Reservoir stability in the process of natural gas hydrate production by depressurization in the Shenhu area of the South China Sea

Abstract

Reservoir stability is a key factor in the production of natural gas hydrate (NGH), and also a prerequisite to ensuring safe and efficient NGH production. However, it has been rarely discussed. To analyze the reservoir stability in the process of NGH production by depressurization in the Shenhu area of the South China Sea, we established a 3D geological model of NGH production by depressurization on the basis of NGH drilling data in this area, which was then discretized by means of nonstructural grid. Then, the mathematical model coupling four fields (i. e. thermal, hydraulic, solid and chemical) was established considering the heat and mass transfer process and sediment transformation process during NGH production. The model was solved by the finite element method together with the nonstructural grid technology, and thus the time-space evolution characteristics of reservoir pore pressure, temperature, NGH saturation and stress in the condition of NGH production by depressurization were determined. Finally, reservoir subsidence, stress distribution and stability in the process of NGH production by depressurization in the Shenhu area were analyzed. The results obtained are as follows. First, the higher the reservoir permeability and the larger the bottom hole pressure drop amplitude are, the larger the subsidence amount and the higher the subsiding speed. Second, as the reservoir pore pressure decreases in the process of production, the effective stress increases and the shear stress near the well increases obviously, resulting in shear damage easily. Third, the increase of effective reservoir stress leads to reservoir subsidence, which mainly occurs in the early stage of NGH production. After the production for 60 days, the maximum reservoir subsidence reached 32 mm and the maximum subsidence of seabed surface was 14 mm. In conclusion, the NGH reservoirs in the Shenhu area of the South China Sea are of low permeability and the effect range of reservoir pressure drop is limited, so the reservoirs would not suffer from shear damage in the sixty-day-production period. © 2018, Natural Gas Industry Journal Agency. All right reserved.

Full text

Research Article
Reservoir stability in the process of natural gas hydrate production by
depressurization in the shenhu area of the south China sea
*,**
Wan Yizhao
a,b,
*, Wu Nengyou
a,b
, Hu Gaowei
a,b
, Xin Xin
c
, Jin Guangrong
d
, Liu Changling
a,b
&
Chen Qiang
a,b
a
MLR Key Laboratory of Gas Hydrate//Qingdao Institute of Marine Geology, Qingdao, Shandong 266071, China
b
Laboratory for Marine Mineral Resources, Qingdao National Laboratory for Marine Science and Technology, Qingdao, Shandong 266071, China
c
College of Environment and Resources, Jilin University, Changchun, Jilin 130012, , China
d
Guangzhou Institute of Energy Conversion, Chinese Academy of Sciences, Guangzhou, Guangdong 510640, China
Received 5 March 2018; accepted 25 April 2018
Available online 7 December 2018
Abstract
Reservoir stability is a key factor in the production of natural gas hydrate (NGH), and also a prerequisite to ensuring safe and efficient NGH
production. However, it has been rarely discussed. To analyze the reservoir stability in the process of NGH production by depressurization in the
Shenhu area of the South China Sea, we established a 3D geological model of NGH production by depressurization on the basis of NGH drilling
data in this area, which was then discretized by means of nonstructural grid. Then, the mathematical model coupling four fields (i.e. thermal,
hydraulic, solid and chemical) was established considering the heat and mass transfer process and sediment transformation process during NGH
production. The model was solved by the finite element method together with the nonstructural grid technology, and thus the time-space
evolution characteristics of reservoir pore pressure, temperature, NGH saturation and stress in the condition of NGH production by depres-
surization were determined. Finally, reservoir subsidence, stress distribution and stability in the process of NGH production by depressurization
in the Shenhu area were analyzed. The results obtained are as follows. First, the higher the reservoir permeability and the larger the bottomhole
pressure drop amplitude are, the larger the subsidence amount and the higher the subsiding speed. Second, as the reservoir pore pressure de-
creases in the process of production, the effective stress increases and the shear stress near the well increases obviously, resulting in shear
damage easily. Third, the increase of effective reservoir stress leads to reservoir subsidence, which mainly occurs in the early stage of NGH
production. After the production for 60 days, the maximum reservoir subsidence reached 32 mm and the maximum subsidence of seabed surface
was 14 mm. In conclusion, the NGH reservoirs in the Shenhu area of the South China Sea are of low permeability and the effect range of
reservoir pressure drop is limited, so the reservoirs would not suffer from shear damage in the sixty-day-production period.
©2018 Sichuan Petroleum Administration. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Keywords: South China sea; Shenhu area; Natural gas hydrate (NGH); Natural gas hydrate production by depressurization; Effective stress; Reservoir stability;
Multi-field coupling numerical simulation
*
Project supported by Marine Geological Survey Program (No: DD20160219), National Key R&D Program of China (No. 2017YFC0307600), financially
supported by Qingdao National Laboratory for Marine Science and Technology (No. QNLM2016ORP0207), Taishan Scholar Special Experts Project (No.
ts201712079) and the Special Funding Project for Post-doctoral Innovation Project of Shandong Province funded the project.
**
This is the English version of the originally published article in Natural Gas Industry (in Chinese), which can be found at https://doi.org/10.3787/j.issn.1000-
0976.2018.04.014.
*Corresponding author.
E-mail address: yizhao_wan@126.com (Wan YZ).
Peer review under responsibility of Sichuan Petroleum Administration.
Available online at www.sciencedirect.com
ScienceDirect
Natural Gas Industry B 5 (2018) 631e643
www.elsevier.com/locate/ngib
https://doi.org/10.1016/j.ngib.2018.11.012
2352-8540/©2018 Sichuan Petroleum Administration. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction
Natural gas hydrate (NGH) is widely found in seabed
sediments and terrestrial permafrost zones, and it is expected
to become an efficient and clean energy to meet energy needs
of mankind for its wide distribution, huge resource quantity
and high energy density [1]. Currently, NGH is mainly pro-
duced by depressurization, heat injection, chemical injection
and carbon dioxide replacement. A few countries [2e6] have
carried out NGH pilot production, such as Japan and China [6]
which extract marine NGHs by depressurization. Typically,
the pilot production of marine NGHs implemented in China on
May 10, 2017 achieved the highest yield, with a cumulative
gas production of 30.9 10
4
m
3
and an average daily gas
production of 5151 m
3
[6]. This pilot production was stable for
60 days.
According to the above-mentioned pilot production prac-
tices worldwide, it is known that the overall NGH production
efficiency is still low, and the depressurization method is the
most effective production method. As to the NGH production
by depressurization, a pressure drop funnel is formed in the
reservoir by reducing bottom hole pressure, and NGH begins
to decompose when the reservoir pressure drops below the
hydrate phase equilibrium pressure. Theoretically, the greater
the bottom hole pressure drop, the larger the influence range of
the pressure drop funnel, and the higher the gas production
rate [7]. However, the reservoir pressure drop will lead to the
increase of effective stress and vertical deformation of the
reservoir. Moreover, the increase of reservoir stress may cause
reservoir instability and destruction due to the poor cementa-
tion and low strength of marine NGH reservoir; NGHs also
cement between reservoir sediment particles, and the decom-
position of hydrates caused by depressurization will reduce the
strength of the reservoir, which further exaggerates the reser-
voir instability. Therefore, reservoir stability is a key issue for
NGH production and a prerequisite for ensuring safe and
efficient NGH production.
According to the latest exploration results, the NGH res-
ervoirs in the Shenhu area in the northern South China Sea are
dominated by clay silt and silty clay. The reservoir, with
sediment size generally less than 20 mm, is a typical pore-filled
NGH reservoir [8,9], and it reflects a low permeability (<10
mD) in in-situ test. Many scholars have investigated the pro-
duction methods and potential of NGH in the Shenhu area by
numerical simulation. Zhang et al. [10,11] established single
horizontal well and double horizontal well models for NGH
production by depressurization þheat injection in the
Shenhu area, and used these models to evaluate the gas
production capacity. Li et al. [12e14] studied the NGH pro-
duction potential of the Shenhu area by the heat
injection þdepressurization method together with the huff and
puff method, and discussed the production effects of different
well types. Su et al. [15] also evaluated the development by
vertical well thermal stimulation in the Shenhu area. In these
numerical simulation studies, NGH production was analyzed
in detail in terms of well types (e.g. vertical well, horizontal
well and double horizontal wells) in the Shenhu area, and the
production methods such as depressurization, heat injection
and huff and puff were simulated.
The scholars around the world have also conducted pre-
liminary studies on reservoir deformation and destruction in
NGH production. Shen et al. [16] integrated the NGH
decomposition effect into the seepage field and the geotech-
nical deformation field, and established a fluid-solid coupling
mathematical model for NGH production; with the model,
they numerically simulated the reservoir stability in the near-
wellbore zone during the production by depressurization.
Cheng et al. [17] combined reservoir subsidence and borehole
stability with the production by depressurization, and estab-
lished a one-dimensional mathematical model of stability
during NGH production by depressurization, which does not
consider the influence of heat transfer process. Sun et al. [18]
established a thermal-hydro-mechanical coupling model that
reflects the deterioration of reservoir mechanical properties
caused by NGH decomposition, redeveloped the solution of
the model based on ABAQUS, and investigated the deforma-
tion and destruction of NGH reservoirs during the production
by thermal method.
However, there are few studies on the mechanical stability
of the NGH reservoirs with clay silt and silty clay in the
Shenhu area [19]. In this paper, based on the drilling data of
NGH production in the Shenhu area of the South China Sea, a
three-dimensional geological model of NGH production by
depressurization is established. According to the multi-field
coupling process of NGH decomposition, heat transfer,
seepage and skeleton mechanical deformation in the reservoir,
the mathematical model and finite element method for the
stability analysis of reservoirs are established. With these
models, the spatialetemporal evolution characteristics of
reservoir pressure, temperature, saturation and stress during
the NGH production by depressurization are obtained, and
then reservoir subsidence, stress distribution and stability are
analyzed.
2. Physical model
In September 2015, the China Geological Survey
completed the third natural gas hydrate drilling voyage
(GMGS3) in the Shenhu area in the northern slope of the
South China Sea. The water depth of the GMGS3-W19 station
was 1273.9 m, and it was determined that there was a NGH
layer with a thickness of about 35 m in the range of
135e170 m below the sea floor. According to the drilling data,
a physical model (Fig. 1) was established. In the horizontal
direction, the model extends 400 m in the xand ydirections
around the wellbore. The top surface of the model is the sea
floor. NGH exists 135 m below the top surface, with a thick-
ness of 35 m. The overlying layer is 135 m thick, and the
underlying layer is 94 m thick, both of which contain no
NGHs. The penetrated well section is 135e162 m, that is, the
28 m interval below the top of the NGH reservoir penetrated
[20].
632 Wan YZ. et al. / Natural Gas Industry B 5 (2018) 631e643

3. Mathematical model and solution
3.1. Basic assumptions
NGH production is a complex heat and mass transfer pro-
cess, including seepage, thermal convection and heat transfer
of multiphase fluids in porous media, chemical reactions of
NGH decomposition, and mechanical deformation of hydrate
deposits. In order to establish a mathematical model
describing the physical process of NGH production, it is
assumed that: ①the hydrate is a pure Type-I hydrate, gas in
the hydrate is 100% methane, and the formation of ice is
ignored; ②the selected control body keeps thermal equilib-
rium locally, that is, the temperature of the sediment particles
and the fluids (gas and liquid) are the same; ③gas and water
flow in the porous media according to the Darcy's law; and ④
hydrate is pore-filled and adheres on to the surface of the
sediment particles, and the hydrate and sediment particles
constitute a continuous composite solid material, which is
jointly stressed, and presents a linear elasticity.
3.2. Mathematical model
Based on the above assumptions, a thermal-hydro-
mechanical-chemical (THMC or four-field) coupling model
is established, which comprehensively considers the thermal
field, gasewater two-phase seepage field, mechanical defor-
mation field of the sediment and chemical field of NGH
decomposition.
3.2.1. Chemical field of NGH decomposition kinetics
The mass conservation equation for NGH decomposition is:
vð4rhShÞ
vt¼mhð1Þ
Based on the Kin-Bishoni kinetic model [21], the gas
production rate at the time of NGH decomposition is:
mg¼kreacMCH4Ars pepgð2Þ
Correspondingly, water production rate and hydrate con-
sumption rate of NGH decomposition are:
mw¼mgNh
MH2O
MCH4
ð3Þ
mh¼mg
Mh
MCH4
ð4Þ
3.2.2. Gasewater two-phase seepage field
Gas and water flow in a porous medium in a hydrate deposit
can be expressed by a continuous equation and Darcy's law,
and finally a model equation for gasewater two-phase seepage
is obtained.
Gas phase:
v4rgSg
vtVKrgKrg
mgVpgþrgg¼mgð5Þ
Water phase:
vð4rwSwÞ
vtVKrwKrw
mw
ðVpwþrwgÞ¼mwð6Þ
The m
g
and m
w
in the right-hand term of Eqs. (5) and (6)
are calculated from the NGH decomposition kinetics model,
which are the key to coupling the seepage field and the
chemical field.
3.2.3. Thermal field
The thermal field during NGH decomposition can be rep-
resented by the energy conservation equation, which considers
the heat transfer and the thermal convection during NGH
decomposition is:
where,
leff ¼ð14Þlsþ4Sglgþ4Swlwþ4Shlh
Fig. 1. Schematic diagram of the GMGS3-W19 station model.
vrw4SwCwTþrg4SgCgTþrh4ShChTþrsð1fÞCsT
vtþVrw4Swv
!w;tCwTþrg4Sgv
!g;tCgT¼VleffVTþQhð7Þ
633Wan YZ. et al. / Natural Gas Industry B 5 (2018) 631e643

3.2.4. Mechanical deformation field of hydrate deposit
The equilibrium differential equation for composite me-
chanical deformation of the hydrate-sediment is:
Vsþrmg¼0ð8Þ
According to the Terzaghi effective stress principle, the
total stress of soil is equal to the sum of the effective stress and
the pore pressure of the composite solid composed of the
hydrate and sediment grain skeletons [22], then:
s¼s0þapeff ð9Þ
where,
peff ¼Sw
SwþSg
pwþSg
SwþSg
pg
According to the linear elasticity assumption, the
stressestrain relationship of the composite solid is:
s0¼2GmεþlmtrðεÞIð10Þ
ε¼1
2Vu
!þVTu
!ð11Þ
Gm¼Em
2ð1þvmÞð12Þ
lm¼Emvm
ð1þvmÞð12vmÞð13Þ
where, tr(ε)represents the diagonal element of ε. Finally the
deformation force field equation with the displacement of the
composite solid as a variable is obtained:
8
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
<
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
:
A1
v2ux
vx2þA3
v2ux
vy2þA3
v2ux
vz2þðA2þA3Þv2uy
vxvyþ
ðA2þA3Þv2uz
vxvzavpeff
vx¼0
A3
v2uy
vx2þA1
v2uy
vy2þA3
v2ux
vz2þðA2þA3Þv2ux
vxvyþ
ðA2þA3Þv2uz
vyvzavpeff
vy¼0
A3
v2uz
vx2þA3
v2uz
vy2þA1
v2uz
vz2þðA2þA3Þv2ux
vxvzþ
ðA2þA3Þv2uy
vyvzavpeff
vz¼0
ð14Þ
where,
A1¼2Gm
1vm
12vm
A2¼2Gm
vm
12vm
A3¼Gm
3.2.5. Auxiliary equation
In addition to the governing equations of physical fields,
auxiliary equations of physical parameters are required.
3.2.5.1. Absolute permeability equation [23].
K¼K0ð1ShÞnð15Þ
3.2.5.2. Capillary force and gasewater two-phase relative
permeability equations [24].
pgpw¼pcð16Þ
Krg ¼KrgoSng
ges ð17Þ
Krw ¼KrwoSnw
wes ð18Þ
where,
Swes ¼Swe Swre
1Swre Sgre
Sges ¼Sge Sgre
1Swre Sgre
Swe ¼Sw
1Sh
Sge ¼Sg
1Sh
Swre ¼Swr
1Sh
Sgre ¼Sgr
1Sh
3.2.5.3. Hydrate phase equilibrium pressure equation [25].
pe¼expA0þA1TþA2T2þA3T3þA4T4þA5T5ð19Þ
3.2.5.4. Hydrate phase change latent heat equation [26].
Qh¼mh
Mh
ðB1B2TÞð20Þ
3.2.5.5. Equation of mechanical properties of hydrate sedi-
ment particle composite solid. The result of triaxial experi-
ment shows that the presence of hydrate increases the strength
of hydrate deposits, but the effect of hydrate on Poisson's ratio
is small [27]. Therefore, assuming the Poisson's ratio is
634 Wan YZ. et al. / Natural Gas Industry B 5 (2018) 631e643

constant, the modulus of elasticity E
m
is represented by the
formula proposed by Santamarina and Ruppel [28]:
Em¼Es0sc
sc0b
þcEhðShÞdð21Þ
The above equations constitute a four-field coupling model
for the NGH production process.
3.3. Finite element solution
Through analysis, the THMC coupling model can be
divided into two subsystems: a flow system containing heat
and mass transfer and a mechanical deformation system. The
two subsystems are solved separately by the finite element
method.
For the flow system, p
g
,S
w
,S
h
and Tare selected as vari-
ables for independent solution. Particularly, S
h
can be directly
obtained by using Eqs. (1) and (4). The finite element equa-
tions of p
g
,S
w
and Tare derived below.
3.3.1. Finite element equation of p
g
The first item of Eq. (5) is expanded:
v4rgSg
vt¼4rg
vSg
vtþrgSg
vð4Þ
vtþ4Sg
vrg
vtð22Þ
According to the gas state equation:
rg¼pg
ZRgTð23Þ
Replace the density in Eq. (5) with pressure, then the
pressure equation of gas phase can be obtained as follows:
4Sg
ZRgT
vpg
vtþ4
ZRgT
vSg
vtþSg
ZRgT
v4
vtpgVKrgKrg
mg
Vpg
¼mg
ð24Þ
The above equation is multiplied by dp
g
and integrated, and
the order is reduced by the Gaussian formula, then:
The pressure is represented by the interpolation function on
the unit:
pg¼Npgeð26Þ
dpg¼Ndpgeð27Þ
Substitute it into the original formula, the finite element
solution equation for the final gas phase pressure is obtained:
Ae_
pgeþQepge¼fgeð28Þ
where,
Ae¼ZN4Sg
ZRgTNTdV
Qe¼ZN4
ZRgT
vSg
vtþSg
ZRgT
v4
vtNTþ
VNTKrgKrg
mg
VNdV
fge¼ZmgNTdVþZqmgNTds
3.3.2. Finite element equation of S
w
Using the capillary force equation, we can get:
VpgVpw¼Vpc¼pc0VSwð29Þ
The water phase pressure in Eq. (6) is expressed by gas
phase pressure and water phase saturation, then
4rw
4Sw
vtþrw
v4
vtSwþVKrwKrw
mw
pc0VSw
VKrwKrw
mw
Vpg¼mw
ð30Þ
The above equation is multiplied by dp
w
and integrated,
and the order is reduced by the Gaussian formula, then:
Z4rw
vSw
vtþrw
v4
vtSwdSwdV
ZKrwKrw
mw
pc0VSwVðdSwÞdVþZKrwKrw
mw
VpgVðdSwÞ¼
ZmwdSwdVþZKrwKrw
mw
vpw
vndSwds
ð31Þ
The saturation is represented by the interpolation function
on the unit:
Sw¼NfSwgeð32Þ
dSw¼NfdSwgeð33Þ
Substitute it into the original formula, then the finite
element solution equation for the final saturation is obtained:
Be_
SweþCefSwge¼ffwgeð34Þ
Z4Sg
ZRgT
vpg
vtþ4
ZRgT
vSg
vtþSg
ZRgT
v4
vtpgdpgdVþZKrgKrg
mg
VpgVdpgdV¼ZmgdpgdVþZKrgKrg
mg
vpg
vndsð25Þ
635Wan YZ. et al. / Natural Gas Industry B 5 (2018) 631e643

where,
Be¼ZN4rwNTdV
Ce¼ZNrw
v4
vtNTdVZKrwKrw
mw
pc0VNVNTdV
ffwge¼ZmgNTdVZKrwKrw
mw
VpgVNTdVþZqmwNTds
3.3.3. Finite element equation of T
Eq. (7) is multiplied by dTand integrated to obtain:
ZðrCÞt
vT
vtþrgCgþrwCwVTVðlVTÞdTdV
¼ZQhþðqmCÞtðTT0ÞdTdVð35Þ
Reduce the order and substitute the interpolation function
T¼NfTge
;dT¼NfdTge, then the finite element solution of
Tis obtained:
De_
TeþEefTge¼ffTgeð36Þ
where,
De¼ZNðrCÞtNTdV
Ee¼ZVNrgCgþrwCwNTdVþZlVN$VNTdV
ffTge¼ZQhNTdVþZlvT
vnNTds
The second subsystem is a mechanical deformation model
and its governing equation is Eq. (14). Take the xdirection as
an example, the equilibrium differential equation in the xdi-
rection is multiplied by du
x
and integrated; after the order is
reduced, then:
ZA1
vux
vxþA2
vuy
vyþA2
vuz
vzpvdux
vxþ
A3
vux
vyþA3
vuy
vxvdux
vyþ
A3
vux
vzþA3
vuz
vxvdux
vzdV¼ZFxduxds
ð37Þ
From the interpolation function, we can obtain:
ux¼Nfuxgeð38Þ
dux¼Nfduxgeð39Þ
The finite element equation in the xdirection is:
½k11fuxgþ½k12uyþ½k13 fuzgþ½k14fpg¼ff1gð40Þ
where,
½k11¼ZA1
vNi
vx
vNj
vxþA3
vNi
vy
vNj
vyþA3
vNi
vz
vNj
vzdV
½k12¼ZA2
vNi
vy
vNj
vxþA3
vNi
vx
vNj
vydV
½k13¼ZA2
vNi
vz
vNj
vxþA3
vNi
vx
vNj
vzdV
½k14¼ZNi
vNj
vxdV
ff1g¼ZFxNjds
Similarly, the finite element equations in the ydirection and
the zdirection can be obtained.
3.4. Model grid division
The physical model shown in Fig. 1 is divided into grids by
a tetrahedral unstructured grid, and the grid diagram is shown
in Fig. 2. For accurate calculation, more dense grids are
divided in the NGH layer, and there are totally 38068 grids.
3.5. Initial conditions and boundary conditions
According to the survey results of the GMGS3-W19 sta-
tion, the temperature at the sea floor is 3.75 C and the
geothermal gradient is 0.045 C/m. According to the reservoir
depth, the initial reservoir temperature is linearly distributed in
longitudinal direction. The pore pressure at the initial moment
of the formation gradually increases with depth, which is
consistent with hydrostatic pressure balance. The initial hy-
drate saturation of the NGH layer is 0.45, and both the over-
lying layer and the underlying layer contain water with a
saturation of 1. The bottomhole pressure remains constant
during production and the outer boundary of the reservoir
follows the constant pressure boundary condition to maintain
the original formation pressure.
The initial in-situ stress can be calculated by the self-
weight of the saturated soil. The water depth at the top sur-
face of the model is 1273.9 m, the converted top pressure is
12.86 MPa, and the sediment density is 2650 kg/m
3
; then, the
formation stress increases vertically at a gradient of 25.97 kPa/
m. As to the boundary conditions of the mechanical defor-
mation field, the upper top surface is the free surface bound-
ary, and the bottom surface of the reservoir is fixed vertically;
the side surface is fixed with the horizontal displacement, that
is, for the side surface perpendicular to x¼0 m and
x¼800 m, the displacement in the xdirection is 0, and for the
side surface perpendicular to y¼0 m and x¼800 m, the
displacement in the ydirection is 0.
3.6. Physical parameters
For this model, the water depth is 1273.9 m, the thickness
of the overlying layer is 135 m, the thickness of the under-
lying layer is 94 m, the bottomhole pressure during
636 Wan YZ. et al. / Natural Gas Industry B 5 (2018) 631e643

production is 5.0 MPa, the initial pressure at the bottom of the
NGH layer is 14.3 MPa, and the initial temperature at the
bottom of the NGH layer is 14.0 C. The thermal and phys-
ical properties, permeability and other parameters are shown
in Tab l e 1 .
The mechanical properties of the hydrate sediments are
shown in Table 2.
4. Model verification
In order to verify the correctness of the numerical model
and the program, we compared the model calculation results
with the results of the Masuda experiment [23]. Masuda
adopted Berea sandstone synthetic hydrate in production by
depressurization, as shown in Fig. 3.
The Berea core, 30 cm long and 5.1 cm in diameter, was
placed in a constant temperature air bath with a temperature of
2.3 C. Its initial temperature (T
i
) was 2.3 C, initial pore
pressure ( p
gi
) was 3.75 MPa, initial hydrate saturation (S
hi
)
was 0.443, water saturation (S
wi
) was 0.351, porosity (4
i
) was
0.182 and permeability (K
i
) was 98 mD. During the experi-
ment, the A-end maintained a constant pressure of 2.84 MPa
as an outlet and the cumulative gas production of the outlet
was recorded.
By using the finite element calculation program estab-
lished in this paper, the above example was calculated to
obtain the cumulative gas production variation at the outlet
with time, and compared with the experimental results, as
shown in Fig. 4. It can be seen that the multi-field coupling
model can be in good agreement with the experimental re-
sults, which proved the validity and reliability of the model
and algorithm.
5. Results and discussion
5.1. Water production and gas production
Fig. 5 shows the variation of water and gas production rate
with time at the bottomhole pressure of 5 MPa. It can be
known that the formation pressure near the wellbore decreases
with the decrease of the bottomhole pressure, which leads to
the decomposition of the hydrate near the wellbore at first, the
water and gas production rate of the wells maintains a high
value and then decrease rapidly. When the water production
increases, the gas production increases gradually. Due to a
faster pressure conduction, the water production can quickly
reach a relatively stable state, and the gas production can also
reach a relatively stable fluctuation state.
5.2. Physical field distribution
Under the assumption of homogeneity of the model, the
predicted spatial variation of pressure, hydrate saturation,
temperature and gas saturation on the 60
th
day of production is
obtained (Fig. 6).
It can be seen from Fig. 6-a that the depressurization area is
mainly concentrated around the wells; the pressure in the well
center is the lowest, about 9 MPa lower than the initial formation
pressure. The horizontal influence range is about 35m around the
well, that is, the interval where the pressure drop ranges from
9 MPa to 0 MPa in the well center is 35 m. Under the production
pressure of 5 MPa, the decomposition zone of hydrate saturation
is restricted to zones around the well, and the decomposition zone
of hydrate is approximately within 2 m from the well in the
horizontal direction. Due to the low permeability of the reservoir,
Fig. 2. Grid division of NGH production model in the Shenhu area.
637Wan YZ. et al. / Natural Gas Industry B 5 (2018) 631e643

the hydrate at the bottom of the reservoir has not completely
decomposed, which has played a certain role in water resistance.
It can be seen from Fig. 6-c that the endothermic effect
during NGH decomposition does not cause a significant
change of temperature spatially on the 60
th
day of production,
the temperature in the well center is reduced by a maximum of
about 4C. In the horizontal direction, the temperature
decrease range is small, which proves that the NGH decom-
position range is small. A part of the methane gas generated by
the NGH decomposition is migrated to the well, and the other
part accumulates in the pore space. Due to the increase of gas
saturation, under the capillary force of the two-phase fluid
flow, the gas cannot be completely produced, so gas saturation
distribution as shown in Fig. 6-d is formed.
5.3. Mechanical response
Two points were selected in the reservoir to monitor the
changes of pore pressure, temperature and stress with time.
The coordinates at the near-wellbore zone and the far-wellbore
zone are (x¼0.3 m, z¼149 m) and (x¼8.1 m, z¼149 m),
respectively.
It can be seen from Fig. 7-a that the pore pressure near the
wellbore reduces rapidly and reaches a stable value due to the
bottomhole depressurization. The pore pressure far away from
wellbore shows a gentle downward trend, and it is much higher
than the pore pressure near the wellbore. The reduction of pore
pressure near the wellbore results in NGH decomposition, and
the NGH decomposition and heat absorption cause its tempera-
ture to decrease (Fig. 7-b). When the decomposition is over, the
temperature gradually rises due to the surrounding heat transfer.
In zones far away from the wellbore, the pore depressurization is
not enough to cause a huge amount of NGH decomposition, so
the temperature remains basically unchanged.
It can be seen from Fig. 7-c that the decrease in pore pressure
causes an increase in the effective principal stress. The effective
stress near the wellbore increases rapidly due to the rapid
decrease of the pore pressure, and finally remains unchanged
(due to the production at constant bottomhole pressure). The pore
pressure in the far-wellbore zone decreases little, and the effec-
tive principal stress increases slowly. The pressure drop makes
the difference between the maximum and minimum principal
stress in the near-wellbore zone larger than that in the far-
wellbore zone, so the increases of shear stress in the near-
wellbore zone ismore obvious than that in the far-wellbore zone.
Fig. 7-d shows the variation of the maximum and minimum
effective principal stresses at the two points with the production
time, i.e. the effective stress path. It can be seen that the stress
state at the two points is outside the rupture line of Mole-
Coulomb shear strength line of sediment [30] at the time
Table 1
Physical parameters of the NGH reservoir in the W19 station of the Shenhu
area.
Parameter Value
Initial hydrate saturation 0.45
Absolute permeability (without hydrate) 2.50 mD
Porosity (without hydrate) 0.48
Permeability decline index 7.00
Gas thermal conductivity 0.06 W/(m$K)
Water thermal conductivity 0.50 W/(m$K)
Hydrate relative permeability model S
wr
¼0.3,S
gr
¼0.05,K
wro
¼0.3,
K
gro
¼0.1, n
g
¼4, n
w
¼4
Coefficient of hydrate phase change
latent heat equation
B1¼3527000, B1¼1050
NGH decomposition rate constant [29] kreac ¼k0
dexpDEa
RT 
k
d
0
3.6 10
4
mol/(m
2
$Pa$s)
DEa
R
9752.73 K
Volume specific surface area of
NGH decomposition [24] Ars ¼4Shffiffiffiffiffiffiffi
43
eff
2K
s;
4eff ¼4ð1ShÞ
Gas specific heat capacity C
g
¼2180 J/(kg$K)
Water specific heat capacity C
w
¼4200 J/(kg$K)
Hydrate specific heat capacity C
h
¼2220 J/(kg$K)
Specific heat capacity of sediment particles C
s
¼2180 J/(kg$K)
Hydrate thermal conductivity 2.0 W/(m$K)
Thermal conductivity of sediment particles 1.0 W/(m$K)
Table 2
Mechanical parameters of hydrate sediments in the Shenhu area.
Parameter Value
Cohesion When S
h
¼0, C
m
¼0.5 MPa
When S
h
¼1, C
m
¼2 MPa and C
m
varies linearly with S
h
Internal friction angle 30
Elasticity modulus When S
h
¼0, E
m
¼95 MPa
When S
h
¼1, E
m
¼670 MPa and E
m
varies linearly with S
h
Poisson ratio 0.15
Biot coefficient a¼1
Fig. 3. Diagram of the Berea core experimental model.
Fig. 4. Comparison of cumulative gas production calculated in this paper and
experimental result. *Note: The Masuda experimental data is from Dr. Shi-
gemi Naganawa (Tokyo University).
638 Wan YZ. et al. / Natural Gas Industry B 5 (2018) 631e643

t¼0, which means that it is already in a compression-stable
state in the sediment history. The stress path shows a rapid
change within 0e1 day near the wellbore and a slow change
after the 1st day, the stress change in the far-wellbore zone lags
behind that near the wellbore. Since the stress paths at both
points appear to deviate from the Moore-Coulomb strength line,
they do not approach or reach the rupture line, indicating that
no damage has occurred. Therefore, the preliminary prediction
results based on this model show that the reservoir will not be
destroyed within 60 days of production.
5.4. Reservoir subsidence
Fig. 8 shows the vertical displacement during the production
process, that is, the production depressurization causes
sedimentation of seafloor sediments. It can be seen from Fig. 8-a
and c that the depressurization in the well causes a funnel-like
spatial subsidence. From the overlooking view (xOy plane), the
subsidence is circularly distributed around the well position, and
the subsidence at the well position is the largest. From the section
diagram (Fig. 8-b&d), it can be seen that the subsidence at the
production section is small, and the vertical displacement above
and below the production section is the largest. Therefore, the
subsidence near the upper part of the production section is the
largest, and the local uplift occurs under the osmotic pressure
below the production section. Under the overlying stress, the
subsidence of soil above the well is larger than the uplift below
the well. Since the soil above the well section is totally settled
under the stress, the influence range of subsidence caused by the
depressurization is greater than that by the pore pressure.
Fig. 5. Variation of water and gas production rate with time at the GMGS3-W19 station.
Fig. 6. Physics field variation around the wellbore after 60 days of pilot production (negative values indicate that they are lower than the initial values).
639Wan YZ. et al. / Natural Gas Industry B 5 (2018) 631e643

Fig. 7. Mechanical response at (x, z)¼(0.3 m, 149 m) and (x, z) ¼(8.1 m, 149 m); and. s
1
0and s
3
0are the corresponding maximum and minimum effective
principal stresses.
Fig. 8. Distribution of subsidence space caused by production section (135e162 m).
640 Wan YZ. et al. / Natural Gas Industry B 5 (2018) 631e643

In Fig. 8, the maximum subsidence at the well location is
0.032 m (or 32 mm) on the 30
th
day, while the subsidence at
the seafloor is about 0.014 m, and the seafloor subsidence
range (>5 mm) has a radius of about 166 m. During the
production, the maximum subsidence of the production well is
0.035 m on the 60
th
day, the seafloor subsidence is about
0.018 m, and the seafloor subsidence range (>5 mm) has a
radius of about 232 m. Subsidence and subsidence range
gradually increase with production.
Fig. 9-a shows the variation of the seafloor subsidence in
the production well (vertical displacement) with time. In the
first 15 days, the subsidence was about 0.009 m (i.e. 9 mm).
Subsequently, as the pore pressure gradually became a state of
dynamic equilibrium in space, the sedimentation rate at the
seafloor decreased. After 60 d, the subsidence gradually
became slow and eventually reached 0.018 m. It can be seen
that the subsidence in the first 15 days accounts for 1/2 of the
total subsidence. Therefore, under the conditions of constant
pressure production, the subsidence mainly occurs in the early
production period.
Fig. 9-b shows the subsidence (vertical displacement) at
different moments in the production well. It can be seen that
stress changes caused by the depressurization are mainly
concentrated around the production well, so large displace-
ment occurs near the depressurization point. Since the bottom
of the depressurization point is fixed and does not move
vertically, the uplift at the bottom of the depressurization point
will be 0 due to the boundary effect. The top of the model is a
free surface that can move freely, and the sediment within a
certain depth range below the seafloor is overall subsidence.
5.5. Influence of permeability on reservoir subsidence
Permeability is a key factor affecting the range of gas and
water migration and pressure. Fig. 10 shows the variation of
seafloor subsidence in the production well with time at
different permeability.
When the permeability is low, the depressurization range is
small, and the decline rate of the seafloor subsidence remains
basically unchanged. When the permeability increases, the
depressurization range increases, and the subsidence occurs at
a higher rate and then at a lower rate, and finally the seafloor
subsidence occurs at a lower rate. For the permeability of 1.0
mD, 2.5 mD and 5.0 mD, based on the subsidence on the 60
th
day, the time required for half sedimentation is 24 d, 15 d and
9.5 d, respectively. As permeability increases, the sedimenta-
tion rate increases, and the time to reach the same subsidence
is advanced.
5.6. Influence of bottomhole pressure on reservoir
subsidence
Bottomhole pressure directly affects the distribution of pore
pressure in the formation, causing changes in the effective
stress of the framework, which in turn affects the subsidence
of the reservoir. Fig. 11 shows the variation of the seafloor
subsidence in the production well with time during the pro-
duction at different pressures. It can be seen that in the early
stage of production, the subsidence under different production
pressures is basically the same; in the stage of stable gas
production, the seafloor subsidence is gradually different. On
the 60
th
day, the subsidence is 0.016 m, 0.018 m and 0.020 m,
respectively, and the time required for half sedimentation is
approximately 15 d. The production pressure is reduced, the
subsidence rate is increased, and the time required to reach the
same subsidence is advanced, but the influence degree is less
than that of permeability.
Fig. 9. Variation of the seafloor subsidence in the production well with time and reservoir burial depth.
Fig. 10. Variation of seafloor subsidence with permeability in the production
well.
641Wan YZ. et al. / Natural Gas Industry B 5 (2018) 631e643

6. Conclusions
1) Based on the drilling data of NGH production in the
Shenhu area of the northern South China Sea, the
physical model of NGH production by depressurization
in a single vertical well is established, and the model is
divided by unstructured grid. Considering the heat
transfer and sediment deformation in the NGH produc-
tion, a four-field (THMC) coupling model is established.
Based on the unstructured grid, the model is solved by
the finite element method to obtain the pore pressure,
temperature, saturation and spatialetemporal evolution
of stress.
2) The permeability of hydrate reservoirs in the Shenhu
area is low, and the influence of pressure drop on the
production by depressurization is limited mainly to the
vicinity of a wellbore, and the decomposition range of
hydrate is also small.
3) During the NGH production process, the decrease of
pore pressure in the reservoir leads to the increase of
effective stress. The increase of effective stress is mainly
concentrated near the wellbore, and the difference be-
tween the maximum and minimum principal stresses
near the wellbore are larger than that in the far-wellbore
zone. Therefore, the shear stress in the near-wellbore
zone is large, so shear damage is likely to occur. The
stress path line near the wellbore after 60-d production is
far away from the Moore-Coulomb strength envelope,
which indicates that the reservoir will not be destroyed
under the assumptions of this model.
4) The effective stress increase results in reservoir subsi-
dence. After 60 days of production, the maximum
reservoir subsidence is 32 mm, and the maximum sea-
floor subsidence is 14 mm; reservoir subsidence mainly
occurs in the early stage of production.
5) Reservoir permeability and bottomhole pressure during
production by depressurization have obvious effects on
reservoir subsidence. The greater the reservoir perme-
ability, the greater the bottomhole pressure drop, the
greater the reservoir subsidence and the faster the
subsidence.
Acknowledgement
We appreciate Dr. Shigemi Naganawa from Tokyo Uni-
versity for providing us the Masuda experimental data.
Symbols
ttime, s
p
w
and p
g
water phase and gas phase pressure, Pa
p
c
capillary pressure, Pa
p
e
hydrate phase equilibrium pressure, Pa
S
h
,S
g
and S
w
hydrate saturation, gas phase and water phase,
respectively
r
h
,r
w
,r
s
and r
g
density of hydrate, water phase, sediment
particles and gas phase, respectively, kg/m
3
r
m
represents the density of complex solids composed of
hydrate and sediment particles, kg/m
3
,rm¼4rhShþ
ð14Þrs
K
reac
NGH decomposition rate constant, mol/(m
2
$Pa$s)
A
rs
surface area of the reservoir NGH decomposition per
unit volume, m
1
m
w
,m
h
and m
g
water production rate of NGH decomposition,
NGH decomposition rate per unit volume, gas pro-
duction rate of NGH decomposition per unit volume
of reservoir, respectively, kg/(m
3
$s)
N
h
number of hydrates
MCH4Moore mass of methane gas, kg/mol
MH2Omolarity of the water phase, kg/mol
M
h
Moore mass of hydrate, kg/mol
K
rg
and K
rw
relative permeability of gas phase and water
phase for gasewater two-phase percolation,
respectively
Kand K
0
sediment porous medium and non-hydrated absolute
permeability, respectively, m
2
K
rgo
and K
rwo
relative permeability of gas phase and oil phase
end points without hydrate
npermeability decline index
n
g
and n
w
relative permeability index of gas phase and water
phase, respectively
S
gr
and S
wr
saturation of gas phase and the water phase at
relative permeability end points
m
g
and m
w
viscosity of methane gas and water phase,
respectively, Pa$s
4porosity of the hydrate deposit
ggravitational acceleration, m/s
2
Treservoir temperature, K
C
w
,C
g
,C
h
and C
s
specific heat capacity of the water phase,
gas phase, hydrate phase and sediment particle phase,
respectively, J/(kg$K)
v
!w;tand v
!g;tvelocity of water phase and the gas phase
relative to the control body, m/s
Fig. 11. Variation of seafloor subsidence with production pressure ( p
w
) in the
production well.
642 Wan YZ. et al. / Natural Gas Industry B 5 (2018) 631e643

l
s
,l
g
,l
w
and l
h
heat transfer coefficients of sediment parti-
cles, methane gas phase, water phase and hydrate,
respectively, W/(m$K)
Q
h
latent heat of phase change of NGH decomposition,
W/m
3
Stotal stress tensor of the hydrate sediment, MPa
seffective stress of the composite solid, MPa
s
c0
reference confining pressure, MPa
aBiot coefficient
εstrain tensor
Iunit vector
u
!deformation and displacement of the sediment solid, m
G
m
and l
m
Lame constant of the composite solid composed of
hydrate and sediment particles
E
m
,E
s0
and E
h
composite solid, non-hydrated sediment under
the reference confining pressure s
c0
, the elastic
modulus of pure hydrate, respectively, MPa
v
m
composite solid Poisson's ratio
bsensitivity coefficient of the sediment elastic modulus
under confining pressure of s
c
chydrate elastic modulus coefficient
dhydrate saturation nonlinear effect coefficientA
0
-A
5
,
B
1
and B
2
represent the constant coefficients
References
[1] Klauda JB &Sandler SI. Global distribution of methane hydrate in ocean
sediment. Energy Fuels 2016;19(2):459e70.
[2] Dubreuil-Boisclair C, Gloaguen E, Bellefleur G &Marcotte D. Non-
Gaussian gas hydrate grade simulation at the Mallik site, Mackenzie
Delta, Canada. Mar Petrol Geol 2012;35(1):20e7.
[3] Uddin M, Wright F, Dallimore S &Coombe D. Gas hydrate dissociations
in Mallik hydrate bearing zones A, B, and C by de-pressurization: effect
of salinity and hydration number in hydrate dissociation. J Nat Gas Sci
Eng 2014;21:40e63.
[4] Schoderbek D, Martin KL, Howard J, Silpngarmlert S &Hester K. North
slope hydrate field trial: CO
2
/CH
4
exchange//OTC arctic technology
conference, 3e5 December. 2012. Houston, Texas, USA, https://doi.org/
10.4043/23725-MS.
[5] Yamamoto K, Terao Y, Fujii T, Terumichi I, Seki M, Matsuzawa M, et al.
Operational overview of the first offshore production test of methane
hydrates in the Eastern Nankai Trough. Offshore Technology Confer-
ence, 5e8 May, Texas, USA.
[6] Chen Huiling &Zhu Xia. The trial production well of gas hydrate from
Shenhu area, South China Sea, shut down on the 60th day of production.
China Land Res Rep 2017.07-10(13):6-6.
[7] Kurihara M, Sato A, Ouchi H, Narita H, Masuda Y, Saeki T, et al. Pre-
diction of gas productivity from eastern Nankai trough methane-hydrate
reservoirs. SPE Reservoir Eval Eng 2009;12(3):477e99.
[8] Liang Jinqiang, Wang Hongbin, Su Xin, Fu Shaoying, Wang Lifeng,
Guo Yiqun, et al. Natural gas hydrate formation conditions and the
associated controlling factors in the northern slope of the South China
Sea. Nat Gas Ind 2014;34(7):128e35.
[9] Yu Xinghe, Wang Jianzhong, Liang Jinqiang, Li Shunli, Zeng Xiaoming,
Sha Zhibin, et al. Depositional accumulation characteristics of gas hy-
drate in the northern continental slope of South China Sea. Acta Petrolei
Sinica 2014;35(2):253e64.
[10] Zhang KN, Moridis GJ, Wu NY, Li XS &Reagan MT. Evaluation of
alternative horizontal well designs for gas production from hydrate de-
posits in the Shenhu area, South China Sea//International Oil and Gas
Conference &Exhibition, 8e10 June. 2010. Beijing, China, https://doi.
org/10.2118/131151-MS.
[11] Hu Litang, Zhang Keni &Gao Tong. Numerical studies of gas production
from gas hydrate zone using heat injection and depres-surization in
Shenhu area, the South China Sea. Geoscience 2011;25(4):675e81.
[12] Li Gang, Li Xiaosen, Zhang KN &Moridis GJ. Numerical simulation of
gas production from hydrate accumulations using a single horizontal well
in Shenhu area, South China Sea. Chin J Geophys 2011;54(9):2325e37.
[13] Li Gang, Li Xiaosen, Qi Chen &Chen Zhaoyang. Numerical simulation
of gas production from gas hydrate zone in Shenhu area, South China
Sea. Hua Hsueh Hsueh Pao 2010;68(11):1083e92.
[14] Li G, Moridis GJ, Zhang KN &Li XS. Evaluation of gas production
potential from marine gas hydrate deposits in Shenhu area of South
China Sea. Energy Fuels 2010;24(11):6018e33.
[15] Su Zheng, He Yong &Wu Nengyou. Numerical simulation on production
potential of hydrate deposits by thermal stimulation. Journal of Tropical
Oceanography 2012;31(5):74e82.
[16] Shen Haichao, Cheng Yuanfang &Hu Xiaoqing. Numerical simulation
of near wellbore reservoir stability during gas NGH production by
depressurization. Petroleum Drilling Techniques 2012;40(2):76e81.
[17] Cheng Jiawang, Su Zheng &WuNengyou. A geomechanicalstability model
analysisof hydrate reservoirfor gas hydrate pro-duction by depressurization.
Advances in New and Renewable Energy 2016;4(1):33e41.
[18] Sun Keming, Wang Tingting, Cheng Zhai &Luo Huiyu. Patterns of
reservoir deformation due to thermal decomposition of natural gas hy-
drates. Special Oil Gas Reservoirs 2017;24(5):91e6.
[19] Sun JX, Zhang L, Ning FL, Lei HW, Liu TL, Hu GW, et al. Production
potential and stability of hydrate-bearing sediments at the site GMGS3-
W19 in the South China Sea: a preliminary feasibility study. Mar Petrol
Geol 2017;86:447e73.
[20] Yang S, Zhang M, Liang J, Lu J, Zhang Z, Melanie H, et al. Preliminary
results of China's third gas hydrate drilling expedition: a critical step
from discovery to development in the South China Sea. Fire in the Ice
2015;15(2):1e5.
[21] Kim HC, Bishnoi PR, Heidemann RA &Rizvi SSH. Kinetics of methane
NGH decomposition. Chem Eng Sci 1985;42(7):1645e53.
[22] Biot MA &Willis DG. The elastic coefficients of the theory of consol-
idation. J Appl Mech 1957;15(2):594e601.
[23] Masuda Y, Naganawa S &Ando S. Numerical calculation of gas pro-
duction performance from reservoirs containing natural gas hydrates.
SPE J 1997;29(3):201e10.
[24] Sun XF &Mohanty KK. Kinetic simulation of methane hydrate forma-
tion and dissociation in porous media. Chem Eng Sci 2006;61(11):
3476e95.
[25] Duan ZH &Sun R. A model to predict phase equilibrium of CH4 and
CO
2
clathrate hydrate in aqueous electrolyte solutions. Am Mineral
2006;91(8/9):1346e54.
[26] Esmaeilzadeh F, Zeighami ME &Fathi J. 1-D modeling of NGH
decomposition in porous media//. Proc World Acad Sci Eng Technol
2013;43:648.
[27] Soga K, Lee SL, Ng MYA &Klar A. Characterization and engineering
properties of methane hydrate soils//proceedings of the 2nd international
workshop on characterization and engineering properties of natural soils,
29 November-1 december 2006. Singapore. London: CRC Press; 2006.
[28] Santamarina JC &Ruppel C. The impact of hydrate saturation on the
mechanical, electrical, and thermal properties of hydrate-bearing
sand, silts, and clay//Proceedings of the 6th International Confer-
ence on Gas Hydrate, 6e10 July. Vancouver: British Columbia,
Canada; 2008.
[29] Clarke M &Bishnoi PR. Determination of the intrinsic rate of gas NGH
decomposition using particle size analysis. Ann N Y Acad Sci
2000;912:556e63.
[30] Rutqvist J, Moridis GJ, Grover T, Silpngarmlert S, Collett TS &
Holdich SA. Coupled multiphase fluid flow and wellbore stability anal-
ysis associated with gas production from oceanic hydrate-bearing sedi-
ments. J Petrol Sci Eng 2012;(92/93):65e81.
643Wan YZ. et al. / Natural Gas Industry B 5 (2018) 631e643