑑Ω
Ω 𝜕𝑥 𝜕𝑦

𝜕𝑁1 1 𝜕𝑁1 1 𝜕𝑁1 2 𝜕𝑁1 2 𝜕𝑁1 3 𝜕𝑁1 3 𝑢1 (34)


[ 𝜕𝑥 𝑁 𝜕𝑦 𝑁 𝜕𝑥 𝑁 𝜕𝑦 𝑁 𝜕𝑥 𝑁 𝜕𝑦 𝑁 ] [ V1 ]
[ ][ ]
[ 𝜕𝑁2 2 2 2 2 2 ] [ 2]
[ 1 𝜕𝑁 1 𝜕𝑁 2 𝜕𝑁 2 𝜕𝑁 3 𝜕𝑁 3 ] [𝑢 ]
=∫ [ 𝑁 𝑁 𝑁 𝑁 𝑁 𝑁 ] [ 2 ] 𝑑Ω.
Ω [ 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 ] [V ]
[ 3 ]
[ 𝜕𝑁 1 𝜕𝑁3 1 𝜕𝑁3 2 𝜕𝑁3 2 𝜕𝑁3 3 𝜕𝑁3 3 ] [𝑢3 ]
𝑁 𝑁 𝑁 𝑁 𝑁 𝑁 3
[ 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 ] [V ]
Journal of Chemistry 7

The second term on the left side of (33) can be expressed as

𝜕𝑁𝑝 𝜕𝑁𝑝 𝜕𝑁𝑝


∫ Δ𝑢∗ 𝑑Ω = ∫ ( Δ𝑢∗ + ΔV∗ ) 𝑑Ω
Ω 𝜕𝑥𝑖 Ω 𝜕𝑥 𝜕𝑦
𝜕𝑁𝑝 𝜕𝑁𝑝
=∫ ( (𝑁1 𝑢1∗ + 𝑁2 𝑢2∗ + 𝑁3 𝑢3∗ ) + (𝑁1 V1∗ + 𝑁2 V2∗ + 𝑁3 V3∗ )) 𝑑Ω
Ω 𝜕𝑥 𝜕𝑦

𝜕𝑁1 1 𝜕𝑁1 1 𝜕𝑁1 2 𝜕𝑁1 2 𝜕𝑁1 3 𝜕𝑁1 3 Δ𝑢1 (35)
[ 𝜕𝑥 𝑁 𝜕𝑦 𝑁 𝜕𝑥 𝑁 𝜕𝑦 𝑁 𝜕𝑥 𝑁 𝜕𝑦 𝑁 ] [ ΔV1∗ ]
[ ][ ]
[ 𝜕𝑁2 2 2 2 2 2 ] [ 2∗ ]
[ 1 𝜕𝑁 1 𝜕𝑁 2 𝜕𝑁 2 𝜕𝑁 3 𝜕𝑁 3 ] [ Δ𝑢 ]
=∫ [ 𝑁 𝑁 𝑁 𝑁 𝑁 𝑁 ] [ 2 ∗ ] 𝑑Ω.
Ω [ 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 ] [ ΔV ]
[ 3 ][ ]
[ 𝜕𝑁 1 𝜕𝑁3 1 𝜕𝑁3 2 𝜕𝑁3 2 𝜕𝑁3 3 𝜕𝑁3 3 ] [Δ𝑢3∗ ]
𝑁 𝑁 𝑁 𝑁 𝑁 𝑁 3∗
[ 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑦 ] [ ΔV ]

The third term on the left side of (33) can be expressed as

𝜕𝑁𝑝 𝜕𝑝 𝜕𝑁𝑝 𝜕𝑝 𝜕𝑁𝑝 𝜕𝑝


∫ 𝑑Ω = ∫ ( + ) 𝑑Ω
Ω 𝜕𝑥𝑖 𝜕𝑥𝑖 Ω 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦
𝜕𝑁𝑝 𝜕 𝜕𝑁𝑝 𝜕
=∫ ( (𝑁1 𝑝1 + 𝑁2 𝑝2 + 𝑁3 𝑝3 ) + (𝑁1 𝑝1 + 𝑁2 𝑝2 + 𝑁3 𝑝3 )) 𝑑Ω
Ω 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 (36)

𝜕𝑁𝑝 𝜕𝑁1 𝜕𝑁𝑝 𝜕𝑁1 𝜕𝑁𝑝 𝜕𝑁2 𝜕𝑁𝑝 𝜕𝑁2 𝜕𝑁𝑝 𝜕𝑁3 𝑝1
𝜕𝑁𝑝 𝜕𝑁3
=∫ [ + + + [
] 𝑝2 ] 𝑑Ω.
Ω 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 3
[𝑝 ]

The fourth term on the left side of (33) can be expressed as

𝜕𝑁𝑝 𝜕Δ𝑝 𝜕𝑁𝑝 𝜕Δ𝑝 𝜕𝑁𝑝 𝜕Δ𝑝


∫ 𝑑Ω = ∫ ( + ) 𝑑Ω
Ω 𝜕𝑥𝑖 𝜕𝑥𝑖 Ω 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦
𝜕𝑁𝑝 𝜕 𝜕𝑁𝑝 𝜕
=∫ ( (𝑁1 Δ𝑝1 + 𝑁2 Δ𝑝2 + 𝑁3 Δ𝑝3 ) + (𝑁1 Δ𝑝1 + 𝑁2 Δ𝑝2 + 𝑁3 Δ𝑝3 )) 𝑑Ω
Ω 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 (37)

𝜕𝑁𝑝 𝜕𝑁1 𝜕𝑁𝑝 𝜕𝑁1 𝜕𝑁𝑝 𝜕𝑁2 𝜕𝑁𝑝 𝜕𝑁2 𝜕𝑁𝑝 𝜕𝑁3 Δ𝑝1
𝜕𝑁𝑝 𝜕𝑁3
=∫ [ + + + [
] Δ𝑝2 ] 𝑑Ω,
Ω 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 3
[Δ𝑝 ]

𝑇
where 𝑃 = [𝑝1 𝑝2 𝑝3 ] ,

1 𝜕𝑝 𝜕Δ𝑝 (38)
𝑓𝑝 = −Δ𝑡 ∫ 𝑁𝑢 [𝑢 + 𝜃1 (Δ𝑢∗ − Δ𝑡 ( + 𝜃2 ))] 𝑑Γ.
Γ 𝜌 𝜕𝑥 𝜕𝑥
𝑇
𝐺 = ∫ (∇𝑁𝑝 ) 𝑁𝑢 𝑑Ω,
Ω

𝑇
𝐻 = ∫ (∇𝑁𝑝 ) ∇𝑁𝑝 𝑑Ω, Thus, (32) can be simplified to
Ω
𝑇
𝑈 = [𝑢1 V1 𝑢2 V2 𝑢3 V3 ] ,
1 1
𝑇 Δ𝑡2 𝜃1 𝜃2 𝐻Δ𝑃 = Δ𝑡 (𝐺𝑈 + 𝜃1 𝐺Δ𝑈∗ − Δ𝑡 𝜃1 𝐻𝑃) − 𝑓𝑝 . (39)
Δ𝑈∗ = [Δ𝑢1∗ ΔV1∗ Δ𝑢2∗ ΔV2∗ Δ𝑢3∗ ΔV3∗ ] , 𝜌 𝜌
8 Journal of Chemistry

1m Δ𝑢𝑖1
𝑁𝑁 𝑁𝑁 𝑁𝑁 [
1 1 1 2
[ 2]
1] 3

= ∫ [𝑁 2 𝑁 1 𝑁 2 𝑁 2 𝑁 2 𝑁 3 ] [ ]
[ 𝑖 ] 𝑑Ω.
Δ𝑢
Ω 3 1 3 2 1 3 [ ]
[𝑁 𝑁 𝑁 𝑁 𝑁 𝑁 ] 3
[Δ𝑢𝑖 ]
u = 0.01 m

(41)

0m 1m

Figure 2: Two-dimensional single-phase free-flow model and The first term on the right side of (40) can be expressed as
Delaunay triangle subdivision.

∫ 𝑁𝑢 Δ𝑢𝑖∗ 𝑑Ω = ∫ 𝑁𝑢 (𝑁1 Δ𝑢𝑖1∗ + 𝑁2 Δ𝑢𝑖2∗ + 𝑁3 Δ𝑢𝑖3∗ ) 𝑑Ω


y Ω Ω

i (Vi ) Δ𝑢𝑖1∗
[ ]
[ 2∗ ]
= ∫ 𝑁𝑝 [𝑁𝑢 𝑁 𝑁𝑢 𝑁 𝑁𝑢 𝑁 ] [Δ𝑢𝑖 ]
1 2 3 [
] 𝑑Ω
Ω [ ]
3∗ (42)
[Δ𝑢𝑖 ]
ui (Ui )
Δ𝑢𝑖1∗
𝑁1 𝑁1 𝑁1 𝑁2 𝑁1 𝑁3 [
[ 2∗ ]
]
[ 2 1 2 2 2 3] [
= ∫ 𝑁 𝑁 𝑁 𝑁 𝑁 𝑁 [Δ𝑢𝑖 ] ] 𝑑Ω.
Ω 3 1 3 2 1 3 [ ]
[𝑁 𝑁 𝑁 𝑁 𝑁 𝑁 ] 3∗
[Δ𝑢𝑖 ]

0
x

Figure 3: A single-grid figure. The second term on the right side of (40) can be expressed as

𝜕𝑝
∫ 𝑁𝑝 𝑑Ω
We use a semi-implicit scheme to solve the incompress- Ω 𝜕𝑥𝑖
ible flow problem, so 0.5 ≤ 𝜃1 , 𝜃2 ≤ 1.
At the completion of this stage, the value of Δ𝑃 is fully 𝜕
determined, and using the characteristic-Galerkin approxi- = ∫ 𝑁𝑝 ( (𝑁1 𝑝1 + 𝑁2 𝑝2 + 𝑁3 𝑝3 )) 𝑑Ω
Ω 𝜕𝑥𝑖 (43)
mation, the weak form of (31) can be expressed as
𝑝1
𝜕𝑁 1
𝜕𝑁 𝜕𝑁 [ 2 ] 3

∫ 𝑁𝑢 Δ𝑢𝑖 𝑑Ω = ∫ 𝑁𝑢 Δ𝑢𝑖∗ 𝑑Ω = ∫ [𝑁𝑝 𝑁𝑝 𝑁𝑝 ][


[ 𝑝2 ]
] 𝑑Ω.
Ω Ω Ω 𝜕𝑥𝑖 𝜕𝑥𝑖 𝜕𝑥𝑖
3
[𝑝 ]
1 𝜕𝑝 𝜕Δ𝑝
− Δ𝑡 ∫ 𝑁𝑝 ( + 𝜃2 ) 𝑑Ω (40)
𝜌 Ω 𝜕𝑥𝑖 𝜕𝑥𝑖

Δ𝑡2 1 𝜕𝑝 𝜕Δ𝑝 The third term on the right side of (40) can be expressed as
− ∫ ∇ ⋅ (𝑁𝑢𝑖 ) ( + 𝜃2 ) 𝑑Ω.
2 𝜌 Ω 𝜕𝑥𝑖 𝜕𝑥𝑖

𝜕Δ𝑝
The left side of (40) can be expressed as ∫ 𝑁𝑝 𝑑Ω
Ω 𝜕𝑥𝑖

𝜕
∫ 𝑁𝑢 Δ𝑢𝑖 𝑑Ω = ∫ 𝑁𝑢 (𝑁1 Δ𝑢𝑖1 + 𝑁2 Δ𝑢𝑖2 + 𝑁3 Δ𝑢𝑖3 ) 𝑑Ω = ∫ 𝑁𝑝 ( (𝑁1 Δ𝑝1 + 𝑁2 Δ𝑝2 + 𝑁3 Δ𝑝3 )) 𝑑Ω
Ω Ω Ω 𝜕𝑥𝑖 (44)
Δ𝑢𝑖1
[ ] Δ𝑝1
[ 2] 𝜕𝑁 1
𝜕𝑁 2
𝜕𝑁 [ 2 ]
[3
= ∫ 𝑁𝑝 [𝑁𝑢 𝑁1 𝑁𝑢 𝑁2 𝑁𝑢 𝑁3 ] [ Δ𝑢 ]
[ 𝑖 ] 𝑑Ω = ∫ [𝑁𝑝 𝑁𝑝 𝑁𝑝 ] Δ𝑝 ]] 𝑑Ω.
Ω [ ] Ω 𝜕𝑥𝑖 𝜕𝑥𝑖 𝜕𝑥𝑖 [
3
3
[Δ𝑢𝑖 ] [Δ𝑝 ]
Journal of Chemistry 9

1 11 1 11
10 10
0.8 9 0.8 9
8 8
7 7
0.6 0.6

(10−3 m/s)

(10−3 m/s)
6 6
5 5
0.4 4 0.4 4
3 3
0.2 2 0.2 2
1 1
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
25 s 50 s

1 11 1 11
10 10
0.8 9 0.8 9
8 8
7 7
0.6 (10−3 m/s) 0.6

(10−3 m/s)
6 6
5 5
0.4 4 0.4 4
3 3
0.2 2 0.2 2
1 1
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
75 s 100 s

Figure 4: Velocity field of the free-flow region at different times.

The fourth term on the right side of (40) can be expressed as 𝜕𝑁𝑝
+ (𝑁1 𝑢1 + 𝑁2 𝑢2 + 𝑁3 𝑢3 )
𝜕𝑥
𝜕
𝜕𝑝 + 𝑁𝑝 (𝑁 V + 𝑁 V + 𝑁3 V3 )
1 1 2 2

∫ ∇ ⋅ (𝑁𝑝 𝑢𝑖 ) 𝑑Ω 𝜕𝑦
Ω 𝜕𝑥𝑖 𝜕𝑁𝑝
𝜕
+ (𝑁1 V1 + 𝑁2 V2 + 𝑁3 V3 )
= ∫ ∇ ⋅ (𝑁𝑝 𝑢𝑖 ) ( (𝑁1 𝑝1 + 𝑁2 𝑝2 + 𝑁3 𝑝3 )) 𝑑Ω 𝜕𝑦
Ω 𝜕𝑥𝑖
𝜕𝑁1 𝜕𝑁𝑝
𝜕𝑁1 𝜕𝑁2 𝜕𝑁3 (45) = (𝑁𝑝 + 𝑁1 ) 𝑢1
= ∫ [∇ ⋅ (𝑁𝑝 𝑢𝑖 ) ∇ ⋅ (𝑁𝑝 𝑢𝑖 ) ∇ ⋅ (𝑁𝑝 𝑢𝑖 ) ] 𝜕𝑥 𝜕𝑥
Ω 𝜕𝑥𝑖 𝜕𝑥𝑖 𝜕𝑥𝑖
𝜕𝑁2 𝜕𝑁𝑝
𝑝1 + (𝑁𝑝 + 𝑁2 ) 𝑢2
[
× 𝑝2 ] 𝑑Ω, 𝜕𝑥 𝜕𝑥
3
[𝑝 ]
𝜕𝑁3 𝜕𝑁𝑝
+ (𝑁𝑝 + 𝑁3 ) 𝑢3
𝜕𝑥 𝜕𝑥
where
𝜕𝑁1 𝜕𝑁𝑝 1
+ (𝑁𝑝 + 𝑁1 )V
𝜕𝑦 𝜕𝑦
𝜕 (𝑁𝑝 𝑢) 𝜕 (𝑁𝑝 V)
∇ ⋅ (𝑁𝑝 𝑢𝑖 ) = + 𝜕𝑁2 𝜕𝑁𝑝 2
𝜕𝑥 𝜕𝑦 + (𝑁𝑝 + 𝑁2 )V
𝜕𝑦 𝜕𝑦
𝜕𝑢 𝜕𝑁𝑝 𝜕V 𝜕𝑁𝑝
= 𝑁𝑝 +𝑢 + 𝑁𝑝 + V 𝜕𝑁3 𝜕𝑁𝑝 3
𝜕𝑥 𝜕𝑥 𝜕𝑦 𝜕𝑦 + (𝑁𝑝 + 𝑁3 )V .
𝜕𝑦 𝜕𝑦
𝜕
= 𝑁𝑝 (𝑁1 𝑢1 + 𝑁2 𝑢2 + 𝑁3 𝑢3 ) (46)
𝜕𝑥
10 Journal of Chemistry

1 1
0.9 0.9
0.8 0.8
0.7 0.7
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
25 s 50 s

1 1
0.9 0.9
0.8 0.8
0.7 0.7
0.6 0.6
0.5 0.5
0.4 0.4
0.3 0.3
0.2 0.2
0.1 0.1
0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
75 s 100 s

Figure 5: Velocity vector streamline diagram of the free-flow region at different times.

The fifth term on the right side of (40) can be expressed as Then, we derive
𝑀𝑢 Δ𝑢𝑖 = 𝑀𝑢 Δ𝑢𝑖∗
𝜕Δ𝑝
∫ ∇ ⋅ (𝑁𝑝 𝑢𝑖 ) 𝑑Ω 1 Δ𝑡 (49)
Ω 𝜕𝑥𝑖 − Δ𝑡 [𝐾𝑝 (𝑃 + 𝜃2 Δ𝑃) + 𝐾𝑔 (𝑃 + 𝜃2 Δ𝑃)] .
𝜌 2
𝜕
= ∫ ∇ ⋅ (𝑁𝑝 𝑢𝑖 ) ( (𝑁1 Δ𝑝1 + 𝑁2 Δ𝑝2 + 𝑁3 Δ𝑝3 )) 𝑑Ω
Ω 𝜕𝑥𝑖 The values of Δ𝑈∗ and Δ𝑃 can be obtained from Steps 1 and
2, and Δ𝑈 can be calculated in Step 3 from (49). Then, 𝑈𝑛+1
𝜕𝑁1 𝜕𝑁2 𝜕𝑁3 (47) can be substituted into Step 1 as an initial value for the next
= ∫ [∇ ⋅ (𝑁𝑝 𝑢𝑖 ) ∇ ⋅ (𝑁𝑝 𝑢𝑖 ) ∇ ⋅ (𝑁𝑝 𝑢𝑖 ) ]
Ω 𝜕𝑥𝑖 𝜕𝑥𝑖 𝜕𝑥𝑖 iterative computation.
Δ𝑝1
[
× Δ𝑝2 ] 𝑑Ω, 4. Case Study and Discussion
3
[Δ𝑝 ]
An example calculation was performed for the initial con-
dition of a cave with a reservoir pressure of 10 MPa, a fluid
where density of 800 Kg/m3 , a fluid viscosity of 1 mPa s, and an axial
inflow velocity of 0.01 m/s. The entrance was in the middle of
𝜕𝑁 the left boundary of the cave, and the exit was opposite to the
𝐾𝑝 = ∫ 𝑁𝑝 𝑑Ω,
Ω 𝜕𝑥𝑖 entrance.
(48) The boundary conditions are nonslip, so that 𝑢 = V =
𝜕𝑁 0. The grid at the cave boundary is refined. The number of
𝐾𝑔 = ∫ ∇ ⋅ (𝑢𝑖 𝑁𝑝 ) 𝑑Ω.
Ω 𝜕𝑥𝑖 nodes is 700, and the number of units is 1282. The influence
Journal of Chemistry 11

1 1.07 1 1.07

1.06 1.06
0.8 0.8
1.05 1.05

0.6 1.04 0.6 1.04

(×107 Pa)

(×107 Pa)
1.03 1.03
0.4 0.4
1.02 1.02

0.2 1.01 0.2 1.01

1.00 1.00
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
25 s 50 s

1 1.07 1 1.07

1.06 1.06
0.8 0.8
1.05 1.05

0.6 1.04 0.6 1.04


(×107 Pa)

(×107 Pa)
1.03 1.03
0.4 0.4
1.02 1.02

0.2 1.01 0.2 1.01

1.00 1.00
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
75 s 100 s

Figure 6: Pressure profile patterns of the free-flow region at different times.

of mass force is ignored, and the physical parameters of the This work provides a theoretical basis for future studies
flow in the reservoir are constant with time. Figure 2 shows a on the FEM approach for multiphase flow and Darcy flow-
two-dimensional single-phase free-flow model and Delaunay free flow in a discrete media network model of a fracture-cave
triangle subdivision. Figure 3 shows a single-grid figure. reservoir; this is an important development in the modeling
Based on the research results above, we used MAT- of fracture-cave flow regimes and should be of considerable
LAB to develop software for numerical simulation of two- use in future research in this field.
dimensional single-phase free flow in a cave. The single-
phase free flow of a large-scale fracture-cave reservoir can
thus be simulated, with the simulation performed on any 5. Conclusions
computer with MATLAB installed. The functions of the
Based on a single-phase free-flow numerical simulation of
software include data input and storage, grid partitioning,
a fracture-cave reservoir, the CBS algorithm was adopted to
simulation of single-phase flow, data output, and graphical
obtain numerical solutions accurately and quickly, in which
plotting. Figures 4, 5, and 6 show outputs of the numerical
each step only requires solution of a system of linear equa-
simulation software.
tions. The velocity field, velocity vector streamline diagram,
Figures 4–6 represent the temporal evolution of the and profile pattern of the free-flow region at different times
velocity field, velocity vector streamlines, and profile patterns were determined, and they demonstrated good conformity
of the free-flow region, respectively. The figures show that two with the actual characteristics of free flow. The results confirm
vortices form at the two corners, beside the entrance to the that the CBS algorithm is applicable to the numerical simula-
cave, which is in agreement with the actual characteristics tion of a discrete media model of a fracture-cave reservoir.
of free flow. Simulations indicate that the CBS algorithm Similarly to the traditional method, we can use statistical
has the advantage of flexibility of grid subdivision, and high information and data on fractures and caves obtained from
efficiency and precision of calculations for the single-phase an oil field to construct a discrete fracture-cave geology
free-flow numerical simulation of fracture-cave carbonate geometry model. Then, we can use the approach in this study
reservoirs. to model realistic single-phase cave free flow in fractured
cave media. This paper provides an effective approach to
12 Journal of Chemistry

developing a fracture-cave reservoir, including development Oil and Gas Conference and Exhibition in China, Beijing, China,
plan design, dynamic prediction, reservoir evaluation, opti- June 2010.
mization of engineering technical problems, and decision- [11] A. F. Gulbransen, V. L. Hauge, and K.-A. Lie, “A multiscale
making. mixed finite-element method for vuggy and naturally-fractured
reservoirs,” in Proceedings of the SPE Reservoir Simulation
Symposium, Paper SPE 119104, pp. 898–908, The Woodlands,
Conflict of Interests Tex, USA, February 2009.
[12] M. S. Jouini and S. Vega, “Elastic properties estimation of car-
The authors declare that there is no conflict of interests bonate samples using X-ray computed tomography images,” in
regarding the publication of this paper. Proceedings of the EAGE Workshop on Applications & Challenges
of Rock Physics for Quantitative Geophysical Interpretation,
Acknowledgments Dubai, UAE, January 2012.
[13] M. D. Jackson, A. H. Muggeridge, M. J. Blunt et al., “Reservoir
Project is supported by the National Science Fund for Distin- modeling for flow simulation using surfaces adaptive unstruc-
guished Young Scholars of China (Grant no. 51125019) and the tured meshes and control-volume-finite-element methods,” in
Natural Science Foundation of China (Grant no. 51374181). Proceedings of the SPE Reservoir Simulation Symposium, Paper
SPE 163633, The Woodlands, Tex, USA, February 2013.
[14] Z. Chen, “An ABAQUS implementation of the XFEM for
References hydraulic fracture problems,” in Paper Presented at the ISRM
International Conference for Effective and Sustainable Hydraulic
[1] C. Pérez-Rosales, A. Mercado-Dı́az, J. Cruz-Hernández, and R. Fracturing, 2013.
Islas-Juárez, “Electric Resistivity of Vuggy Fractured Media,” in
[15] Y. Cinar, S. Rahman, and N. G. Doonechaly, “A new finite-
Proceedings of the SPE International Petroleum Conference and
element numerical model for analyzing transient pressure
Exhibithion in Mexico, Paper SPE 74340, pp. 11–23, Villaher-
response of naturally-fractured reservoirs,” in Proceedings of the
mosa, Mexico, February 2002.
SPE Annual Technical Conference and Exhibition, New Orleans,
[2] Z. Chen, Finite Element Methods and their Applications, vol. 5, La, USA, September-October 2013.
Springer, Berlin, Germany, 2005.
[16] Z. J. Luo and F. Zeng, “Finite element numerical simulation of
[3] V. Dalen, “Simplified finite-element models for reservoir flow land subsidence and groundwater exploitation based on visco-
problems,” Society of Petroleum Engineers Journal, vol. 19, no. 5, elastic-plastic Biot’s consolidation theory,” Journal of Hydrody-
pp. 333–343, 1979. namics, vol. 23, no. 5, pp. 615–624, 2011.
[4] J. G. Kim and M. D. Deo, “Finite element, discrete-fracture [17] W. Lai and A. A. Khan, “Time stepping in discontinuous
model for multiphase flow in porous media,” AIChE Journal, Galerkin method,” Journal of Hydrodynamics Series B, vol. 25,
vol. 46, no. 6, pp. 1120–1130, 2000. no. 3, pp. 321–329, 2013.
[5] J. Cai, X. Hu, D. C. Standnes, and L. You, “An analytical [18] Y. L. Zhou, H. W. Tang, and X. H. Liu, “A split-characteristic
model for spontaneous imbibition in fractal porous media finite element model for 1-D unsteady flows,” Journal of Hydro-
including gravity,” Colloids and Surfaces A: Physicochemical and dynamics, vol. 19, no. 1, pp. 54–61, 2007.
Engineering Aspects, vol. 414, pp. 228–233, 2012. [19] T. Liu, H. Xu, and X. Qiu, “A combination method of mixed
[6] J. Cai, E. Perfect, C.-L. Cheng, and X. Hu, “Generalized multiscale finite-element and laplace transform for flow in
modeling of spontaneous imbibition based on hagen-poiseuille a dual-permeability system,” ISRN Applied Mathematics, vol.
flow in tortuous capillaries with variably shaped apertures,” 2012, Article ID 202893, 10 pages, 2012.
Langmuir, vol. 30, no. 18, pp. 5142–5151, 2014. [20] D. Gupta, L. Kumar, and B. Singh, “Finite element solution
[7] Q. Liu, T. Stone, G. Han, R. Marsden, and G. Shaw, “Coupled of unsteady mixed convection flow of micropolar fluid over a
stress and fluid flow using a finite element method in a porous shrinking sheet,” The Scientific World Journal, vol. 2014,
commercial reservoir simulator,” in SPE Asia Pacific Oil and Article ID 362351, 11 pages, 2014.
Gas Conference and Exhibition, Perth, Australia, October 2004, [21] J. Yao, Z. Huang, Y. Li, C. Wang, and X. Lv, “Discrete fracture-
Paper SPE 88616. vug network model for modeling fluid flow in fractured vuggy
[8] M. Sheng, G. Li, S. N. Shah, and X. Jin, “Extended finite element porous media,” in Proceedings of the International Oil and Gas
modeling of multi-scale flow in fractured shale gas reservoirs,” Conference and Exhibition in China, Beijing, China, June 2010,
in Proceedings of the SPE Annual Technical Conference and Paper SPE 130287.
Exhibition (ATCE ’12), Paper SPE 159919, San Antonio, Tex, [22] Z. Chen, G. Huan, and Y. Ma, Computational Methods for
USA, October 2012. Multiphase Flows in Porous Media, vol. 2, SIAM, Philadelphia,
[9] Y. Wei, T. V. Carlos, A. Ridvan, A. D. Saleh, A. T. Abdullah, Pa, USA, 2006.
and E. Haluk, “Interpretation of frequency-dependent dual- [23] Z.-Y. Liu and H.-Z. Chen, “Discontinuous Galerkin immersed
laterolog measurements acquired in Middle-East carbonate finite volume element method for anisotropic flow models
reservoirs using a second-order finite-element method,” in in porous medium,” Abstract and Applied Analysis, vol. 2014,
Proceedings of the 48th Annual Logging Symposium, Society of Article ID 520404, 10 pages, 2014.
Petrophysicists and Well Log Analysts (SPWLA), Austin, Tex, [24] Z. Li, T. Lin, and X. Wu, “New Cartesian grid methods
USA, June 2007. for interface problems using the finite element formulation,”
[10] G. Qin, L. Bi, P. Popov, Y. Efendiev, and M. Espedal, “An Numerische Mathematik, vol. 96, no. 1, pp. 61–98, 2003.
efficient upscaling procedure based on Stokes-Brinkman model [25] Z. Li, “The immersed interface method using a finite element
and discrete fracture network method for naturally fractured formulation,” Applied Numerical Mathematics, vol. 27, no. 3, pp.
carbonate karst reservoirs,” in Proceedings of the International 253–267, 1998.
Journal of Chemistry 13

[26] Y. Gong, B. Li, and Z. Li, “Immersed-interface finite-element


methods for elliptic interface problems with nonhomogeneous
jump conditions,” SIAM Journal on Numerical Analysis, vol. 46,
no. 1, pp. 472–495, 2007.
[27] Y.-L. Zhao, L.-H. Zhang, J.-X. Luo, and B.-N. Zhang, “Perfor-
mance of fractured horizontal well with stimulated reservoir
volume in unconventional gas reservoir,” Journal of Hydrology,
vol. 512, pp. 447–456, 2014.
[28] J. Cai and B. Yu, “A discussion of the effect of tortuosity on
the capillary imbibition in porous media,” Transport in Porous
Media, vol. 89, no. 2, pp. 251–263, 2011.
[29] J. Donea and A. Huerta, Finite Element Methods for Flow
Problems, John Wiley & Sons, New York, NY, USA, 2003.
[30] J. N. Reddy and D. K. Gartling, The Finite Element Method in
Heat Transfer and Fluid Dynamics, CRC Press, New York, NY,
USA, 2010.
Hindawi Publishing Corporation
Journal of Chemistry
Volume 2015, Article ID 604103, 9 pages
http://dx.doi.org/10.1155/2015/604103

Research Article
Compositional Modeling for Optimum Design of
Water-Alternating CO2-LPG EOR under Complicated
Wettability Conditions

Jinhyung Cho, Sung Soo Park, Moon Sik Jeong, and Kun Sang Lee
Department of Natural Resources and Environmental Engineering, Hanyang University, 222 Wangsimni-ro,
Seongdong-gu, Seoul 133-791, Republic of Korea

Correspondence should be addressed to Kun Sang Lee; kunslee@hanyang.ac.kr

Received 11 November 2014; Revised 20 December 2014; Accepted 24 December 2014

Academic Editor: Jianchao Cai

Copyright © 2015 Jinhyung Cho et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The addition of LPG to the CO2 stream leads to minimum miscible pressure (MMP) reduction that causes more oil swelling and
interfacial tension reduction compared to CO2 EOR, resulting in improved oil recovery. Numerical study based on compositional
simulation has been performed to examine the injectivity efficiency and transport behavior of water-alternating CO2 -LPG EOR.
Based on oil, CO2 , and LPG prices, optimum LPG concentration and composition were designed for different wettability conditions.
Results from this study indicate how injected LPG mole fraction and butane content in LPG affect lowering of interfacial tension.
Interfacial tension reduction by supplement of LPG components leads to miscible condition causing more enhanced oil recovery.
The maximum enhancement of oil recovery for oil-wet reservoir is 50% which is greater than 22% for water-wet reservoir. According
to the result of net present value (NPV) analysis at designated oil, CO2 , propane, and butane prices, the optimal injected LPG mole
fraction and composition exist for maximum NPV. At the case of maximum NPV for oil-wet reservoir, the LPG fraction is about
25% in which compositions of propane and butane are 37% and 63%, respectively. For water-wet reservoir, the LPG fraction is 20%
and compositions of propane and butane are 0% and 100%.

1. Introduction immiscible flood are similar to miscible flood, but one major
disadvantage is the limited solubility of CO2 in oil, resulting
CO2 injection has been found to be an efficient method in the restricted swelling effect and viscosity reduction.
for oil recovery worldwide through a miscible or an immis- Injected CO2 and reservoir oil can be miscible by con-
cible displacement process. Mechanism of CO2 enhanced tinuous contact. At the fore-end of injected fluid, CO2 is
oil recovery (EOR) is divided into two different processes, persistently contacted with fresh oil following flow direction,
miscible flood and immiscible flood. Although miscible gas and they are eventually miscible by the vaporizing-gas drive
injection is a widely applied EOR process, it can be only process. In contrast, at the back-end of injected CO2 , near
applied when the reservoir pressure is higher than minimum injection well, reservoir oil is continuously contacted with
miscible pressure (MMP). The main process of miscible fresh CO2 that causes the miscible state by the condensing-
gas injection is displacement efficiency improvement by oil gas drive process [2]. CO2 miscible flood making high
viscosity reduction and swelling effect to reduce residual oil enhanced oil recovery effect has a limit that it can be only
saturation. When reservoir pressure is higher than MMP, applied when the reservoir pressure is higher than MMP.
the injected CO2 and reservoir oil are completely miscible It can be settled by the application of CO2 -LPG EOR that
and the displacement efficiency can be enhanced by zero is able to lower MMP less than that from the application
interfacial tension [1]. Immiscible flood is usually applied of only CO2 EOR. The addition of alkane solvents to the
when reservoir pressure is insufficient to miscible flood or CO2 injection generally accelerates swelling oil, reducing oil
reservoir oil contains many heavy components. The effects of viscosity and decreasing the interfacial tension that can lead
2 Journal of Chemistry

Light Table 1: Modelled fluid composition for Weyburn oil.


component
Solvent J1 Components Mole fraction
N2 0.0207
CO2 0.0074
H2 S 0.0012
CH4 0.0749
J2 C2 H6 0.0422
C3 H8 0.0785
J3
i-C4 to n-C4 0.0655
i-C5 to n-C5 0.0459
I1 I2 C6+ 0.6637
Heavy Intermediate Total 1
component component

Figure 1: Phase ternary diagram for reservoir oil and injection gas
relation [5].
Teklu et al. [10] showed MMP reduction effect by CO2 -
LPG flood in various reservoir scenarios using simulation
model, but it focused on only the relationship between pore
to better performance in enhancing oil recovery [3]. The confinement, permeability, and MMP in shale reservoirs. It
effects of CO2 -LPG injection are verified by the experiment also does not make connection between MMP reduction
[4]. and oil recovery in consideration of gas transport in porous
Figure 1 indicates the ternary diagram of phase behavior media. Recent studies based on the modeling of spontaneous
of reservoir oil and injected solvent. J and I signify injected imbibition also indicated that transport properties of oil are
fluid and reservoir oil. In the inner area of the ternary affected by wettability condition [11, 12]. For this reason,
diagram curve, two phases of the reservoir fluid exist. In case different wettability conditions are applied for analyzing the
of J1 -I2 , only CO2 is injected into reservoir oil I2 . J1 and I2 performance of CO2 -LPG EOR.
cannot be miscible at the first contact because the line passes It has been identified that CO2 -LPG flood is an effective
through the two-phase area. However, they arrive at miscible method for MMP reduction causing oil recovery enhance-
condition by multiple contact miscibility process. The J1 -I1 ment through many experimental studies. Compositional
line lies on the two-phase territory and both points J1 and model for CO2 -LPG EOR is necessary to investigate how gas
I1 are located in the same side on the basis of limiting tie transport affects MMP reduction and oil recovery enhance-
line. Therefore, J1 -I1 cannot be miscible by first and multiple
ment. In this research, compositional fluid and multiphase
contact miscibility process. At J2 -I1 and J3 -I1 cases, first
simulation models are developed and injected LPG mole
and miscible contact miscibility process is available. By the
fraction and composition are optimized based on recent oil,
addition of LPG to the CO2 stream, the location of injected
solvent is moved from J1 to J2 or J3 depending on the amount CO2 , propane, and butane prices for maximum net present
of injected LPG. It makes miscible condition from J1 -I1 case value (NPV).
that was not supposed to be miscible.
To improve sweep efficiency, WAG (water-alternating-
gas) process is applied to CO2 -LPG EOR method in this 2. Numerical Simulation
research. At the same WAG condition, injected LPG amount 2.1. Fluid Modeling. Fluid data of Weyburn reservoir is
and composition are the variables considered in the study. referred for NPV based solvent injection simulation. Wey-
Many experimental researches about the effects of LPG burn reservoir, located in southeast Saskatchewan and oper-
and impurities on MMP with oil have been actively devel-
ated by PanCanadian Petroleum Ltd., has reached its eco-
oped [6]. Several established researches demonstrate the
nomic limit of production by waterflooding. The reservoir is
MMP reduction and oil recovery enhancement by CO2 -LPG
a target for CO2 miscible flooding to enhance oil recovery.
EOR through only experimental ways [7, 8], but numerical
approach to analyze the effectiveness of CO2 -LPG EOR was The oil composition is shown in Table 1 and comparison
not included. Shokir [9] developed ACE algorithm model to between computed fluid model properties and actual fluid
analyze the effects of impurities on MMP between injected data of Weyburn reservoir is given in Table 2 [13]. Oil gravity,
fluid and oil, but it could not explain how lower MMP affects formation volume factor, and gas-oil ratio are calculated
oil recovery. Talbi et al. [3] conducted experimental research through regression process to match separator experimental
on oil swelling, viscosity reduction, interfacial tension reduc- data. Saturation pressure is also computed by regression
tion, and oil recovery improvement that resulted from inject- process. Details of the calculation techniques for saturation
ing solvents into CO2 . However, if it is applied to field pressure can be found in [14]. Acceptable match of computed
scale, it should be time consuming, so reservoir simulation properties from fluid model and Weyburn’s data increases
model for CO2 -LPG flood is positively necessary. Recently, reliability of the fluid model for compositional simulation.
Journal of Chemistry 3

Table 2: Comparison between properties of fluid model and where mixing rules are used for multicomponent fugacity
Weyburn data. expression as follows:
Parameters Fluid model Weyburn 𝑁 𝑁
Saturation pressure (psi) 688 713 𝐴 = ∑ ∑ 𝑦𝑖 𝑦𝑗 𝐴 𝑖𝑗 ,
𝑖=1 𝑗=1
Oil gravity (∘ API) 47 31
Formation volume factor (bbl/STB) 1.11 1.12 𝑁
(7)
Gas-oil ratio (SCF/STB) 166 32 𝐵 = ∑ 𝑦𝑖 𝐵𝑖 ,
Minimum miscibility pressure (psi) 1,996 2,059 𝑖=1

𝐴 𝑖𝑗 = (1 − 𝑘𝑖𝑗 ) √𝐴 𝑖 𝐴 𝑗 ,
Phase behavior of fluid model was determined by Peng- where 𝑘𝑖𝑗 is binary-interaction parameters.
Robinson EOS [15] with the reservoir oil and injected fluid Multiple mixing cell method [17] was applied to fluid
composition. The PR EOS is given by model to estimate MMP between injected CO2 and reservoir
oil. Multiple mixing cell method follows the order below.
𝑅𝑇 𝑎
𝑝= − , (1) (1) Specify the reservoir temperature and an initial pres-
V − 𝑏 V (V + 𝑏) + 𝑏 (V − 𝑏)
sure.
or in terms of Z factor, (2) Calculate the tie-line length for each pressure step by
using the equation below:
𝑍3 − (1 − 𝐵) 𝑍2 + (𝐴 − 3𝐵2 − 2𝐵) 𝑍 − (𝐴𝐵 − 𝐵2 − 𝐵3 ) = 0,
𝑁𝑐
(2) 2
𝑇𝐿 = √ ∑ (𝑥𝑖 − 𝑦𝑖 ) , (8)
𝑖=1
and 𝑍𝑐 = 0.3074.
The EOS constants for pure components are given by where 𝑁𝑐 is the number of components and 𝑥𝑖
and 𝑦𝑖 are liquid and gas equilibrium compositions,
𝑝
𝐴=𝑎 , respectively.
(𝑅𝑇)2
(3) Draw a tie-line length graph as a function of pressures.
𝑝 (4) Perform a multiple-parameter regression of the min-
𝐵=𝑏 ,
𝑅𝑇 imum tie-line lengths to determine the exponent
𝑅2 𝑇𝑐2 𝑛 in 𝑇𝐿𝑛 = 𝑎𝑃 + 𝑏 (power-law extrapolation).
𝑎 = Ω𝑜𝑎 𝛼, (3) These parameters are determined when correlation
𝑝𝑐 coefficient exceeds 0.999.
𝑅𝑇𝑐 (5) Determine the MMP when the power-law extrapola-
𝑏 = Ω𝑜𝑏 , tion gives zero of minimum tie-line length.
𝑝𝑐
2 After generating the fluid model which has approximate
𝛼 = [1 + 𝑚 (1 − √𝑇𝑟 )] , MMP to Weyburn fluid, MMPs were computed between
oil and LPGs. The composition of LPG is propane 63%
and butane 37%, and the calculated MMPs are indicated
where Ω𝑜𝑎 = 0.45724, Ω𝑜𝑏 = 0.07780, and
in Table 3. MMPs of LPG (composition: propane 100% and
butane 0%) mole fraction 20% and 25% are 1,747 psi and
𝑚 = 0.37464 + 1.54226𝜔 − 0.26992𝜔2 . (4) 1,614 psi.

Robinson and Peng [16] proposed a modified 𝑚 for 2.2. Interfacial Tension Calculations. The equation for cal-
heavier components (𝜔 > 0.49) as follows: culating interfacial tension in multicomponent systems is as
follows [18]:
𝑚 = 0.3796 + 1.485𝜔 − 0.1644𝜔2 + 0.01667𝜔3 . (5)
𝑛𝑐
𝜎1/4 = ∑ 𝑝𝑎𝑟𝑖 (𝑥𝑖 𝜌𝐿 − 𝑦𝑖 𝜌𝑔 ) , (9)
Fugacity expressions are given by 𝑖=1

ln 𝜙𝑖 = 𝑍 − 1 − ln (𝑍 − 𝐵) where 𝜎 is the interfacial tension between liquid and gas


phases (dyne/cm) and 𝜌𝐿 and 𝜌𝑔 are molar densities of liquid
𝐴 𝐵 2 𝑁 𝑍 + (1 + √2) 𝐵 and gas phases (mole/cm3 ), respectively. The parachor (𝑝𝑎𝑟𝑖 )
− ( 𝑖 − ∑ 𝑦𝑖 𝐴 𝑖𝑗 ) ln [ ], is defined as follows:
2√2𝐵 𝐵 𝐴 𝑗=1 𝑍 − (1 − √2) 𝐵
(6) 𝑝𝑎𝑟𝑖 = 𝜉CN𝑖 , (10)
4 Journal of Chemistry

Table 3: MMP estimates for injection gas according to LPG mole 1


fraction.
LPG mole fraction (%) MMP (psi)
0.8
0 1,996
5 1,995

Relative permeability
10 1,825 0.6
15 1,820
20 1,412
25 1,354 0.4
30 1,046

Waterflooding 0.2
330 ft

330 ft
0
0 0.5 1
Water saturation
20 ft krw (oil-wet) krw (water-wet)
kro (oil-wet) kro (water-wet)

Producer Figure 3: Relative permeability curves for different wettability


conditions.

Figure 2: 3D view of simulation model.


CO2

where CO2 + LPG 10%

CO2 + LPG 20%


40, CN ≤ 12,
𝜉={ (11) CO2 + LPG 30%
40.3, CN > 12,

and CN is the carbon number of the components 𝑖.

2.3. Reservoir Modeling. The reservoir model was assumed


as 2D model which is discretized into 33 × 33 × 1 grid Limiting
blocks. Each grid block has dimension as 10 ft × 10 ft × tie line
20 ft as shown in Figure 2. The model size is general one
injector-one producer scale of 10-acre five-spot model [19]. C5+ Oil C3 -C4
This simulation study utilized homogeneous 2D areal model
not considering heterogeneity and gas overriding effect. Figure 4: Ternary diagram for CO2 /LPG/Oil system at reservoir
Without these effects, oil recovery can be governed only pressure 1,500 psi.
by displacement efficiency from LPG addition and can be
expected near 100% [20].
Contact angle which is a determinant for wettability is
different residual oil saturation and mobility [21]. The relative
defined by Young’s equation as follows:
permeability curves are predicted by simulations. Simulation
𝜎𝑜𝑠 − 𝜎𝑤𝑠 methods to predict relative permeability are already verified
cos 𝜃 = , (12) by previous studies [22, 23] and similar water relative curve
𝜎𝑜𝑤
can be found. Residual oil saturations of oil- and water-wet
where 𝜎𝑜𝑠 , 𝜎𝑤𝑠 , and 𝜎𝑜𝑤 are oil-solid, water-solid, and oil- reservoirs are 18% and 15%, respectively.
water interfacial tensions. As indicated in the above equation, After waterflooding for three years, water-alternating
if 𝜎𝑜𝑠 is greater than 𝜎𝑤𝑠 , 𝜃 is smaller than 90∘ , so the CO2 EOR and CO2 -LPG EOR were applied to water- and oil-
reservoir rock exhibits water-wet solid. The inverse case is wet reservoirs for ten years. WAG cycle of CO2 and CO2 -LPG
oil-wet condition. Water-wet and oil-wet reservoirs have the EOR is 1 : 1, and one cycle period is 6 months. Production
constant porosity and isotropic permeability is also assumed. pressure is 1,500 psi which is within a limitation of miscible
Reservoir initial conditions are shown in Table 4. The poros- condition by first or multiple contact miscibility process
ity, permeability, and relative permeability were gained from when added LPG concentration is larger than 20% (Figure 4).
the same reservoir, and two different relative permeability Injected LPG mole fraction and composition are indicated in
curves (Figure 3) are used in this simulation for establishing Table 5.
Journal of Chemistry 5

Table 4: Properties of reservoir rock and fluids. Table 6: Economic parameters for optimal design.

Properties Values Parameters Values


Depth (ft) 4,000 Oil ($/bbl) 80
Pressure (psi) 2,000 CO2 ($/ton) 80
Temperature (∘ F) 145 Propane ($/ton) 800
Permeability (md) 122 Butane ($/ton) 850
Porosity (%) 24 Water injection ($/bbl) 0.25
Oil saturation (𝑆𝑜 ) 0.64 Produced water handling ($/bbl) 1.5
Water saturation (𝑆𝑤 ) 0.36

Table 5: Operating conditions and injection design parameters. 3. Results and Discussion

Properties Values 3.1. Oil Production. The aim of this study is to confirm the
effectiveness of water-alternating CO2 -LPG EOR process in
Producing pressure at bottom
hole (psi)
1,500 oil recovery for different reservoirs. The performance of CO2 -
LPG injection process has been compared with that of CO2
Total injection (PV) 1.5
WAG process. LPG is composed of 63% propane and 37%
Period (years) 10 butane. Results of oil recovery with various LPG concentra-
WAG ratio 1:1 tions are indicated in Figure 5. Increased oil recoveries for
Injected LPG mole fraction oil- and water-wet reservoirs by CO2 -LPG flood are 46% and
0, 10, 15, 20, 25, and 30
(%) 22%. For both wettability conditions, the higher LPG mole
Injected LPG composition 100 : 0, 63 : 37, 37 : 63, and fraction is injected, the more oil is produced. However, signif-
(propane : butane) 0 : 100 icant differences are not found if LPG mole fraction is greater
than 25%. The tendency is also identified by experimental
results in the literature [7]. To detect the influence of LPG
2.4. Net Present Value. The NPV of a time series of cash flows composition in the injected fluid, oil recoveries with different
is defined as the sum of the present values. NPV considering ratio of propane and butane (LPG 15%) are shown in Figure 6.
prices of oil, CO2 , propane, and butane and costs of water Figure 6 shows that the higher fraction of butane causes more
injection and produced water handling is calculated by the enhanced oil recovery. Increments of oil recovery for oil-
following equation [24]: and water-wet reservoirs are 25% and 15% as compared with
CO2 EOR. This phenomenon was already revealed by the
𝑇 experimental study and it was explained that the result is
𝑅𝑡
NPV = ∑ , (13) because of higher mole weight [27].
𝑡=1 (1 + 𝐼)𝑡 When reservoir oil and injected gas are miscible, gas
saturation decreases further than immiscible condition
where 𝑇 is total production period (day), 𝑡 is time, 𝑅𝑡 is net (Figure 7). Injected gas reached production well at around
profit at time 𝑡, and 𝐼 is daily discount rate. 𝐼 is estimated by 2004, so the gas saturation of WAG CO2 case increased
yearly discount rate as abruptly. However, in case of WAG CO2 + LPG 30%, the gas
saturation did not increase even though injected gas already
𝐼 = 𝑒ln(1+Yearly discount rate)/365 − 1, (14) reached production well. It indicates that miscible condition
reduces gas saturation. The reduction of gas saturation
where yearly discount rate is 10% and 𝑅𝑡 is defined by the causes a decrease in gas relative permeability (Figure 8). As
difference between the profit from oil production and total both gas saturation and relative permeability decline, liquid
investment costs at time 𝑡: saturation and relative permeability increase, which leads to
the enhancement of oil recovery.
𝑅𝑡 = 𝑄𝑜 𝑃𝑜 − (𝑄CO2 𝑃CO2 + 𝑄C3 𝑃C3 The addition of LPG to CO2 stream is more effective
(15) to lower interfacial tension between oil and gas phases. In
+𝑄C4 𝑃C4 + 𝑄𝑤1 𝑃𝑤1 + 𝑄𝑤2 𝑃𝑤2 ) , particular, if the reservoir is in miscibility condition, inter-
facial tension reaches zero [1]. As shown in Figure 9(a), the
where 𝑄𝑜 , 𝑄CO2 , 𝑄C3 , 𝑄C4 , 𝑄𝑤1 , and 𝑄𝑤2 are oil production swept zone is left in nonzero interfacial area, so reservoir is
rate (bbl/day), CO2 injection rate (lb/day), propane injection not in miscible condition by only CO2 injection. In contrast,
rate (lb/day), butane injection rate (lb/day), water injection the addition of LPG to CO2 stream as 20% brings the swept
rate (bbl/day), and water production rate (bbl/day). Param- area into zero interfacial zone indicating miscible condition
eters 𝑃𝑜 , 𝑃CO2 , 𝑃C3 , 𝑃C4 , 𝑃𝑤1 , and 𝑃𝑤2 are oil price ($/bbl), (Figure 9(b)). Zero interfacial tension indicates that oil and
CO2 price ($/lb), propane price ($/lb), butane price ($/lb), gas become single-phase, so it flows easier than two-phase
water injection cost ($/bbl), and produced water handling fluid.
cost ($/bbl). All values of parameters for NPV calculation are If more butane content is injected than propane, more oil
shown in Table 6 [25, 26]. recovery is expected because of its higher molecular weight.
6 Journal of Chemistry

100
100
90
90
80
80
70

Oil recovery factor (%)


70
Oil recovery factor (%)

60 60

50 50

40 40

30 30

20 20

10 10

0 0

1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009

Time (date) 2010 Time (date)

WAG CO2 WAG CO2 + LPG 20% WAG CO2 WAG CO2 + LPG 20%
WAG CO2 + LPG 10% WAG CO2 + LPG 25% WAG CO2 + LPG 10% WAG CO2 + LPG 25%
WAG CO2 + LPG 15% WAG CO2 + LPG 30% WAG CO2 + LPG 15% WAG CO2 + LPG 30%

(a) Oil-wet (b) Water-wet

Figure 5: Oil recovery factors with LPG mole fraction of injection gas for different wettability conditions (composition of LPG: propane 63%,
butane 37%).

100
100
90
90
80
80
Oil recovery factor (%)

70
70
Oil recovery factor (%)

60
60
50
50
40
40
30
30
20
20
10
10
0
0
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010

Time (date)
Time (date)
WAG CO2 WAG CO2 + LPG (37 : 63) WAG CO2 WAG CO2 + LPG (37 : 63)
WAG CO2 + LPG (100 : 0) WAG CO2 + LPG (0 : 100) WAG CO2 + LPG (100 : 0) WAG CO2 + LPG (0 : 100)
WAG CO2 + LPG (63 : 37) WAG CO2 + LPG (63 : 37)

(a) Oil-wet (b) Water-wet

Figure 6: Oil recovery factors with LPG composition for different wettability conditions (LPG mole fraction of injection gas: 15%).

It was proved that butane is much more effective in MMP in reservoir is shown in Figure 10. Figure 10 indicates the
reduction [27]. The addition of alkane solvents to the CO2 oil saturation when LPG mole fraction is 25% for oil-wet
stream accelerates the process of reducing oil viscosity; thus, reservoir after 6 months from the end of waterflooding. In
it leads to higher oil recovery [4]. To compare the aspect case of 100% propane, oil saturation near injection well is zero
of oil recovery by injected LPG composition, oil saturation because of immaculate expulsion and it is 0.5 at the fore-end
Journal of Chemistry 7

0.18 0.06

0.16
0.05
0.14

Gas relative permeability


0.12 0.04
Gas saturation

0.1
0.03
0.08

0.06 0.02

0.04
0.01
0.02

0 0
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010

1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
Time (date) Time (date)
WAG CO2 WAG CO2
WAG CO2 + LPG 30% WAG CO2 + LPG 30%

Figure 7: Gas saturation with LPG mole fraction of injection gas Figure 8: Gas relative permeability with LPG mole fraction of
near production well for oil-wet condition. injection gas near production well for oil-wet condition.

of injected fluid (Figure 10(a)). In case of 100% butane, zero Table 7: Maximum NPV improvements with LPG mole fraction
zone of oil saturation is more widespread with near wellbore and composition for oil-wet reservoir (base case: $2,652,887).
as a center. Furthermore, oil saturation at the fore-end is 0.7
which is higher than that in the case of 100% propane because Maximum NPV improvements (%)
more oil is displaced from wider area (Figure 10(b)). LPG LPG LPG LPG LPG
Tables 7 and 8 indicate the amount of increase in maxi- 10% 15% 20% 25% 30%
mum NPV after waterflooding for different wettability con- Propane 100%
10.7 11.0 11.8 12.8 20.0
ditions. NPVs are calculated according to LPG concentration Butane 0%
and composition. The maximum NPV increments by CO2 Propane 63%
WAG are 12% and 13% for oil- and water-wet reservoirs. 9.3 10.3 13.0 24.1 19.2
Butane 37%
As shown in Tables 7 and 8, the maximum value is 24.1% Propane 37%
(LPG 25%: propane 63%, butane 37%) for oil-wet reservoir 9.0 10.9 18.4 19.2 15.4
Butane 63%
and 17.0% (LPG 20%: propane 0%, butane 100%) for water- Propane 0%
wet reservoir. When LPG mole fraction is less than 15% and 9.2 13.8 22.5 17.3 13.0
Butane 100%
20% (propane 100%), maximum increase in NPV is less than
CO2 WAG. These cases are in immiscible condition, so oil
recovery is not high compared to the economic feasibility of
LPG. Injected fluid and reservoir oil are in miscible condition, Table 8: Maximum NPV improvements with LPG mole fraction
and maximum NPV improvements are higher than those of and composition for water-wet reservoir (base case: $3,387,572).
CO2 WAG cases. Maximum NPV increment by CO2 -LPG
EOR occurred for two different wettability conditions, but the Maximum NPV improvements (%)
effect in oil-wet reservoir is better than in water-wet reservoir LPG LPG LPG LPG LPG
because of higher residual oil saturation after waterflooding. 10% 15% 20% 25% 30%
Propane 100%
12.2 12.0 12.3 12.6 15.5
Butane 0%
4. Conclusions Propane 63%
11.8 12.2 13.1 16.0 16.7
Butane 37%
In this study, water-alternating CO2 -LPG EOR simulation
model was developed. To examine the efficiency of CO2 - Propane 37%
11.8 12.5 15.6 16.7 14.2
Butane 63%
LPG EOR considering oil, CO2 , and LPG prices, extensive
simulations have been performed for different wettability Propane 0%
11.7 13.6 17.0 15.4 12.4
Butane 100%
conditions and the following conclusions have been drawn.
8 Journal of Chemistry

8.3 8.3
7.5 7.5
6.6 6.6
5.8 5.8
5.0 5.0
4.2 4.2
3.3 3.3
2.5 2.5
1.7 1.7
0.8 0.8
0.0 0.0
(a) CO2 100% (b) LPG 20%

Figure 9: Interfacial tension between oil and gas phases with injected gas mole fraction after six months of gas injection (LPG composition:
63% propane and 37% butane).

0.82 0.82
0.77 0.77
0.73 0.73
0.68 0.68
0.64 0.64
0.59 0.59
0.55 0.55
0.50 0.50
0.46 0.46
0.41 0.41
0.36 0.36
0.32 0.32
0.27 0.27
0.23 0.23
0.18 0.18
0.14 0.14
0.09 0.09
0.05 0.05
0.00 0.00
(a) 100% propane (b) 100% butane

Figure 10: Oil saturation with LPG composition after six months of gas injection (LPG volume fraction of injection gas: 15%).

(1) When LPG concentration is 30% and composition of than CO2 WAG process. When LPG concentration
butane is 100%, oil recovery increased by 46% and is higher than 20% (miscible condition), maximum
25% for oil-wet reservoir. When LPG concentration is NPV improved and optimum LPG concentration and
30% and butane composition is 100%, the maximum composition exist for maximum NPV improvement.
increasing amounts are 22% and 15% in case of water-
wet reservoir. As injected LPG concentration and
butane composition increased, significantly enhanced (3) CO2 -LPG EOR can be applicable in low pressure
oil recovery was observed from the reduction of MMP reservoirs that CO2 is not miscible. LPG addition to
and interfacial tension. Oil recovery for different CO2 stream can appreciably improve oil recovery by
wettability by CO2 -LPG EOR has become close to zero interfacial tension bringing miscible condition.
100%. Moreover, the optimization of LPG concentration
and composition is absolutely necessary for economic
(2) When LPG concentration is 25% and butane com- feasibility. The necessity of optimization is required
position is 37%, maximum NPV improvement is more in oil-wet reservoir due to better performance
24.1% for oil-wet reservoir. When LPG concentration of displacement efficiency.
is 20% and butane composition is 100%, maximum
NPV improvement is 17.0% for water-wet reservoir.
For both oil- and water-wet reservoirs, when LPG Conflict of Interests
concentrations are 10%, 15%, 20% (propane 100%),
and 25% (propane 100%), the reservoir condition is The authors declare that there is no conflict of interests
immiscible and maximum NPV increment is lower regarding the publication of this paper.
Journal of Chemistry 9

Acknowledgment [15] D.-Y. Peng and D. B. Robinson, “A new two-constant equation of


state,” Industrial and Engineering Chemistry Fundamentals, vol.
This work was supported by the Energy Efficiency & 15, no. 1, pp. 59–64, 1976.
Resources Core Technology Program of the Korea Institute [16] D. B. Robinson and D. Y. Peng, “The characterization of
of Energy Technology Evaluation and Planning (KETEP) the heptanes and heavier fractions,” Research Report 28, Gas
granted financial resource from the Ministry of Trade, Indus- Processors Association, Tulsa, Okla, USA, 1978.
try & Energy, Republic of Korea (no. 20122010200060). [17] K. Ahmadi and R. T. Johns, “Multiple-mixing-cell method for
MMP calculations,” SPE Journal, vol. 16, no. 4, pp. 733–742, 2011.
References [18] R. C. Reid, J. M. Prausnitz, and T. K. Sherwood, The Properties
of Gases and Liquids, McGraw-Hill, New York, NY, USA, 1977.
[1] D. Makimura, M. Kunieda, Y. Liang, T. Matsuoka, S. Takahashi, [19] G. P. Willhite, Waterflooding, Society of Petroleum Engineers,
and H. Okabe, “Application of molecular simulations to CO2 - Richardson, Tex, USA, 1986.
enhanced oil recovery: phase equilibria and interfacial phenom- [20] G. F. Teletzke, P. D. Patel, and A. L. Chen, “Methodology for
ena,” Society of Petroleum Engineers, vol. 18, no. 2, pp. 319–330, miscible gas injection EOR screening,” in Proceedings of the SPE
2013. International Improved Oil Recovery Conference in Asia Pacific
[2] F. I. Stalkup Jr., Miscible Displacement, Society of Petroleum (IIORC ’05), SPE 97650, pp. 315–325, Kuala Lumpur, Malaysia,
Engineers of AIME, Dallas, Tex, USA, 2nd edition, 1983. December 2005.
[3] K. Talbi, T. M. V. Kaiser, and B. B. Maini, “Experimental [21] M. Delshad, N. F. Najafabadi, G. A. Anderson, G. A. Pope,
investigation of CO2 -based VAPEX for recovery of heavy oils and K. Sepehrnoori, “Modeling wettability alteration in natu-
and bitumen,” Journal of Canadian Petroleum Technology, vol. rally fractured reservoirs,” in Proceedings of the 15th SPE/DOE
47, no. 4, pp. 29–36, 2008. Improved Oil Recovery Symposium, vol. 2 of SPE 100081, Tulsa,
[4] H. Li, S. Zheng, and D. Yang, “Enhanced swelling effect and Okla, USA, April 2006.
viscosity reduction of solvent(s)/CO2 /heavy-oil systems,” SPE [22] B. Xiao, J. Fan, and F. Ding, “Prediction of relative permeability
Journal, vol. 18, no. 4, pp. 695–707, 2013. of unsaturated porous media based on fractal theory and Monte
[5] L. W. Lake, Enhanced Oil Recovery, Society of Petroleum Carlo simulation,” Energy and Fuels, vol. 26, no. 11, pp. 6971–
Engineers, Richardson, Tex, USA, 2010. 6978, 2012.
[6] R. B. Alston, G. P. Kokolis, and C. F. James, “CO2 minimum [23] B. Xiao, J. Fan, and F. Ding, “A fractal analytical model for the
miscibility pressure: a correlation for impure CO2 streams and permeabilities of fibrous gas diffusion layer in proton exchange
live oil systems,” Society of Petroleum Engineers Journal, vol. 25, membrane fuel cells,” Electrochimica Acta, vol. 134, pp. 222–231,
no. 2, pp. 268–274, 1985. 2014.
[7] N. Kumar and W. D. V. Gonten, “An investigation of oil recovery [24] S. Salem and T. Moawad, “Economic study of miscible CO2
by injecting CO2 and LPG mixtures,” in Proceedings of the 48th flooding in a mature waterflooded oil reservoir,” in Proceedings
Annual Fall Meeting of the Society of Petroleum Engineers of of the SPE Saudi Arabia Section Annual Technical Symposium
AIME, SPE 4581, Las Vegas, Nev, USA, September 1973. and Exhibition, SPE 168064, Al-Khobar, Saudi Arabia, May
[8] P. Y. Zhang, S. Huang, S. Sayegh, and X. L. Zhou, “Effect of CO2 2013.
impurities on gas-injection EOR processes,” in Proceedings of [25] C. L. Liao, X. W. Liao, X. L. Zhao et al., “Study on enhanced oil
the SPE/DOE Symposium on Improved Oil Recovery, SPE 89477, recovery technology in low permeability heterogeneous reser-
Tulsa, Okla, USA, April 2004. voir by water-alternate-gas of CO2 flooding,” in Proceedings of
[9] E. M. E. Shokir, “CO2 -oil minimum miscibility pressure model the SPE Asia Pacific Oil and Gas Conference and Exhibition, SPE
for impure and pure CO2 streams,” in Proceedings of the Offshore 165907, Jakarta, Indonesia, October 2013.
Mediterranean Conference and Exhibition (OMC ’07), Ravenna, [26] OPIS Europe LPG Report, 2014, http://www.opisnet.com.
Italy, March 2007. [27] R. S. Metcalfe, “effects of impurities on minimum miscibility
[10] T. W. Teklu, N. Alharthy, H. Kazemi, X. Yin, and R. M. Graves, pressures and minimum enrichment levels for CO2 and rich-gas
“Hydrocarbon and non-hydrocarbon gas miscibility with light displa cements,” SPE Journal, vol. 22, no. 2, pp. 219–225, 1982.
oil in shale reservoirs,” in Proceedings of the SPE Improved Oil
Recovery Symposium, SPE 169123, Tulsa, Okla, USA, April 2014.
[11] J. Cai, X. Hu, D. C. Standnes, and L. You, “An analytical
model for spontaneous imbibition in fractal porous media
including gravity,” Colloids and Surfaces A: Physicochemical and
Engineering Aspects, vol. 414, pp. 228–233, 2012.
[12] J. Cai, E. Perfect, C.-L. Cheng, and X. Hu, “Generalized
modeling of spontaneous imbibition based on hagen-poiseuille
flow in tortuous capillaries with variably shaped apertures,”
Langmuir, vol. 30, no. 18, pp. 5142–5151, 2014.
[13] R. K. Srivastava, S. S. Huang, and M. Dong, “Laboratory
investigation of Weyburn CO2 miscible flooding,” Journal of
Canadian Petroleum Technology, vol. 39, no. 2, pp. 41–51, 2000.
[14] L. X. Nghiem, Y.-K. Li, and R. A. Heidemann, “Application
of the tangent plane criterion to saturation pressure and
temperature computations,” Fluid Phase Equilibria, vol. 21, no.
1-2, pp. 39–60, 1985.
Hindawi Publishing Corporation
Journal of Chemistry
Volume 2015, Article ID 685871, 6 pages
http://dx.doi.org/10.1155/2015/685871

Research Article
Heat Transport and Water Permeability during Cracking of
the Landfill Compacted Clay Cover

Haijun Lu,1 Jianjun Liu,1,2 Yuan Li,2 and Yiqie Dong1


1
Institute of Poromechanics, Wuhan Polytechnic University, Wuhan 430023, China
2
School of Geoscience and Technology, Southwest Petroleum University, Chengdu 610500, China

Correspondence should be addressed to Jianjun Liu; liujj0906@163.com

Received 27 July 2014; Revised 1 October 2014; Accepted 9 October 2014

Academic Editor: Peng Xu

Copyright © 2015 Haijun Lu et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The heat-moisture transport through the compacted clay was observed in laboratory. The hydraulic conductivity of cracked clay
under wetting-drying cycles was also investigated. At the early phase of heating, the temperature of soil columns rose fast and
moisture decreased dramatically; after this phase, the temperature rose at a lower speed and moisture loss stabilized gradually. The
moisture content of compacted clay at 25 cm depth decayed to 0. The crack intensity factor (CIF) of compacted clay was 0.043 and
0.097; the crack depth was about 6.5 cm and 8.2 cm at 50∘ C and 60∘ C, respectively. The hydraulic conductivity of compacted clay
was within 8.3 × 10−7 to 1.5 × 10−5 cm/s after four wetting-drying cycles. This value was 2∼3 orders of magnitude higher than that
of uncracked clay.

1. Introduction matric potentials become very small, the compacted clay


shrinks horizontally and crack occurs [5–7]. For compacted
In order to reduce the generation of landfill leachate, a cover clay cover systems with very low hydraulic conductivity,
system, as an important part of the antiseepage structure, moisture will be slowly transported through landfill cover
must be set when landfill reaches its maximum reserve. systems [8]. However, rainwater can directly leak into landfill
Compacted clay is recommended to be used for building if crack occurs in the compacted clay layer. Cracking leads to
the cover system structure and the hydraulic conductivity of the changes of hydraulic conductivity (from 10−7 to 10−4 cm/s)
the clay must be less than 1 × 10−7 cm/s [1]. However, under and thermal conductivity of compacted clay [9, 10] and
the internal and external environmental effects, crack of the ultimately influences the heat-moisture transport in the cover
compacted clay cover system always occurs, which provides system.
a convenient path for the infiltration of rainwater and then To investigate the surface cracking of soil columns and
results in a sharp increase in landfill leachate which may take a heat-moisture transport, the tests of heat-moisture transport
severe toll on the ecological environment around the landfill. through soil columns during cracking were conducted by
Large amounts of heat are released by the degradation the monitoring system of cracking. The permeability of com-
of solid waste, causing the rise of temperature in landfill. pacted clay during wetting-drying cycles was measured with
The temperature inside the landfill can go up to 70∘ C [2]. the PN3230M environmental soil flexible wall permeameter.
Due to the temperature dependence of vapor density and the
temperature gradient between the landfill and surrounding 2. Materials and Methods
atmospheric environment, vapor diffuses upward and out of
the compacted clay cover. This strongly enhances the down- 2.1. Materials. The clay used in the tests was taken from a
ward transport of liquid water by gravity [3, 4]. Therefore, construction site in Wuhan City of China. Basic physical
vapor diffusion leads to the decrease in matric potential, caus- properties and chemical composition of the clay are shown
ing the shrinkage and desiccation of compacted clay. When in Tables 1 and 2.
2 Journal of Chemistry

Table 1: The physical properties of the clay.

Particle size distribution (%)


𝜌𝑑 max (g/cm3 ) 𝑊opt (%) 𝑊𝐿 (%) 𝑊𝑝 (%) 𝐼𝑝 (%)
>0.05 mm 0.05∼0.005 mm 0.005∼0.002 mm <0.002 mm
1.65 19 48.5 26.2 22.3 12 32 45 11

Digital microscope camera

40 cm
0

Compacted soil
Z

Temperature sensor

30 cm
Data acquisition system Humidity sensor

Temperature-controlled
heating plate 1 2

Figure 1: The monitoring system of cracking and heat-moisture transport in compacted clay.

Table 2: The chemical composition of the clay. The samples were placed in a vacuum saturator can for 48 hr
before permeability tests and then were placed in a dryer at
SiO2 AL2 O3 Fe2 O3 CaO MgO K2 O Na2 O 50∘ C for 72 hr after permeability tests. The aforementioned
58.42 25.23 0.24 0.51 0.12 5.32 2.67 steps were repeated for four wetting-drying cycles.

2.2. Methods. The soil and moisture were mixed evenly, put 3. Results and Discussion
into soil columns, and then compacted layer by layer. The
compacted soil layer was formed at the maximum dry density 3.1. Heat-Moisture Transport of Cracked Clay. Temperature
in the column. The height and diameter of the soil columns variation at different depths of soil columns showed similar
were 30 cm and 40 cm, respectively. The bottoms of the soil patterns at the heating source of 50∘ C or 60∘ C (Figure 2).
columns were placed directly on the temperature-controlled Early in the heating, temperature at different depths rose
heating plate (50∘ C and 60∘ C), surrounded by the insulation quickly but then the growth trend attenuated significantly and
pad to prevent the dissemination of heat. Temperature and eventually stabilized. The temperature was close to 50∘ C or
humidity sensors were placed in the depth of 5 cm, 15 cm, 60∘ C at the depth of 25 cm after 22 hr of heating, indicating
and 25 cm of the soil columns, monitoring temperature and that the column bottom had cracks by now. The heat transport
moisture changes of compacted clay. A digital microscope in soil columns was fast and able to achieve stability in a
camera was placed 50 cm above the soil column to observe short time. Since the compacted clay layer can store heat,
the surface cracking (Figure 1). transport of heat in soil columns was significantly delayed
The PN3230M environmental soil flexible wall perme- with increasing distance from the heating source.
ameter, made by American GEOEQUIP, was used to test the The soil moisture loss at various soil depths were given
permeability of the compacted clay under different wetting- in Figures 3 and 4. The soil moisture loss near the bottom
drying cycles. The permeameter can prevent the sidewall was highly sensitive to heat loading, and the effect was
leakage that might be caused by the specimen’s volume significantly attenuated and time lagged along with increasing
change during the wetting-drying cycles. The specimen was distance from the heating source. The testing results of the
initially sized at a diameter of 5 cm and a height of 5 cm for heating source of 60∘ C indicated that soil moisture loss
permeability tests. All samples were made by pressing soil of was very rapid within the first 20 hr and slowed down
specific quantity into the cutting ring of the mentioned size. during 20 to 384 hr. Within 20 hr, at the depth of 25 cm,
Journal of Chemistry 3

55
60
50
55
45
Temperature (∘ C)

Temperature (∘ C)
50 T = 60∘ C
40 45

35 40

30 35

T = 50∘ C 30
25
25
0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400
Time (h) Time (h)
D = 25 cm D = 25 cm
D = 15 cm D = 15 cm
D = 5 cm D = 5 cm
(a) (b)

Figure 2: The temperature of compacted clay changes with time.

22 32
20
28
18 T = 50∘ C
16 24
Moisture content (%)

Moisture content (%)

14 20
12
10 16
8 12
6
4
8 T = 60∘ C
2 4
0
0
−2
0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400
Time (h) Time (h)
D = 25 cm D = 25 cm
D = 15 cm D = 15 cm
D = 5 cm D = 5 cm
(a) (b)

Figure 3: The moisture content of compacted clay changes with time.

the soil moisture content decreased from 0.16 to 0.038. The CIF of compacted clay under heating source of 50∘ C or 60∘ C
temperature gradient between the clay layer and the heat is shown in Figure 5. Compacted clay in the first 3-4 days
source was great at the early heating, resulting in fast moisture showed no obvious cracks. Subsequently, the CIF increased
transmission. However, the soil moisture content at the depth dramatically from 0.0025 to 0.028 or from 0.0042 to 0.036.
of 25 cm decreased gradually from 0.038 to 0 during 20 to Then the CIF increased slowly to 0.043 and 0.097.
384 hr, owing to the lower temperature gradient. A kind of sticky black tracer was injected into soil
The cracking of compacted clay was characterized by columns to observe the crack depth. Soil material was colored
the cracking intensity factor (CIF). The CIF value was the in black at the tracer flow path. The crack depth was deter-
ratio of the cracking area (𝐴 𝑐 ) to the total area (𝐴 𝑡 ) [11–13]. mined by the soil color change. After 4 hr, the soil column was
Comparing the cracked and uncracked part of the surface, the cut along the cracking direction from top to bottom. It can
former was obviously dim. The CIF was calculated by using be concluded that the surface crack depth of soils was about
the digital microscopy image analysis software. The surface 6.5 cm and 8.2 cm at 50∘ C and 60∘ C, respectively.
4 Journal of Chemistry

22 32
20 28
18
16 24

Moisture content (%)


Moisture content (%)

14 20
12
16
10
8 12
6
8
4
2 4 T = 60∘ C
T = 50∘ C
0 0
−2
20 25 30 35 40 45 50 28 32 36 40 44 48 52 56 60
Temperature (∘ C) Temperature (∘ C)
D = 25 cm D = 25 cm
D = 15 cm D = 15 cm
D = 5 cm D = 5 cm
(a) (b)

Figure 4: The moisture content of compacted clay changes with temperature.

0.10
0.09
0.08
0.07
0.06
0.05
CIF

0.04
0.03
0.02
0.01
0.00
−0.01
0 2 4 6 8 10 12 14 16 18
Time (d)
T = 50∘ C
T = 60∘ C

Figure 5: CIF of compacted clay under temperature gradient.

3.2. Hydraulic Conductivity of Cracked Clay. As shown in the hydraulic conductivity of compacted clay samples is close
Figures 6 and 7, the influence of wetting-drying cycles on to equilibrium after three wetting-drying cycles. The cracks
hydraulic conductivity of the compacted clay is remarkable. presented in the clay surface provided a path for water
After four wetting-drying cycles, under the dry density (𝜌 = seepage after wetting-drying cycles.
1.65 g/cm3 ) and moulding moisture content (𝑤0 = 0.25),
the hydraulic conductivity of the compacted clay showed
a significant increase, with its value jumping from 3.6 ×
4. Conclusions
10−8 cm/s to 1.5 × 10−5 cm/s. The hydraulic conductivity of The main purpose of the study was to observe the heat-
the compacted clay (𝜌 = 1.65 g/cm3 , 𝑤0 = 0.21) increased moisture transport and water permeability of compacted clay
from 2.5 × 10−8 cm/s to 8.3 × 10−7 cm/s. The hydraulic during cracking of compacted clay. To fulfill this purpose,
conductivity of other samples is two orders of magnitude a series of tests of heat-moisture transport through soil
higher than that of uncracked soils. It was also observed that columns during cracking were performed, and the hydraulic
Journal of Chemistry 5

1E − 5 1E − 5
Hydraulic conductivity (cm/s)

Hydraulic conductivity (cm/s)


1E − 6 1E − 6

1E − 7 1E − 7

1E − 8 1E − 8
0 1 2 3 4 0 1 2 3 4
Number of wetting-drying cycles Number of wetting-drying cycles
𝜌= 1.50 g/cm3 w = 0.21 𝜌= 1.65 g/cm3 w = 0.17
𝜌= 1.55 g/cm3 w = 0.21 𝜌= 1.55 g/cm3 w = 0.19
𝜌= 1.60 g/cm3 w = 0.21 𝜌= 1.65 g/cm3 w = 0.21
𝜌= 1.65 g/cm3 w = 0.21 𝜌= 1.65 g/cm3 w = 0.23
𝜌= 1.70 g/cm3 w = 0.21 𝜌= 1.65 g/cm3 w = 0.25
(a) (b)

Figure 6: The hydraulic conductivity of compacted clay during wetting-drying cycles.

1E − 5 1E − 5
Hydraulic conductivity (cm/s)

Hydraulic conductivity (cm/s)

1E − 6 1E − 6

1E − 7 1E − 7

1E − 8 1E − 8
1.50 1.55 1.60 1.65 1.70 0.16 0.18 0.20 0.22 0.24 0.26
Soil compaction (g/cm3 ) Moulding moisture content
Before wetting-drying cycles Before wetting-drying cycles
After wetting-drying cycles After wetting-drying cycles
(a) (b)

Figure 7: The hydraulic conductivity of compacted clay before and after wetting-drying cycles.

conductivity of soil under different wetting-drying cycles was This value increased by 2∼3 orders of magnitude compared
also investigated. to that of uncracked clay, suggesting that the compacted clay
In the early phase of heating, the temperature of soil completely lost its antiseepage capability.
columns rose fast and the moisture loss was serious. The
speed of temperature increased and then attenuated signifi- Conflict of Interests
cantly, with a long period of slow dehydration. After heating
for 16 consecutive days at 50∘ C and 60∘ C, at the depth of The authors declare that there is no conflict of interests
25 cm, the temperature increased to 48.5∘ C and 59.1∘ C and regarding the publication of this paper.
the soil moisture content decreased to 0. Cracks appeared at
the surface of soil columns. The CIF was 0.043 and 0.097 and Acknowledgments
the crack depth was about 6.5 cm and 8.2 cm, respectively.
After four wetting-drying cycles, the hydraulic conductiv- The authors would like to express their great appreciation
ity of compacted clay was within 8.3 × 10−7 to 1.5 × 10−5 cm/s. of the funding provided by the National Natural Science
6 Journal of Chemistry

Foundation of China (11002102 and 51474168) and Wuhan


Youth Chenguang Program of Science and Technology
(201271031419).

References
[1] China Architecture and Building Press, “Technical code for
municipal solid waste sanitary landfill,” CJJ 17-2004, China
Architecture and Building Press, Beijing, China, 2004, (Chi-
nese).
[2] R. Klein, T. Baumann, E. Kahapka, and R. Niessner, “Temper-
ature development in a modern municipal solid waste incin-
eration (MSWI) bottom ash landfill with regard to sustainable
waste management,” Journal of Hazardous Materials, vol. 83, no.
3, pp. 265–280, 2001.
[3] P. Döll, “Desiccation of mineral liners below landfills with
heat generation,” Journal of Geotechnical and Geoenvironmental
Engineering, vol. 123, no. 11, pp. 1001–1009, 1997.
[4] Y. Zhou and R. K. Rowe, “Development of a technique for
modelling clay liner desiccation,” International Journal for
Numerical and Analytical Methods in Geomechanics, vol. 27, no.
6, pp. 473–493, 2003.
[5] P. J. Hewitt and L. K. Philip, “Problems of clay desiccation in
composite lining systems,” Engineering Geology, vol. 53, no. 2,
pp. 107–113, 1999.
[6] N. Lecocq and N. Vandewalle, “Dynamics of crack opening in
a one-dimensional desiccation experiment,” Physica A, vol. 321,
no. 3-4, pp. 431–441, 2003.
[7] C. Basnett and R. Bruner, “Clay desiccation of a single-
composite liner system,” in Proceedings of the Geosynthetics’93
Conference, pp. 1329–1334, Vancouver, Canada, 1993.
[8] H. Lu, M. Luan, and J. Zhang, “Transport of Cr(VI) through clay
liners containing activated carbon or acid-activated bentonite,”
Applied Clay Science, vol. 50, no. 1, pp. 99–105, 2010.
[9] S. Sadek, S. Ghanimeh, and M. El-Fadel, “Predicted perfor-
mance of clay-barrier landfill covers in arid and semi-arid
environments,” Waste Management, vol. 27, no. 4, pp. 572–583,
2007.
[10] C. Bradley, M. Mosugu, and J. Gerrard, “Seasonal dynamics of
soil-water pressure in a cracking clay soil,” Catena, vol. 69, no.
3, pp. 253–263, 2007.
[11] H. Mi, Kinematic wave formulation for flow through macrop-
orous soil [Ph.D. thesis], Department of Civil and Environmen-
tal Engineering, Wayne State University, Detroit, Mich, USA,
1995.
[12] C. J. Miller, H. Mi, and N. Yesiller, “Experimental analysis of
desiccation crack propagation in clay liners,” Journal of the
American Water Resources Association, vol. 34, no. 3, pp. 677–
686, 1998.
[13] N. Yesiller, C. J. Miller, G. Inci, and K. Yaldo, “Desiccation
and cracking behavior of three compacted landfill liner soils,”
Engineering Geology, vol. 57, no. 1-2, pp. 105–121, 2000.
Hindawi Publishing Corporation
Journal of Chemistry
Volume 2015, Article ID 592951, 11 pages
http://dx.doi.org/10.1155/2015/592951

Research Article
Modeling Wettability Variation during
Long-Term Water Flooding

Renyi Cao,1 Changwei Sun,1 and Y. Zee Ma2


1
School of Petroleum Engineering, China University of Petroleum, Beijing 102245, China
2
Schlumberger, Denver, CO 80202, USA

Correspondence should be addressed to Renyi Cao; caorenyi@gmail.com

Received 30 September 2014; Revised 20 January 2015; Accepted 18 March 2015

Academic Editor: Cengiz Soykan

Copyright © 2015 Renyi Cao et al. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Surface property of rock affects oil recovery during water flooding. Oil-wet polar substances adsorbed on the surface of the rock will
gradually be desorbed during water flooding, and original reservoir wettability will change towards water-wet, and the change will
reduce the residual oil saturation and improve the oil displacement efficiency. However there is a lack of an accurate description
of wettability alternation model during long-term water flooding and it will lead to difficulties in history match and unreliable
forecasts using reservoir simulators. This paper summarizes the mechanism of wettability variation and characterizes the adsorption
of polar substance during long-term water flooding from injecting water or aquifer and relates the residual oil saturation and relative
permeability to the polar substance adsorbed on clay and pore volumes of flooding water. A mathematical model is presented to
simulate the long-term water flooding and the model is validated with experimental results. The simulation results of long-term
water flooding are also discussed.

1. Introduction film during the deposition process [6, 7], as shown in


Figure 1(a). Thus, the reservoir rocks are water-wet. During
Wettability is an important factor that affects oil recovery in the accumulation process, the partial polar substances in
water flooding. In the past, oil reservoirs were commonly crude oil penetrate water film to adsorb on rock surface
interpreted to be strongly water-wet, because water phase through van der Waals, electrostatic, hydrogen bonding, and
is generally the initial fluid contact with the reservoir rock. acid-base [8–11], and the rock with the polar substances
However, some researchers found that many reservoir rocks changes to oil-wet, as shown in Figure 1(b). Note, however,
were not strongly water-wet [1, 2]. The latter results were that if the partial polar substances do not penetrate water film,
obtained without considering the natural surfactant in crude the rock will be water-wet. During long-term water flooding,
oil, such as asphalt and paraffin substances, which are easily the polar substances can be washed out and desorbed from
absorbed on the solid-liquid interface and change reservoir the rock surface; the wettability may change to water-wet, as
rocks to oil-wet [3]. Kusakov [4] found that the rupture shown in Figure 1(c).
of water film could lead to crude oil directly contacting Wettability may affect almost all the variables in core
the quartz surface and the surface would become oil-wet analysis, including capillary pressure, relative permeability,
instead of water-wet. Schmid [5] claimed that wettability oil recovery, and EOR. Morrow [12] confirmed that a higher
could be changed and strong water-wet rock could become water flooding recovery could be obtained in strongly water-
weak water-wet after contact with crude oil. wet core through one-dimensional water flooding experi-
For the sandstone reservoirs, the wettability changes ments. The one-dimensional displacement experiments with
during deposition process, accumulation process, and water high pore volume (PV) of injected water [13] showed that
flooding process (Figure 1). During deposition process, the the changes of wettability and pore structure during long-
pores are saturated with water to form a layer of water term water flooding reduced the critical capillary number
2 Journal of Chemistry

Oil
Oil

Water Polar sub Water Polar sub Water

Clay

Water film Water film Water film


Rock Rock Rock

(a) Deposition process (b) Accumulation process (c) Water-flooding process

Figure 1: Wettability variation during reservoir formation process.

and the residual oil saturation. When the injected water was 2. Mathematical Models for
5000 PV of rock core, the oil displacement efficiency of cores Long-Term Water Flooding
was about 57%; when the injected water was 1000 PV or
even 10000 PV, the oil displacement efficiency was increased 2.1. Conditions. The following conditions are assumed in our
(near 80%) [13]. Therefore, under different water flooding mathematical modeling for long-term water flooding.
conditions, oil displacement efficiency of the rock is not a
constant. (1) Temperature of reservoir is constant.
The wettability changes during long-term water flooding, (2) Reservoir fluids are oil and water phases.
and this could affect the residual oil saturation and oil (3) Components include water, oil, polar substance (AO),
recovery. A number of studies showed that, during low- clay, Na+ , Ca2+ , Mg2+ , Fe3+ , OH− , and other anions.
salinity water flooding, the ions could exchange between (4) No phase transition process is in reservoir.
injecting water and rock, and it could lead to adsorption of
divalent ions and minerals dissolution, change the wettability (5) Porosity and permeability don’t change during long-
to water-wet, and enhance oil recovery [14–20]; it could also term water flooding.
lead to changes of relative permeability [21–25]. Recently, (6) Gravity is not considered.
mechanism on wettability alternation, such as surfactant
flooding, gas displacement, and low salinity water flooding 2.2. Model of Variation for Polar Substances on the Rock
has been studied [16, 20, 26–30]. Surface. The amount of polar substances adsorbed on the
The main factor affecting wettability is the amount of rock surface mainly is determined by the polar substances
polar substances adsorbed on the clay surface. In order to adsorbed on the clay and the content of clay in rock. The
model the wettability variation, we should first evaluate the content of polar substances adsorbed on the rock can be
amount of polar substances adsorbed on the surface of clay. described by
Five factors that impact the polar adsorption and desorption
should be considered, and these include the concentration of AO AO clay
𝐶rock = 𝐶clay ⋅ 𝐶rock , (1)
polar substances in crude oil and on rock surface [31], salinity
and pH of injecting water and formation water [32, 33], where 𝐶rock
AO
—concentration of polar substances (AO)
clay contents in rock [28], and flow rate of water through
adsorbed on the rock; 𝐶clay
AO
—concentration of polar
the pore [31]. Yet, an accurate description of wettability clay
alternation model during long-term water flooding from substances (AO) adsorbed on the clay; and 𝐶rock —concentra-
injecting well or aquifer is lacking, which will lead to failure tion of clay in the rock.
in history matching and unreliable forecasts using numerical The clay could fall off from the rock surface by the water
simulation [34, 35]. flushing under long-term water flooding, and this process can
The purpose of this paper includes: (1) characterizing be described by
the adsorption of polar substance during long-term water DESORPTION
clay clay
flooding from injecting water or aquifer and then relate the 𝐶rock 󳨀󳨀󳨀󳨀󳨀󳨀󳨀󳨀󳨀󳨀→ 𝐶water , (2)
residual oil saturation and relative permeability to the polar
clay
substance adsorbed on clay and pore volumes of flooding where 𝐶water —concentration of clay in the aqueous phase
water and (2) building a mathematical model to simulate the (water).
long-term water flooding. The simulation results of long-term Assuming that the process for the clay desorption is
water flooding are also discussed in this paper. nonlinear and the desorption rate is a function of the aqueous
Journal of Chemistry 3

phase flow rate 𝜐𝑤 , the Langmuir adsorption law can be used by reaction in phase 𝑙 with unit volume. 𝑄𝑘 𝐹𝑆 is output rate of
to characterize it: component 𝑘 solid-phase under mechanical force (long-term
clay clay
erosion) conditions with unit volume. For clay component, it
𝑑𝐶rock 𝑐rock can be determined by (3). 𝑄𝑘 𝑇𝑆 is output rate of ion balance
= 𝐾1 ⋅ 𝜐𝑤 ⋅ , (3)
𝑑𝑡 clay
1 + 𝑏𝑐rock for component 𝑘 on rock surface with unit volume.
The output rate of polar substances and clay components
clay can be determined by
where 𝑑𝐶rock /𝑑𝑡—desorption rate; 𝐾1 , 𝑏—clay desorption
constant; and 𝜐𝑤 —flow rate of water phase. 𝑛𝑝
𝑇 exm 𝑘
Equation (3) implies the influence of total volume of 𝑄𝑘 = ∑𝑆𝑙 ⋅ 𝐾𝑘𝑙 𝐶rock , (8)
water flushing the rock surface during the long-term water 𝑙=1
flooding, because the rate of clay desorption in rock is related
clay
to the flow rate of water phase. 𝑐rock
The equilibrium between the concentration of polar 𝑄𝑘 𝐹𝑆 = 𝐾1 ⋅ 𝜐𝑤 ⋅ clay
, (9)
1 + 𝑏𝑐rock
substances in the oil phase and the concentration of polar
substances in clay mineral surface is expressed by 𝑄𝑘 𝑇𝑆 = 𝐾exm 𝐶rock
𝑘
, (10)
AO EQUILIBRIUM
AO
𝐶clay ←󳨀󳨀󳨀󳨀󳨀󳨀󳨀󳨀󳨀→ 𝐶oil , (4) where 𝐾𝑘𝑙exm
= 𝑑𝐶𝑙𝑘 /𝑑𝑡 is the output rate of component 𝑘 in
the liquid phase 𝑙 and 𝐾1 is deposition rate of component 𝑘
where 𝐶oil
AO
—concentration of the polar substance (AO) in the on the rock surface; consider
oil phase.
This equilibrium can be represented by the exchange ∑ [Na+ ] [Ca2+ ] [Mg2+ ] [Fe3+ ]
𝐾exm = 𝑘2 . (11)
equilibrium equation with a coefficient, 𝜅exm , which is the [OH− ]
equilibrium constant for the exchange between polar sub-
stances on the clay surface and the oil phase: The aqueous-phase pressure is obtained by an overall
mass balance on volume components [36]. Because of the
𝐶clay
AO
following condition
𝜅exm = . (5)
𝐶oil
AO
𝑛cv
∑ 𝐶𝑘𝑙 = 1, (12)
Formation water salinity and pH influence adsorption of 𝑘=1
polar substances on the rock surface [18, 28]. The equilib-
rium constant 𝜅exm can be characterized as the relationship the pressure equation in terms of the reference phase pressure
between the total concentration of the metal cations in (phase 1) is
formation water and the concentration of hydroxide ions:
𝜕𝑝1
𝜙𝐶𝑡 ⋅ + ∇𝐾 ⋅ 𝜆 𝑟𝑇𝑐 ⋅ ∇𝑝1
+ 2+
∑ [Na ] [Ca ] [Mg ] [Fe ] ⋅ ⋅ ⋅ 2+ 3+ 𝜕𝑡
𝜅exm = 𝐾2 , (6) 𝑛𝑝 𝑛𝑝
[OH− ]
= −∇∑𝐾 ⋅ 𝜆 rec ⋅ ∇ℎ + ∇ ⋅ ∑𝐾 ⋅ 𝜆 𝑟𝑇𝑐 ⋅ ∇𝑝𝑐𝑙1 (13)
𝑙=1 𝑙=1
where 𝐾2 —balance factor; [Na+ ][Ca2+ ][Mg2+ ][Fe3+ ]—the
concentration of sodium ion in the aqueous phase. 𝑛𝑝
+ ∑ (𝑄𝑘 + 𝑄𝑘 𝑇 + 𝑄𝑘 𝐹𝑆 + 𝑄𝑘 𝑇𝑆 ) ,
𝑙=1
2.3. Mass Conservation Equations and Pressure Equations.
The continuity of mass for component 𝑘 in association with 𝑛 𝑛
where 𝜆 rec = (𝑘𝑟𝑙 /𝜇𝑙 )⋅∑𝑘=1
cv
⋅𝜌𝑘 ⋅𝐶𝑘𝑙 ; 𝜆 𝑟𝑇𝑐 = ∑𝑙=1
𝑝
𝜆 rec is the total
Darcy’s law is expressed as a function of the overall volume of
relative mobility with the correction for fluid compressibility.
component per unit pore volume (𝐶𝑘 ): The total compressibility 𝐶𝑡 is the volume-weighted sum
𝑛𝑝 of the rock or soil matrix (𝐶𝑟 ) and component compressibil-
𝜕 ̃𝑘𝑙 )] ities (𝐶𝑘0 ) are
(𝜙𝐶𝑘 ⋅ 𝜌𝑘 ) + ∇ [∑𝜌𝑘 (𝐶𝑘𝑙 ⋅ 𝑈𝑙 − 𝐷
𝜕𝑡 𝑙=1 (7) 𝑛cv
𝑇 𝐹𝑆 𝑇𝑆 ̃𝑘 ,
𝐶𝑡 = 𝐶𝑟 + ∑ 𝐶𝑘0 ⋅ 𝐶 (14)
= 𝑄𝑘 + 𝑄𝑘 + 𝑄𝑘 + 𝑄𝑘 ,
𝑘=1

where 𝑛cv —total number of volume-occupying components where 𝜙 = 𝜙𝑅 ⋅ [1 + 𝐶𝑟 (𝑝ref − 𝑝ref0 )].
(including water, oil, and polar substances); 𝑛𝑝 —number of
phases; 𝐶𝑘 —adsorbed concentration of component 𝑘, 𝐶𝑘 = 2.4. Relative Permeability Model. During long-term water
𝑛cv ̂ 𝑛𝑝
1−[∑𝑘=1 𝐶𝑘 ] ∑𝑙=1 ̂𝑘 , 𝑘 = 1, 2, . . . , 𝑛cv ; 𝜌𝑘 —density of
𝑆𝑙 ×𝐶𝑘𝑙 + 𝐶 flooding, the wettability changes with the reduction of polar
pure component 𝑘; 𝑄𝑘 —injection/output rate of component substances on the rock surface, and the residual oil saturation
𝑘 with Unit volume; and 𝑄𝑘 𝑇 —output rate of component 𝑘 and displacement oil efficiency will change as well. Since the
4 Journal of Chemistry

After high PV,


high water-wet
After low PV,
Cmin
initial wettability
kro Cmax
krw
kro
krw krw
kro

(a) (b) (c)


Figure 2: Relative permeabilities determined at different pore volumes of water injected.

polar substances on the rock surface is not constant, which The relative permeabilities for phase l are assumed to be
depends on the time of displacement and flow rate of water unique functions of their respective saturations only [37],
phase, the residual oil saturation is a function of flow rate described as
and time of water flooding (the cumulative volumes of water
𝑛
displacement). 𝐾𝑟𝑙 = 𝐾𝑟𝑙end (𝑆𝑙∗ ) , (17)
Since the variation of wettability is reflected by the relative
permeabilities and residual oil saturation, different relative
where 𝑆𝑙∗ —the normalized saturations for phase 𝑙, 𝑆𝑤

= ((𝑆𝑤 −
permeabilities (which could be measured with cores after ∗ ∗
𝑆𝑤𝑐 )/(1 − 𝑆𝑤𝑐 − 𝑆or )), 𝑆𝑜 = 1 − 𝑆𝑤 ; and 𝑛—the exponent
different volumes of water displacement, Figure 2) can be
determined by fitting the data.
used to calculate the residual oil saturation and displacement
oil efficiency [28].
The relative permeability curve at low PV (initial wettabil- 3. Numerical Models
ity, Figure 2(a)) of injected water reflects the initial wettability
AO max Based on (7), the polar substances and clay only dissolve in
of rock with maximum contents (𝐶rock represented by
𝐶max ) of polar substances adsorbed on rock; the relative water phase, and the governing equations can be described as
permeability curve at high PV (high water-wet at ultimate follows:
water displacement, Figure 2(c)) of injected water reflects
the initial wettability of rock with minimum contents (𝐶min ) ̃𝑘,𝑤 )
𝜕 (𝜑𝜆 𝑘 𝜌𝑘 𝐶
clay
of polar substances adsorbed on rock. We can use 𝐶rock 𝜕𝑡
to interpolate the relative permeabilities and residual oil 𝜕 𝜕𝐶𝑘,𝑤
saturation (Figure 2(b)) at different PV of injected water. + [𝜌𝑘 (𝐶𝑘,𝑤 𝑢𝑥𝑤 − 𝜑𝑆𝑤 (𝐾𝑥𝑥𝑘𝑤
𝜕𝑥 𝜕𝑥
Given relative permeability curves at 𝐶max and 𝐶min ,
we can determine the maximum and minimum residual 𝜕𝐶𝑘,𝑤
+ 𝐾𝑥𝑦𝑘𝑤
oil saturation, and the residual oil saturation 𝑆or can be 𝜕𝑦
calculated with the linear interpolation methods [27]: 𝜕𝐶𝑘,𝑤
+ 𝐾𝑥𝑧𝑘𝑤 ))]
𝜕𝑧
𝐶 − 𝐶min
𝑆or = 𝑆or max − (𝑆 max − 𝑆or min ) , (15) 𝜕 𝜕𝐶𝑘,𝑤
𝐶max − 𝐶min or + [𝜌𝑘 (𝐶𝑘,𝑤 𝑢𝑦𝑤 − 𝜑𝑆𝑤 (𝐾𝑦𝑥𝑘𝑤
𝜕𝑦 𝜕𝑥
where 𝑆or max —the residual oil saturation at 𝐶max (low PV, 𝜕𝐶𝑘,𝑤 (18)
initial wettability); 𝑆or max —the residual oil saturation at 𝐶min + 𝐾𝑦𝑦𝑘𝑤
(high PV, high water-wet wettability). 𝜕𝑦
The endpoints of relative permeabilities are computed 𝜕𝐶𝑘,𝑤
+ 𝐾𝑦𝑧𝑘𝑤 ))]
with a linear interpolation between the given input values at 𝜕𝑧
𝐶max and 𝐶min , 𝜕𝐶𝑘,𝑤
𝜕
+ [𝜌𝑘 (𝐶𝑘,𝑤 𝑢𝑧𝑤 − 𝜑𝑆𝑤 (𝐾𝑧𝑥𝑘𝑤
𝑆or − 𝑆or min 𝜕𝑧 𝜕𝑥
max
𝐾𝑟𝑙end = 𝐾𝑟𝑙end − (𝐾max − 𝐾𝑟𝑙end
min
), (16) 𝜕𝐶𝑘,𝑤
𝑆or max − 𝑆or min 𝑟𝑙end + 𝐾𝑧𝑦𝑘𝑤
𝜕𝑦
where 𝐾𝑟𝑙end —the endpoint of relative permeability for phase 𝜕𝐶𝑘,𝑤
max
𝑙 at 𝑆or ; 𝐾𝑟𝑙end —the endpoint of relative permeability for phase + 𝐾𝑧𝑧𝑘𝑤 ))]
max min 𝜕𝑧
𝑙 at 𝑆or ; and 𝐾𝑟𝑙end —the endpoint of relative permeability
for phase 𝑙 at 𝑆or .min = 𝑅𝑘 𝑘 = 1, 2, . . . , 𝑛cv ,
Journal of Chemistry 5

where Table 1: Parameters of the numerical model.


𝑛
Parameters Values
̂𝑚 ) 𝑆𝑤 ,
𝜆 𝑘 = (1 − ∑ 𝐶
Porosity 0.25
𝑚=1
Permeability/mD 1000
𝐾𝑥𝑥𝑘𝑤 , 𝐾𝑥𝑦𝑘𝑤 , 𝐾𝑥𝑧𝑘𝑤 (19) Grid numbers 100
[ ]
[𝐾𝑦𝑥𝑘𝑤 , 𝐾𝑦𝑦𝑘𝑤 , 𝐾𝑦𝑧𝑘𝑤 ] , Length/cm 7.5
[ ]
𝐾
[ 𝑧𝑥𝑘𝑤 , 𝐾𝑧𝑦𝑘𝑤 , 𝐾𝑧𝑧𝑘𝑤 ]
𝑛
is the diffusion tensor of component 𝑘, 𝑅𝑘 = 𝑄𝑘 +𝑄𝑘 𝑇 +𝑄𝑘 𝐹𝑆 + Δ𝑦Δ𝑧 (𝐾𝜆 𝑟𝑡𝑐 )
𝑖±1/2
𝑄𝑘 𝑇𝑆 . 𝑇𝑡𝑤,𝑖±1/2 = ,
Δ𝑥𝑖±1/2
If the water phase is selected as the reference phase, 𝑛
with the assumption that the component of oil cannot be Δ𝑥Δ𝑧 (𝐾𝜆 𝑟𝑡𝑐 )
𝑗±1/2
dissolved in water phase, the pressure equation in water phase 𝑇𝑡𝑤,𝑗±1/2 = ,
is expressed as Δ𝑦𝑗±1/2
𝑛
𝑛
𝜕𝑝 cv Δ𝑦Δ𝑧 (𝐾𝜆 𝑟𝑡𝑐 )
𝜙𝐶𝑡 𝑤 − ∇ (𝐾𝜆 𝑟𝑡𝑐 ∇𝑝𝑤 ) = ∑ 𝑅𝑘 . (20) 𝑇𝑡𝑤,𝑚±1/2 =
𝑚±1/2
.
𝜕𝑡 𝑘=1 Δ𝑧𝑚±1/2
The oil phase pressure equations can be computed by (22)
adding the capillary pressure between phases. The implicit
form to process the pressure of water phase 𝑃𝑤 and the explicit Thus, (21) can be reformulated to
form to process the coefficients can be used:
𝑛
(𝜑𝐶𝑡 𝑉) 𝑛+1 𝑛
𝑛+1 𝑛 (𝑃𝑤,𝑖 − 𝑃𝑤,𝑖 )
𝑛 (𝑃𝑤,𝑖 − 𝑃𝑤,𝑖 ) Δ𝑡
(𝜑𝐶𝑡 𝑉)
Δ𝑡 𝑛+1
− {𝑇𝑡𝑤,𝑖+1/2 (𝑃𝑤,𝑖+1 𝑛+1
− 𝑃𝑤,𝑖 )
𝑛+1 𝑛+1
𝑛 𝑃𝑤,𝑖+1 − 𝑃𝑤,𝑖 𝑛+1 𝑛+1
− Δ𝑦Δ𝑧 [(𝐾𝜆 𝑟𝑡𝑐 ) + 𝑇𝑡𝑤,𝑖−1/2 (𝑃𝑤,𝑖−1 − 𝑃𝑤,𝑖 )
𝑖+1/2 Δ𝑥𝑖+1/2
𝑛+1 𝑛+1
𝑛+1 𝑛+1 + 𝑇𝑡𝑤,𝑗+1/2 (𝑃𝑤,𝑗+1 − 𝑃𝑤,𝑗 )
𝑛 𝑃𝑤,𝑖 − 𝑃𝑤,𝑖−1
− (𝐾𝜆 𝑟𝑡𝑐 ) ] 𝑛+1 𝑛+1 (23)
𝑖−1/2 Δ𝑥𝑖−1/2 + 𝑇𝑡𝑤,𝑗−1/2 (𝑃𝑤,𝑗−1 − 𝑃𝑤,𝑗 )
𝑛+1 𝑛+1 𝑛+1 𝑛+1
𝑛 𝑃𝑤,𝑗+1 − 𝑃𝑤,𝑗 + 𝑇𝑡𝑤,𝑚+1/2 (𝑃𝑤,𝑚+1 − 𝑃𝑤,𝑚 )
− Δ𝑥Δ𝑧 [(𝐾𝜆 𝑟𝑡𝑐 ) 𝑛+1 𝑛+1
𝑗+1/2 Δ𝑦𝑗+1/2 + 𝑇𝑡𝑤,𝑚−1/2 (𝑃𝑤,𝑚−1 − 𝑃𝑤,𝑚 )}
𝑛+1 𝑛+1 (21)
𝑛 𝑃𝑤,𝑗 − 𝑃𝑤,𝑗−1 𝑁cv
− (𝐾𝜆 𝑟𝑡𝑐 ) ] = 𝑉 ∑ 𝑅𝑘 .
𝑗−1/2 Δ𝑦𝑗−1/2 𝑘=1
𝑛+1 𝑛+1
𝑛 𝑃𝑤,𝑚+1 − 𝑃𝑤,𝑚
− Δ𝑥Δ𝑦 [(𝐾𝜆 𝑟𝑡𝑐 ) Equation (23) includes seven pairs of diagonal matrix for the
𝑚+1/2 Δ𝑧𝑚+1/2 three-dimensional space:
𝑛+1 𝑛+1
𝑛 𝑃𝑤,𝑚 − 𝑃𝑤,𝑚−1
− (𝐾𝜆 𝑟𝑡𝑐 ) ] 𝑎𝑖,𝑗,𝑘 𝑃𝑖,𝑗,𝑘−1 + 𝑏𝑖,𝑗,𝑘 𝑃𝑖,𝑗−1,𝑘 + 𝑐𝑖,𝑗,𝑘 𝑃𝑖−1,𝑗,𝑘 + 𝑑𝑖,𝑗,𝑘 𝑃𝑖,𝑗,𝑘
𝑚−1/2 Δ𝑧𝑚−1/2 (24)
+ 𝑒𝑖,𝑗,𝑘 𝑃𝑖−1,𝑗,𝑘 + 𝑓𝑖,𝑗,𝑘 𝑃𝑖,𝑗+1,𝑘 + 𝑔𝑖,𝑗,𝑘+1 𝑃𝑖,𝑗,𝑘−1 = ℎ𝑖,𝑗,𝑘 .
𝑁cv
= 𝑉 ∑ 𝑅𝑘 .
𝑘=1
Based on the UTCHEM simulator (UTCHEM, Version
6.1, 1999), we use the incomplete LU decomposition method
Define to precondition (24) and use the conjugate gradient method
𝑛 to solve (24). In order to improve the accuracy in the
Δ𝑦Δ𝑧 (𝐾𝜆 𝑟𝑙𝑐 ) front of displacement, we choose the third-order format
𝑖±1/2
𝑇𝑙,𝑖±1/2 = , TVD method [38] as the difference scheme to process the
Δ𝑥𝑖±1/2
𝑛 coefficients.
Δ𝑥Δ𝑧 (𝐾𝜆 𝑟𝑙𝑐 )
𝑗±1/2
𝑇𝑙,𝑗±1/2 = , 4. Results and Discussion
Δ𝑦𝑗±1/2
𝑛
Δ𝑦Δ𝑧 (𝐾𝜆 𝑟𝑙𝑐 ) We use the model to simulate a one-dimensional formation;
𝑚±1/2 the parameters are shown in Table 1. The parameters of polar
𝑇𝑙,𝑚±1/2 = ,
Δ𝑧𝑚±1/2 substances are referenced from the experiments [39].
6 Journal of Chemistry

Table 2: Parameters of experimental cores.

Core number Lengths (mm) Diameter (mm) Dry weight (g) Porosity (%) Gas-measured
permeability (×10−3 𝜇m2 )
2-1 25.3 39.5 33.3 23.1 2535
2-2 25.1 42.5 35.7 25.7 2513

Table 3: Ion content and salinity of brine.

Ions Na+
Ca+ Mg2+ K+ Cl− HCO3 − SO4 2−
Concentration (mg/L) 11.224 768 187 15.0 19.211 7.6 72

Table 4: Displacement experiment data.

Injected rate Viscosity Interfacial tension Sweep rate Displacement Residual oil Capillary
Core number
(mL/min) (cp) (mN/m) (10−5 m/s) Efficiency (%) saturation (%) number
2-1 1.00 5.3 1.1𝐸 − 01 2.15 65.4 25.5 1.0𝐸 − 03
2-2 1.00 5.3 2.3𝐸 − 02 2.72 74.9 18.8 6.3𝐸 − 03

100
P
P
Oil displacement efficiency (%)

P 80
Core holder

Hand pump 60
Brine

Intermediate
Oil

container 40

20

Pump 0
0.1 1 10 100 1000 10000 100000
Injected pore volume (PV)
Figure 3: Setups for experimental displacement. Result from Ji et al.
Experiment by this paper

Figure 4: Comparison of the result with experiment by Ji et al. [13].


4.1. Experiments and Verification In Ji’s experiment, the core with 2413 mD permeability was from
Daqing Oilfield of China and the injected water reached 30,000 PV.
4.1.1. Experimental Materials and Conditions. The cores used
in the experimental material were sealed cores from the
reservoir formation (Table 2).
Brine was prepared to simulate the formation water of shown in Figure 5(a), and the pressure and flow rate
1080 mg/L salinity (Table 3). Oil with 5.3 cp of viscosity at of oil and water were recorded (Table 4).
80∘ C was a mixture of crude oil and light oil.
The temperature was kept at 80∘ C for all the tests. (3) After 30 PV of water flooding, 2000 PV of brine
The brine was injected by a constant-velocity pump of was continuously injected and the oil displacement
1.0 mL/min. efficiency was also continuously recorded (Figure 4).
(4) The core was saturated with oil again, and then the
4.1.2. Experimental Setup and Procedures. The displacement relative permeability was determined with 30 PV of
setup is shown in Figure 3 and the procedures include the brine, as shown in Figure 5(b).
following.

(1) The core was saturated with brine and oil, and then 4.1.3. Experimental Results and Verification. In order to
core was socked to recover the initial wettability for compare the result with another study, the displacement
15 days. experiment was compared with results by Ji et al. [13]
shown in Figure 4. Due to the different cores in experiments,
(2) The relative permeability was determined with 30 PV only trends of increasing of oil displacement efficiency were
of brine injected with the nonsteady method, as compared, and the results show that the trends were the same
Journal of Chemistry 7

1 1

0.8 0.8

Relative permeability
Relative permeability

0.6 0.6

0.4 0.4

0.2 0.2

0 0
0 20 40 60 80 100 0 20 40 60 80 100
Water saturation (%) Water saturation (%)

Kro for core number 2-1 at 30 PV Kro for core number 2-2 at 2000 PV
Krw for core number 2-1 at 30 PV Krw for core number 2-2 at 2000 PV

(a) Relative permeabilities determined at 30 PV (b) Relative permeabilities determined at 2000 PV

Figure 5: Relative permeabilities determined at 30 PV and 2000 PV.


Oil displacement efficiency\residual

100 100
Oil displacement efficiency\residual

80 80
oil saturation (%)
oil saturation (%)

60 60

40 40

20 20

0 0
0 1 10 100 0 1 10 100 1000 10000
Injected pore volume (PV) Injected pore volume (PV)

Oil displacement efficiency Oil displacement efficiency


Residual oil saturation Residual oil saturation
Water cut Water cut

Figure 6: Oil displacement efficiency, residual oil saturation, and Figure 7: Oil displacement efficiency, residual oil saturation, and
water cut with semilogarithmic injected pore volume for 30 PV. water cut with injected pore volume for 2000 PV.

for the experiment results of this paper and the results by Ji et


al. [13]. experiment of core number 2-2, oil displacement efficiency
The relative permeability curves at 𝐶max and 𝐶min used could almost reach 80% while the brine was injected for more
in simulation are shown in Figures 5(a) and 5(b), and than 2000 PV as a high PV, shown in Figure 7. Based on
they represent the conditions with initial wettability and the experiments of core number 2-1 and number 2-2, the oil
high water-wet wettability. The simulation result shows that displacement can be enhanced as the brine injected increases.
the changing trend of oil displacement efficiency (𝐸𝑑 ) and The water displacement experiment data is shown in Table 4.
residual oil saturation (𝑆or ) can match the experimental result Notice that many possible factors can influence the
well. matching result between the simulation and experiment,
In core number 2-1, brine was injected in the pore for and a better match can be obtained through adjusting the
30 PV, the oil displacement was nearly 70%, and residual oil parameters of adsorption function (9) during the simulation
saturation had a value of 25%, shown in Figure 6. From the (Figure 8).
8 Journal of Chemistry

Oil displacement efficiency (%)


100 100

Residual oil saturation (%)


80 80

60 60

40 40

20 20

0 0
0 1 10 100 1000 10000
Injected pore volume

Experiment
Model

Figure 8: Oil displacement efficiency, residual oil saturation from experiment, and model during long-term water flooding.

INJTR PRODN INJTR PRODN

10 20 30 40 50 60 70 80 90 100 110 10 20 30 40 50 60 70 80 90 100 110


(a) The initial condition (b) Water flooding multiples reach 30 PV

INJTR PRODN INJTR PRODN

10 20 30 40 50 60 70 80 90 100 110 10 20 30 40 50 60 70 80 90 100 110


(c) Water flooding multiples reach 1000 PV (d) Water flooding multiples reach 2000 PV

Figure 9: Adsorption contents of clay under the condition of different water flooding multiples (gmole/m3 ).

INJTR PRODN INJTR PRODN

10 20 30 40 50 60 70 80 90 100 110 10 20 30 40 50 60 70 80 90 100 110


(a) The initial condition (b) Water flooding multiples reach 30 PV

INJTR PRODN INJTR PRODN

10 20 30 40 50 60 70 80 90 100 110 10 20 30 40 50 60 70 80 90 100 110


(c) Water flooding multiples reach 1000 PV (d) Water flooding multiples reach 2000 PV

Figure 10: Adsorption contents of AO under the condition of different water flooding multiples (gmole/m3 ).

clay
4.2. Change of Polar Substances during Long-Term Water variations of 𝐶rock
AO
and 𝐶rock are shown in Figure 11. 𝐶rock
AO
and
Flooding. The new model can simulate the ion exchange clay
𝐶rock decrease with the increasing PV of injected brine, and
chemical reaction by defining different conditions of ion the decreasing rate is high between 30 PV and 500 PV during
exchange and clay adsorption properties, and we can obtain clay
AO
water flooding. When above 500 PV, the 𝐶rock AO
and 𝐶rock
the contents of polar substances (𝐶rock ) and clay on the rock change slowly, which means that same reduction of polar
clay
(𝐶rock ) at different PV of injected brine (Figures 9 and 10). The substances needs more PV of brine injected at the high water
Journal of Chemistry 9

100 100
Oil displacement efficiency (%)

Oil displacement efficiency (%)


80 80

60 60

40 40

20 20

0 0
0 10 20 30 40 50 60 0 5000 10000 15000 20000
Injected pore volume Injected pore volume

Polar substances content (low PV) Polar substances content (high PV)
(a) When flooding multiples reach 50 PV (b) When flooding multiples reach 2000 PV

Figure 11: Oil displacement efficiency under the condition of different water flooding multiples.

100 100 𝐸𝑑 could reach 75%. Therefore, improving water volume can
Oil displacement efficiency (%)

significantly improve oil displacement efficiency.


Residual oil saturation (%)

80 80 Most of conventional reservoir simulators ignore the


desorption of polar substances. Here we set 𝐾1 to be equal to
60 60
0 (3) to simulate the condition that the desorption is ignored,
clay
40 40 which implies that the desorption rate 𝑑𝐶rock /𝑑𝑡 is equal
to zero. The predicted 𝐸𝑑 while ignoring the desorption is
20 20 lower. This is because 𝑆or is a constant when the desorption
of polar substances is ignored, and only one permeability
0 0 curve is used in the simulation model. When the desorption
1.E − 06

1.E − 04

1.E − 02

1.E + 00

1.E + 02

1.E + 04

1.E + 06

of polar substances is taken into account, the wettability


alternates to water-wet and the permeability curves under
Injected pore volume high PV of water injected are used in simulation model.
Thus, the gradual reduction of 𝑆or during long-term water
Oil displacement efficiency (with desorption)
flooding is reflected in simulation, and the calculated 𝐸𝑑 is
Residual oil saturation (with desorption)
Oil displacement efficiency (without desorption)
more accurate.
Residual oil saturation (without desorption)

Figure 12: Relationship between water flooding multiples and oil


5. Conclusions
displacement efficiency within different adsorption conditions. We draw the following conclusions from our study.
(1) The wettability of oil reservoir changes to water-wet
during long-term water flooding, and the residual
cut stage. This is because the equilibrium of clay adsorption is oil saturation gradually reduces and oil displacement
relatively stable at low water cut, and great washing will break efficiency increases. Therefore, the alternation of wet-
the equilibrium of polar substances adsorbed on clay under tability should be considered during simulation of
high water cut and water saturation above 30 PV. long-term water flooding; otherwise it will lead to an
incorrect history match and unreliable forecasting.
4.3. Change of Residual Oil Saturation during Long-Term (2) The desorption of polar substances from the rock
Water Flooding. The polar substances adsorbed on rock during long-term water flooding is the mechanism of
surface play an important role in the wettability; the residual wettability variation. Since the wettability is reflected
oil saturation (𝑆or ) could be reduced with the wettability by the relative permeability curves in simulation
alternating to water-wet. The simulation results for 𝑆or and model, the key point to characterize the process of
𝐸𝑑 during a long-term water flooding are shown in Figure 12, long-term water flooding is to describe the rela-
in which one case considers and the other ignores the tionship between relative permeability curves and
desorption of polar substances. adsorption of polar substances on rock. The factors
When considering the desorption of polar substances, 𝐸𝑑 for wettability variation include polar substances in
at low PV (<50 PV) (Figure 11) can reach 60%; during 50– crude oil, salinity, and pH of brine, clay in rocks, and
2000 PV 𝐸𝑑 continues to increase to 70%, up to 10000 PV, and flow rate during long-term water flooding.
10 Journal of Chemistry

(3) An efficient approach for long-term water flooding [3] M. O. Denekas, C. C. Mattax, and G. T. Davis, “Effect of crude
simulation is achieved through building a water-oil oil components on rock wettability,” Transactions of the AIME,
two-phase multicomponent simulation model. This vol. 216, p. 330, 1959.
model describes most important physical phenomena [4] M. M. Kusakov, Research in Surface Forces, edited by B. V.
during long-term water flooding, including polar Deryagin, Consultants Bureau, New York, NY, USA, 1963.
substances desorption, residual oil saturation reduc- [5] C. Schmid, “The wettability of petroleum rocks and results of
tion, and variation of relative permeabilities. The experiments to study effects of variations in wettability of core
model is solved through an implicit format to process samples,” Erdoel Kohle, vol. 17, no. 8, p. 605, 1964.
the pressure and through an explicit format to process [6] W. G. Anderson, “Wettability literaure survey—part 1:
the coefficients; the third-order format TVD method rock/oil/brine in teractions and the effects of core handling on
wettabilty,” Journal of Petroleum Technology, vol. 38, no. 11, pp.
was selected as the difference scheme to process the
1125–1144, 1986.
coefficients and to improve the accuracy in the front
[7] J. S. Buckley, Y. Liu, X. Xie, and N. R. Morrow, “Asphaltenes and
of displacement. crude oil wetting—the effect of oil composition,” in Proceedings
(4) The simulation results have a good match to experi- of the 10th Symposium on Improved Oil Recovery, pp. 205–220,
mental results. The results have shown that the reduc- April 1996.
tion of polar substances occurs under the condition of [8] L. Cuiec, “Rock/crude-oil interactions and wettability: an
high water cut and high water saturation, the residual attempt to understand their interrelation,” in Proceedings of the
oil saturation can be gradually reduced by long-term SPE Annual Technical Conference and Exhibition, SPE-13211-MS,
water flooding, and high oil recovery can be obtained Houston, Tex, USA, September 1984.
by increasing the volume of injected water. [9] J. S. Buckley, Y. Liu, and S. Monsterleet, “Mechanisms of wetting
alteration by crude oils,” SPE Journal, vol. 3, no. 1, pp. 54–61,
(5) For the field scale, before history match of simulation, 1998.
a set of configuration parameters should be turned [10] R. S. H. Al-Maamari and J. S. Buckley, “Asphaltene precipitation
according to one-dimensional experiments, such as and alteration of wetting: the potential for wettability changes
ion exchange, clay and polar substances adsorption- during oil production,” SPE Reservoir Evaluation & Engineering,
desorption, and relative permeability interpolation vol. 6, no. 4, pp. 210–214, 2003.
sets. All these turned parameters could be used in [11] A. Graue, B. G. Viksund, and B. A. Baldwin, “Reproducible
each grid of map brown field for numerical simula- wettability alteration of low-permeable outcrop chalk,” SPE
tion. Then, the polar substances could be calculated Reservoir Evaluation & Engineering, vol. 2, no. 2, pp. 134–140,
for each grid and this value will be used in the 1999.
relative permeability interpolation process. However, [12] N. R. Morrow, “Physics and thermodynamics of capillary
there are some limitations for residual oil reduction action in porous media,” Journal of Industrial and Engineering
in this paper, such as the variation of porosity, and Chemistry, vol. 62, no. 1, pp. 43–50, 1970.
permeability should also be considered. We could [13] S. Ji, C. Tian, C. Shi, J. Ye, Z. Zhang, and X. Fu, “New
define the relation of porosity and permeability with understanding on water-oil displacement efficiency in a high
water-cut stage,” Petroleum Exploration and Development, vol.
PV number to deal with this variation.
39, no. 3, pp. 362–370, 2012.
[14] P. P. Jadhunandan and N. R. Morrow, “Effect of wettability
Conflict of Interests on waterflood recovery for crude-oil/brine/rock systems,” SPE
Reservoir Engineering, vol. 10, no. 1, pp. 40–46, 1995.
The authors declare that there is no conflict of interests [15] G. Q. Tang and N. R. Morrow, “Salinity, temperature, oil
regarding the publication of this paper. composition and oil recovery by waterflooding,” SPE Reservoir
Engineering, vol. 12, no. 4, pp. 269–276, 1997.
Acknowledgments [16] G.-Q. Tang and N. R. Morrow, “Influence of brine composition
and fines migration on crude oil/brine/rock interactions and oil
The authors gratefully acknowledge that the project was par- recovery,” Journal of Petroleum Science and Engineering, vol. 24,
tially funded by the National Natural Science Foundation of no. 2–4, pp. 99–111, 1999.
China (E040351304220), Beijing Natural Science foundation [17] G. Q. Tang and N. R. Morrow, “Oil recovery by waterflooding
(no. 3144033), Specialized Research Fund for the Doctoral and imbibition—invading brine cation valency and salinity,” in
Proceedings of the International Symposium of the Society of Core
Program of Higher Education (no. 20130007120014), and Sci-
Analysis, Paper SCA-9911, Golden, Colo, USA, August 1999.
ence Foundation of China University of Petroleum, Beijing.
[18] A. Lager, K. J. Webb, I. R. Collins, and D. M. Richmond,
“LoSalTM enhanced oil recovery: evidence of enhanced oil
References recovery at the reservoir scale,” in Proceedings of the SPE/DOE
Improved Oil Recovery Symposium, paper SPE 113976, Tulsa,
[1] P. G. Nutting, Some Physical and Chemical Properties of Reser- Okla, USA, April 2008.
voir Rocks Bearing on the Accumulation and Discharge of Oil, [19] A. Lager, K. J. Webb, I. R. Collins, and D. M. Richmond,
vol. 12, AAPG, 1934. “LoSalTM enhanced oil recovery: evidence of enhanced oil
[2] R. Leach, O. Wagner, H. Wood, and C. Harpke, “A laboratory recovery at the reservoir scale,” in Proceedings of the SPE/DOE
and field study of wettability adjustment in water flooding,” Improved Oil Recovery Symposium, Paper SPE 113976, Tulsa,
Journal of Petroleum Technology, vol. 14, no. 2, pp. 206–212, 2013. Okla, USA, April 2008.
Journal of Chemistry 11

[20] P. Vledder, J. C. Fonseca, T. Wells, I. Gonzalez, and D. Ligthelm, Journal of Petroleum Science and Engineering, vol. 104, pp. 27–
“Low salinity water flooding: proof of wettability alteration on a 37, 2013.
field wide scale,” in Proceedings of the SPE Improved Oil Recovery [36] M. Delshad, G. A. Pope, and K. Sepehrnoori, “A compositional
Symposium, Paper SPE 129564, pp. 200–209, Tulsa, Okla, USA, simulator for modeling surfactant enhanced aquifer remedia-
April 2010. tion, 1 Formulation,” Journal of Contaminant Hydrology, vol. 23,
[21] K. J. Webb, C. J. J. Black, and Al-Ajeel, “Low salinity oil recovery- no. 4, pp. 303–327, 1996.
log-inject-log,” in Proceedings of the SPE/DOE Symposium on [37] M. Delshad, G. A. Pope, and L. W. Lake, “Two-and three-
Improved Oil Recovery, Paper SPE 89379, Tulsa, Okla, USA, phase relative permeabilities of micellar fluids,” SPE Formation
April 2004. Evaluation, vol. 2, no. 3, pp. 327–337, 1987.
[22] K. J. Webb, C. J. J. Black, and I. J. Edmonds, “Low salinity [38] J. Liu, M. Delshad, G. A. Pope, and K. Sepehrnoori, “Application
oil recovery: the role of reservoir condition corefloods,” in of higher-order flux-limited methods in compositional simula-
Proceedings of the 13th EAGE Symposium on Improved Oil tion,” Transport in Porous Media, vol. 16, no. 1, pp. 1–29, 1994.
Recovery, Paper C18, Budapest, Hungary, April 2005. [39] H. J. Hill, “Cation exchange in chemical flooding: part 3—
[23] K. J. Webb, A. Lager, and C. J. Black, “Comparison of high/low experimental,” Society of Petroleum Engineers Journal, vol. 18,
salinity water/oil relative permeability,” in Proceedings of the no. 6, pp. 445–456, 1978.
International Symposium of the Society of Core Analysts, Abu
Dhabi, UAE, October–November 2008.
[24] S. M. Rivet, Coreflooding oil displacements with low salinity brine
[M.S. thesis], University of Texas at Austin, 2009.
[25] I. Fjelde, S. V. Asen, and A. Omekeh, “Low salinity water
flooding experiments and interpretation by simulations,” in
Proceedings of the 18th SPE Improved Oil Recovery Symposium,
Paper SPE 154142, Tulsa, Okla, USA, April 2012.
[26] J. Zhang, W. Yan, and Q. Liu, “Field practice of production
with very high water cut under water flooding,” Petroleum
Exploration and Development, vol. 13, no. 1, pp. 50–54, 1986.
[27] Y. S. Wu and B. Bai, “Efficient simulation for low salinity water-
flooding in porous and fractured reservoirs,” in Proceedings of
the SPE Reservoir Simulation Symposium, SPE-118830-MS, The
Woodlands, Tex, USA, February 2009.
[28] C. T. Q. Dang, L. X. Nghiem, Z. J. Chen, and Q. P. Nguyen,
“Modeling low salinity waterflooding: ion exchange, geochem-
istry and wettability alteration,” in Proceedings of the SPE
Annual Technical Conference and Exhibition, SPE-166447-MS,
New Orleans, La, USA, September–October 2013.
[29] A. V. Ryazanov, K. S. Sorbie, and M. I. J. van Dijke, “Structure
of residual oil as a function of wettability using pore-network
modelling,” Advances in Water Resources, vol. 63, pp. 11–21, 2014.
[30] S. T. Taqvi, A. Almansoori, and G. Bassioni, “Modeling the
impact of wettability alterations on calcium carbonate system
for crude oil and asphaltenic solutions,” Industrial & Engineer-
ing Chemistry Research, vol. 53, no. 12, pp. 4773–4777, 2014.
[31] A. Skauge, S. Standal, S. O. Boe, and A. M. Blokhus, “Effects of
organic acids and bases, and oil composition on wettability,” in
Proceedings of the SPE Annual Technical Conference, 1999.
[32] C. E. Brown and E. L. Neustadter, “The wettability of oil/water/
silica systems with reference to oil recovery,” Journal of Cana-
dian Petroleum Technology, vol. 19, no. 3, pp. 100–110, 1980.
[33] P. Zhang, M. T. Tweheyo, and T. Austad, “Wettability alteration
and improved oil recovery by spontaneous imbibition of seawa-
ter into chalk: impact of the potential determining ions Ca2+ ,
Mg2+ , and SO2−4 ,” Colloids and Surfaces A: Physicochemical and
Engineering Aspects, vol. 301, no. 1–3, pp. 199–208, 2007.
[34] C. Maschio, A. C. Vidal, and D. J. Schiozer, “A framework to
integrate history matching and geostatistical modeling using
genetic algorithm and direct search methods,” Journal of
Petroleum Science and Engineering, vol. 63, no. 1–4, pp. 34–42,
2008.
[35] M. K. Zahoor and M. N. Derahman, “New approach for
improved history matching while incorporating wettability
variations in a sandstone reservoir—Field implementation,”
Hindawi Publishing Corporation
Journal of Chemistry
Volume 2015, Article ID 961256, 10 pages
http://dx.doi.org/10.1155/2015/961256

Research Article
Numerical Study of the Influence of Cavity on Immiscible Liquid
Transport in Varied-Wettability Fractures

Zhi Dou,1,2 Zhifang Zhou,2 Yefei Tan,3 and Yanzhang Zhou4


1
Ministry of Education Key Laboratory of Integrated Regulation and Resource Development on Shallow Lakes,
Hohai University, Nanjing 210098, China
2
School of Earth Science and Engineering, Hohai University, Nanjing 210098, China
3
Department of Hydraulic Engineering, Nanjing Hydraulic Research Institute, Nanjing 210029, China
4
Department of Geotechnical Engineering, Nanjing Hydraulic Research Institute, Nanjing 210029, China

Correspondence should be addressed to Zhi Dou; dz.uriah@gmail.com

Received 23 September 2014; Revised 22 December 2014; Accepted 22 December 2014

Academic Editor: Agus Sasmito

Copyright © 2015 Zhi Dou et al. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Field evidence indicates that cavities often occur in fractured rocks, especially in a Karst region. Once the immiscible liquid flows
into the cavity, the cavity has the immiscible liquid entrapped and results in a low recovery ratio. In this paper, the immiscible
liquid transport in cavity-fractures was simulated by Lattice Boltzmann Method (LBM). The interfacial and surface tensions were
incorporated by Multicomponent Shan-Chen (MCSC) model. Three various fracture positions were generated to investigate the
influence on the irreducible nonwetting phase saturation and displacement time. The influences of fracture aperture and wettability
on the immiscible liquid transport were discussed and analyzed. It was found that the cavity resulted in a long displacement
time. Increasing the fracture aperture with the corresponding decrease in displacement pressure led to the long displacement
time. This consequently decreased the irreducible nonwetting phase saturation. The fracture positions had a significant effect on
the displacement time and irreducible saturation. The distribution of the irreducible nonwetting phase was strongly dependent
on wettability and fracture position. Furthermore, this study demonstrated that the LBM was very effective in simulating the
immiscible two-phase flow in the cavity-fracture.

1. Introduction to investigate the effect of fluids properties on nonequilib-


rium capillarity effects and found that the immiscible two-
The problems of immiscible two-phase flows in fractures phase flow displacement mechanism is highly related to the
are frequently encountered in environmental protection viscosity ratio and capillary number. The research on the
and petroleum industries, such as enhanced oil recovery, influence of capillary number showed that the displacement
geological sequestration of carbon dioxide, and polluted time and irreducible saturation are strongly dependent on
groundwater remediation. Research on the influence of fluid capillary number [5]. Dou et al. [6] investigated the influence
properties on immiscible two-phase flow has been ongoing of wettability on interfacial area during the immiscible liquid
for decades and has shown that fluid properties (e.g., the invasion into a rough fracture and the results showed that
capillary number, bond number, viscosity ratio, saturation, the wettability also affects the displacement mechanism and
and wettability) are very important for immiscible two-phase leads to the variation of interfacial area. In fractures, once
flows [1–3]. However, in practice, the immiscible two-phase the displacement pressure (𝑃th = 2𝛾 cos 𝜃/𝑒, where 𝛾 is the
flow in fracture is not only controlled by fluid properties but interfacial tension, 𝜃 is the contact angle, and 𝑒 is the local
also influenced by fracture structures. aperture in fractures) overcomes the threshold pressure, one
Past research dealing with immiscible two-phase flow fluid phase can be displaced by another fluid phase. The dis-
has focused primarily on fluid properties. Joekar-Niasar and placement pressure varies with the aperture in fractures [7, 8].
Hassanizadeh [4] presented a dynamic pore-network model The aperture distributions in fractures and the influence on
2 Journal of Chemistry

(see Figure 1). Varied apertures in fracture always associate


with a rough fracture wall. An extreme aperture distribution
Inlet fracture Outlet fracture can be found in cavity-fracture. Due to the displacement pres-
Cavity sure depended on the local aperture, the outlet fracture may
play a significant role on displacement rate and irreducible
saturation. Cai et al. [14] proposed a generalized model of
spontaneous imbibition and the results indicated that the
variably shaped apertures have an important effect on the
spontaneous imbibition.
Modeling immiscible two-phase flow in fractures is a
frontier challenge for petroleum resource science. Work by
various researchers has presented different numerical meth-
Cavity
ods to couple fluid flow with other phenomenon. Kordilla
et al. [15] used the SPH model to simulate surface tension
Inlet fracture Outlet fracture
dominated flow on smooth fracture surfaces under various
flow conditions. Yang et al. [16] proposed an adaptive circle-
Symmetrical fracture fitting approach to calculate the in-plane interfacial curvature
during the drainage in fractures. Raeesi and Piri [17] used a
3D mixed-wet random pore-scale network model to investi-
gate the impacts of wettability and trapping on the capillary
Outlet fracture pressure-saturation-interfacial area relationship during two-
phase drainage and imbibition processes. However, it is still
Cavity difficult for those methods to consider the factors (e.g., inter-
Inlet fracture
facial tension, surface tension, wettability, and viscosity) that
influence the fluid flow. In recent years, the lattice Boltzmann
method (LBM) has attracted much attention because of its
Asymmetrical fracture simplicity and accuracy in handling complex immiscible two-
phase flow. LBM is capable of modeling the fluid/fluid and
Figure 1: Illustration of cavity-fracture. fluid/solid interactions. There are several available models in
LBM for the immiscible two-phase flow simulation, such as
the chromodynamics model, the pseudopotential model, free
flow have been a topic of considerable interest for many years, energy model, and He-Shan-Doolen model.
because an aperture distribution is essential to the studies of In this paper, the immiscible two-phase flow in cavity-
the mechanical and transport properties of rock fractures. fracture was simulated using lattice Boltzmann method.
Grasmueck et al. [9] used Ground-Penetrating Radar (GPR) The interfacial and surface tensions were incorporated by
in Karst region and found that there are some cavities in Multicomponent Shan-Chen model. Three various positions
vertical fractures. However, only a few researches have inves- of the inlet and outlet fractures (so-called “lower case,”
tigated the immiscible two-phase flow in cavity-fractures. “middle case,” and “upper case”) were generated to investigate
Zhou et al. [10] used smoothed particle hydrodynamics (SPH) the influence on irreducible non-wetting phase saturation
to study the oil displacement in cavity-fractures and found and displacement time. The effect of the aperture of the inlet
that the ultimate oil recovery is determined by the height and outlet fractures on the immiscible liquid transport was
of the fracture. A further study reported by Zhou et al. [11] discussed and analyzed in detail. In addition, by adjusting
investigated the influences of the height of initial two-phase the interfacial and surface tensions, the wettability of the
interface on the oil displacement. The lid driven cavity flow cavity-fracture wall varied from strongly water-wet to weakly
is commonly used as a benchmark problem in the field of water-wet and then the effect of the wettability was investi-
traditional computational fluid dynamics (CFD). Patel et al. gated.
[12] presented the two-sided cubic lid-driven cavity flow
with parallel lids moving in opposite direction at a large 2. Methodology
Reynolds number (Re = 1200). This study showed that the
dissipative nature of the flow inside the cavity increases with 2.1. LBM Fundamentals and Multicomponent Shan-Chen
increase in the size of corner eddies. Cai et al. [13] used a (MCSC) Model. The MCSC model is based on the basic
partial transfixion single fracture to investigate the influence lattice Boltzmann equations. In this study, two distribution
of cavity flow on miscible solute transport and they found that functions were used to represent wetting phase and non-
the porous media in cavity have a significant effect on flow wetting phase in cavity-fractures, respectively. The lattice
regime. Bhatnagar-Gross-Krook (LBGK) model was implemented for
The well-known fractured karst reservoir is composed of the collision term approximation in each evolution equation.
many cavity-fractures. Those cavities are much bigger than The LBGK model is most widely used model due to its
the fractures that connect with them. For the typical cavity- simplicity and accuracy in solving the collision term approx-
fracture, the cavity is settled between inlet and outlet fractures imation [18, 19]. The evolution equation of each distribution
Journal of Chemistry 3

function satisfies the basic lattice Boltzmann equation of a the typical MCSC model, the total force includes the fluid-
single relaxation time model, fluid cohesive force F𝜎𝑐 and the fluid-solid adhesive force F𝜎ads .
In (6), F𝜎 = F𝜎𝑐 +F𝜎ads . The initial values of particle distribution
𝑓𝑖𝜎 (X + e𝑖 Δ𝑡, 𝑡 + Δ𝑡) function of each component 𝜎 are usually assumed as
(1) calculated local equilibrium density distribution for the same
Δ𝑡 𝜎 𝜎(eq)
= 𝑓𝑖𝜎 (X, 𝑡) − [𝑓 (X, 𝑡) − 𝑓𝑖 (X, 𝑡)] , component based on initial macroscopic fluid density and
𝜏𝜎 𝑖 velocity.
where 𝑓𝑖𝜎 (X, 𝑡) is the fluid particle distribution function in the The fluid-fluid cohesive force is calculated in (8) based on
𝑖th velocity e𝑖 at (X, 𝑡) and 𝜎 denotes either the wetting phase the presence of the other fluid in neighboring lattice sites,
or nonwetting phase. 𝜏𝜎 is the dimensionless relaxation time
F𝜎𝑐 (X, 𝑡) = −𝐺𝑐 𝜌𝜎 (X, 𝑡) ∑ 𝜔𝑖 𝜌𝜎 (X + e𝑖 Δ𝑡, 𝑡) e𝑖 ,
which is related to the kinematic viscosity as 𝜐𝜎 = 𝑐𝑠2 (𝜏𝜎 − (8)
𝑖
0.5Δ𝑡). For the D2Q9 model, the nine discrete velocities are
given by where the 𝜎 and 𝜎 denote the wetting phase and the
nonwetting phase, respectively, and 𝐺𝑐 is a parameter that
[e0 , e1 , e2 , e3 , e4 , e5 , e6 , e7 , e8 ] controls the strength of fluid-fluid interaction.
(2) The fluid-solid adhesive force, F𝜎ads (X, 𝑡), is calculated in
0 1 0 −1 0 1 −1 −1 1
= 𝑐[ ]. (9) based on the presence of the solid phase at neighboring
0 0 1 0 −1 1 1 −1 −1
lattice sites,
𝜎(eq)
In (1), 𝑓𝑖 (X, 𝑡) is the local equilibrium distribution which F𝜎𝑐 (X, 𝑡) = −𝐺ads
𝜎
𝜌𝜎 (X, 𝑡) ∑ 𝜔𝑖 𝑠 (X + e𝑖 Δ𝑡, 𝑡) e𝑖 , (9)
is a specially discretized Maxwell-Boltzmann function with a 𝑖
Taylor expansion and defined as
where 𝑠(X + e𝑖 Δ𝑡, 𝑡) is an indicator function that is 1 and 0
𝜎
𝜎(eq) 2 for a solid and a fluid lattice sites, respectively, and 𝐺ads is a
e𝑖 ⋅ u𝜎(eq) (e𝑖 ⋅ u )
(X, 𝑡) = 𝜔𝑖 𝜌𝜎 [1 +
𝜎(eq) parameter that controls the strength between each fluid and a
𝑓𝑖 +
𝑐𝑠2 2𝑐𝑠4 wall. Due to the incorporating the fluid-fluid and fluid-solid
[
(3) interaction force, the actual whole fluid velocity V and the
2
(u𝜎(eq) ) whole fluid pressure 𝑃(𝑥) [20] are defined as
− ],
2𝑐𝑠2 ∑𝜎 F𝜎
] V=u+ ,
2 ∑𝜎 𝜌𝜎
where 𝑐𝑠 is the lattice sound speed and 𝑐𝑠 = 𝑐/√3 with 𝑐 = (10)
Δ𝑥/Δ𝑡 being the ratio of lattice spacing Δ𝑥 and time step Δ𝑡. 𝐺
𝑃 (𝑥) = 𝑐𝑠2 ⋅ ∑ 𝜌 (𝑥) + 𝑐 𝜌𝜎 (X) 𝜌𝜎 (X) .
In (3), the weights 𝜔𝑖 in the D2Q9 model are given by 𝜎 3

4 The fluid-fluid and fluid-solid interaction forces in MCSC


{
{ , 𝑖 = 0, model can be easily fitted to obtain the interfacial tension
{
{ 9
{
{1 and surface tension. The interfacial tension between two
𝜔𝑖 = { , 𝑖 = 1, 2, 3, 4, (4)
{
{ 9 immiscible fluids is determined according to Laplace’s Law
{
{ (Δ𝑝 = 𝛾/𝑅, where 𝛾 is the interfacial tension, 𝑅 is the mean
{1
, 𝑖 = 5, 6, 7, 8. interface curvature, and Δ𝑝 is a pressure difference between
{ 36
the inside and outside of bubbles or droplets). In MCSC
The macroscopic fluid density, macroscopic velocity, and the model, different interfacial tensions between two immiscible
velocity common to the various components are computed phases can be obtained by adjusting the coefficient 𝐺𝑐 [21].
from In this paper, the contact angle for a water droplet was
changed from 0 to 90 degrees, implying the solid surface
𝜌𝜎 = ∑ 𝑓𝑖𝜎 , (5) ranged from strongly water-wet to weakly water-wet. The
𝑖
solid surface was assumed to be hydrophilic. However, if
𝜏𝜎 F𝜎 the contact angle is greater than 90 degree, the solid surface
u𝜎(eq) = u + Δu = u + , (6) becomes hydrophobic.
𝜌𝜎
∑𝜎 (1/𝜏𝜎 ) ∑𝑖 𝑓𝑖𝜎 e𝑖 2.2. Boundary Condition. Many boundary conditions are
u= . (7)
∑𝜎 (1/𝜏𝜎 ) 𝜌𝜎 available in the literatures [22–24] for the MCSC model, such
as the nonslip boundary, the velocity and pressure boundary,
It should be noted that u is the common averaged velocity of and the convection-diffusion boundary. In the MCSC model,
all the fluid components without any additional force. In (6), any lattice site in the computational domain represents either
Δu is a change in velocity field due to the additional force. a solid site or a fluid site. For the solid sites, the half-way
In the MCSC model, the effect of total force is incorpo- bounce-back algorithm collision step is implemented instead
rated through adding acceleration into the velocity field. In of the collision step to provide a nonslip wall boundary
4 Journal of Chemistry

Y
Wetting phase b
a

Nonwetting phase (0, 0) X

Wetting phase

Flow direction

Figure 2: Schematic of immiscible two-phase flow in a 2D smooth fracture.

condition. In order to simulate the displacement process respectively. Figure 3 shows that the results of LB simulation
of the immiscible two-phase flow, the pressure and veloc- and analytical solution. The MCSC model simulation result is
ity conditions can be implemented. The pressure reservoir very consistent with the analytical solution.
boundary condition is applied at inlet and outlet fractures. Moreover, to test the grid independence for MCSC model,
The immiscible two-phase flow under this pressure boundary three different lattice grid resolutions were generated. For
condition can be assumed as a quasiequilibrium process, the three different lattice grid resolutions, the body force,
where the capillary force dominated the flow behavior. kinematic viscosity, contact angle, and boundary conditions
In this work, the algorithms for MCSC model were were constant. The initial region of the wetting phase was
developed as a C++ code. All the variables are in lattice units set as 𝑎 = 10 lu in the different lattice grid resolutions.
such that one lattice unit (Δ𝑥) is defined as 1 lu, and one Figure 3 shows the velocity distributions for the 200×101 lu2 ,
time step (Δ𝑡) is defined as 1 ts. The densities of both the 600 × 301 lu2 , and 1200 × 601 lu2 lattice grid resolutions. The
wetting and nonwetting phase were set to 1. The effect of the MCSC model simulation results are very consistent with the
gravitational field is neglected. analytical solution under the different lattice grid resolutions.
It means that the MCSC model solution is not gird sensitive.
2.3. Model Validation: Immiscible Two-Phase Flows in a 2D In this study, the sets of grid were chosen depending on the
Smooth Fracture. For the immiscible two-phase flows in capacity of computer. In Figure 3, it is found that although
fractures, the wetting-phase covers and moves along the the body force is constant, the maxium velocity varies with
fracture wall, while the nonwetting phase flows in the center the lattice grid resolution. This is because the fact that the
of fracture (see Figure 2). body force is applied in each lattice grid. As the lattice grid
To validate the MCSC model, the velocity distributions resolution increases, the resultant body force increases.
of the immiscible two-phase flow in a 2D smooth fracture
was investigated. A computational domain of 200 × 101 lu2 3. Results and Discussion
representing the 2D smooth fracture was used for model
validation of the MSCS model implementation. The period- 3.1. Effect of Inlet and Outlet Fractures. To investigate the
ical boundary condition was applied to the outlet and inlet effect of inlet and outlet fractures on the irreducible nonwet-
of the fracture. A constant body force 𝐺 = 8 × 10−8 was ting phase saturation and displacement time, the immiscible
added to each phase. The viscosity ratio has an effect on the liquid transport in cavity-fracture with various fracture posi-
velocity distribution of the nonwetting phase and wetting tions and apertures was simulated by the MCSC model. In
phase [21]. Thus, the kinematic viscosities of nonwetting this paper, the cavity was assumed to be a simple 100×100 lu2
phase and wetting phase were 1.667 lu2 /ts and 0.1667 lu2 /ts, square that previously was occupied by the nonwetting phase.
respectively. The contact angle between the nonwetting phase The inlet and outlet fractures had the smooth and paralleled
and wetting phase was set to 𝜃 = 45∘ . fracture walls and symmetrically distributed on the right and
Assuming the wetting phase flows along the fracture in left sides of the cavity, respectively. The cavity aperture was
the region 𝑎 < |𝑥| < 𝑏, the wetting phase and nonwet- set to 98 lu. For all the simulations, the maximum wetting
ting phase saturations are defined as 𝑆𝑤 = (𝑏−𝑎)/𝑏 and 𝑆nw = phase velocity was 𝑢 = 0.039 lu/ts by adjusting the pressure
𝑎/𝑏, respectively. The analytical solutions for velocity distri- boundary condition and the kinematic viscosities of both
bution can be derived by solving the Navier-Stokes equations: phases were 0.1667 lu2 /ts. Thus, the dimensionless Re number
was less than 23, indicating that there was a laminar flow in
Δ𝑝 (𝑏2 − 𝑥2 ) 𝐺 (𝑏2 − 𝑥2 ) the cavity-fracture [25].
𝑢 (𝑥) = = 𝑎 < |𝑥| < 𝑏, Firstly, the processes of the wetting phase displacing
2𝐿𝜇𝑤 2𝜐𝑤 𝜌𝑤
(11) the nonwetting phase (imbibition process) are simulated
𝐺 (𝑏2 − 𝑎2 ) 𝐺 (𝑎2 − 𝑥2 ) with three various positions of the symmetrical inlet and
𝑢 (𝑥) = + 0 < |𝑥| < 𝑎, outlet fractures. The so-called “lower case,” “middle case,”
2𝜐𝑤 𝜌𝑤 2𝜐nw 𝜌nw
and “upper case” for the three various fracture positions
where 𝜐𝑤 , 𝜐nw , 𝜌𝑤 , and 𝜌nw are the kinematic viscosities can be found in Figures 4 and 5, respectively. For the three
and densities of the wetting phase and nonwetting phase, various fracture positions, the apertures are equal to 5 lu.
Journal of Chemistry 5

1.2 × 10−4 1.4 × 10−3

1.0 × 10−4 1.2 × 10−3

1.0 × 10−3
8.0 × 10−5
8.0 × 10−4

Ux
6.0 × 10−5
Ux

6.0 × 10−4
−5
4.0 × 10
4.0 × 10−4
2.0 × 10−5
2.0 × 10−4

0.0 0.0
0 20 40 60 80 100 0 50 100 150 200 250 300
Lattice unit Lattice unit

MCSC LBM simulation MCSC LBM simulation


Analytical solution Analytical solution
(a) 200 × 101 lu2 grid resolution (b) 600 × 301 lu2 grid resolution

6.0 × 10−3

5.0 × 10−3

4.0 × 10−3

3.0 × 10−3
Ux

2.0 × 10−3

1.0 × 10−3

0.0
0 100 200 300 400 500 600
Lattice unit

MCSC LBM simulation


Analytical solution
(c) 1200 × 601 lu2 grid resolution

Figure 3: Comparison of MCSC model results and analytical solution for the velocity distribution at 𝑥 = 10 lu in the 2D smooth fracture
with three different lattice grid resolutions.

Furthermore, in order to investigate the effect of the fracture pressure for the inlet and outlet fracture, the wetting phase
aperture on immiscible liquid transport, three apertures for invades the cavity. However, the displacement pressure is
inlet and outlet fractures are set as 𝑒1 = 5 lu, 𝑒2 = 7 lu, strongly dependent on the contact angle and aperture. The
and 𝑒3 = 9 lu. The contact angle in this section is equal to effect of contact angle and aperture on the displacement will
12∘ by adjusting the interfacial and surface tensions, implying be investigated in the following section.
that the solid surface is weakly water-wet. If the nonwetting In Figure 4, the wetting phase invades into the cavity
saturation change over 5000 ts is less than 0.1%, it is assumed from the inlet fracture and there exists an apparent interface
that the process of the wetting phase invasion has finished between the nonwetting phase and wetting phase due to
and that the nonwetting phase saturation is equal to the the interfacial tension. In this case, as the capillary pressure
irreducible saturation. This computation takes about five increases with time, the interface between nonwetting phase
hours to run on the Intel CPU (3.2 GHz) computer for a and wetting phase moves to the outlet fracture. This process is
computational domain of 162 × 102 lu2 . similar to the imbibition process in porous media. Thus, the
As shown in Figure 4, the cavity is initially saturated by process of the wetting phase displacing the nonwetting phase
the wetting phase. Due to the pressure boundary condition is strongly dependent on the interfacial, surface tensions,
of the inlet and outlet fracture, the flow direction in cavity- and fracture structure. As more wetting phase was injected
fracture is from left to right. The capillary pressure increases into the inlet fracture, the nonwetting phase in the cavity
with time. Once the capillary pressure overcomes the entry decreased. It is found that the nonwetting phase still remains
6 Journal of Chemistry

(a) 𝑇 = 0 ts (b) 𝑇 = 80000 ts

(c) 𝑇 = 145000 ts

Figure 4: The displacement process in the cavity-fracture for the lower case (blue domain represents solid phase, red domain represents
nonwetting phase, and green domain represents wetting phase).

(a) Middle case, 𝑇 = 133000 ts (b) Upper case, 𝑇 = 126000 ts

Figure 5: The irreducible nonwetting phase distributions in cavity-fractures.

and becomes residual, when the displacement reaches the an important effect on the displacement time. Many studies
final steady-state. For this case, the irreducible nonwetting [10, 12] indicate that the cavity leads to the long displacement
phase saturation is equal to 𝑆ir = 0.49. From Figure 5, it can time. However, in Figure 6, it is found that both of the
be seen that the position of the inlet and outlet fracture has aperture and fracture position have an important effect on the
an effect on the irreducible nonwetting phase saturation. The displacement time. For each case, as the aperture increases,
irreducible nonwetting phase saturations for the middle and the displacement time becomes shorter. This is because the
upper case are 𝑆ir = 0.43 and 𝑆ir = 0.38, respectively. Thus, the fact that the capillary pressure increases with time for the
position of the inlet and outlet fracture becomes close to the pressure boundary condition on the inlet and outlet fracture.
middle of the cavity, and the irreducible saturations decreases. Depending on the definition of the displacement pressure,
The effect of aperture on the irreducible nonwetting phase a larger aperture results in a smaller displacement pressure.
saturation is investigated. Other two apertures (7 lu and 9 lu) This consequently decreases the displacement time. On the
for the fractures are employed, respectively. Figure 6 shows other hand, although the domain of the cavity for all the
the distributions of the irreducible nonwetting phase for the simulations is constant, the ratio of the aperture to cavity
various positions of the inlet and outlet fractures. It can be decreases due to the decreased aperture. Thus, the process
seen that the aperture has little effect on the distributions of the wetting phase displacing the nonwetting phase takes
of the irreducible nonwetting phase for the same position less time because of the relative large volume capacity of the
of the inlet and outlet fractures. However, the aperture has cavity. For the upper case, the position of the inlet and outlet
Journal of Chemistry 7

Lower case Middle case Upper case

(a) 𝑒 = 7 lu, 𝑇 = 113000 ts (b) 𝑒 = 7 lu, 𝑇 = 107000 ts (c) 𝑒 = 7 lu, 𝑇 = 99300 ts

(d) 𝑒 = 9 lu, 𝑇 = 112000 ts (e) 𝑒 = 9 lu, 𝑇 = 98000 ts (f) 𝑒 = 9 lu, 𝑇 = 94000 ts

Figure 6: The distributions of the irreducible nonwetting phase for the various fracture positions and apertures with weakly water-wet fracture
wall.

fracture is far away from the nonslip boundary of the cavity. 0.50
As a result, the fracture position moves far away from the
Irreducible nonwetting phase saturation

0.48
middle of the cavity, and the displacement time increases. 0.46
Furthermore, in Figure 6, due to the smaller displacement
pressure, it can be seen that the small nonwetting phase 0.44
droplet remains at the lower right corner of the cavity. This 0.42
also increases the irreducible nonwetting phase saturation. 0.40
However, as the fracture position moves to the bottom of
0.38
the cavity, the small nonwetting phase droplet disappears.
Comparing Figures 5 and 6, it is found that the aperture 0.36
effects the occurrence of the small nonwetting phase droplet. 0.34
In Figure 5, the nonwetting phase is continuous for three 0.32
various fracture positions. As the aperture increases more
0.30
than 5 lu, the nonwetting phases in the middle and upper 4 5 6 7 8 9 10
cases are discontinuous. The small nonwetting phase droplets Aperture (lu)
are captured in the lower right corner of the cavity. This
is mainly due to the increased aperture and varied fracture Upper case Lower case
position, and the displacement pressure decreases. Middle case
Figure 7 shows that both of aperture and fracture position
Figure 7: The relationship between the irreducible nonwetting
have an important effect on the irreducible non-wetting phase phase saturation and aperture for three various cases.
saturation. From Figure 7, it can be seen that as the fracture
position becomes close to the middle position of the cavity,
the irreducible nonwetting phase saturation decreases. This
result is consistent with the experimental study reported by Figure 8 shows the distributions of the irreducible non-
Zhou et al. [10]. wetting phase in various cavity-fractures with the strongly
water-wet walls. For the three various case, there are different
3.2. Effect of Wettability. The wettability plays an important apertures and fracture positions. Comparing to the cases with
role on the immiscible two-phase flow and is investigated the weakly water-wet fracture walls (Figures 4 and 5), it is
in the following section. By adjusting the surface tension found that the irreducible nonwetting phase droplets tend
and keeping the interfacial tension constant, the solid surface to be circles. This is due to the increased surface tension.
becomes strongly water-wet (the contact angle is equal to Thus, the nonwetting phase tends to move away from the
85∘ ). The domain of the cavity is the same as that in Section strongly water-wet fractures wall and the wetting phase is
3.1. more easier to wet the fracture wall. For the three various
8 Journal of Chemistry

Lower case Middle case Upper case

(a) 𝑇 = 182400 ts, 𝑆ir = 0.34, 𝑒 = 5 lu (b) 𝑇 = 163800 ts, 𝑆ir = 0.28, 𝑒 = 5 lu (c) 𝑇 = 156200 ts, 𝑆ir = 0.19, 𝑒 = 5 lu

(d) 𝑇 = 140000 ts, 𝑆ir = 0.65, 𝑒 = 7 lu (e) 𝑇 = 129400 ts, 𝑆ir = 0.21, 𝑒 = 7 lu (f) 𝑇 = 112400 ts, 𝑆ir = 0.15, 𝑒 = 7 lu

(g) 𝑇 = 132400 ts, 𝑆ir = 0.61, 𝑒 = 9 lu (h) 𝑇 = 119000 ts, 𝑆ir = 0.48, 𝑒 = 9 lu (i) 𝑇 = 102400 ts, 𝑆ir = 0.14, 𝑒 = 9 lu

Figure 8: The distributions of the irreducible nonwetting phase for the various fracture positions and apertures with strongly water-wet
fracture wall.

cases in Figure 8, the displacement time decreases as the high. Furthermore, due to the increased contact angle and
aperture increases and the fracture position becomes close the variation of aperture from cavity to fracture, the increased
to the middle of the cavity, which is similar to the result capillary pressure changed the displacement mechanism. The
with the weakly water-wet fracture walls. In Figure 8, it is outlet fracture is important for the displacement process.
found that when the fracture wall is strongly water wet, The increased capillary pressure is actually caused by the
the irreducible nonwetting phase saturation decreases as the decreased nonwetting phase pressure and the increased
fracture position becomes close to the middle of the cavity. wetting phase pressure. Once the wetting phase occupied the
This is also similar to the result with the weakly water wet outlet fracture, the nonwetting phase in the cavity will remain
fracture walls. For the lower case in Figure 8, the irreducible as the irreducible nonwetting phase. However, this process is
nonwetting phase saturation increases sharply when the strongly dependent on the wettability and fracture aperture.
fracture aperture is more than 5 lu. This result is different
from the lower case in Figures 5 and 6. The increased contact 4. Conclusions
angle leads to not only the increased displacement pressure
but also the weak cohesive force between the nonwetting In this paper, a numerical model based on the lattice Boltz-
phase and fracture wall. The weak cohesive force however mann method has been developed to study the immiscible
results in the increment of the irreducible nonwetting phase. two-phase flow in the cavity-fracture. The effects of fracture
In fact, the weak cohesive force is in favor of the movement of apertures and positions on the irreducible nonwetting phase
the nonwetting phase. Thus, the irreducible nonwetting phase saturation and displacement time were discussed in detail. By
saturations with the strongly water-wet fracture walls are adjusting the interfacial and surface tensions, the immiscible
generally lower than that with the weakly water-wet fracture two-phase flow in the cavity-fracture with the strongly water-
walls. For the lower case, the fracture position is close to wet and the weakly water-wet fracture walls was simulated.
the bottom of the cavity, where the flow velocity is relatively This study demonstrated that the LBM was very effective
Journal of Chemistry 9

in simulating the immiscible two-phase flow in the cavity- Acknowledgments


fracture.
Although the computational domain of cavity was con- The study is financially supported by the National Natural
stant, the cavity resulted in a long displacement time. For the Science Foundation of China (Grant nos. 41402217, 41202177,
pressure boundary condition on the inlet and outlet fracture, and 51109139) and the Natural Science Foundation of Jiangsu
the capillary pressure in cavity-fractures increased with time. (Grant nos. BK2011110 and BK2012814). The authors thank the
The displacement pressure decreased as the fracture aperture two anonymous reviewers for their suggestions in improving
increased. This consequently decreased the irreducible non- the paper.
wetting phase saturation. The fracture positions had a signif-
icant effect on the displacement time and irreducible satura- References
tion. The distribution of the irreducible nonwetting phase was
strongly dependent on wettability and fracture position. [1] J. Cai, X. Hu, D. C. Standnes, and L. You, “An analytical
model for spontaneous imbibition in fractal porous media
including gravity,” Colloids and Surfaces A: Physicochemical and
Nomenclature Engineering Aspects, vol. 414, pp. 228–233, 2012.
𝑐𝑠 : Lattice sound speed [2] Y.-L. Wang, L. Ma, B.-J. Bai, G.-C. Jiang, J.-F. Jin, and Z.-B.
𝑒𝑖 : 𝑖th velocity Wang, “Wettability alteration of sandstone by chemical treat-
ments,” Journal of Chemistry, vol. 2013, Article ID 845031, 8
𝑒: Local aperture in fractures
pages, 2013.
𝑓𝑖𝜎 (X, 𝑡): Fluid particle distribution function
𝜎(eq)
[3] I. Cichowska-Kopczyńska, M. Joskowska, B. Dębski, J. Łuczak,
𝑓𝑖 (X, 𝑡): Local equilibrium distribution function and R. Aranowski, “Influence of ionic liquid structure on
F𝜎𝑐 : Fluid-fluid cohesive force supported ionic liquid membranes effectiveness in carbon
F𝜎ads : Fluid-solid adhesive force dioxide/methane separation,” Journal of Chemistry, vol. 2013,
Article ID 980689, 10 pages, 2013.
𝐺𝑐 : Parameter that controls the strength of
fluid-fluid interaction [4] V. Joekar-Niasar and S. M. Hassanizadeh, “Effect of fluids prop-
𝜎 erties on non-equilibrium capillarity effects: dynamic pore-
𝐺ads : Parameter that controls the strength network modeling,” International Journal of Multiphase Flow,
between each fluid and a wall vol. 37, no. 2, pp. 198–214, 2011.
𝐺: Body force [5] Z. Dou and Z.-F. Zhou, “Numerical study of non-uniqueness of
𝜃: Contact angle the factors influencing relative permeability in heterogeneous
Δ𝑝: Pressure difference between the inside and porous media by lattice Boltzmann method,” International
outside of bubbles or droplets Journal of Heat and Fluid Flow, vol. 42, pp. 23–32, 2013.
𝑃th : Displacement pressure [6] Z. Dou, Z. Zhou, and B. E. Sleep, “Influence of wettability on
𝑃(𝑥): Whole fluid pressure interfacial area during immiscible liquid invasion into a 3D self-
affine rough fracture: lattice Boltzmann simulations,” Advances
𝜌𝜎 : Macroscopic fluid density
in Water Resources, vol. 61, pp. 1–11, 2013.
𝜌𝑤 : Wetting phase density
[7] J. Cai and B. Yu, “A discussion of the effect of tortuosity on
𝜌nw : Nonwetting phase density the capillary imbibition in porous media,” Transport in Porous
𝛾: Interfacial tension Media, vol. 89, no. 2, pp. 251–263, 2011.
𝑠(X + e𝑖 Δ𝑡, 𝑡): Indicator function [8] J. Cai, B. Yu, M. Zou, and L. Luo, “Fractal characterization of
𝑆𝑤 : Wetting phase saturation spontaneous co-current imbibition in porous media,” Energy &
𝑆nw : Nonwetting phase saturation Fuels, vol. 24, no. 3, pp. 1860–1867, 2010.
Δ𝑡: Time step [9] M. Grasmueck, M. C. Quintà, K. Pomar, and G. P. Eberli,
𝜏𝜎 : Dimensionless relaxation time “Diffraction imaging of sub-vertical fractures and karst with
full-resolution 3D Ground-Penetrating Radar,” Geophysical
u: Common averaged velocity of all the fluid Prospecting, vol. 61, no. 5, pp. 907–918, 2013.
components
[10] G. Zhou, Z. Chen, W. Ge, and J. Li, “SPH simulation of oil dis-
Δu: Change in velocity field placement in cavity-fracture structures,” Chemical Engineering
u𝜎(eq) : Macroscopic velocity Science, vol. 65, no. 11, pp. 3363–3371, 2010.
V: Actual whole fluid velocity [11] G. Zhou, W. Ge, B. Li et al., “SPH simulation of selective
𝜐nw : Nonwetting phase kinematic viscosity withdrawal from microcavity,” Microfluidics and Nanofluidics,
𝜐𝑤 : Wetting phase kinematic viscosity vol. 15, no. 4, pp. 481–490, 2013.
𝜐𝜎 : Kinematic viscosity [12] D. K. Patel, M. K. Das, and S. Roy, “LES of turbulent flow
in a cubical cavity with two parallel lids moving in opposite
𝜔𝑖 : Weights direction,” International Journal of Heat and Mass Transfer, vol.
Δ𝑥: Lattice spacing. 72, pp. 37–49, 2014.
[13] J.-L. Cai, Z.-F. Zhou, and Y. Huang, “Laboratory experiments on
Conflict of Interests solute transport in a partial transfixion single fracture,” Journal
of Hydrodynamics Series B, vol. 23, no. 5, pp. 570–579, 2011.
The authors declare that there is no conflict of interests [14] J. Cai, E. Perfect, C.-L. Cheng, and X. Hu, “Generalized mod-
regarding the publication of this paper. eling of spontaneous imbibition based on hagen-poiseuille flow
10 Journal of Chemistry

in tortuous capillaries with variably shaped apertures,” Lan-


gmuir, vol. 30, no. 18, pp. 5142–5151, 2014.
[15] J. Kordilla, A. M. Tartakovsky, and T. Geyer, “A smoothed parti-
cle hydrodynamics model for droplet and film flow on smooth
and rough fracture surfaces,” Advances in Water Resources, vol.
59, pp. 1–14, 2013.
[16] Z. Yang, A. Niemi, F. Fagerlund, and T. Illangasekare, “A gen-
eralized approach for estimation of in-plane curvature in inva-
sion percolation models for drainage in fractures,” Water
Resources Research, vol. 48, no. 9, Article ID W09507, 2012.
[17] B. Raeesi and M. Piri, “The effects of wettability and trapping
on relationships between interfacial area, capillary pressure and
saturation in porous media: a pore-scale network modeling
approach,” Journal of Hydrology, vol. 376, no. 3-4, pp. 337–352,
2009.
[18] H. Huang, J.-J. Huang, and X.-Y. Lu, “Study of immiscible dis-
placements in porous media using a color-gradient-based
multiphase lattice Boltzmann method,” Computers & Fluids, vol.
93, pp. 164–172, 2014.
[19] Q. Chen and X. B. Zhang, “A novel less dissipation finite-dif-
ference lattice boltzmann scheme for compressible flows,” Inter-
national Journal of Modern Physics C, vol. 23, no. 11, Article ID
1250074, 2012.
[20] H. Huang, D. T. Thorne, M. G. Schaap, and M. C. Sukop, “Pro-
posed approximation for contact angles in Shan-and-Chen-
type multicomponent multiphase lattice Boltzmann models,”
Physical Review E—Statistical, Nonlinear, and Soft Matter
Physics, vol. 76, no. 6, Article ID 066701, 2007.
[21] A. Ghassemi and A. Pak, “Numerical study of factors influ-
encing relative permeabilities of two immiscible fluids flowing
through porous media using lattice Boltzmann method,” Jour-
nal of Petroleum Science and Engineering, vol. 77, no. 1, pp. 135–
145, 2011.
[22] T. Zhang, B. Shi, Z. Guo, Z. Chai, and J. Lu, “General bounce-
back scheme for concentration boundary condition in the
lattice-Boltzmann method,” Physical Review E, vol. 85, Article
ID 016701, 2012.
[23] Q. Chen, X. Zhang, and J. Zhang, “Improved treatments for
general boundary conditions in the lattice Boltzmann method
for convection-diffusion and heat transfer processes,” Physical
Review E: Statistical, Nonlinear, and Soft Matter Physics, vol. 88,
no. 3, Article ID 033304, 2013.
[24] Z. Dou and Z.-F. Zhou, “Lattice Boltzmann simulation of
solute transport in a single rough fracture,” Water Science and
Engineering, vol. 7, pp. 277–287, 2014.
[25] H. Huang and X.-Y. Lu, “Relative permeabilities and coupling
effects in steady-state gas-liquid flow in porous media: a lattice
Boltzmann study,” Physics of Fluids, vol. 21, no. 9, Article ID
092104, 2009.
Hindawi Publishing Corporation
Journal of Chemistry
Volume 2015, Article ID 706312, 10 pages
http://dx.doi.org/10.1155/2015/706312

Research Article
Study on Skin Factor and Productivity of Horizontal Well after
Acidizing with Nonuniform Damage

Lin Yuan, Xiao-Ping Li, Xiao-Hua Tan, and Lie-Hui Zhang


State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum University, Chengdu 610500, China

Correspondence should be addressed to Xiao-Hua Tan; xiaohua-tan@163.com

Received 16 October 2014; Revised 3 December 2014; Accepted 5 December 2014

Academic Editor: Agus Sasmito

Copyright © 2015 Lin Yuan et al. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Horizontal well (HW) was divided into several elements after acidizing. In each element, there would exist a composite zone that
was made up of damage zone (DZ) and acidizing zone (AZ). Local skin factor after acidizing was used to describe the resistance
in each composite zone. The models for local skin factor and productivity of HW were created using the method of equivalent
filtration resistance and displacement between two similar flow modes. According to the solution of the models, type curves of
skin factor and production-increasing ratio after acidizing were illustrated, and the effect of different parameters in DZ and AZ on
distribution of skin factor and production-increasing ratio after acidizing were discussed. The present model has a significant guide
on the practice of acidizing technology.

1. Introduction the invasion depth of drilling fluid and completion fluid,


which also simultaneously considered the skin factor, flow
Numerous researchers have modeled the skin factor and efficiency of the well, and production loss in order to predict
productivity of horizontal well (HW) by different methods the extent of damage to the formation.
and, furthermore, the flow in the wellbore has also been Cai et al. [8, 9] proposed analytical fractal model to
considered. analyze the depth of extraneous fluid invasion, where the
Frick and Economides [1] presented a mathematical tortuosity of capillaries and wettability effect [10] are taken
model of skin factor for a HW. In their model, the anisotropy into account. Their models show that the impregnation depth
of the permeability and the shape of a truncated elliptical cone follows a scaling law of time where the exponent is a function
in damage zone (DZ), with the larger DZ near the vertical of fractal tortuosity dimension rather than the classical
section of HW, were taken into account. constant 0.5. Except for the methods that are mentioned
For anisotropic medium, Permadi et al. [2–5] pointed that above, the fractal theory could also be used in the study on
the shape of damage region was circular near the well and skin factor for HW [11–13].
elliptical far from the well and presented a new analytical Yildiz et al. [14, 15] considered the problem of steady
model for damage skin factor and the resulting reservoir filtration flow of an incompressible single-phase fluid near the
inflow for a horizontal well, which can be more accurate than HW with varying permeability in DZ. In this model, the effect
previous model for skin factor and productivity of HW. of damage ratio, the size and shape of DZ, and the geometric
Parn-anurak and Engler [6] put forward a new model to parameters of the permeability profile on productivity of HW
describe the distribution of the skin factor along the wellbore were investigated. Based on the research, the effect of varying
of a HW, in which the DZ had a cylindrical-conical shape; permeability in the DZ could also be taken into account on
meanwhile, the penetration rate, the length of HW, and the skin factor and productivity of HW [16].
anisotropy ratio were taken into account. Taking into account a negative skin, Nie et al. [17] consid-
Ye et al. [7] proposed a method for evaluating formation ered a set of complex boundary conditions and built a com-
damage from a drilled HW. Their model could calculate prehensive semianalytical model for a HW in homogeneous
2 Journal of Chemistry

z z
z z

= − +
y y y y

Composite area near the (I) Original damage zone (II) Damage zone with the (III) Acidizing zone
wellbore after acidizing same size as acidizing zone

Figure 1: Division of elliptic composite zone.

or dual- or triple-porosity reservoir, which performed well section, but due to the nonuniform distribution of radius of
against the real field data. DZ and acidizing zone (AZ), we should use an infinitesimal
An and Wu [18] built a semianalytical model for predict- method to get the local skin factor of HW after acidizing,
ing production rate of complex wells in anisotropic reservoirs eventually obtaining the total skin factor.
considering damage to the formation due to the invasion
of drilling and well completion, which were expressed by 2.1.1. Model for Local Skin Factor. In order to study the local
a skin factor. Their model could fully take into account skin factor of HW, firstly we should select an infinitesimal
the characteristics of formation damage and its impact on section at any location along the horizontal section, as is
productivity of HW. shown in Figure 1, so it can be considered as a point sink in
The methods mentioned above had discussed the estab- elliptic composite zone, but there is no accurate method to
lishment, solution, and applications in actual production solve this problem so far, so we can divide the elliptic com-
process of models of skin factor for HW in detail; however, posite zone into combination of three homogeneous zones
the models can be just for the HW with no stimulation and then solve the filtration problem in every homogeneous
treatment, while in most cases, a suitable acid would be used zone, respectively, and finally use the superposition theorem
to remove the formation damage [19, 20], so a new model of to solve the filtration problem in composite zone.
skin factor for HW after acidizing will be more suitable to Section I in Figure 1 is the original DZ near the wellbore;
actual production conditions. In this paper, DZ distributing in this section, we can consider it as a vertical well in the
nonuniformly along the wellbore after being treated by acid elliptic supply boundary, so its seepage differential equation
is simulated using a series of elliptic composite areas, which and boundary conditions can be described as follows, respec-
can be the most appropriate approach for actual shape of tively:
DZ after acidizing. Meanwhile, the flows in porous media
𝜕2 Φ (𝑥) 𝜕2 Φ (𝑥)
and wellbore are coupled to calculate the productivity of HW + = 0;
after acidizing. A comprehensive model is then presented by 𝜕2 𝑦 𝜕2 𝑧
introducing the nonuniform DZ to calculate the skin factor 𝑦2 𝑧2
2
+ 2 = 1, Φ (𝑥) = Φ𝑑 (𝑥) ; (1)
and productivity of HW after acidizing. Factors influencing 𝑑ℎ (𝑥) 𝑑V (𝑥)
the distribution of skin factor and productivity of HW after 𝜕Φ (𝑥)
acidizing are analyzed. The model can provide a new solution ∮ 𝑑𝑠 = 𝑑𝑞 (𝑥) ,
𝜕𝑛
to calculate the skin factor and productivity of HW after
acidizing in empirical application. where Φ𝑑 is potential in the DZ’s boundary, 𝑑ℎ is horizontal
radius of DZ, 𝑑V is vertical radius of DZ, and 𝑞 is the
2. Coupling Model between production of HW.
The solution of the problem as (1) shows is very difficult;
Reservoir and Wellbore there is no analytical solution so far, but approximate solution
2.1. Flow Model in the Reservoir. When the HW is treated can be used, namely, the method of displacement between
by acidizing, the DZ near the wellbore will be improved, two similar flow modes [22]; that is to say, we can consider the
which leads to the decrease of skin factor, while the zone filtration problem of a point sink in elliptic supply boundary
as that between two paralleled isopiestic lines, which have a
far from the wellbore is the same as that before acidizing, so
distance of 2𝐻(𝑥), as Figure 2 illustrates.
the study on HW after acidizing will be focused on the flow
Its seepage differential equation and boundary conditions
characteristic near the wellbore. In anisotropic reservoir, due
can be turned into
to the permeability difference between horizontal plane and
vertical plane, at any location along the horizontal section, 𝜕2 Φ (𝑥) 𝜕2 Φ (𝑥)
+ = 0;
near the wellbore, there exists an elliptic DZ; moreover, when 𝜕2 𝑦 𝜕2 𝑧
the HW is treated by acidizing, the isopiestic lines near the 𝑧 = ±𝐻 (𝑥) , Φ (𝑥) = Φ1 (𝑥) ; (2)
wellbore are still a cluster of concentric elliptical arcs [21]; that
is to say, after acidizing, near the horizontal wellbore, there 𝜕Φ (𝑥)
∮ 𝑑𝑠 = 𝑑𝑞 (𝑥) ,
will be a composite zone at any location along the horizontal 𝜕𝑛
Journal of Chemistry 3

Elliptic boundary Constant pressure boundary z H(x)

y
Point source

Constant pressure boundary −H(x)

Figure 2: Method of displacement between two similar flow modes.

where Φ1 is potential in constant pressure boundary and 10


2𝐻(𝑥) is the distance between two constant pressure lines.

H(x)/d (x)
8
Using the potential superposition principles, the analyti- 6
cal solution of (2) can be got as follows [23]: 4
2
0
0.92𝑑𝑞 (𝑥) ch (𝜋𝑦/2𝐻(𝑥)) − cos (𝜋𝑧/2𝐻(𝑥)) 0 1 2 3 4 5 6 7 8
Φ(𝑥) = ln
𝑑𝑥 ch (𝜋𝑦/2𝐻(𝑥)) + cos (𝜋𝑧/2𝐻(𝑥)) dh (x)/d (x)

Figure 3: Relationship curve between 𝐻(𝑥)/𝑑V (𝑥) and 𝑑ℎ (𝑥)/𝑑V (𝑥).


+ Φ1 (𝑥) .
(3)
Coupling (5) and (6),
Equation (3) is deduced based on the SI system of basic
unit in oil reservoir; we can improve it into SI system of mine 𝑝12 (𝑥) = 𝑝𝑑2 (𝑥)
unit in gas reservoir, so (4) shows the distribution of potential
near a point sink in the elliptic supply boundary: 1.291 × 10−3 𝑑𝑞𝑠𝑐 (𝑥) ⋅ 𝑇𝜇𝑔 𝑍 𝜋𝑑V (𝑥)
− ln(tan )
𝐾𝑑V ⋅ 𝑑𝑥 4𝐻 (𝑥)
2
1.291 × 10−3 𝑑𝑞𝑠𝑐 (𝑥) ⋅ 𝑇𝜇𝑔 𝑍 (7)
𝑝 (𝑥) =
2𝐾𝑑V ⋅ 𝑑𝑥
𝜋𝑑V (𝑥) 𝜋𝑑 (𝑥)
tan = th ℎ . (8)
4𝐻 (𝑥) 4𝐻 (𝑥)
ch (𝜋𝑦/2𝐻(𝑥)) − cos (𝜋𝑧/2𝐻(𝑥))
× ln + 𝑝12 (𝑥),
ch (𝜋𝑦/2𝐻(𝑥)) + cos (𝜋𝑧/2𝐻(𝑥)) Due to the reason that, through (8), we cannot get the
(4) analytical solution of 𝐻(𝑥), we can study its change rule by
graphing method; Figure 3 is the relationship curve between
where 𝑞𝑠𝑐 is the production of horizontal gas well, 𝑇 is the 𝐻(𝑥)/𝑑V (𝑥) and 𝑑ℎ (𝑥)/𝑑V (𝑥).
gas reservoir temperature, 𝜇𝑔 is the gas viscosity, 𝑍 is the gas From Figure 3, it is interesting to find that when the value
deviation factor, and 𝐾𝑑V is the vertical permeability of DZ. of 𝑑ℎ (𝑥)/𝑑V (𝑥) is more than 1.5, the value of 𝐻(𝑥)/𝑑V (𝑥) will
Because points [𝑑ℎ (𝑥), 0] and [0, 𝑑V (𝑥)] are in the same tend to 1, due to the effect of permeability anisotropy; the
isopotential line, substituting the plane of the two points into value of 𝑑ℎ (𝑥)/𝑑V (𝑥) is always more than 1.5, so in this paper,
(4), respectively, we can make an assumption that 𝐻(𝑥) is similar to 𝑑V (𝑥).
Substituting (7) into (4),
1.291 × 10−3 𝑑𝑞𝑠𝑐 (𝑥) 𝑇𝜇𝑔 𝑍
𝑝𝑑2 (𝑥) = 1.291 × 10−3 𝑑𝑞𝑠𝑐 (𝑥) ⋅ 𝑇𝜇𝑔 𝑍
2𝐾𝑑V ⋅ 𝑑𝑥 𝑝2 (𝑥) = 𝑝𝑑2 (𝑥) +
4𝐾𝑑V ⋅ 𝑑𝑥
(5) (9)
ch(𝜋𝑑ℎ (𝑥)/2𝐻(𝑥)) − 1 ch (𝜋𝑦/2𝐻(𝑥)) − cos (𝜋𝑧/2𝐻(𝑥))
× ln + 𝑝12 (𝑥), × ln .
ch(𝜋𝑑ℎ (𝑥)/2𝐻(𝑥)) + 1 ch (𝜋𝑦/2𝐻(𝑥)) + cos (𝜋𝑧/2𝐻(𝑥))

Because the isopiestic line near the wellbore is circular, we


1.291 × 10−3 𝑑𝑞𝑠𝑐 (𝑥) 𝑇𝜇𝑔 𝑍
𝑝𝑑2 (𝑥) = can take any point (0, 𝑟𝑤 ). At this point, the pressure is 𝑝𝑤𝑓 (𝑥):
2𝐾𝑑V ⋅ 𝑑𝑥
(6)
2
1.291 × 10−3 𝑑𝑞𝑠𝑐 (𝑥) ⋅ 𝑇𝜇𝑔 𝑍
1 − cos (𝜋𝑑V (𝑥)/2𝐻(𝑥)) 𝑝𝑤𝑓 (𝑥) = 𝑝𝑑2 (𝑥) +
× ln + 𝑝12 (𝑥), 𝐾𝑑V ⋅ 𝑑𝑥
1 + cos (𝜋𝑑V (𝑥)/2𝐻(𝑥)) (10)
𝜋𝑟𝑤
× ln (tan ),
where 𝑝𝑑 is the pressure of DZ’s boundary. 4𝐻 (𝑥)
4 Journal of Chemistry

where 𝑝𝑤𝑓 (𝑥) is pressure at any location along the wellbore dv.max
and 𝑟𝑤 is the wellbore radius. dv.min
av.max
Equation (10) can be transformed as
av.min
Δ𝑝𝑑2 (𝑥) = 𝑝𝑑2 (𝑥) − 𝑝𝑤𝑓
2
(𝑥)

1.291 × 10−3 𝑑𝑞𝑠𝑐 (𝑥) ⋅ 𝑇𝜇𝑔 𝑍 𝜋𝑟𝑤


= ln (cot ).
𝐾𝑑V ⋅ 𝑑𝑥 4𝐻 (𝑥)
(11)

According to the formula of vertical well’s skin factor [23], Figure 4: Distribution of DZ and AZ along the wellbore.

1.291 × 10−3 ⋅ 𝑞𝑠𝑐 𝑇𝜇𝑔 𝑍


Δ𝑝𝑠2 = 𝑆, (12)
𝐾ℎ ℎ So the formula of radius of DZ, 𝑑V (𝑥), and AZ, 𝑎V (𝑥), can
be written:
where 𝐾ℎ is the permeability of formation, ℎ is the gas
𝑑V. max − 𝑑V. min
reservoir thickness, and 𝑆 is the skin factor of vertical well. 𝑑V (𝑥) = (𝐿 − 𝑥) + 𝑟𝑤 + 𝑑V. min ,
We can make an assumption that the anisotropy before 𝐿
(16)
and after acidizing is constant; namely, 𝛽 = (𝐾ℎ/𝐾V )1/2 = 𝑎 − 𝑎V. min
𝑎V (𝑥) = V. max (𝐿 − 𝑥) + 𝑟𝑤 + 𝑎V. min ,
(𝐾𝑑ℎ /𝐾𝑑V )1/2 = (𝐾𝑎ℎ /𝐾𝑎V )1/2 ; comparing (11) with (12), the 𝐿
skin factor of section I can be written as follows: where 𝑑V. max is the vertical maximum distance of mud
𝐾ℎ 𝜋𝑟𝑤 immersion, 𝑑V. min is the vertical minimum distance of mud
𝑆1 (𝑥) = ( − 1) ln [cot ]. (13) immersion, 𝑎V. max is the vertical maximum distance of acid-
𝐾𝑑ℎ 4𝑑V (𝑥)
rock reaction, 𝑎V. min is the vertical minimum distance of acid-
Identical method can be used to get the skin factor of rock reaction, and 𝐿 is the horizontal wellbore length.
sections II and III; namely, The deduction above is based on the factor that the
distance of acid-rock reaction is less than the distance of mud
𝐾ℎ 𝜋𝑟𝑤 immersion; if the distance of acid-rock reaction is more than
𝑆2 (𝑥) = ( − 1) ln [cot ], the distance of mud immersion, we can use the same method
𝐾𝑑ℎ 4𝑎V (𝑥)
(14) to calculate the skin factor, which has the same formula as
𝐾 𝜋𝑟𝑤 (15), only 𝑎V (𝑥) is more than 𝑑V (𝑥), so there are no details here.
𝑆3 (𝑥) = ( ℎ − 1) ln [cot ],
𝐾𝑎ℎ 4𝑎V (𝑥)
2.1.2. Calculation of Total Skin Factor. In the 𝑦-𝑧 plane at
where 𝐾𝑑ℎ is the horizontal permeability of DZ, 𝐾𝑎ℎ is the any location of HW, we select an infinitesimal section whose
horizontal permeability of AZ, and 𝑎V is the vertical radius of length is 𝑑𝑥; the production formula can be described as
AZ.
So the local skin factor of HW along the wellbore after 774.6𝐾V Δ𝑝2 𝑑𝑥
acidizing will be written as follows: 𝑑𝑞 = , (17)
𝑇𝜇𝑔 𝑍 ln [ℎ/ (2𝑟𝑤 )] + 𝑆 (𝑥)
𝑆 (𝑥) = 𝑆1 (𝑥) − 𝑆2 (𝑥) + 𝑆3 (𝑥) where 𝐾V is the vertical permeability of gas reservoir.
So the productivity formula of total HW is given as
𝐾ℎ cot (𝜋𝑟𝑤 /4𝑑V (𝑥))
=( − 1) ln [ ]
𝐾𝑑ℎ cot (𝜋𝑟𝑤 /4𝑎V (𝑥)) (15) 𝐿 774.6𝐾V Δ𝑝2 𝐿 𝑑𝑥
𝑞 = ∫ 𝑑𝑞 𝑑𝑥 = ∫ .
0 𝑇𝜇𝑍 0 ln [ℎ/ (2𝑟𝑤 )] + 𝑆 (𝑥)
𝐾ℎ 𝜋𝑟𝑤
+( − 1) ln [cot ]. (18)
𝐾𝑎ℎ 4𝑎V (𝑥)
Under normal conditions, the productivity formula is
If we want to calculate the distribution of local skin factor described considering skin factor as follows:
of HW after acidizing, the formula about vertical radius of
DZ and AZ must be known. During drilling and acidizing, 774.6𝐾V 𝐿Δ𝑝2
the heel has the maximum radius of DZ and AZ with the 𝑞= . (19)
𝑇𝜇𝑔 𝑍 (ln (ℎ/2𝑟𝑤 ) + 𝑆𝑒𝑞 )
reason of longest time exposing to fluid, namely, 𝑑V. max and
𝑎V. max , while the toe has the least time exposing to fluid, so we Comparing (18) and (19), the skin factor in isotropic
consider it as minimum damage, namely 𝑑V. min and 𝑎V. min . formation can be got from the following model:
Based on the Frick-Economides’ research on skin factor of
HW, we can make an assumption that the vertical radius of 𝐿 ℎ
DZ and AZ decreases linearly from heel to toe along the 𝑆𝑒𝑞 = 𝐿
− ln . (20)
∫0 (𝑑𝑥/ (ln [ℎ/ (2𝑟𝑤 )] + 𝑆 (𝑥))) 2𝑟𝑤
wellbore, as Figure 4 illustrates.
Journal of Chemistry 5

The symbol “−” at the right side of (24) indicates that


fluid’s flow direction is opposite to the 𝑥 plane. For the reser-
voir with infinite outer boundary, the boundary conditions
x=0 x=L can be described as follows:
lim 𝑝2 (𝑥) = 𝑝𝑒2 ,
𝑥 → ∞ 𝑤𝑓
(25)
2 2
lim 𝑝𝑤𝑓 (𝑥) = 𝑝𝑤𝑓 .
𝑥→0
Figure 5: Coupling model between reservoir and wellbore.
Transforming the boundary condition as (25) shows into
that related to flow rate,
If the anisotropy is taken into account, (20) will be 𝑑𝑄 (𝑥)
lim = 0,
changed to 𝑥→∞ 𝑑𝑥
(26)
𝑑𝑄 (𝑥)
𝐿 𝛽ℎ lim = −𝐽ℎ (𝑝𝑒2 − 𝑝𝑤𝑓
2
).
𝑆𝑒𝑞 = − ln . 𝑥 → 0 𝑑𝑥
𝐿 (𝛽 + 1) 𝑟𝑤
∫0 (𝑑𝑥/ (ln[𝛽ℎ/(𝛽 + 1)𝑟𝑤 ] + 𝑆 (𝑥)))
Typically, the pressure gradient in the horizontal wellbore
(21) can be written as
2
𝑑𝑝𝑤𝑓 (𝑥) 𝛾𝑔 𝑍𝑇
2.1.3. Productivity Formula of Horizontal Well after Acidizing. = 9.066 × 10−12 𝜆 𝑄2 (𝑥), (27)
The formula of skin factor after acidizing, namely, (21), can 𝑑𝑥 𝐷5
be added on Joshi’s productivity formula [24] to get the
where 𝜆 is the friction coefficient, 𝐷 is the wellbore diameter,
productivity index of per unit length of HW:
𝛾𝑔 is the gas relative density, and 𝑄(𝑥) is the total rate at any
location of horizontal wellbore.
2 In fully turbulent flow zone, friction coefficient 𝜆 is always
[ 𝑎 + √𝑎 − (𝐿/2) ]
2
774.6𝐾ℎ ℎ
𝐽ℎ = ( ) (ln[ ] constant for a specific slotted liner or screen pipe:
(𝑇𝜇𝑔 𝑍𝐿) 𝐿/2
[ ] 1
(22) 𝜆= , (28)
−1 4𝑙𝑔2 [𝜀/ (3.7𝐷)]
𝛽ℎ 𝛽ℎ
+ (ln + 𝑆𝑒𝑞 )) , where 𝜀 is the wellbore coarse degree.
𝐿 2𝑟𝑤
2.3. Solution Method for Productivity Model. Taking the
where 𝑎 is a semimajor axis of elliptic drainage area. derivative of (22) and (23), respectively,
So from formation to wellbore, the flow rate of per unit 2
length of HW can be written as 𝑑𝑞ℎ (𝑥) 𝑑𝑝𝑤𝑓 (𝑥)
+ 𝐽ℎ = 0, (29)
𝑑𝑥 𝑑𝑥
𝑞ℎ (𝑥) = 𝐽ℎ [𝑝𝑒2 − 𝑝𝑤𝑓
2
(𝑥)] , (23) 𝑑2 𝑄 (𝑥) 𝑑𝑞ℎ (𝑥)
+ = 0. (30)
𝑑𝑥2 𝑑𝑥
where 𝑝𝑒 is the driving pressure of gas reservoir.
We can derive from (29) and (30) that
2
2.2. Flow in the Wellbore. Because of the pressure drop in the 𝑑2 𝑄 (𝑥) 𝑑𝑝𝑤𝑓 (𝑥)
− 𝐽 = 0. (31)
wellbore [25–30], the flow rate is changeable at any location 2 ℎ
𝑑𝑥 𝑑𝑥
along the wellbore; meanwhile, the total rate in the wellbore
is increasing gradually. When the HW is treated by acidizing, Substituting (27) and (28) into (31), the following model
the production is greatly increased, so an assumption can be can be obtained:
made that the condition of fluid in the wellbore is turbulent 𝛾𝑔 𝑍𝑇
𝑑2 𝑄 (𝑥)
flow, for the barefoot well completion, the oil flow from the 2
− 9.066 × 10−20 𝐽ℎ 𝜆 5 𝑄2 (𝑥) = 0. (32)
formation to the wellbore uniformly, as is shown in Figure 5. 𝑑𝑥 𝐷
Based on the volume conservation between reservoir and Combining boundary conditions (26), the analytic solu-
wellbore, the relationship between flow rate change in the tion when the HW’s length tends to reach infinity can be got:
wellbore and that from formation to wellbore can be written
−1/3 −2
as 𝜇𝑔 𝐷 𝐶 3
𝑄(𝑥) = 55.56 × [√ 1 𝑥 + (√ 𝐶) ] ,
𝑑𝑄 (𝑥) 𝛾𝑔 6 2𝐶1 2
= −𝑞ℎ (𝑥) . (24) [ ]
𝑑𝑥 (33)
6 Journal of Chemistry

5.5 9
8
4.5

Local skin factor S(x)


7
Local skin factor S(x)

6
3.5 5
4
2.5
3
2
1.5
1
0
0.5 0 100 200 300 400 500 600 700 800 900 1000
0 100 200 300 400 500 600 700 800 900 1000
Any location along the horizontal section x (m)
Any location along the horizontal section x (m)
dv.max = 2.4 m, dv.min = 0.30 m
Kdh = 0.05 mD Kdh = 0.20 mD
dv.max = 2.8 m, dv.min = 0.35 m
Kdh = 0.10 mD Kdh = 0.25 mD
dv.max = 3.2 m, dv.min = 0.40 m
Kdh = 0.15 mD Kdh = 0.30 mD
dv.max = 1.2 m, dv.min = 0.15 m
dv.max = 1.6 m, dv.min = 0.20 m
Figure 6: Effect of DZ’s permeability (𝐾𝑑ℎ ) on skin factor (𝐿 =
dv.max = 2.0 m, dv.min = 0.25 m
1000 m, 𝑟𝑤 = 0.1 m, 𝜀 = 0.00016 m, 𝛾𝑔 = 0.56, 𝜇𝑔 = 0.023 mPa⋅s,
𝑍 = 0.9, 𝑇 = 366 K, ℎ = 10 m, 𝑝𝑤𝑓 = 20 MPa, 𝐾ℎ = 0.6 mD, Figure 7: Effect of distance of mud immersion (𝑑V. max , 𝑑V. min ) on
𝐾V = 0.1 mD, 𝐾𝑎ℎ = 1.2 mD, 𝑑V. max = 1.6 m, 𝑑V. min = 0.2 m, skin factor (𝐿 = 1000 m, 𝑟𝑤 = 0.1 m, 𝜀 = 0.00016 m, 𝛾𝑔 = 0.56,
𝑎V. max = 1.2 m, and 𝑎V. min = 0.15 m). 𝜇𝑔 = 0.023 mPa⋅s, 𝑍 = 0.9, 𝑇 = 366 K, ℎ = 10 m, 𝑝𝑤𝑓 = 20 MPa,
𝐾ℎ = 0.6 mD, 𝐾V = 0.1 mD, 𝐾𝑎ℎ = 1.2 mD, 𝐾𝑑ℎ = 0.1 mD, 𝑎V. max =
1.2 m, and 𝑎V. min = 0.15 m).
where 𝐶1 = 5.04 × 10−18 𝜆𝐽ℎ 𝑍𝑇𝜇𝑔 /𝐷4 and 𝐶2 = 1.8 ×
10−2 𝐽ℎ 𝛾𝑔 Δ𝑝2 /(𝜇𝑔 𝐷).
In (33), 𝑄(𝑥) indicates the total production that 𝑥 >
𝑥󸀠 , where 𝑥󸀠 means any location along the wellbore, so the Figure 7, when the maximum and minimum distance of mud
production of HW can be described by immersion are rising from (1.2 m, 0.15 m) to (1.6 m, 0.20 m),
respectively, the local skin factor increases significantly, but
𝑄 = 𝑄(0) − 𝑄(𝐿). (34) when they increase from (1.6 m, 0.20 m) to (3.2 m, 0.40 m),
the local skin factor’s range of variation is less than the above
Namely, mentioned case; that is to say, the local skin factor after
2/3 acidizing is increasing as the advancing of distance of mud
3 𝜇𝑔 𝐷
𝑄 = 55.56 [√ 𝐶] immersion, but the tendency is changed gradually; this shows
𝛾𝑔 2𝐶1 2 that if the DZ has a long distance of mud immersion near
−2
(35) the wellbore, the acidizing treatment has a small contribution
−1/3
𝜇𝑔 𝐷 to decrease the skin factor, so an in-depth acidification
− 55.56 [√ 𝐶1 𝐿 + (√ 3 𝐶2 ) ] . technology should be taken into account in order to increase
𝛾𝑔 6 2𝐶1 the distance of acid-rock reaction largely and improve the DZ
[ ]
in the great degree.
3. Results and Discussion Figure 8 provides the comparison of distribution of local
skin factor along the wellbore under different permeability of
3.1. Sensitivity Analysis on Skin Factor of Horizontal Well after AZ. As demonstrated in Figure 8, the permeability of AZ does
Acidizing. Figure 6 shows the distribution of local skin factor not change the distribution trend of local skin factor along the
along the wellbore under different DZ’s permeability. From wellbore; it also attains maximum at the heel, while it attains
Figure 6, we can obviously find that the local skin factor minimum at the toe. When the permeability of AZ ranges
decreases from the heel to the toe; meanwhile, the local from 0.8 mD to 1.0 mD, the local skin factor decreases sharply,
skin factor decreases as the permeability of DZ section is while, from 1.0 mD to 1.8 mD, the value of local skin factor
increased, but the tendency of increasing is not constant, as only has fewer changes; namely, the local skin factor decreases
Figure 6 shows. Namely, when the DZ’s permeability is rising by the increasing of permeability of AZ, but the tendency is
from 0.05 mD to 0.10 mD, the local skin factor’s variation more and more gentle, so the conclusion can be drawn when
range is about 1.6, but when the DZ’s permeability is changing the permeability of AZ reaches a certain extent; if we want
from 0.10 mD to 0.30 mD, the local skin factor decreases to decrease the skin factor through increasing permeability
by only 1, which demonstrates that, for the DZ with high of AZ, it will have little contribution; any other stimulation
permeability, it is difficult to obtain smaller skin factor, so treatments must be taken.
any other stimulation treatment must be taken to decrease Comparing the curves of distribution of local skin factor
the skin factor. under different distance of acid-rock reaction as illustrated in
Figure 7 compares the distribution of local skin factor Figure 9, it is obvious to find that when the distance of acid-
under different mud immersion’s distance. As is shown in rock reaction ranged from (0.8 m, 0.10 m) to (1.0 m, 0.125 m),
Journal of Chemistry 7

4 7

Production-increasing ratio
3.5 6
Local skin factor S(x)

5
3
4
2.5
3
2 2

1.5 1
0 100 200 300 400 500 600 700 800 900 1000 0 100 200 300 400 500 600 700 800 900 1000
Any location along the horizontal section x (m) Horizontal well length L (m)

Kah = 0.8 mD Kah = 1.4 mD Kdh = 0.05 mD Kdh = 0.20 mD


Kah = 1.0 mD Kah = 1.6 mD Kdh = 0.10 mD Kdh = 0.25 mD
Kah = 1.2 mD Kah = 1.8 mD Kdh = 0.15 mD Kdh = 0.30 mD

Figure 8: Effect of AZ’s permeability (𝐾𝑎ℎ ) on skin factor (𝐿 = Figure 10: Effect of DZ’s permeability (𝐾𝑑ℎ ) on production-
1000 m, 𝑟𝑤 = 0.1 m, 𝜀 = 0.00016 m, 𝛾𝑔 = 0.56, 𝜇𝑔 = 0.023 mPa⋅s, increasing ratio (𝐿 = 1000 m, 𝑟𝑤 = 0.1 m, 𝜀 = 0.00016 m, 𝛾𝑔 = 0.56,
𝑍 = 0.9, 𝑇 = 366 K, ℎ = 10 m, 𝑝𝑤𝑓 = 20 MPa, 𝐾ℎ = 0.6 mD, 𝜇𝑔 = 0.023 mPa⋅s, 𝑍 = 0.9, 𝑇 = 366 K, ℎ = 10 m, 𝑝𝑤𝑓 = 20 MPa,
𝐾V = 0.1 mD, 𝐾𝑑ℎ = 0.1 mD, 𝑑V. max = 1.6 m, 𝑑V. min = 0.2 m, 𝐾ℎ = 0.6 mD, 𝐾V = 0.1 mD, 𝐾𝑎ℎ = 1.2 mD, 𝑑V. max = 1.6 m,
𝑎V. max = 1.2 m, and 𝑎V. min = 0.15 m). 𝑑V. min = 0.2 m, 𝑎V. max = 1.2 m, and 𝑎V. min = 0.15 m).

7 3.2. Sensitivity Analysis Study on Production-Increasing Ratio


6 of Horizontal Well after Acidizing. The relationship curves
between production-increasing ratio and HW length at
Local skin factor S(x)

5
4
different DZ’s permeabilities are calculated and presented in
Figure 10. As is shown in Figure 10, when the DZ’s perme-
3
ability is constant, the production-increasing ratio increases
2 as the HW length is increased, but when the HW length
1 reaches a certain extent, the production-increasing ratio
0 becomes constant; meanwhile, the production-increasing
−1
ratio decreases as the permeability of DZ is increased, but
0 100 200 300 400 500 600 700 800 900 1000 the tendency will become more and more gentle; that is
Any location along the horizontal section x (m) to say, the lower the DZ’s permeability is, the higher the
production-increasing ratio will be. So when the distance of
av.max = 0.8 m, av.min = 0.10 m
av.max = 1.0 m, av.min = 0.125 m
mud immersion is constant, if the permeability of DZ is low,
av.max = 1.2 m, av.min = 0.15 m the acidizing effect will be better, which is also the reason why
av.max = 1.4 m, av.min = 0.175 m acidizing had been widely used in developing the gas and oil
av.max = 1.6 m, av.min = 0.20 m field.
av.max = 1.8 m, av.min = 0.225 m Figure 11 shows the relationship between production-
increasing ratio and HW length under different distances of
Figure 9: Effect of distance of acid-rock reaction (𝑎V. max , 𝑎V. min ) on
mud immersion. From Figure 11, it is obvious to find that by
skin factor (𝐿 = 1000 m, 𝑟𝑤 = 0.1 m, 𝜀 = 0.00016 m, 𝛾𝑔 = 0.56,
𝜇𝑔 = 0.023 mPa⋅s, 𝑍 = 0.9, 𝑇 = 366 K, ℎ = 10 m, 𝑝𝑤𝑓 = 20 MPa,
the increasing of distance of mud immersion, the production-
𝐾ℎ = 0.6 mD, 𝐾V = 0.1 mD, 𝐾𝑎ℎ = 1.2 mD, 𝐾𝑑ℎ = 0.1 mD, 𝑑V. max = increasing ratio decreases; this is because when the distance
1.6 m, and 𝑑V. min = 0.2 m). of acid-rock reaction is constant, the deeper the distance of
mud immersion is, the smaller the scale transformation will
be, so for the HW with deep DZ, acidizing treatment will be
ineffective and fracturing can be a good choice to stimulate,
respectively, the local skin factor decreases sharply, but when which not only can improve the DZ to a great degree, but also
the distance of acid-rock reaction increases from (1.0 m, can increase the vertical permeability, eventually improving
0.125 m) to (1.8 m, 0.225 m), the local skin factor only has a the fluidity of oil near the wellbore.
small increase; namely, the local skin factor decreases by the Figure 12 compares the relationship curves between
increasing of distance of acid-rock reaction, but it will reach production-increasing ratio and HW length under different
a constant when the distance has become a certain degree, permeability of AZ. It illustrates that by the increasing of
so this will reveal that, during HW’s acidizing treatment, permeability of AZ, the production-increasing ratio also
reasonable acid type should be chosen, which can not only increases, but the extent will be more and more gentle,
increase the distance of acid-rock reaction, but also meet the eventually tending to a constant value. This is because when
demand of economic benefit. the permeability of DZ is constant, if the permeability of AZ
8 Journal of Chemistry

7 11
10

Production-increasing ratio
Production-increasing ratio

6 9
8
5
7
4 6
5
3 4
3
2
2
1 1
0 100 200 300 400 500 600 700 800 900 1000 0 100 200 300 400 500 600 700 800 900 1000
Horizontal well length L (m) Horizontal well length L (m)

dv.max = 1.2 m, dv.min = 0.15 m av.max = 0.8 m, av.min = 0.10 m


dv.max = 1.6 m, dv.min = 0.20 m av.max = 1.0 m, av.min = 0.125 m
dv.max = 2.0 m, dv.min = 0.25 m av.max = 1.2 m, av.min = 0.15 m
dv.max = 2.4 m, dv.min = 0.30 m av.max = 1.4 m, av.min = 0.175 m
dv.max = 2.8 m, dv.min = 0.35 m av.max = 1.6 m, av.min = 0.20 m
dv.max = 3.2 m, dv.min = 0.40 m av.max = 1.8 m, av.min = 0.225 m

Figure 11: Effect of distance of mud immersion (𝑑V. max , 𝑑V. min ) Figure 13: Effect of distance of acid-rock reaction (𝑎V. max , 𝑎V. min )
on production-increasing ratio (𝐿 = 1000 m, 𝑟𝑤 = 0.1 m, 𝜀 = on production-increasing ratio (𝐿 = 1000 m, 𝑟𝑤 = 0.1 m, 𝜀 =
0.00016 m, 𝛾𝑔 = 0.56, 𝜇𝑔 = 0.023 mPa⋅s, 𝑍 = 0.9, 𝑇 = 366 K, 0.00016 m, 𝛾𝑔 = 0.56, 𝜇𝑔 = 0.023 mPa⋅s, 𝑍 = 0.9, 𝑇 = 366 K,
ℎ = 10 m, 𝑝𝑤𝑓 = 20 MPa, 𝐾ℎ = 0.6 mD, 𝐾V = 0.1 mD, 𝐾𝑎ℎ = 1.2 mD, ℎ = 10 m, 𝑝𝑤𝑓 = 20 MPa, 𝐾ℎ = 0.6 mD, 𝐾V = 0.1 mD, 𝐾𝑎ℎ = 1.2 mD,
𝐾𝑑ℎ = 0.1 mD, 𝑎V. max = 1.2 m, and 𝑎V. min = 0.15 m). 𝐾𝑑ℎ = 0.1 mD, 𝑑V. max = 1.6 m, and 𝑑V. min = 0.2 m).

5
the production-increasing ratio increases sharply by the
Production-increasing ratio

4 increasing of distance of acid-rock reaction; this is because


the bigger the distance of acid-rock reaction is, the greater
the improvement degree (including DZ and no DZ) will
3
be, eventually enhancing the production-increasing ratio. So
during acidizing of HW, retarded acid will be a good choice
2 to increase the distance of acid-rock reaction and content to
the high permeability in AZ.
1
0 100 200 300 400 500 600 700 800 900 1000
Horizontal well length L (m) 4. Conclusions
Kah = 0.8 mD Kah = 1.4 mD The model for skin factor and productivity of HW after
Kah = 1.0 mD Kah = 1.6 mD acidizing is established and solved, during which HW is
Kah = 1.2 mD Kah = 1.8 mD
divided into many segments; at any segment, the method
Figure 12: Effect of AZ’s permeability (𝐾𝑎ℎ ) on production- of displacement between two similar flow modes is used.
increasing ratio (𝐿 = 1000 m, 𝑟𝑤 = 0.1 m, 𝜀 = 0.00016 m, 𝛾𝑔 = 0.56, Type curves of skin factor and production-increasing ratio
𝜇𝑔 = 0.023 mPa⋅s, 𝑍 = 0.9, 𝑇 = 366 K, ℎ = 10 m, 𝑝𝑤𝑓 = 20 MPa, are illustrated, and the factors affecting skin factor and
𝐾ℎ = 0.6 mD, 𝐾V = 0.1 mD, 𝐾𝑑ℎ = 0.1 mD, 𝑑V. max = 1.6 m, production-increasing ratio are analyzed. Analysis results
𝑑V. min = 0.2 m, 𝑎V. max = 1.2 m, and 𝑎V. min = 0.15 m). show after acidizing as the increasing of the distance of
mud immersion, local skin factor of HW increases, while
as the increasing of DZ’s permeability, AZ’s permeability
is higher, it will indicate that acidizing fluid improves the and distance of acid-rock reaction, it decreases, production-
DZ perfectly, but when the permeability of AZ reaches high increasing ratio increases as the increasing of HW length,
levels, the contribution of acidizing to production-increasing AZ’s permeability and distance of acid-rock reaction, while it
ratio will become small, so during application of acidizing, decreases as the increasing of DZ’s permeability and distance
a reasonable acid type and an acid concentration should be of mud immersion.
taken into account in order to get an optimal effect.
Figure 13 compares the relationship curves between
production-increasing ratio and HW length under differ-
Nomenclature
ent distances of acid-rock reaction. As demonstrated in 𝑎: Semimajor axis of elliptic drainage area, m
Figure 13, when the distance of mud immersion is constant, 𝑎V : Vertical radius of acidizing zone, m
Journal of Chemistry 9

𝑎V. max : Vertical maximum distance of acid-rock Acknowledgments


reaction, m
𝑎V. min : Vertical minimum distance of acid-rock The authors are grateful for financial support from the
reaction, m National Science Fund for Distinguished Young Scholars of
𝛽: Anisotropy coefficient China (Grant no. 51125019) and the 2014 Australia China Nat-
𝐷: Wellbore diameter, m ural Gas Technology Partnership Fund Top Up Scholarships.
𝑑ℎ : Horizontal radius of damage zone, m
𝑑V : Vertical radius of damage zone, m References
𝑑V. max : Vertical maximum distance of mud
immersion, m [1] T. P. Frick and M. J. Economides, “Horizontal well damage
𝑑V. min : Vertical minimum distance of mud characterization and removal,” SPE Production & Facilities, vol.
immersion, m 8, no. 1, pp. 15–22, 1993.
𝜆: Friction coefficient [2] P. Permadi and W. Wibowo, “Effects of non-uniform skin
𝐻: Distance between two constant pressure factor distribution on horizontal well inflow performance,” in
lines, m Proceedings of SPE European Damage Conference, pp. 1–6, 2001.
ℎ: Gas reservoir thickness, m [3] K. Furui, D. Zhu, and A. D. Hill, “A rigorous formation damage
𝐽ℎ : Gas productivity index per unit, skin factor and reservoir inflow model for a horizontal well,”
SPE Production & Facilities, vol. 18, no. 3, pp. 151–157, 2003.
m3 /(d⋅MPa⋅M)
𝐾ℎ : Permeability of formation, mD [4] T. Yildiz, “Discussion of ‘A rigorous formation damage skin
𝐾V : Vertical permeability of gas reservoir, mD factor and reservoir inflow model for a horizontal well’,” SPE
Production & Facilities, vol. 19, no. 2, pp. 105–106, 2004.
𝐾𝑎V : Vertical permeability of acidizing zone,
mD [5] K. Furui, D. Zhu, and A. D. Hill, “Authors’ reply to ‘Discussion
𝐾𝑑V : Vertical permeability of damage zone, mD of a rigorous formation damage skin factor and reservoir inflow
model for a horizontal well’,” SPE Production & Facilities, vol. 19,
𝐾𝑑ℎ : Horizontal permeability of damage zone,
no. 2, pp. 107–108, 2004.
mD
[6] S. Parn-anurak and T. W. Engler, “Modeling of fluid filtration
𝐾𝑎ℎ : Horizontal permeability of acidizing zone,
and near-wellbore damage along a horizontal well,” Journal of
mD
Petroleum Science and Engineering, vol. 46, no. 3, pp. 149–160,
𝐿: Horizontal well length, m 2005.
𝑝𝑒 : Driving pressure of gas reservoir, MPa
[7] Y. Ye, J. Yan, S. Zou, S. Wang, and R. Lu, “A new laboratory
𝑝𝑑 : Pressure of damage zone’s boundary, MPa method for evaluating formation damage in fractured carbon-
𝑝𝑤𝑓 : Pressure of heel, MPa ate reservoirs,” Petroleum Science, vol. 5, no. 1, pp. 45–51, 2008.
𝑝𝑤𝑓 (𝑥): Pressure at any location along the [8] J. Cai, B. Yu, M. Zou, and M. Mei, “Fractal analysis of
wellbore, MPa invasion depth of extraneous fluids in porous media,” Chemical
Φ𝑑 : Potential in the damage zone’s boundary, Engineering Science, vol. 65, no. 18, pp. 5178–5186, 2010.
10−6 m2 /s [9] J. Cai and B. Yu, “A discussion of the effect of tortuosity on
Φ1 : Potential in constant pressure boundary, the capillary imbibition in porous media,” Transport in Porous
10−6 m2 /s Media, vol. 89, no. 2, pp. 251–263, 2011.
Δ𝑝𝑠 : Pressure drop caused by skin factor, MPa [10] J. Cai, E. Perfect, C.-L. Cheng, and X. Hu, “Generalized
Δ𝑝: Production pressure drop of vertical well, modeling of spontaneous imbibition based on hagen-poiseuille
MPa flow in tortuous capillaries with variably shaped apertures,”
𝑞ℎ : Production of per unit length, m3 /d Langmuir, vol. 30, no. 18, pp. 5142–5151, 2014.
𝑞𝑠𝑐 : Production of horizontal gas well, m3 /d [11] X. H. Tan and X. P. Li, “Transient flow model and pressure
𝑄: Production of horizontal well, m3 /d dynamic features of tree-shaped fractal reservoirs,” Journal of
𝑟𝑤 : Wellbore radius, m Hydrodynamics B, vol. 26, no. 4, pp. 654–663, 2014.
𝑆: Skin factor of vertical well [12] X. H. Tan, X. P. Li, L. H. Zhang et al., “Analysis of transient
𝑆(𝑥): Local skin factor of horizontal well flow and starting pressure gradient of power-law fluid in fractal
𝑆𝑒𝑞 : Total skin factor of horizontal well porous media,” International Journal of Modern Physics C, vol.
26, no. 4, Article ID 1550045, 2015.
𝛾𝑔 : Gas relative density
𝜀: Wellbore coarse degree, m [13] X.-H. Tan, X.-P. Li, J.-Y. Liu, G.-D. Zhang, and L.-H. Zhang,
“Analysis of permeability for transient two-phase flow in fractal
𝑇: Gas reservoir temperature, K
porous media,” Journal of Applied Physics, vol. 115, no. 11, Article
𝜇𝑔 : Gas viscosity, mPa⋅s ID 113502, 2014.
𝑍: Gas deviation factor.
[14] T. Yildiz, “Equivalent skin factor for non-uniformly damage
horizontal and multilateral wells,” in Proceedings of SPE Annual
Technical Conference and Exhibition, pp. 1–17, 2008.
Conflict of Interests [15] I. V. Voronich, L. A. Gaidukov, and N. N. Mikhailov, “Fluid
filtration to a horizontal well with variation in the parameters
The authors declare that there is no conflict of interests of the damage zone,” Journal of Applied Mechani