Sie sind auf Seite 1von 20

IPTC-19288-MS

A Reservoir and Geomechanical Coupling Simulation Method: Case Studies


in Shale Gas and CBM Reservoir

Xuanhe Tang, State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest Petroleum
University; Haiyan Zhu, State Key Laboratory of Oil and Gas Reservoir Geology and Exploitation, Southwest
Petroleum University & State Key Laboratory for Geomechanics and Deep Underground Engineering, China
University of Mining & Technology; Qingyou Liu and Yujia Song, State Key Laboratory of Oil and Gas Reservoir
Geology and Exploitation, Southwest Petroleum University

Copyright 2019, International Petroleum Technology Conference

This paper was prepared for presentation at the International Petroleum Technology Conference held in Beijing, China, 26 – 28 March 2019.

This paper was selected for presentation by an IPTC Programme Committee following review of information contained in an abstract submitted by the author(s).
Contents of the paper, as presented, have not been reviewed by the International Petroleum Technology Conference and are subject to correction by the author(s). The
material, as presented, does not necessarily reflect any position of the International Petroleum Technology Conference, its officers, or members. Papers presented at
IPTC are subject to publication review by Sponsor Society Committees of IPTC. Electronic reproduction, distribution, or storage of any part of this paper for commercial
purposes without the written consent of the International Petroleum Technology Conference is prohibited. Permission to reproduce in print is restricted to an abstract of
not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of where and by whom the paper was presented.
Write Librarian, IPTC, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax +1-972-952-9435.

Abstract
To investigate the time-lapse, three-dimensional (so-called four dimensional/4D) stress during production/
injection, a 4D multi-physical modeling method is proposed. A finite difference method (FDM) reservoir
simulator is used to couple thermal-hydrological-chemical (THC) processes, while a finite element method
(FEM) geomechanical simulator takes on the role of a thermal-hydrological-mechanical (THM) coupling
calculator. Heterogeneity and anisotropy of the reservoir flow and geomechanical properties as well as the
permeability stress-sensitivity can be considered in modelling based on field and experimental data. In order
to couple the flow model with the geomechanical model, an improved interface (coupling) Python code is
provided to communicate data between the finite difference (FD) and finite element (FE) grids. Ultimately,
this method is applied to analyze the stress and poro-elastic parameters evolution of hydraulic fractured
Sichuan Basin shale gas reservoir and Qinshui Basin coalbed methane (CBM) reservoir in production.
Keywords: Multi-physical modeling, 4D stress, Interface code, Heterogeneity and anisotropy, Shale gas
reservoir and CBM reservoir

Introduction
To quantitatively describe stress and reservoir properties field evolution during production/injection,
dynamic stress modeling methods were widely developed for years. By coupling a geomechanical simulator
and a reservoir simulator, researchers investigated the rock deformation, formation compaction/subsidence
and porous parameters change (Teufel and Rhett, 1991; Gutierrez, 1994; Gutierrez and Makurat, 1997;
Settari and Walters, 1999) with different coupling method (Onaisi and Samier et al., 2002; Samier and Onaisi
et al., 2006). Moreover, formation mechanical anisotropy (Sinha and Kostek, 1996; Prioul and Bakulin et al.,
2004) and stress-sensitivity (Shovkun and Espinoza, 2017) have been considered in numerical simulations.
With commercial simulators have been widely used to build flow-geomechanical model and have proved
2 IPTC-19288-MS

insightful, FEM/FDM numerical solutions are the most popular for full coupling (Bower and Zyvoloski,
1997; Li and Ito et al., 2009; Kolditz and Görke et al., 2012), partial coupling (Rutqvist and Tsang, 2002;
Liu and Kvamme, 2008; Taron and Elsworth et al., 2009; Fei and Li et al., 2015) and one-way coupling
(Olden and Pickup et al., 2012; Nakaten and Kempka, 2014). The effect of discontinuities and natural
fractures is more difficult to rigorously represent in these coupling solutions/methods, although there are
some documented cases including using Eclipse or CMG with Abaqus, TOUGHReact with FLAC3D and
FEMH with Abaqus.
This paper presents a 4D multi-physical modeling method by improving the previous method (Zhu and
Tang et al., 2018) to investigate the stress and reservoir properties evolution. In the improved edition, the
geometry of the overburden formation was neglected while the heterogeneous vertical stress distribution
was applied on the reservoir as an alternative. More importantly, except for the properties mapping during
coupling process, the interference code was integrated with the functions of 3D properties initialization for
the geomechanical model and the stress initialization testing. Besides, the FL2 platform with four fracturing
treated wells in Fuling shale gas reservoir was performed with this improved modelling method. In addition,
the more comprehensive data were utilized to investigate the geomechanics evolution during Qinshui Basin
CBM reservoir drainage. These supplementary data includes the calibrated well logs and the longer period
well production.
In this paper, CMG or Eclipse was selected as a reservoir simulator which can conduct THC coupling
process, while Abaqus serves as the geomechanical simulator which can conduct THM coupling process.
To establish the geomechanical and complete a partial (sequential) coupling process, a Python interface
code bridges the properties from geological model to the geomechanical model, as well as communicates
the data between an FD grid in the reservoir simulator and an FE grid in the geomechanical simulator. With
a set of subroutines, heterogeneity and anisotropy of the reservoir flow and geomechanical properties as
well as the permeability stress-sensitivity can be considered in modelling based on field and experimental
data. Using this modeling method, a 4D multi-physics model for Sichuan Basin shale gas reservoir (China)
and a 4D multi-physics model for Qinshui Basin CBM reservoir are developed to discuss the evolution of
stress and poro-elastic parameters.

Schema for a dynamic multi-physical modeling method

Reservoir FDM model. Partial coupling is adopted. This means that, in one calculation step, the
THC coupling process is fully coupled within the flow FDM model independently considering the THM
coupling results. THM coupling follows the fully coupling mode within the FEM geomechanical model,
independently considering THC coupling. Thus, the flow model for a shale gas reservoir should be first
established in a reservoir simulator. Since the THMC coupling is only partially implemented (soft coupling),
THC coupling cannot be independently simulated from the initial production/injection time to the expected
end time. The simulation schema is illustrated in the following sections.
Geomechanical FEM model.
IPTC-19288-MS 3

Figure 1—New interference code for geomechanical modeling and coupling process

Geometry, mesh and properties. Following the development of THC coupling flow model, a THM
coupling geomechanical model is constructed using the Abaqus platform. As is shown in Fig. 1, the reservoir
geometry in the geomechanical model is simply constructed with conventional CAD software (e.g. Pro-E)
and imported into Abaqus. In the previous modelling method, the shape of grid can be altered as hexahedron
or tetrahedron to overcome the non-convergence in simulation dependent on the discontinuities. And the
geometry with the discontinuities can be the same as the geologic model used to develop the reservoir
FDM model (in CMG or Eclipse), while the overburdens can be considered in the geometry for the load of
overburden pressure in geomechanical model. Using a customized interpolation code, the geomechanical
properties profiles of each well are assigned to the geomechanical mesh model.

Figure 2—Tetrahedral grid for discontinuities (Zhu and Tang et al., 2018)

The previous method was applied to construct the geomechanical properties model by interpolating the
geomechanical properties profiles from individual wells into the 3D geomechanical mesh model (Zhu and
Tang et al., 2018). However, there are several problems in the previous method: (i) since the properties of
the overburdens may not be measured in most cases, the overburden pressure in the numerical model may
be largely different with the real overburden pressure; (ii) although the geomechanical properties model
interpolated from individual wells can describe the properties distribution, the properties in any 3D location
may be inconsistent between the reservoir model and geomechanical model, because interpolation strategies
are different between geomechanical model and the geological which is consistent with reservoir model.
The interference code was improved to modify the 3D isotropic geomechanical modelling method: (i)
only the geometry of reservoir is constructed for geomechanical mesh model, and the overburden pressure
4 IPTC-19288-MS

can be ignored, because the overburden pressure effect can be replaced by the vertical stress as an alternative
(Zoback, 2007), which can be applied on the geomechanical model immediately. On the other hand, without
the overburdens, the mesh refinement is preferred and practical to obtain a more accurate strain for the
smaller sized grid, as shown in Fig. 3; (ii) the properties are calibrated with individual well properties
profiles in geological model, and then mapped to geomechanical mesh model to build the 3D isotropic
geomechanical model.

Figure 3—Geometry modelling strategy comparison and improvement

Assumptions of isotropy in a formation may not be realistic except conventional sand formation.
Anisotropy tests should be conducted to acquire the rock mechanical parameters in different directions
of cores, such as acoustic emission and uniaxial/tri-axial compression tests. An anisotropic elastic-plastic
tensor matrix can be implemented to represent transversely isotropic constitutive behavior. Anisotropic
parameters can be assigned with field variables in Abaqus. Meanwhile, simplified analytical models
with one-way coupling for the permeability stress sensitivity in the flow model can't describe the actual
geomechanical effect on the permeability (Law and Bert Van Der Meer et al., 2005). To represent stress
sensitivity in the THMC coupling simulation, user-defined FORTRAN-subroutines are called in Abaqus to
describe stress dependent porosity and the permeability.
Since the stress distributions are mapped from the geological model, and properties in geological
model are interpolated from the individual wells profiles without the constraint of sedimentary facies, the
stress distributions may be inaccurate due to the trial interpolation strategies. Thus, a stress equilibrium
simulation should be conducted for the 3D actual geomechanical model. If the stress equilibrium simulation
is convergent and the volumetric strain is less than 1.0 × 10−6, the stress distributions can be regarded as the
reasonable initial stress (Shen and Standifird, 2017). If the stress equilibrium simulation is non-convergent
or the volumetric strain is much larger than 1.0 × 10−6, another interpolation strategies must be tried to test
the stress equilibrium.
Multi-physical coupling process. The coupling process follows a sequential coupling which means the
flow model and geomechanical model are implemented separately, and the parameters within these two
models are updated at the end of each calculation step for the next calculation step (Rutqvist, 2011). With the
interface code, properties data are communicated between the flow and the geomechanical models. The data
communicating process between the FD grid in the flow model and the FE grid in the geomechanical model
are shown in Fig. 4. In every coupling step, the data interface involves three generic procedures: Create
an adaptive spherical search zone centered on one node in the non-updated model, and search the nearest
node(s) in the updated model; Interpolate the properties from the node(s) in the updated model to the node
in un-updated model using the nearest neighbor method or the linearly interpolation method depends on the
grid density difference between the updated model and the non-updated model; Perform the interpolation
over all nodes in the non-updated model.
IPTC-19288-MS 5

Figure 4—Data communicating between the FD grid in the flow model and the FE grid in the geomechanical model

A 4D multi physical modeling case-FL2 platform in the Sichuan basin


Modeling process
Overview. The FL2 platform accommodates four horizontal wells (FL2-1HF, FL2-2HF, FL2-3HF,
FL2-4HF), and is located in the Sichuan Basin shale gas reservoir, China. This laminated gas shale reservoir
is located in the Lower Silurian Longmaxi-Upper Ordovician Wufeng formation. There is poor development
of natural fractures in this normal fault stress regime. The vertical depth to the target formation varies
from 1641.14 m to 2045.54 m. The shale sequence is about 160 m thick. A geologic model was developed
from previous petro-physical evaluations of the four wells. Layering is evident as is lateral variation and
heterogeneity. The FL2 platform data collected includes gas and water production rates, and casing pressure
which covers productivity testing period and gas production period. Forward production predictions have
been carried out for 60 months (from January 2015 to January 2020) into the future. The well logging data
of four wells were interpolated to the geological model with the constraint of sedimentary facies, part of
the 3D distributions of physical and geomechanical properties are shown in Fig. 5. As there might be a
overturning, high stress region appears in the shallow area while the low stress region appears in the deeper
area in the same layer. Besides, a strike-slip fault mechanism (the vertical stress is the middle principal
stress) shows in the reservoir mainly affected by tectonism.
6 IPTC-19288-MS

Figure 5—Physical and geomechanical properties distributions of FL2

Modelling. Based on the calibrated geological and well logging data, a flow model for the area drained by
wells from the FL2 pad (referred to as the global model) was prepared in Eclipse. Moreover, the hydraulic
fractures were inversed in Petrel dependent by microseismic monitoring results and mapped to Eclipse, as is
shown in Fig. 6, the fractures length of this four wells ranges from 145 m to 156 m, and the fractures height
ranges from 28 m to 37 m. The 3D geological model was used to build a global geomechanical model in the
FL2 platform vicinity. The dimension of this domain are 5616 m×2355 m×402 m. A 10 nodes displacement-
pore-temperature element (C3D10MPT) was used to mesh the model in Abaqus.
IPTC-19288-MS 7

Figure 6—Hydraulic fractures of FL2 described with permeability

Fluid transport and storage, rock mechanical and thermal properties are required. Geomechanical
properties profiles for the four wells are interpolated into calibrate the geological properties model, and then
the properties from the geological model are attributed to the grid with the interference code. The stress
equilibrium was simulated to test the initial stress distribution of geomechanical model (Fig. 7)

Figure 7—Stress initialized geomechanical model of FL2

Transversely isotropy and stress-dependent permeability. In this shale reservoir, the laminated character
precludes treating the shale as an isotropic elastic material. Recognizing significant bedding in this reservoir,
the anisotropic (transversely isotropic) characteristics should be considered in the THM simulations
(Baumgärtel, 1979). The transverse isotropy intensity coefficient for Young's modulus rE, and transverse
isotropy intensity coefficient for Poisson's ratio rv are based on the experimental result of the reference (Zhu
and Tang et al., 2018). Besides, the relationships between porosity/permeability and volumetric strain-εv,
can be expressed as (Zhu and Zhao et al., 2015) and incorporated into a user-defined subroutine associated
with the geomechanical model.

(1)

(2)

where ϕs is the porosity of shale, εv is the volumetric strain, ϕs0 is the initial shale porosity, where ks is the
permeability of shale, and ks0 is the initial permeability of shale.
8 IPTC-19288-MS

Coupled simulation results and discussion


Variations in the reservoir. Fig. 8 shows the effect on the pore pressure distribution of shale gas
production in the period of 60 producing months after fracturing. Fig. 8(a) shows that the initial pore
pressure distribution declines from north-east to south-west. After producing for 60 months, the depression
cones occur around each well, shown as Fig. 8(b), and the hydraulic fractures evidently contributes to the
domain increase of production effect. Besides, the greatest pore pressure drops is located at the second stage
(Stage-2) of FL2-3HF. Fig. 9 shows the compaction distribution of shale gas production in the period of 60
producing months after fracturing, the largest compactions are not located at trajectory for any wells but
at the side where there is no other well.

Figure 8—Effect of 60 months production on pore pressure distribution of FL2

Figure 9—Compactions after 60 producing months of FL2


IPTC-19288-MS 9

Fig. 10 and Fig. 11 show the effect on the minimum/maximum horizontal effective stress distribution
of shale gas production in the period of 60 producing months after fracturing. Fig. 10(a) and Fig. 11(a)
show that the initial minimum/maximum horizontal effective stress distribution declines from north-east to
south-west resulted from the formation overturning. After producing for 60 months, the depression cones
occur around each well, shown as Fig. 10(b) and Fig. 11(b), and it can be seen that the stress changes are
mainly controlled by the pore pressure. Fig. 10(c) and Fig. 11(c) show the stress orientation distribution
after production, and the dramatic re-orientations appear around the wellbore and that mean the stress
reorientations are mainly control by the pore pressure.

Figure 10—Minimum horizontal effective stress after 60 producing months of FL2

Figure 11—Maximum horizontal effective stress after 60 producing months of FL2


10 IPTC-19288-MS

Stress evolutions in depletion. The comparative simulations were conducted with isotropic material and
transversely isotropic material. The rock mechanical parameters used for the isotropic material were those
measured parallel to the shale bedding. As is shown in Fig. 8 to Fig. 11, the greatest pore pressure drop
is located in the stage-2 of hydraulic fracturing in the well of FL2-3HF, and the discussions are conducted
based on the evolution of FL2-3HF stage-2.
Fig. 12 shows the transient values of the minimum and maximum horizontal effective stresses with pore
pressure depletion near well FL2-3HF. The pore pressure decline is decreasing from 35.15 MPa to 19.06
MPa with time and the minimum and maximum horizontal effective stresses increase with the pore pressure
drop. The minimum horizontal effective stress increases from 25.36 MPa to 28.08 MPa in the transversely
isotropic material and reaches 27.57 MPa in the isotropic material. The maximum horizontal effective stress
increases from 29.15 MPa to 35.76 MPa in the transversely isotropic material and to 34.88 MPa in isotropic
material. Since the elastic modulus parallel to the bedding is greater than the elastic modulus perpendicular
to the bedding, deformation is preferentially impeded in the direction parallel to the bedding, and the
minimum/maximum horizontal effective stress changes for a transversely isotropic material will be greater
than for an isotropic material. In addition, for the transversely isotropic material, the increase in minimum/
maximum horizontal effective stress is 11% and 23%, respectively. For an isotropic material, the increase
in the minimum/maximum horizontal effective stresses over 60 months is 9% and 20%, respectively. For
both transversely isotropic and isotropic materials, the increase in the minimum horizontal effective stress
is greater than that of the maximum horizontal effective stress, and that means that stress re-orientation
occurs during depletion.

Figure 12—Effect of rock mechanical anisotropy of FL2

Since the hydraulic fractures of this model were considered in the reservoir simulation, the effect
of fracturing on the flow-geomechanical coupling process should be discussed. Fig. 13(a) shows the
relationship between the minimum horizontal effective stress and the pore pressure at the area near the
wellbore and the east half-fracture tips of FL2-3HF stage-2 fracture. Moreover, the trajectory of FL2-3HF is
paralleled to FL2-4HF which was fracturing treated and produced in the same period of FL2-3HF, thus the
effect of paralleling fractured wells producing should be discussed as well. Fig. 13(b) shows relationship
between the minimum horizontal effective stress and the pore pressure at the area near the wellbore, the
west half-fracture tips of FL2-3HF stage-2 fracture, the center area of FL2-3HF and FL2-4HF and the same
distance area to the east of FL2-3HF.
IPTC-19288-MS 11

Figure 13—Stress comparison at different areas of FL2

According to the Fig. 13(a), the pore pressure drops 16.10 MPa near the wellbore, while the pore
pressure drops 13.48 MPa near east half-fracture tips of FL2-3HF stage-2 fracture. Besides, the minimum
horizontal effective stress increases by 2.72 MPa near the wellbore, while minimum horizontal effective
stress increases by 2.47 MPa near east half-fracture tips of FL2-3HF stage-2fracture. Whether the pore
pressure or the effective stress, the geomechanical changes near the wellbore is greater than near east
halffracture tips of FL2-3HF stage-2 fracture, because the half-fracture tips is just one of the producing
effect contribution to the changes near the wellbore.
According to the Fig. 13(b), the minimum horizontal effective stress increases by 2.72 MPa with the pore
pressure drops 16.10 MPa near the wellbore, the minimum horizontal effective stress increases by 2.53 MPa
with the pore pressure drops 13.71 MPa near west half-fracture tips of FL2-3HF stage-2 fracture, and the
minimum horizontal effective stress increases by 1.44 MPa with the pore pressure drops 4.89 MPa at the
center area of FL2-3HF and FL2-4HF. The difference of the effective stress increase among the three area
show the same change trend. It can be concluded that, decreasing with the distance away from FL2-3HF,
the pore pressure drops less and the effective stress increase less.
Besides, by comparing the difference between the center area of FL2-3HF and FL2-4HF and the same
distance area to the east of FL2-3HF, the minimum horizontal effective stress increases by 1.44 MPa with the
pore pressure drops 4.89 MPa at the center area of FL2-3HF and FL2-4HF, while the minimum horizontal
effective stress increases by 0.52 MPa with the pore pressure drops 2.35 MPa at the same distance area to
the east of FL2-3HF. It can be conclude that the parallel well (FL2-4HF) have a considerable effect on the
geomechanical changes.
Permeability evolutions. The transverse isotropy is more realistic to describe the shale rock mechanical,
hence, only the effect of fracturing on the flow-geomechanical coupling process and the effect of fractured
wells paralleling are discussed.
Fig. 14(a) shows the relationship between normalized permeability and the pore pressure at the area
near the wellbore, the east half-fracture tips of FL2-3HF stage-2 fracture. The normalized permeability
at the area near the wellbore decreases to 0.502 while decreases to 0.630 near the east half-fracture tips
of FL2-3HF stage-2 fracture. Fig. 14(b) shows the relationship between normalized permeability and the
minimum horizontal effective stress at the area near the wellbore, the west half-fracture tips of FL2-3HF
stage-2 fracture, and the center area of FL2-3HF and FL2-4HF. Comparing the distance effect away from
the wellbore, The normalized permeability at the area near the wellbore decreases to 0.502, the normalized
permeability near the west half-fracture tips of FL2-3HF stage-2 fracture decreases to 0.601, and the
normalized permeability at the center area of FL2-3HF and FL2-4HF decreases to 0.799. Meanwhile, the
12 IPTC-19288-MS

normalized permeability decreases at the center area of FL2-3HF and FL2-4HF is larger than the decrease
of same distance area to the east of FL2-3HF which is only 0.103.

Figure 14—Permeability evolution with pore pressure at different areas of FL2

Fig. 15(a) shows the relationship between normalized permeability and pore pressure at the area near the
wellbore, the east half-fracture tips of FL2-3HF stage-2 fracture. Fig. 15(b) shows the relationship between
normalized permeability and the minimum horizontal effective stress at the area near the wellbore, the west
half-fracture tips of FL2-3HF stage-2 fracture, and the center area of FL2-3HF and FL2-4HF. Based on the
Fig. 15, the effect of stress increase at different region on the permeability can be compared.

Figure 15—Permeability evolution with effective stress at different areas of FL2

A 4D multi physical modeling case-Shouyang A2 region in the Qinshui basin


Modeling process
Overview. Shouyang NYZ-A2 area locates in the middle of Shanxi province and the northern part of the
Qinshui Basin. The altitude of this area ranges from 980.0 m to 1342.1 m. The altitude of the western and
northern part is higher than that of the eastern and southern parts, shown as Fig. 16. 15# coalbed of this area
are confirmed as the main commercial production layer. The A2 area is the main producing area in the NYZ
area with totally 165 wells drilled and 127 wells stimulated, as presented in Fig. 17. According to the wells
sites of 23 gas producing wells, most of them locate in the central-northern part of the NYZ-A2 area.
IPTC-19288-MS 13

Figure 16—Geological cross section of Shouyang CBM NYZ area

Figure 17—Wells location in the 15# coalbed (bottom structure contours)

Modelling. Since the production of most wells was less than 50 m3/day or even 10 m3/day, the flow model
was established in CMG with the geological sub-model of the major producing which covers about 39 wells,
shown as Fig. 18. The production data of 36 months, including casing pressure, bottomhole pressure, gas
production rate and water production rate of 23 gas producing wells and 16 water drainage wells, were
employed for the THM coupling simulation after the geomechanics model establishment.
14 IPTC-19288-MS

Figure 18—The major producing area

Fig. 19 shows the computational domain of the NYZ-A2 major producing area, which covers 3375
m × 1875 m in plane and 311 m to 519 m in vertical direction. The vertical depth of 15# coalbed
ranges from 435.15 m to 926.66 m with a thickness of 3.77 m to 6.08 m. The roof (with a thickness
of 20-33 m) and the floor (with a thickness of 3.41-5.29 m) of the 15# coalbed are the limestone and
mudstone respectively. The geomechanical mesh model was established by the geometry and meshes
of the geological model. Considering that the thickness of coalbed is too thin for tetrahedral grid, the
element type of the geomechanical model was assigned as the C3D8PT with 8 nodes and hexahedron
geometry which can conduct the THM coupling calculation. The initial physical properties fields (porosity,
permeability, saturation) and geomechanical properties fields (Young's modulus, Poisson's ratio, pore
pressure, temperature and in-situ stresses) were obtained from the geological model (Fig. 20). Then initial
properties are bridged to the geomechanical model by the interface code as well as the initial boundary
conditions. The bottom and four lateral boundaries are restricted with all displacements and rotations. With
the 3D geomechanical model, the initial stresses fields were tested by stress equilibrium simulation. (Fig.
21).

Figure 19—NYZ-A2 major producing area geometry model

Figure 20—NYZ-A2 major producing area geological model


IPTC-19288-MS 15

Figure 21—The initial stress field of the NYZ-A2 major producing area (minimum horizontal stress)

Anisotropy and stress-dependent permeability. The orthotropic intensity coefficient of the Young's
modulus and Poisson's ratio in the x direction & the y direction are considered and the stress-sensitive
permeability model are implemented in the geomechanical model by the subroutines as the following
equations which was verified by the laboratory experiments (Zhu and Tang et al., 2018).
The equations of the porosity and permeability is:

(3)

where ϕ is the coal porosity, a is the Biot's coefficient, ϕ0 is the initial coal porosity, Kc is the bulk modulus,
p0 is the pore pressure, p is the pore pressure, and the mean effective stress .

(4)

where k is the permeability of coal, and k0 is the initial permeability of coal.

Coupled simulation results and discussion


Variations in the reservoir. Fig. 22 shows the effect on the pore pressure distribution of CBM production
in the period of 36 producing months. After producing for 36 months, the depression cones occur
around each well. According to the production data of all the wells in this region, six wells (SYNY-109,
SYNY-125, SYNY-136, SYNY-161, SYNY-173, SYNY-187) with the similar high gas production rate and
low water production rate among these producing wells were preferred for re-fracturing operation, and
correspondingly, it shows the larger pore pressure drop at these six wells. Fig. 23 shows the compaction
distribution of CBM production in the period of 36 producing months, the largest compactions are located
at region where the producing wells concentrated.

Figure 22—Pore pressure after 36 producing months of NYZ-A2 major producing area
16 IPTC-19288-MS

Figure 23—Compaction after 36 producing months of NYZ-A2 major producing area

Fig. 24 and Fig. 25 show the effect on the minimum/maximum horizontal effective stress distribution of
CBM production in the period of 36 producing months. Fig. 24(a) and Fig. 25(a) show that the increase of
minimum/maximum horizontal effective stress are mainly affected by the pore pressure drop and the well
location concentration simultaneously. Fig. 24(b) and Fig. 25(b) show the minimum/maximum horizontal
effective stress vectors distribution, it can be seen that the stress re-orientation appears at the region where
the producing wells concentrated, especially for the maximum horizontal effective stress, and that is the
major controlling factor of stress re-orientation

Figure 24—Minimum horizontal effective stress after 36 producing months of NYZ-A2 major producing area

Figure 25—Maximum horizontal effective stress after 36 producing months of NYZ-A2 major producing area

Stress evolution in production and drainage. Considering the rock mechanical anisotropy of the coalbed,
the minimum horizontal effective stress and the maximum horizontal effective stress changes of the
six producing wells (SYNY-109, SYNY-125, SYNY-136, SYNY-161, SYNY-173, SYNY-187) during 36
months' production are illustrated respectively in Fig. 26 and Fig. 27 to discuss the heterogeneity of the
coalbed. The pore pressure drops of these six wells range from 2.59 MPa to 2.94 MPa, and their minimum/
maximum horizontal effective stress increase with the pore pressure decrease. Comparing that the largest
IPTC-19288-MS 17

pore pressure drop is 2.94 MPa which appears at the wellbore of SYNY-173, but the largest minimum
horizontal effective stress increase is 0.94 MPa which appears at the wellbore of SYNY-187 and the largest
maximum horizontal effective stress increase is 1.19 MPa which appears at the wellbore of SYNY-125.
As differences still exist on the stress evolution of each well, it can be concluded that the initial stresses
and stresses evolution differences are caused by the geomechanical heterogeneity and producing wells
concentration of this CBM reservoir. Based on the differences among the stress change rate of these six
wells, the intensities of the stress reorientation are distinct with each other owing to the heterogeneity of
different wells including reservoir depth/thickness, pore parameters and rock mechanics parameters.

Figure 26—Minimum horizontal effective stress evolution in depletion of NYZ-A2 major producing area

Figure 27—Maximum horizontal effective stress evolution in depletion of NYZ-A2 major producing area

Permeability evolution with the change of mean effective stress. The volumetric strain of the coalbed
changes with the variation of stress condition, affecting the pores, fractures and cleats in coalbed. The
porosity and permeability change with the re-construction of pore medium. The relationship between
permeability and volumetric strain, and the relationship between permeability and the effective stress can
reveal the mechanism of permeability stress-sensitivity during depletion.
Since the change of the mean effective stress of the well SYNY-136 is 0.274 MPa which is the
lowest among the six wells mentioned before, SYNY-136 is preferred to discuss the sensitivity of
porosity and permeability in CBM drainage. Fig. 28 and Fig. 29 show the effects of coal anisotropy
and heterogeneity on the permeability evolution. During the CBM drainage, the coal around the well
SYNY-136 is under compression with the effective stress increasing. Correspondingly, the normalized
porosity/permeabilities in three directions respectively decrease to 0.436/0.190 (perpendicular to face cleat),
18 IPTC-19288-MS

0.390/0.152 (perpendicular to butt cleat) and 0.544/0.296 (perpendicular to bedding). The coupling model
embedded with the permeability model proposed in this paper presents a high permeability stress-sensitivity.

Figure 28—Change of normalized permeability with volumetric strain at SYNY-136

Figure 29—Change of normalized permeability with mean effective stress at SYNY-136

Table 1 shows the relationship between the mean effective stress change and normalized permeabilities
in different directions after 36 months' production at six wells.

Table 1—Mean effective stress change with the normalized permeabilities

Mean effective Normalized permeability Normalized permeability Normalized permeability


Wells
stress (MPa) perpendicular to face cleat perpendicular to butt cleat perpendicular to bedding

SYNY-109 0.821 0.150 0.067 0.236


SYNY-125 0.863 0.149 0.068 0.238
SYNY-136 0.263 0.190 0.152 0.296
SYNY-161 0.374 0.169 0.147 0.225
SYNY-173 0.651 0.079 0.051 0.139
SYNY-187 0.615 0.087 0.076 0.191

Although the six wells have the similar gas production rate and water drainage rate, as more high
gas production rate wells crowd in the area where SYNY-109 and SYNY-125 locates in, the inter-well
interference is merged in this area and the effective stresses at SYNY-109 and SYNY-125 increase more
obviously than other four wells, the corresponding permeabilities decrease more sharply. Besides, the mean
IPTC-19288-MS 19

effective stress and normalized permeability at the SYNY-109 and SYNY-125 behave the similar evolutions
which results from almost the same buried depth, drainage and production rate, and flowgeomechanical
conditions of the two wells. The normalized permeability in the direction of butt cleat shows the highest
stress-sensitivity, and the permeability stress-sensitivity in the direction of face cleat is higher than in the
direction of bedding, which shows that the cleats have a larger influence than the bedding on the permeability
change.

Conclusions
1. In the case of the Sichuan Basin simulations, after 60 months of production, the transversely
isotropic intensity has a moderate effect on how effective stresses, porosity/permeability, and reservoir
compaction evolve.
2. The hydraulic fractures contributes a lot to the pore pressure drops and the effective stress increases,
and whether the pore pressure or the effective stress, the geomechanical changes near the wellbore is
greater than near fracture tips; Besides, the effect of paralleling fractured wells producing contributes
to the pore pressure drops and the effective stress increases dependent with the distance away from
the wellbore.
3. The anisotropy and heterogeneity of formation have considerable effects on the stress evolution and
permeability evolution during depletion in Qinshui Basin CBM reservoir.
4. The permeability stress-sensitivity of the Shouyang 15# coalbed is very high in different directions,
and it appear the strongest sensitivity in the direction perpendicular to butt cleat.

Acknowledgements
This work was funded by the National Natural Science Foundation of China (No. 51874253), the China
Postdoctoral Science Foundation (No. 2018T110142) and the Science and Technology Project of Sichuan
Province (No. 2018FZ0069).

References
Baumgärtel, H., 1979. Thirring, W., Lehrbuch der Mathematischen Physik, Band 1: Klassische Dynamische Systeme.
Journal of Applied Mathematics and Mechanics, 59, 490–490.
Bower, K. M. and G. Zyvoloski, 1997. A numerical model for thermo-hydro-mechanical coupling in fractured rock.
International Journal of Rock Mechanics and Mining Sciences, 34, 1201–1211.
Fei, W. B., Q. Li, X. C. Wei, R. R. Song, M. Jing and X. C. Li, 2015. Interaction analysis for CO2 geological storage and
underground coal mining in Ordos Basin, China. Engineering Geology, 196, 194–209.
Gutierrez, M., 1994. Fully coupled analysis of reservoir compaction and subsidence. the European Petroleum Conference.
London, U. K, Society of Petroleum Engineers, 339–347.
Gutierrez, M. and A. Makurat, 1997. Coupled HTM modelling of cold water injection in fractured hydrocarbon reservoirs.
International Journal of Rock Mechanics and Mining Sciences, 34, 1–15.
Kolditz, O., U. Görke, H. Shao and W. Wang Eds. 2012. Thermo-hydro-mechanical-chemical processes in porous media:
benchmarks and examples. Lecture Notes in Computational Science and Engineering, Springer Berlin Heidelberg.
Law, D. H. S., L. G. H. Bert Van Der Meer and W. D. Bill Gunter, 2005. Comparison of numerical simulators for
greenhouse gas storage in coalbeds, part IV: History match of field micropilot test data. Greenhouse Gas Control
Technologies 7. Oxford, Elsevier Science Ltd, 2239–2242.
Li, Q., K. Ito, Z. Wu, C. S. Lowry and S. P. Loheide II, 2009. COMSOL Multiphysics: A novel approach to ground water
modeling. Ground Water, 47, 480–487.
Liu, S. and B. Kvamme, 2008. Improved newton Raphson method - an effective tool in solving flow-mechanic-chemistry
equations of CO2 storage in saline aquifers. American conference on applied mathematics. Harvard, Massachusetts,
USA.
Nakaten, B. and T. Kempka, 2014. Workflow for fast and efficient integration of Petrel-based fault models into coupled
hydro-mechanical TOUGH2-MP - FLAC3D simulations of CO2 storage. Energy Procedia, 63, 3576–3581.
20 IPTC-19288-MS

Olden, P., G. Pickup, M. Jin, E. Mackay, S. Hamilton, J. Somerville and A. Todd, 2012. Use of rock mechanics
laboratory data in geomechanical modelling to increase confidence in CO2 geological storage. International Journal
of Greenhouse Gas Control, 11, 304–315.
Onaisi, A., P. Samier, N. Koutsabeloulis and P. Longuemare, 2002. Management of stress sensitive reservoirs using
two coupled stress-reservoir simulation tools : ECL2VIS and ATH2VIS. 10th Abu Dhabi International Petroleum
Exhibition and Conference. Abu Dhabi, Society of Petroleum Engineers.
Prioul, R., A. Bakulin and V. Bakulin, 2004. Nonlinear rock physics model for estimation of 3D subsurface stress in
anisotropic formations: Theory and laboratory verification. GEOPHYSICS, 69, 415–425.
Rutqvist, J., 2011. Status of the TOUGH-FLAC simulator and recent applications related to coupled fluid flow and crustal
deformations. Computers & Geosciences, 37, 739–750.
Rutqvist, J. and C. Tsang, 2002. A study of caprock hydromechanical changes associated with CO2-injection into a brine
formation. Environmental Geology, 42, 296–305.
Samier, P., A. Onaisi and G. Fontaine, 2006. Comparisons of uncoupled and various coupling techniques for practical
field examples. SPE Journal, 89–102.
Settari, A. T. and D. A. Walters, 1999. Advances in coupled geomechanical and reservoir modeling with applications to
reservoir compaction. 1999 SPE Reservoir Simulation Symposium. Houston, TX, Society of Petroleum Engineers.
Shen, X. and W. Standifird, 2017. Numerical Simulation in Hydraulic Fracturing: Multiphysics Theory and Applications,
CRC Press.
Shovkun, I. and D. N. Espinoza, 2017. Coupled fluid flow-geomechanics simulation in stress-sensitive coal and shale
reservoirs: Impact of desorption-induced stresses, shear failure, and fines migration. Fuel, 195, 260–272.
Sinha, B. and S. Kostek, 1996. Stress-induced azimuthal anisotropy in borehole flexural waves. GEOPHYSICS, 61,
1899–1907.
Taron, J., D. Elsworth and K. Min, 2009. Numerical simulation of thermal-hydrologic-mechanical-chemical processes in
deformable, fractured porous media. International Journal of Rock Mechanics and Mining Sciences, 46, 842–854.
Teufel, L. W. and D. W. Rhett, 1991. Geomechanical evidence for shear failure of chalk during production of the Ekofisk
Field. the 66th Annual Technical Conference and Exhibition of the Society of Petroleum Engineers. Dallas. TX, Society
of Petroleum Engineers.
Zhu, H., X. Tang, Q. Liu, K. Li, J. Xiao, S. Jiang and J. D. McLennan, 2018. 4D multi-physical stress modelling during
shale gas production: A case study of Sichuan Basin shale gas reservoir, China. Journal of Petroleum Science and
Engineering, 167, 929–943.
Zhu, H., X. Tang, Q. Liu, S. Liu, B. Zhang, S. Jiang and J. D. McLennan, 2018. Permeability stress-sensitivity in 4D flow-
geomechanical coupling of Shouyang CBM reservoir, Qinshui Basin, China. Fuel, 232, 817–832.
Zhu, H., X. Zhao, J. Guo, X. Jin, F. An, Y. Wang and X. Lai, 2015. Coupled flow-stress-damage simulation of deviated-
wellbore fracturing in hard-rock. Journal of Natural Gas Science and Engineering, 26, 711–724.
Zoback, M. D., 2007. Reservoir Geomechanics. Cambridge, United Kingdom, Cambridge University Press.