Out PDF
Out PDF
Out PDF
A Thesis in
by
Suntichai Silpngarmlert
Doctor of Philosophy
May 2007
UMI Number: 3266205
Turgay Ertekin
Program Chair and Professor of Petroleum and Natural Gas Engineering
George E. Trimble Chair in Earth and Mineral Sciences
Chair of Graduate Program
Thesis Co-Advisor
Co-Chair of Committee
Luis Ayala
Assistant Professor of Petroleum and Natural Gas Engineering
Thesis Co-Advisor
Co-Chair of Committee
Michael Adewumi
Professor of Petroleum and Natural Gas Engineering
Quentin E. and Louise L. Wood University Endowed Fellow in Petroleum
and Natural Gas Engineering
Zuleima Karpyn
Assistant Professor of Petroleum and Natural Gas Engineering
Derek Elsworth
Professor of Energy and Geo-Environmental Engineering
a two phase, free-gas and water, zone. A Class 1 hydrate reservoir is more preferable than
class 2 and class 3 hydrate accumulations because a small change of pressure and
technique are studied. In this work, the production characteristics and efficiency from
pressure production scheme, both gas and water production rates exponentially decrease
with time. However, for methane-hydrate reservoirs, gas production rate exponentially
declines with time whereas water production rate increases with time because methane
examined. The simulation results indicate that the moving well completion location
strategy provides better gas production performance than the fixed completion location
strategy) is at the middle of free-gas zone. Due to the effects of hydrate saturation on
formation permeability, one should not complete a well in the hydrate zone.
iii
The effect of well spacing on the production efficiency is also investigated. As
expected, smaller well-spacing system yields more total gas production and it can
dissociate gas-hydrate more rapidly than the larger well-spacing system. However, the
number of wells increases when the well-spacing decreases resulting in the increase of
the capital investment of the project. Based on this study, when the well-spacing
increased about 100 percent (from 45.0 acres to 74.38 acres) the cumulative gas
production decreased about 8.4 percent at 1,000 days of production. Therefore, once the
similar simulation study for a particular reservoir has been performed, the optimum well
multiple-well systems, starting all production wells at the same time provides faster
hydrate dissociation. However, based on this study, starting production wells at different
times yields more produced gas (about 10 percent by volume) even though less gas-
iv
TABLE OF CONTENTS
LIST OF FIGURES………………………………………………………………....vii
LIST OF TABLES…………………………………………………………………..xii
NOMENCLATURE…………………………………………………………….......xv
ACKNOWLEDGEMENTS……………………………………………………….xxii
CHAPTER 1: Introduction…………..……………………………………………......1
v
5.3 Fluid Property Calculations…..…………………………….....................81
BIBLIOGRAPHY…………………………………….……………………….......175
vi
LIST OF FIGURES
STOMP-HYD simulator…………………………………………………....40
Figure 2-11: Types of the reviewed hydrate reservoir simulation models …………......44
Figure 5-2: The possible calculated conditions at a new iteration level ……………….75
Figure 6-3: The effects of grid block dimensions (Δx and Δy)
viii
Figure 6-13: Structure of a conventional gas and gas-hydrate reservoirs…....………..123
methane-hydrate reservoirs………...………………………………..…..129
Figure 6-23: Cumulative produced gas during 1,500 days of operation ……….….…139
ix
Figure 6-26: System configuration of the production characteristics study ………….144
Figure 6-27: Gas production from one well for different well-spacing systems……...145
Figure 6-28: Cumulative dissociated gas and percent hydrate recovery for
Figure 6-29: Hydrate recovery of different well-spacing at 750 and 1,080 days……..148
Figure 6-30: Cumulative gas production and hydrate recovery of the three
Figure 6-31: Well structure in the methane-hydrate reservoir used in this study.…….151
Figure 6-32: Hydrate recovery and cumulative dissociated gas for each case.……….152
Figure 6-36: Aqueous phase saturation of well block layer for Case (a).…….……….159
Figure 6-37: Aqueous phase saturation of well block layer for Case (b).…….……….160
Figure 6-38: Aqueous phase saturation of well block layer for Case (c).…….……….161
Figure 6-39: Aqueous phase saturation of well block layer for Case (d).…….………162
Figure 6-40: Hydrate saturation of the initial dissociation front layer for Case (a)…...163
Figure 6-41: Hydrate saturation of the initial dissociation front layer for Case (b)…...164
Figure 6-42: Hydrate saturation of the initial dissociation front layer for Case (c)…...165
Figure 6-43: Hydrate saturation of the initial dissociation front layer for Case (d)…...166
x
Figure 6-45: Hydrate recovery and cumulative produced gas of Cases 1 and 2………170
Figure B-2: The comparison when water flow rate was specified at 50 STB/day…….186
Figure B-3: The comparison when gas flow rate was specified at 4 MMSCF/day…....186
Figure B-4: The comparison when sand face pressure was specified at 500 psia ….....187
xi
LIST OF TABLES
Table 2-7: The equations used in Sun and Mohanty’s work ……………...……………27
Table 2-12: Hydrate formation and decomposition rate models used in this work ….....42
Table 4-5: Empirical constants in Kamath’s correlation for methane hydrate …………62
Table 4-6: Empirical constants in Makogon’s correlation for methane hydrate ……….63
Table 4-7: Empirical constants in Moridis’ correlation for methane hydrate ………….64
xii
Table 5-1: Sets of primary variables for each phase appearance case ………………….73
Table 6-5: Parameter values for permeability and capillary pressure calculations …….99
Table 6-8: Initial values for the initialization of the lower section ……..………….….103
Table 6-9: Results from the initialization of the lower section …..……..………….….103
Table 6-10: Initial values for the initialization of the upper section.……..……………104
Table 6-12: Initial values for the initialization of the entire reservoir …..………….…105
Table 6-13: Results from the initialization of the entire reservoir …..…..…………….106
Table 6-14: Initial conditions of the reservoir used in this study ..…..…..……….…...107
Table 6-16: Characteristics and properties of the reservoir in Burshears’ work .……..118
Table 6-19: Cumulative produced gases for three different well spacings ………..….150
xiii
Table 6-20: Production schedules used in this study.………………………………….151
Table 6-21: Cumulative Gas to Water Ratio at 1,000 days for each case …………….169
Table B-1: Reservoir and grid block properties for model validation …………..…….185
xiv
NOMENCLATURE
a = shape parameter
cH = concentration of hydrate
cp = average hydrate cap heat capacity including rock, hydrate, water, and gas
E = activation energy
EB = energy balance
xv
EH = gas hydrate formation factor
FE = heat flux
g = gravitational acceleration
Hl = enthalpy of phase l
Hm = Henry constant
k = permeability at φ
xvi
K = three-phase (water-hydrate-vapor) equilibrium value
k0 = permeability at φ 0
k = average permeability
Pe = dissociation pressure
xvii
Qdis = heat of hydrate dissociation
RH = hydrate recovery
rw = wellbore radius
Sa = aqueous saturation
Sg = gas saturation
Sh = hydrate saturation
SI = ice saturation
xviii
Ss = solid (hydrate and/or ice) saturation
Sw = water saturation
T = temperature
Vb = elementary volume
μa = aqueous viscosity
μg = gas viscosity
xix
Δc pw = heat capacity difference between water (or ice) and empty-hydrate lattice
Δhwo = reference enthalpy difference between water (or ice) and empty-hydrate
lattice
ρr = rock density
φ = rock porosity
φe = effective porosity
xx
vl = velocity of phase l
ω = acentric factor
xxi
ACKNOWLEDGEMENTS
The author expresses his sincere appreciation and gratitude to his academic advisors, Dr.
Turgay Ertekin and Dr. Luis Ayala, for their professional guidance. The author also
would like to extend his gratitude to Dr. Derek Elsworth, Dr. Michael Adewumi, and Dr.
Zuleima Karpyn for their interest in serving as committee members, and to Irene
Gullapalli for her suggestion. Their suggestions, insights, and contribution to this work
The author expresses his sincere appreciation to his fellow graduate students, friends and
family especially his sister, Nuntawan Silpngarmlers, and his mother, Siriporn
xxii
CHAPTER 1
INTRODUCTION
The demand of natural gas as a clean source of energy is rapidly increasing in the
world. Natural gas hydrates are drawing attention worldwide as an unconventional source
of energy because of the vast availability of this resource of energy and the foreseen
increase in the demand of natural gas[1]. The estimates on the amount of natural gas being
trapped within gas hydrate deposits vary within the limits of 1015 to 1019 standard m3. It
has been assessed that the current world energy consumption could be sustained for 180
years only by recovering one tenth of this class of trapped gas [1].
crystalline water structure which traps the hydrocarbons. From a technical point of
view, all potential schemes of gas production from hydrate reservoirs need to achieve
• bring the pressure and temperature around hydrate particles outside the
hydrate-stability zone
wells
have been suggested[2]: (1) depressurization, (2) thermal stimulation, and (3) inhibitor
injection.
1
There are a number of challenges associated with producing gas from hydrate
reservoirs. The gas hydrate phase is immobile and impairs flow conditions within the
reservoir. This could slow down the depressurization rate within the reservoir and
consequently reduce the rate of hydrate dissociation. Another issue is that the sensible
energy within the reservoir may not be enough to provoke hydrate dissociation at an
economical rate. The temperature drop due to dissociation process could slow down the
dissociation rate. Additionally, after dissociation, released free water may freeze and end
up as ice which obstructs the flow in porous media (even without freezing, the free water
The complete numerical modeling of this type of reservoir is still at its developing
stages. The reservoir modeling of this type of reservoir is still not clear because its
dynamics and the role of parameters that control these dynamics have not been fully
understood. Therefore, the need of the development as a tool that can further our
understanding of the complication of this type of reservoirs and provide guidance for
2
CHAPTER 2
LITERATURE REVIEW
The history of natural gas hydrates has evolved over three major periods as
described below[3]:
Natural gas hydrates were first discovered by Sir Humphrey Davy in the early
1800s. Throughout the remaining of the century, scientists attempted to identify all of the
describe the compounds, their compositions, and physical properties. Since then, much
work has been done for cataloging the various molecules that could co-exist as host
(lattice-formers) and guests (hydrate-formers), and the various conditions at which each
variety was stable. However, during this period, the subject of hydrates remained largely
responsible for blocking natural gas transmission lines, particularly at temperatures above
freezing point of water. This discovery was crucial in bringing more attention to gas
hydrates, and thereafter led to the regulation of the water content in natural gas
3
transmission lines. Many workers including Hammerschmidt (1939), Deaton and Frost
(1946), Bond and Russell (1949), Kobayashi et al. (1951), and Woolfolk (1952) studied
the effects of inhibitors, such as several chloride salts, on hydrate formation. In order to
prevent and alleviate the hydrate appearance problem, methanol became one of the most
popular inhibitors because it becomes concentrated in the free water phase after being
In the late 1960s, the global view of clathrate science began to change
subsurface sediments in the giant gas fields of the Western Siberia basin. Thereafter,
hydrate was also found in shallow, sub-permafrost sediments on the North Slope of
Alaska. In 1972, hydrate cores and a full well log suite were retrieved at the Arco-Exxon
NW Eileen well number 2 in West Prudhoe Bay, Alaska. A limited production test was
also conducted, which recovered 92% methane at the surface from a hydrate bearing
sandstone at an approximate depth of 2,200 feet. The estimated production rate (~4
mcf/d) was clearly not economical, and there were no further tests in the region.
pressure conditions necessary for hydrate formation should exist extensively around the
globe, not only in permafrost regions, but also under deep oceans. In 1974, Soviet
scientists recovered large hydrate nodules from the floor of the Black Sea. Then, in the
early 1980s, the research vessel Glomar Challenger traveled the globe collecting cores of
ocean bottom sediments as part of a renewed round of Deep Sea Drilling Project tests.
4
Many of the samples found chemical evidence for hydrate. However, one core taken off
the coast of Guatemala included a one-meter long core composed almost entirely of
methane hydrate.
Class 1 Accumulation: This class of accumulation comprises two layers: the hydrate
interval (often exhibiting very low permeability due to the presence of large hydrate
saturation in pore space), and an underlying two-phase fluid zone with free gas. In this
class of accumulation, the bottom of hydrate stability zone (BHSZ) typically coincides
with the bottom of hydrate interval because pressure increases with depth and
temperature also increases with depth (under sea floor) due to geothermal energy. Thus,
hydrate phase becomes unstable below a particular level (called the BHSZ) under sea
Class 2 Accumulation: This class of hydrate accumulation features two zones: a hydrate-
bearing interval, and overlying a mobile water zone with no free gas (e.g., an aquifer).
5
sea surface
Temperature profile
sea floor
Temperature
sea surface
Temperature profile
sea floor
water zone
Temperature
the hydrate interval, and is characterized by the absence of an underlying zone of mobile
fluids. The conditions for class 3 hydrate accumulation is shown in Figure 2-3.
6
sea surface
Temperature profile
sea floor
Temperature
In classes 2 and 3 hydrate accumulations, the entire hydrate interval may be well
within the hydrate stability zone. This means that the bottom of hydrate interval does not
The desirability of class 2 and class 3 hydrate accumulations as for gas production
target is less well defined than for class 1 deposits[15]. This is because the bottom of
hydrate stability zone (BHSZ) in class 1 deposits typically coincides with the bottom of
hydrate interval. Therefore, it requires only small changes in pressure and temperature to
induce dissociation process. Whereas the bottom of hydrate interval in class 2 and class 3
typically be within the hydrate stability zone. Consequently, it requires a larger pressure
7
2.3 Production Methods for Gas Hydrate Reservoirs
crystalline water structure which traps hydrocarbon gas molecules inside the crystalline
cavities. From a technical point of view, three potential gas production schemes for gas-
2.3.1 Depressurization
In this method, the gas phase pressure around hydrate lattice is reduced across the
hydrate stability zone. The depressurization can be achieved by removing gas and water
from the reservoir through production wells. Gas-hydrates in the reservoir may not
immediately start dissociating at the early stage of the production because reservoir
pressure is still higher than the hydrate dissociation pressure at the reservoir temperature.
is likely to be observed. So, the path of pressure and temperature changes for the
In thermal stimulation technique, a hot fluid such as hot water is injected into the
when the reservoir conditions cross the stability boundary. Typically, average reservoir
pressure decreases during the production. Thus, the change of reservoir conditions for
8
thermal stimulation is likely to follow the path shown in Figure 2-5. This production
technique has a disadvantage as a significant portion of the energy introduced into the
system is lost in the injection path and surroundings. Therefore, only a fraction of
Pressure
Ti , Pi
Temperature
Pressure
Ti , Pi
Thermal Stimulation
Temperature
9
2.3.3 Inhibitor Injection
The objective of this technique is to move the stability boundary of the reservoir
boundary due to inhibitor injection is shown in Figure 2-6. However, this technique is
due to the large amounts of inhibitor utilized. Moreover, penetration of inhibitors may
Pressure
Ti , Pi
Inhibitor Injection
Temperature
10
2.4 Review of Hydrate Reservoir Simulation Models
During the past decades, gas-hydrates started to appear as a research topic within
the agenda of researchers. Gas hydrate reservoirs are considered as the new potential
energy sources for the next century as they represent a potentially enormous supply of
natural gas. Since the production mechanisms from gas-hydrate reservoirs are not well
explore the feasibility of various production schemes. In the next section, some of these
The goal of this study was to estimate the contribution of gas-hydrates to the total
gas production of the reservoir. This study examines the dissociation of methane hydrates
in a square reservoir of uniform thickness. The hydrate and gas zones were stratified (50
ft thick for each zone) but were included in the same media. The hydrate zone was
assumed to be above the gas zone and the hydrate zone was impermeable (Figure 2-7).
The gas flow occurred only in the portion of reservoir containing gas.
Additionally, the hydrate dissociation occurred only at the hydrate-gas interface which
was at a uniform depth throughout the area extent of the reservoir. The model developed
in this study was a single-phase (gas) flow simulator and water production from hydrate
11
3,330 ft
Hydrate Zone
Gas Zone 50 ft
3,330 ft
In this study, heat for hydrate dissociation was provided from the sensible heat of
the hydrate reservoir. Furthermore, heat flow to hydrate-gas interface was allowed by
conduction only.
During gas production, the hydrate-gas interface moved toward the surface due to
hydrate dissociation. However, it was assumed that the interface still remained at a
uniform although changing depth. The temperature at the interface was assumed to be the
that point. The equations used in this model are listed in Table 2-1.
The amount of gas obtained from hydrates was compared to the total amount of
gas production from this reservoir for a period of 1,000 days. The results showed that the
percentage of produced gas coming from hydrate phase increased with time and it
contributed about 20 to 30 percent of the total gas production. Moreover, the hydrate-gas
interface became colder with time resulting in a greater heat flux to hydrate zone and
12
Table 2-1: Equations used in Holder et al.’s work
Equations
⎛ K (∇P) ⎞ d (φ / B)
Pressure distribution ∇⎜⎜ ⎟⎟ + Q =
⎝ μB ⎠ dt
1 ∂T
Temperature distribution ∇ 2T =
α ∂t
Hydrates dissociation ΔH D ⎛ • ⎞
q ∂T ⎜ mH ⎟
= k = = ⎜
2 ∂z 2 ⎜ A ⎟⎟
⎝ ⎠
production schemes for gas reservoirs which were in contact with a gas-hydrate cap. The
production of gas results in gas depressurization at the hydrate-gas interface. The primary
objective of the model was to determine whether or not the depressurization was capable
of initiating and sustaining hydrate dissociation at a practical rate. The second objective
of the study was to determine the contribution of gas from dissociated hydrates to total
gas production. The effect of water from hydrate dissociation on the production, such as
the increase of water-gas ratio and the decrease of relative permeability to gas, was also
investigated.
13
In this study, gas was produced from a single well located in the center of
reservoir. The well was completed throughout the gas zone. The gas production rate was
constant at a user-specified value. Besides, in free-gas zone, gas flow was radial toward
the well. The temperature at any point on the hydrate-gas interface was assumed to be the
equilibrium temperature for hydrate dissociation at a local pressure. Moreover, heat flow
from the interior of hydrate cap and the gas reservoir to the hydrate-gas interface was
possible by conduction only. In this model, the hydrate-gas interface moved upward
toward the Earth’s surface (due to hydrate dissociation) and the variation in interface
depth was smoothened out and an average interface depth was used to determine the
volume of the gas reservoir. The hydrate dissociation took place only at the hydrate-gas
interface and the interior of hydrate zone was assumed to be impermeable. The equations
For a wide range of realistic variation studies, the simulator indicated that hydrate
can be dissociated without an external energy source, i.e., the sensible heat of the
reservoir provided the necessary energy for hydrate dissociation. The water involved in
hydrate dissociation did not result in an undesirable high producing water-gas ratio and
did not impair gas flow to a notable degree. It was concluded that depressurization was a
feasible method for dissociating gas-hydrates found at depth greater than 2,300 ft (701 m)
14
Table 2-2: Equations used in Burshears et al.’s work
Equations
⎛ k k rw ⎞ d (φS w / Bw )
∇⎜⎜ (∇Pw − ρ w gz )⎟⎟ + Qw =
⎝ μ w Bw ⎠ dt
Pressure distribution
⎛ k k rg ⎞ d (φS g / B g )
∇⎜
⎜μ B
(∇Pg − ρ g gz )⎟ + Q g
⎟
=
⎝ g g ⎠ dt
Saturation relationship Sw + Sg = 1
Capillary Pressure
Pw + Pc = Pg
relationships
1 ∂T
Temperature Distribution ∇ 2T =
α ∂t
⎡ b⎤
Pe = exp ⎢a + ⎥ ( pure gas )
⎣ T⎦
Dissociation pressure
⎛P ⎞
ln⎜⎜ e ⎟⎟ = ∑ (A x
i i + Bi xi2 ) (mixed gas )
⎝ Po ⎠
HD = c + dT ( pure gas )
⎛ H ⎞
∑ (A x )
Dissociation enthalpy
ln⎜⎜ D ⎟⎟ = i i + Bi xi2 (mixed gas )
⎝ H DO ⎠
15
2.4.3 Model 3, Yousif, Abass, and Selim (1991) [7]
simulate the process of gas production from Berea sandstone samples containing methane
In all models previously proposed, flow equations in a hydrate zone were not
considered as a complement to the equation of mass and energy. One purpose of this
for both mass and momentum for each phase (gas, water, and hydrate) in the porous
medium. The equations used in the model are listed in Table 2-3.
The model was validated using the data collected in laboratory experiments. It
was found that the absolute permeability, relative permeability, and reaction rate constant
must be varied in order to obtain a satisfactory match of the model to the available
experimental data. For example, the relative permeability curve for water was shifted
upward about 20% to make the system more permeable to water. The relative
permeability curve for gas was shifted upward 3% to make the system slightly more
permeable to gas. Furthermore, the dissociation reaction rate constant (kd) was decreased
16
Table 2-3: Equations used in Yousif et al.’s work
Equations
∂ ⎛ ρ w kk rw ∂Pw ⎞ • ∂
⎜⎜ ⎟⎟ + m w = (φρ w S w ) ( water )
∂x ⎝ μ w ∂x ⎠ ∂t
∂ ⎛⎜ ρ g kk rg ∂Pg ⎞⎟ ∂
(φρ g S g )
•
Flow equations + mg = ( gas )
∂x ⎜⎝ μ g ∂x ⎟⎠ ∂t
• ∂
− mH = (φρ H S H ) (hydrate)
∂t
Saturation relationship Sw + Sg + SH = 1
Capillary pressure
Pc ( S w ) = Pg − Pw
relationships
•
m g = k d As (Pe − P )
Hydrate dissociation (
As = φ wg
3
/ 2k )1/ 2
• • Mg
mg = mH
NH M w + M g
17
2.4.4 Model 4, Chuang, Goodarz, and Duane (2001) [8]
analysis. The temperature and pressure distributions in porous layer of methane hydrate
and in the gas region were evaluated for different well pressures and reservoir
temperatures. In this study, it was assumed that the hydrate decomposition in a porous
medium does not occur in the entire volume, but takes place in a narrow region that can
be treated as a surface, the so-called decomposition front. This moving front separates the
volume of the reservoir into two zones with different phases. In the near-well zone, only
natural gas and water exist, whereas only the solid hydrate and natural gas exist in the
zone further away from the well (Figure 2-8). For a one-dimensional model, the
distribution of pressure in the layer was described using an analog of the classical Stefan
problem for melting. There were several important assumptions involved in this study.
One was that the pressure and temperature at any point on the decomposition front were
the equilibrium pressure and temperature for dissociation of methane hydrate. The
hydrate reservoir was also assumed to be porous and contain free natural gas. As the
dissociation front moves outward, heat must be supplied to the front because of the
dissociation, the movement of released water in the porous medium was assumed to be
negligible. The governing equations used in the model are listed in Table 2-4. The
production rate equation listed in the table was developed from the first three equations
using linearization technique. It should be noted here that the model used in this study
18
neglects the heat conduction in the entire reservoir. Thus, the energy balance at the
Impermeable rock
well
Hydrate + NG
Hydrate + NG
NG + water
Impermeable rock
19
In this work, authors were trying to find answers to two questions: (1) Can natural
gas be produced by depressurization through drilling a well into a hydrate reservoir? (2)
What are the parameters that control the natural gas production rate? In particular, is the
The results presented in this study guided the authors towards the following
conclusions:
• Under favorable conditions, natural gas can be produced from hydrate
• The natural gas production rate was controlled by the physical and thermal
• The required heat for hydrate dissociation could be supplied by the natural
gas convection from the exterior (hydrate and natural gas) portions of the
field.
linearized model), and they depend only on the reservoir conditions and
• For a fixed reservoir pressure and temperature, the well output decreases
and the motion of the decomposition front slows as the well pressure
increases.
• For fixed reservoir and well pressures, the gas production rate decreases
20
• For homogenous hydrate reservoir, the reservoir permeability significantly
affects the rate of convective heat transfer and consequently the rate of
• For a fixed porosity, the reservoir with higher permeability has higher
production rate and the decomposition front penetrates faster into the
presented as a new module for the TOUGH2 general-purpose simulator for multi-
component, multiphase fluid and heat flow in the subsurface. EOSHYDR2 can model
the non-isothermal gas release, phase behavior and flow of fluids and heat under typical
conditions of common natural gas hydrate deposits (i.e., in the permafrost and in deep
ocean sediments).
and dissociation. This model accounts for up to four phases (gas phase, liquid phase, ice
phase, and hydrate phase) and up to nine components. The mass and energy balance
equations and all correlations used in this work are listed in Table 2-5.
21
Table 2-5: Equations used in Moridis’ work
Equations
d
dt v∫n ∫F
M k dV = k
⋅ n dΓ + ∫ q k dV
Γn vn
Flow Equations
Mw =
β
∑ (φ ρ β S β X β )
≡ A ,G , I
w
+ M Hw
M m ,ν =
β
∑ (φ ρ β S β X β ν )
≡ A ,G , I
w,
+ M Hw,ν
ΔE k
q k = χ k Wk Ak k k exp( )( f ek (T ) − f k ), k ≡ m, v
RT
Hydrate Dissociation
ΔE k
qw = ∑
k ≡ m ,v
χ k N k Wk Ak k k exp(
RT
)( f ek (T ) − f k )
Dissociation Pressure ⎛ e2 ⎞
Pe (kPa) = exp ⎜⎜ e1 − ⎟
⎝ T ( K ) ⎟⎠
•
Dissociation Rate m g = k d As (Pe − P )
(Optional)
d (ln P)
Heat of Dissociation ΔH 0 = zRT 2
dT
F θ = − {(1 − φ ) K R + φ[ S H K H + S A K A + S G K G + S I K I ] }∇T
Heat flux + ∑ hβ Fβ
β ≡ A ,G
22
In this study, results from four test problems that explore different mechanisms
and strategies for production from methane hydrate accumulations were presented. The
results indicate that methane production from reservoirs containing methane hydrates can
improved when coupled with depressurization and the use of hydrate inhibitors.
developed to simulate two regimes of gas production from sediments containing methane
and the flow controlled regimes. In their work, authors implemented the finite difference
scheme and equations; they were implicit in water and gas saturation, pressure, and
temperature, and explicit in hydrate saturation. The equations used in this work are listed
in Table 2-6. The model could be used to fit laboratory-scale experimental data, but the
dissociation rate constant, the multiphase flow parameters and the heat transfer
• The parameter, Rτ, proposed in this work determines whether the process is
23
controlled if the permeability is high enough. The production rate does not change
• Gas production rate from a linear reservoir is more sensitive to the heat transfer
coefficient with the surrounding than the longitudinal heat conduction coefficient,
in 1-D simulations.
24
Table 2-6: Governing equations used in Sun et al.’s work
Equations
∂ ∂
(φ ρ g S g ) + ( ρ gν g ) = m& g
∂t ∂x
∂ ∂
Flow Equations (φ ρ w S w ) + ( ρ wν w ) = m& w
∂t ∂x
d
(φ ρ h S h ) = − m& h
dt
⎛ 9459 ⎞
Dissociation Pressure Pe = 1.15 exp ⎜ 49.3185 − ⎟
⎝ T ⎠
nw
⎛ s w /( s w + s g ) − S wirr ⎞
k rw = ⎜ ⎟
⎜ 1− S ⎟
⎝ wirr − S gr ⎠
Relative Permeability
ng
⎛ s g /( s w + s g ) − S wirr ⎞
k rg = ⎜ ⎟
⎜ 1− S ⎟
⎝ wirr − S gr ⎠
− nc
Pc ⎛ s w /( s w + s g ) − S wirr ⎞
Capillary Pressure = h ( s ) = ⎜⎜ ⎟⎟
1 − S swirr
c w
Pce ⎝ ⎠
∂
∂t
[
φ ( ρ g S g H g + ρ w S w H w + ρ h S h H h ) + (1 − φ ) ρ S H S ]
Energy Balance
∂ ∂ ⎛ ∂T ⎞
+ ( ρ gν g H g + ρ wν w H w ) = ⎜λ ⎟ + q
∂x ∂x ⎝ ∂x ⎠
25
2.4.7 Model 7, Sun and Mohanty (2005) [10]
In this work, a methane hydrate reservoir simulator was developed based on the
assumption of equilibrium phase transition. Two components (methane and water) and
four phases (hydrate, gas, aqueous, and ice) are considered in the proposed model. Mass
transport including two-phase flow and the molecular diffusions, and heat transfer
through convection and conduction are included in the mathematical model. The
governing equations are discretized using finite difference method and are solved with
the Newton-Raphson method in a fully implicit manner. The equations used in this work
In this work, three examples were presented about the hydrate accumulation in
offshore sediments and gas production from hydrate reservoir using depressurization and
sediments are consistent with the predictions by previous researchers, and the results of
gas production simulation can be readily used in future economical feasibility analysis of
26
Table 2-7: Equations used in Sun and Mohanty’s work
Equations
∂ ⎡ ⎤
⎢φ
∂t ⎣
∑ ρ j S j w mj ⎥ + ∇ ⋅ F m = q m (methane)
j = H ,G , A , I ⎦
Mass Balance
∂ ⎡ ⎤
⎢φ
∂t ⎣
∑ ρ j S j w wj ⎥ + ∇ ⋅ F w = q w ( water )
j = H ,G , A , I ⎦
∂ ⎡ ⎤
Energy Balance ⎢φ
∂t ⎣
∑ ρ j S jU j + (1 − φ ) ρ RU R ⎥ + ∇ ⋅ F E = q E
j = H ,G , A , I ⎦
φ e k0
Capillary Pressure Pc = Pce S ( ) e* − nc
A where Pce = Pce 0
φ0 k
Local Absolute 2β
k φ e ⎛ φ e (1 − φ 0 ) ⎞
= ⎜ ⎟ where φ e = φ ( S g + S a )
Permeability k0 φ 0 ⎜⎝ φ 0 (1 − φ e ) ⎟⎠
k rG = k rG
0
( S Ge* ) nG
Relative Permeability
k rA = k rA
0
( S Ae* ) n A
27
2.4.8 Model 8, Moridis (2005) [11]
In this work, the developed simulator TOUGH-FX was used. The simulator
includes both an equilibrium and kinetic model of hydrate formation and dissociation.
The main mathematical equations shown in Table 2-5 (see section 2.4.5) are still used in
this model. Unlike the previous developed model, the effect of hydrate saturation on
capillary pressure between free gas and aqueous phases is also taken into account in this
simulator. The additional equations used in this model are listed in Table 2-8.
The main objectives of this study were to evaluate the production potential of
the factors and conditions affecting it. Two different systems were studied. The first
system has hydrate and water in the upper hydrate zone and is hence referred to as a
Class 1W deposit. The second system has hydrate and free gas in the upper hydrate zone
• Large volume of gas is readily produced at high production rate from class 1
• Wellbore heating of the entire production interval and substantial part of the
• In class 1 deposits, up to 65% of the production rate and 45% of the cumulative
28
Table 2-8: Additional equations in the modified Moridis’ model (2005)
Equations
k 1
[
= k ra ( S a = 1 − S s ) + k rg ( S g = 1 − S s )
k0 2
]
Intrinsic permeability or
n
k ⎛ φ − φc ⎞
= ⎜⎜ ⎟⎟
k 0 ⎝ φ0 − φc ⎠
k0 φ
Pcap ( S a ) = Pcap , 0 ⋅
k φo
or
Capillary pressure
Pcap ( S aA , S H ) = H ( S H ) ⋅ F ⋅ Pe ⋅ ( S aA ) c
interface that first appears at the top of the hydrate interval and advances
downward.
monotonically increases with time. Up to 75% of the rate of gas production and
54% of the cumulative gas production are replenished by gas released from
29
hydrate dissociation. Class 1G deposits are slower to respond than Class 1W
reaction. However, the temperature drops are mild indicating that th risk of
• Water production remains very low during the long-term gas production from
complex processes that play a critical role in production from gas hydrate systems.
The new models to describe the effect of the presence of hydrates on the
lensing which is analogous to the process called frost heave. A stronger capillary
suction in the regions with larger hydrate content, due to smaller effective pore
sizes, causes water to be drawn toward regions with larger hydrate saturation
This results in the development of banded patterns with alternating layers of large
30
2.4.9 Model 9, Hong and Pooladi-Darvish’s (2005) [12]
In this work, a 2D cylindrical simulator for gas hydrate reservoirs was developed.
The model includes the equations for two phase (gas-water) flow, conductive and
convective heat transfer, and intrinsic kinetics of hydrate decomposition. The developed
simulator was used to study a hydrate reservoir where the hydrate-bearing layer overlies a
free-gas zone (class 1 hydrate accumulation). A production well was drilled and
completed in the middle of free-gas zone. In this study, the impact of the overlying
hydrate in improving production performance of the underlying gas reservoir and the
effect of various parameters on gas production behavior were examined. The equations
used in this model are listed in Table 2-9. The conclusions from this examination are
summarized below:
• The presence of gas hydrate on top of a free gas reservoir provides a significant
• A larger thermal conductivity yields more heat transfer from the cap- and base-
• Heat conduction and sensible heat of hydrate zone plays dominant role in gas
generation by depressurization.
31
Table 2-9: Equations used in Hong and Pooladi-Darvish’s work
Equations
r ∂
− ∇ ⋅ (ρ l u lD ) + q& ml + g& l = (φρ l S l )
∂t
Flow Equation where
r kk ∂Φ l
u lD = − rl (l = g , w ; D = r , z )
μ l ∂D
Energy balance
∂
+ Q& in =
∂t
[
(1 − φ ) ρ RU R + φ ( S H ρ H U H + S w ρ wU w + S g ρ g U g ) ]
where subscript D = r , z
g& g = k d M g Adec ( f e − f g )
Mw
g& w = g& g N H
Mg
MH
g& H = − g& g
Kinetics Equations Mg
where :
⎛−E⎞
k d = k d0 exp⎜ ⎟
⎝ RT ⎠
Adec = φ S H AHS
• The dissociation front remains at equilibrium through out the simulation which
• Lower wellbore pressure can increase hydrate dissociation and gas production
rates. However, reducing too much bottom-hole pressure may cause the reservoir
32
temperature to fall below the freezing point leading to the formation of ice which
In this work, the first tank-type material balance model for a hydrate-capped gas
reservoir, Class 1 hydrate accumulations, has been developed. The material balance
equation is developed by analytically and simultaneously solving the mass and energy
balance equations. The solutions of the equations provided the average reservoir pressure
and the released gas from hydrate dissociation as a function of cumulative produced gas,
assumptions:
• One mole of hydrate decomposes to one mole of methane gas and NH moles of
• The expansions of water and rock are modeled using average compressibility
• The porosity and initial saturations are uniform throughout the reservoir.
33
These equations are nonlinear and require numerical technique to obtain solutions.
Therefore, to obtain a model using analytical techniques, the following assumptions are
made:
that the pressure and temperature in the system are instantaneously uniform.
negligible.
formations (cap and base rocks) remain constant during the production period.
For energy balance calculation, the sensible heat of rocks and fluids surrounding
hydrate in porous media and heat conducted from the cap and base rocks are the two
sources of heat available for hydrate dissociation. The governing equation for heat
transfer from the cap rock is determined by conservation of energy using Fourier’s law of
heat conduction[14]. The equations used in this model are listed in Table 2-10.
34
Table 2-10: Equations used in Gerami and Pooladi-Darvish’s work
Equations
Material balance P (t ) p ⎛ G p (t ) − G g (t ) ⎞⎟
= i ⎜1 −
Z (t ) z i ⎜⎝ Gf ⎟
⎠
(Simplified)
dT ⎛ ΔH ⋅ ρ H ⎞ dG g ⎛ ρ ⋅ c p ⋅ Vb ⎞ dTse
Energy balance = ⎜⎜ ⎟⎟ + ⎜⎜ ⎟⎟
dz z =0 ⎝ k cr ⋅ A ⋅ E H ⎠ dt ⎝ k cr ⋅ A ⎠ dt
⎛ β ⎞
Dissociation Pressure Pe = exp ⎜⎜ λ − ⎟⎟
⎝ Tse ⎠
Temperature gradient ∂T t
= 2 ⋅ b(t )
at hydrate interface ∂z z =0 π αr
EH A H ρ c p ⎛ 4 ρ r c pr α r ⎞
Gas dissociation rate q g (t ) = b(t ) ⎜ t + 1⎟
ρ H ΔH ⎜ ρc H π ⎟
⎝ p ⎠
ρ cp ⎛ 8ρ r c pr α r ⎞
Hydrate recovery Rh (t ) = b(t ) ⎜ t + 1⎟ ⋅ t
φ S H ρ H ΔH ⎜ 3ρ c H π ⎟
⎝ p ⎠
−1
⎛ 1 − ( p oe Z i ) /( pi Z oe ) ⎞
⎜ Gf ⎟
⎜ Ti − Toe ⎟
Temperature parameter b(t ) = q w ⎜ ⎟
⎜ + E H A H ρ c p ⎛⎜ 8 ⎞⎟
⎟
k cr c pr ρ r t + 1 ⎟
⎜ ρ H ΔH ⎜⎝ 3ρ c p H π ⎟
⎝ ⎠⎠
⎡ q E A ⎛ 8 t avg ⎞⎤
(Ti − Toe ) ⎢ w − H ⎜ k cr ρ r c pr + Hρc p ⎟⎥
Initial free gas-in-place ⎢ bmatch ρ H ΔH ⎜ 3 π ⎟⎥
⎣ ⎝ ⎠⎦
Gf =
1 − ( poe Z i ) /( pi Z oe )
1 ⎡⎛ q 1 − ( p oei Z i ) /( p i Z oe ) ⎞ ρ H ΔH 8 t avg ⎤
Hydrate cap thickness H= ⎢⎜⎜ w − Gf ⎟⎟ − k cr ρ r c pr ⎥
ρ cp ⎢⎣⎝ bmatch Ti − Toe ⎠ EH A 3 π ⎥⎦
35
In this study, the developed model was used to study (using its forward solution
mode) the effects of important parameters (including reservoir porosity, rock thermal
conductivity, production rate, hydrate zone thickness, free-gas zone thickness, reservoir
characteristics of hydrate reservoirs. The initial free gas-in-place, Gf, and the initial
thickness of the hydrate cap, H, can be determined using the inverse solution mode of the
model.
• The temperature in hydrate zone drops in a close-to-linear manner and the rate of
• The dissociation rate of hydrate phase in the reservoir strongly depends on the
used for hydrate-capped gas reservoirs if it is modified for the average reservoir
being examined for commercial gas production from gas hydrate reservoirs. In this study,
36
an alternative of gas production from gas hydrate reservoirs using CO2 injection is
favored over CH4 in hydrate, 2) the heat released by formation of CO2-hydrate is 20%
greater than the hat needed for dissociating CH4-hydrate, 3) refilling pore space with
CO2 from atmosphere while simultaneously producing clean-burning natural gas. This
study focused on the evaluation of a set of optimum parameters for methane recovery
Four mass conservation equations and one energy conservation equation are used
in STOMP simulator to describe the subsurface hydrate system where the conserved mass
components are water, CH4, CO2, and salt (NaCl). Aqueous, liquid CO2, and free gas are
assumed to be mobile phases whereas hydrate, ice, precipitated salt, and host porous
media are assumed to be immobile phases in the model. The equations used in this model
37
Table 2-11: Equations used in STOMP-HYD simulator
Equations
∂⎛ ⎞
Energy conservation ⎜ ∑
⎜ (φρ γ sγ u γ ) + (1 − φ ) ρ s u s ⎟⎟ = − ∑ (∇(hγ Fγ ) )
∂t ⎝ γ =l , g ,n ,h ,i , p ⎠ γ =l , g ,n
∂⎛ ζ ⎞
⎜ ∑
⎜
∂t ⎝ γ =l , g ,n ,h ,i , p
(φρ γ s γ ω γ (
) ⎟⎟ = − ∑ ∇(ωγζ Fγ ) − ∑ ∇( J γζ )) ( )
Mass conservation ⎠ γ =l , g , n γ =l , g
diffusion-dispersive ρ γ k rγ k i
Fγ = − (∇Pγ + ρ γ g z g ) where γ = l , g , n
flux and advective
μγ
Mζ ζ
Diffusive mass flux
J ζ
γ = − φτ γ ρ γ sγ γ
Dγ ∇χ γζ ( ) for γ = l and ζ = w, a, o, s
M
for γ = g and ζ = w, a, o
More information about STOMP simulator and its constitutive equations can be
The study was initially carried out on the simple 1-D reservoir models to support
the hypothesis of CO2-microemulsion injection for methane recovery from the gas
hydrate reservoirs. In this step, the effects of various parameters such as CO2-
formation rock were examined. In this case, CO2-microemulsion injection was considered
38
from the West boundary of the system whereas the methane was produced from the East
boundary of the system. The conclusions from this examination are summarized below:
• Nearly pure methane was produced which is consistent with the observations
• The results indicate that higher hydraulic conductivity results in much faster
From the results from 1-D simulation study, it indicates the feasibility of using
CO2-microemulsion injection for methane recovery. The more complex 2-D reservoir
models were then examined. In this case, the range of CO2-microemulsion temperature
and concentration for optimizing methane production were determined. Similarly, CO2-
microemulsion injection was considered from the West boundary of the system whereas
the methane was produced from the East boundary of the system as shown in Figure 2-9.
39
Injected CO2-slurry Produced gas
10
y
5
x
1 2 3 4 5 6 7 8 9 10
1 z
28 oC.
40
2.4.12 Model 12, Uddin, Coombe, and Gunter’s Model [17]
This work describes a new unified kinetic model, couple with a compositional
thermal reservoir simulator (CMG STARS), that can simulate the dynamics of CH4 and
CO2 hydrates formation and decomposition in a geological formation. The kinetic model
contains two mass transfer equations: 1) formation equation transfers gas and water into
hydrate, and 2) decomposition equation transfers hydrate into free gas and water.
In this work, the proposed model was evaluated in two case studies. In case 1, a
single well natural hydrate reservoir was examined for studying the kinetics of CH4 and
CO2 hydrates formation and decomposition In this case, the a detailed parameter
sensitivity analysis was performed. In case 2, a multi-well reservoir was examined for
studying the unified kinetic model to demonstrate the flexibility of CO2 sequestration in a
natural hydrate reservoir with potential enhancement of methane recovery. In this case,
limitations of the numerical model to simulate the CO2 sequestration processes in gas
hydrate reservoirs were identified and suggestions for future model development were
recommended.
The unified kinetic model which can handle CH4 and CO2 hydrates formation
and/or decomposition has been developed in this work. In this model, a system of first
order rate equations (kinetic equations for water, methane, CO2, CH4 hydrate, and CO2
hydrate) was formulated for the kinetics of hydrate formation and decomposition. The
41
Table 2-12: Hydrate formation and decomposition rate models used in
Uddin et al.’s work
Equations
Rate of hydrate dc H E ⎛ 1 ⎞
= A ⋅ exp(− )(φ S a ρ a )(φ S H ρ H )( y i p g )⎜⎜1 − ⎟⎟
formation dt Form RT ⎝ K ( p, T ) ⎠
Rate of hydrate dc H E ⎛ 1 ⎞
= B ⋅ (1 + φ S H )⋅ exp(− )(φ S a ρ a )( y i p g )⎜⎜1 − ⎟⎟
decomposition
dt Decomp RT ⎝ K ( p, T ) ⎠
• The effect of kinetic rate constants to the CH4 and CO2 hydrates decomposition
significant. The rapid pressure drawdown in high permeable system results a large
• The hydrate formation with gas injection reduces the effective porosity and
permeability around the injection well. The continuation of gas injection could
• The unified gas hydrate model effectively describes the dynamics of gas hydrate
42
hydrodynamics conditions. The CH4 and CO2 hydrates can form and are stable
under specific ranges of elevated gas pressure and temperature conditions. This
The overall genealogy of the gas hydrate reservoir simulation models reviewed in
this thesis are shown in Figure 2-10, Figure 2-11, and Table 2-13.
43
Note that one of the objectives of this study is to develop a numerical simulation
model of gas hydrate reservoir. Therefore, the literature review of this study focused
44
Table 2-13: Model comparison
Models Coordinate system Model type Field scale Heat of conduction Heat of convection Kinetics model Equilibrium model Production scheme
Holder, G. and Angert 3-D Cartesian Numerical Y Y N N Y Depressurization
Burshears, O’Brien, Malone 3-D Cartesian Numerical Y Y N N Y Depressurization
Yousif, Abass, Selim 1-D Cartesian Numerical N N N Y N Depressurization
Chuang, Goodarz, and Duane 1-D Cartesian Analytical Y N N N Y Depressurization
Moridis 3-D Cylindrical/Cartesian Numerical Y Y Y Y Y Depressurization/Thermal stimulation
Sun, Nanchary, Mohanty 1-D Cartesian Numerical Y Y Y Y N Depressurization
Sun, Mohanty 3-D Cartesian Numerical Y Y Y N Y Depressurization
Moridis (Modified) 3-D Cylindrical/Cartesian Numerical Y Y Y Y Y Depressurization/Thermal stimulation
Hong, Pooladi-Darvish 2-D Cylindrical Numerical Y Y Y Y N Depressurization
Gerami, Pooladi-Darvish Non-dimensional Analytical Y Y N N Y Depressurization
Phale, Zhu 3-D Cartesian Numerical Y Y Y N Y CO2 sequestration
Uddin, Coombe, Gunter 3-D Cartesian Numerical Y Y Y Y N CO2 sequestration
This study 3-D Cartesian Numerical Y Y Y N Y Depressurization
45
CHAPTER 3
PROBLEM STATEMENT
hydrate equilibrium model to be coupled with the formulation of fluid flow in porous
media. The methane-hydrate equilibrium model predicts whether methane hydrate phase
is stable at any particular pressure and temperature values. In addition, the dissociation
of methane hydrate is an endothermic reaction. For this reason, the system cannot be
There are three main production methods for methane-hydrate reservoir which are
constraint for all recovery methods is their economic implication. The thermal
stimulation method has the terrible flaw of the tremendous amount the associated energy
losses taken by the surroundings and the injection systems. Because of this, only a
fraction of the injected energy finally meets the hydrate; and from this fraction, part of it
is taken by the matrix of the rock. Moreover, penetration of hot fluids or inhibitors may
depressurization has stood as the most viable recovery method from an economic
prospective.
temperature is likely to be observed because some energy is used for the dissociation of
46
methane hydrate. The decrease of reservoir pressure will start the dissociation of methane
hydrate in the reservoir when reservoir pressure becomes equal to or lower than the
dissociation pressure of methane hydrate, the pressure which methane hydrates become
hydrate exponentially decreases with temperature drop. If the decrease of the dissociation
reservoir is faster than the decrease of reservoir pressure, the dissociation of methane
hydrate will not occur. Therefore, the changes of temperature and pressure of the system
All the reservoir simulation studies of gas-hydrate reservoirs which have been
done so far focused on the feasibility study of gas production (by depressurization and
heat stimulation) from gas-hydrate reservoirs. In this study, the effects of production
parameters such as well-penetrating location and well spacing on the gas production
efficiency are examined. This work mainly focuses on a constant bottom-hole pressure
production scheme.
well bore model has not been incorporated into the simulator resulting in the limitations
of the simulator to be used for studying gas production characteristics from different
production strategies. For a multilayer well with constant gas flow rate specification case,
the gas flow rate of each layer must be specified in the TOUGH FX simulator and these
specified flow rates do not change for the entire simulation which is not accurate. The
incorporation of wellbore model improves the accuracy of the simulation results. In the
47
case that a wellbore model is incorporated in the model, the gas and aqueous flow rates of
each layer change corresponding to the mobilities of aqueous and free gas phases in each
layer, and the total gas flow rate is equal to the specification value. Moreover, the study
production schemes is developed in this work and it is used for exploring production
48
CHAPTER 4
water and suitably sized gas molecules. Water (host) molecules form unstable lattice
structures, upon hydrogen bonding, with several cavities. Gases (guest) molecules
can occupy the lattice cavities and when the minimum numbers of cavities are
occupied, the crystalline structure becomes stable even at temperatures well above the
The two common forms of gas hydrates known as structure I (sI) and II (sII) have
been investigated using X-ray diffraction methods by Von Stackelberg and Müller
(1954). They found that the unit cell of hydrate structure I is a 12 oA cube consisting of
46 water molecules which has two cavity types: small and large cavities. There are two
small cavities and six large cavities in a unit cell of hydrates structure I. The small cavity
suggested the nomenclature description ( nimi ), where ni is the number of edges in the I
type of face and mi is the number of faces of i type. Thus the small cavity for hydrate
structure I is denoted 512 because it has twelve pentagonal and two hexagonal faces,
whereas the large cavities for this structure is denoted 51262 because it has twelve
49
pentagonal and two hexagonal faces. The smaller cavities are almost spherical, whereas
The unit cell of structure II (sII), which is a 17.3 oA cube with 136 water
molecules, also contains two types of cavities. The 16 small cavities are distorted
pentagonal dodecahedra (512), and the 8 large cavities are hexadecahedra (51264) having 4
hexagonal faces and twelve pentagonal faces. The latter cavities are almost spherical in
shape.
requires both large and small molecules to stabilize the structure. They proposed that the
new structure is known as structure H (sH). According to Ripmeester, the unit cell
consists of 34 water molecules forming a hexagonal lattice. This structure has three
different types of cavities. It consists of three 512 cavities which is common to all known
hydrate structures, two new 12 faces 435663 cavities and one new large 51268 cavity. The
435663 cavity has three square faces, six pentagonal faces, and three hexagonal faces. The
51268 cavity has twelve pentagonal faces and eight hexagonal faces. The first two cavities
accommodate the small molecules. The large cavities can accommodate even larger
molecules, so that the molecules in the size range of 7.5 to 8.6 oA can potentially form
this type of gas hydrate structure. The cavities in each hydrate structure are shown in
Figure 4-1.
50
Figure 4-1: Structures of cavity types in hydrate structures[18]
In all hydrate structures, each cavity can contain at most one guest molecule.
occupied by more than one guest molecule. The comparison between the sizes of cavities
regions) such as cyclopropane and trimethylene oxide, which have diameters of 5.8 oA
and 6.1 oA, respectively, may form either structure I or structure II.
51
Cavities Occupied
3 oA
No Hydrates
Ar
512 + 51264
Structure II
4 oA Kr
N2
O2
CH4
Xe, H2S
Structure I 512 + 51262
5 oA CO2
C2H6
Structure I 51262
c-C3H6
o
6 A (CH2)3O
C3H8 51264
iso-C4H10 Structure II
o
7 A
n-C4H10
Structure H
o
8 A
Figure 4-2: Comparison between guest molecule and cavity sizes[3]
52
4.2 Hydrate Phase Equilibrium Models
used for calculating the rate of gas and water released from the dissociation reaction.
methods. The first method uses the statistical thermodynamic concept to set up the
model. Another method uses the regression analysis to set up the models from
available experimental data. In this chapter, some empirical correlations and the
The model used in this study was based on classical statistical thermodynamics
proposed by Van der Waals and Plateeuw[19] in 1959. The method for predicting
hydrate and water rich (or ice) phases. Using μβ, the chemical potential of an unoccupied
hydrate lattice, as the reference state, the equilibrium condition can be written as:
μ β (T , P) − μ w (T , P) = μ β (T , P) − μ H (T , P) 4.1
Δμ w = Δμ H 4.2
where μ w is the chemical potential of water in water rich (or ice) phase
53
4.2.1.1 Calculation of Δμw
⎛ Δμ ⎞ ⎛ Δh ⎞ ⎛ ΔV ⎞
d ⎜ w ⎟ = − ⎜ w2 ⎟ dT + ⎜ w ⎟ dP
⎝ RT ⎠ ⎝ RT ⎠ ⎝ RT ⎠ 4.3
where Δhw is the enthalpy difference between water (or ice) and empty-
hydrate lattice
ΔVw is the volume difference between water (or ice) and empty-
hydrate lattice
Δμ w Δμ wo T
⎛ Δh ⎞
P
⎛ ΔVw ⎞
= − ∫ ⎜ w2 ⎟ dT + ∫ ⎜⎝ ⎟ dP − ln ( x w γ w )
RT RTo To ⎝ RT ⎠ 0
RT ⎠ 4.4
The first term on the right hand side of the above equation represents the chemical
potential difference between the theoretical empty hydrate and liquid water (or ice) at its
reference state. The second term accounts for the change in chemical potential difference
due to temperature at zero pressure. The third term accounts for the change in chemical
54
potential difference due to pressure. The last term accounts for the solubility of gas in
water rich phase. The activity coefficient of water rich phase accounts for the non-
idealities of the water rich phase. The activity coefficient of water rich phase was
assumed to be 1.0 when only gas and water systems were studied because the gas
T
Δhw = Δhwo + ∫ Δc
T0
pw dT
4.5
where Δhwo is the reference enthalpy difference between water (or ice) and
empty-hydrate lattice
empty-hydrate lattice
Δc pw = Δc opw + b (T − T0 ) 4.6
b is a constant
The values of Δμ wo , Δh pw
o
, and Δc pw used in this study are listed in Table 4-1.
The value of ΔVw in equation (4.4) is a constant and it depends on the type of phase
55
Table 4-1: Values of Δμ wo , Δh pw
o
, and Δc pw used in this study[3]
Parameters T0 = 273.15 K
Δμ (J/mol)
o
w
1220
T > T0 -4,300
Δ h wo (J/mol-K)
T ≤ T0 1270
T > T0 -38.12 + 0.141 (T- T0)
Δc pw (J/mol-K)
T ≤ T0 3.20 + 0.121 (T- T0)
The value of mole fraction of water in water rich phase in the last term in equation
nc
xw = 1 − ∑x
k =1
k 4.7
The variable xk is the mole fraction of dissolved gas component k in water rich
expression[3]:
fk
xk =
H k ⋅ exp( PVk∞ / RT ) 4.8
Here Vk∞ is the volume of gas component k in water at infinite dilution. The value
H k0 H1 H2 H3
− ln ( H k ) = + k + k ln (T ) + k T 4.9
R RT R R
The units of Hk and T in this equation are “atm” and “Kelvin”, respectively. The
56
Table 4-2: Henry’s constants and Vk∞ values[3]
Component Ho H1 H2 H3 V∞ (cm3/mol)
Methane -365.183 18,016.7 49.7554 -0.000285 32
Ethane -533.392 26,565.0 74.6240 -0.004573 32
Propane -628.866 31,638.4 88.0808 0.000000 32
i-Butane 190.082 -4,913.0 -34.5102 0.000000 32
n-Butane -639.209 32,785.7 89.1483 0.000000 32
H2S -297.158 16,347.7 40.2024 0.002571 32
N2 -327.850 16,757.6 42.8400 0.016765 32
CO2 -317.658 17,371.2 43.0607 -0.002191 32
The value of ΔμH can be calculated from the following equation purposed by Van
n nc
Δμ H = − RT ∑ν j ln (1 − ∑ y ij )
j i
4.10
Here yij is the fractional occupancy of the cages defined as the fraction of the
cavity type “j” by the guest molecule “i”. The parameter ν j is the number of cavity type
“j” per water molecule in the hydrate structure. For structure I of hydrate lattice, there
are two cavity types (small and large cavities). The value of ν 1 (small cavity) and ν 2
(large cavity) for structure I of hydrate lattice are 1/23 and 3/23, respectively. The
parameters “nc” and “n” in this equation are the number of guest components and the
number of cavity types in the hydrate structure. The value of yij is calculated from the
57
C ij f i
y ij = nc
1 + ∑ C kj f k 4.11
k
where Cij is the Langmuir constant of hydrate former “i” in cavity type “j”
In this study, the Peng-Robinson Equation of State is used for evaluating the gas
phase fugacity of guest component. Van der Waals and Platteeuw[19] suggested the use of
Lennard-Jones potential to represent the interaction between enclathrated gas and hydrate
satisfactory only for guest components which are small spherical molecules. The use of
Kihara Potential function is recommended[3] since it gives better results for larger poly-
atomic and rod-like molecules. Van der Waals and Platteeuw assumed that only nearest
neighbor water molecules had an effect on the energy of the enclathrated gas molecule[19].
The Kihara parameter values given by Sloan[3] are used in this study. The values of
Kihara parameters for methane are listed in Table 4-3. The values of radius (R) and
coordination number (Z) are listed in Table 4-4. The Langmuir constant for component
R − ai
4π ⎛ Wij (r ) ⎞ 2
C ij =
kT ∫
0
exp⎜⎜ −
⎝ kT ⎟⎠
⎟ r dr
4.12
where Wij is the cell potential for guest component “i” in cavity “j”
58
⎡ σ 12 a σ6 a ⎤
Wij (r ) = 2 Z j ε i ⎢ 11i (δ ij10 + i δ ij11 ) − 5i (δ ij4 + i δ ij5 ) ⎥
⎣⎢ R j r Rj Rj r Rj ⎦⎥ 4.13
where,
1⎡ r a r a ⎤
δ ijm = ⎢(1 − − i ) − m − (1 + − i ) −m ⎥
m ⎣⎢ Rj Rj Rj Rj ⎦⎥ 4.14
Superscript m in equation (4.14) is equal to 4, 5, 10, and 11. Rj and Zj are the
cavity radius and coordination number for cavity type “j”, respectively. Parameters σ i ,
ai , and ε i are the Kihara parameters for guest component “i”, and r is the distance of the
F ( P, T ) = Δμ w − Δμ H = 0 4.15
⎡ Δμ wo T ⎛ Δhw ⎞ ⎤
F ( P, T ) = ⎢ − ∫ ⎜ 2 ⎟ dT ⎥ T + ΔVw P − RT ln ( x wγ w )
⎣⎢ To To ⎝ T ⎠ ⎥⎦
n nc
+ RT ∑ν j ln (1 − ∑ y ij ) = 0
j i
4.16
In this study, equation (4.16) is used to calculate the hydrate association pressure
59
F ( P, T )
P k +1 = P k −
F ' ( P, T ) 4.17
where F ' ( P, T ) is the derivative of F(P,T) with respect to pressure. The solution
The comparison between the dissociation pressures obtained from the statistical
60
Methane Hydrate Equilibrium (Lw-H-V)
100000
Pressure (kPa)
1000
10
175 195 215 235 255 275 295 315
Temperature (K)
Figure 4-3: Methane hydrate dissociation pressure from statistical thermodynamic model and experimental data[3]
61
4.2.2 Empirical Correlations
developed during the past decades. Some of these models are summarized below:
⎛ e ⎞
Pe = exp ⎜ e1 + 2 ⎟
⎝ T ⎠ 4.18
Here Pe is the dissociation pressure (kPa), T is temperature (K), and e1 and e2 are
empirical constants. The values of empirical constants in the equation for methane
62
4.2.2.2 Makogon’s correlation[23]
log10 Pe = a (T − T0 ) + b (T − T0 ) 2 + c 4.19
composition. The units of Pe and T are Pascal and Kelvin, respectively. The values of
equation:
ln Pe = a T 5 + b T 4 + c T 3 + d T 2 + e T + f 4.20
The units of Pe and T in this equation are MPa and K, respectively. The
63
Table 4-7: Empirical constants in Moridis’ correlation for methane hydrate[24]
Temperature (K)
Empirical constants
150 K ≤ T ≤ 273.15 K 273.15 < T ≤ 315 K
-10
a 1.09882180475307x10 8.86065316687571x10-8
b -1.03669656828834x10-7 -1.30465829788791x10-4
c 3.85413985900724x10-5 7.67559117787059x10-2
d -7.27291427030502x10-3 -2.25540264493806x10 1
e 7.76302133739303x10-1 3.31018213397926x10 3
f -4.38921173434628x10 1 -1.94138504464560x10 5
64
Methane Hydrate Equilibrium (Lw-H-V)
100000
Pressure (kPa)
1000
10
175 185 195 205 215 225 235 245 255 265 275 285 295 305 315
Temperature (K)
Figure 4-4: Methane hydrate dissociation pressures from various empirical correlations and experimental data[3]
65
CHAPTER 5
MODEL DEVELOPMENT
developed using mass and energy conservation equations for the three-dimensional
x
y
z
In this study, the system was modeled as a four-phase (gas, water, ice, and hydrate
phases) system where only two components (methane and water) were to be found at any
point in the system. The amount of methane dissolved in aqueous phase and the amount
of water vapor in free gas phase were taken into account. However, ice phase was
assumed to be methane-free. Only aqueous and free gas phases were mobile. The phase
66
change between aqueous and ice phase were assumed to be instantaneous once freezing
conditions were attained. Rocks were assumed to be rigid. In this model, the dissociation
of methane hydrate was assumed to be instantaneous. The model can handle phase
Methane
∂ ⎡ Ax k x k rg ⎛ ∂Φ g ⎞ A k k ⎛ ∂Φ a ⎞⎤
⎢ ym ρ g ⎜⎜ ⎟⎟ + x m ρ a x x ra ⎜ ⎟⎥ Δx
∂x ⎣⎢ μg ⎝ ∂x ⎠ μ a ⎝ ∂x ⎠⎦⎥
∂ ⎡ Ay k y k rg ⎛ ∂Φ g ⎞ Ak k⎞⎤ ⎛ ∂Φ a
+ ⎢ ym ρ g ⎜⎜ ⎟⎟ + x m ρ a y y ra
⎟⎟⎥ Δy⎜⎜
∂y ⎢⎣ μg ⎝ ∂y ⎠ μa ⎠⎥⎦ ⎝ ∂y
∂ ⎡ Az k z k rg ⎛ ∂Φ g ⎞ A k k ⎛ ∂Φ a ⎞⎤ Qm
+ ⎢ ym ρ g ⎜⎜ ⎟⎟ + x m ρ a z z ra ⎜ ⎟⎥ Δz +
∂z ⎢⎣ μ g ⎝ ∂z ⎠ μ a ⎝ ∂z ⎠⎥⎦ 5.615
⎛ V ⎞ ∂
= ⎜ b ⎟ ( y mφ S g ρ g + x m φ S a ρ a + N mφ S H ρ H ) 5.1
⎝ 5.615 ⎠ ∂t
Water
∂ ⎡ Ax k x k rg ⎛ ∂Φ g ⎞ ⎞⎤
A k k ⎛ ∂Φ a
⎢(1 − y m ) ρ g ⎜⎜ ⎟⎟ + (1 − x m ) ρ a x x ra ⎜
⎟⎥ Δx
∂x ⎢⎣ μg ⎝ ∂x ⎠ μ a ⎝ ∂x
⎠⎥⎦
∂ ⎡ Ay k y k rg ⎛ ∂Φ g ⎞ A k k ⎛ ∂Φ a ⎞⎤
+ ⎢(1 − y m ) ρ g ⎜⎜ ⎟⎟ + (1 − x m ) ρ a y y ra ⎜⎜ ⎟⎥ Δy
∂y ⎣⎢ μ g ⎝ ∂y ⎠ μ a ⎝ ∂y ⎟⎠⎦⎥
∂ ⎡ Az k z k rg ⎛ ∂Φ g ⎞ A k k ⎛ ∂Φ a ⎞⎤ Qw
+ ⎢(1 − y m ) ρ g ⎜⎜ ⎟⎟ + (1 − x m ) ρ a z z ra ⎜ ⎟⎥ Δz +
∂z ⎣⎢ μg ⎝ ∂z ⎠ μ a ⎝ ∂z ⎠⎥⎦ 5.615
⎛ V ⎞ ∂
= ⎜ b ⎟ ((1 − y m ) φ S g ρ g + (1 − x m ) φ S a ρ a + φ S I ρ I + N wφ S H ρ H ) 5.2
⎝ 5.615 ⎠ ∂t
The parameters ρ g and ρ a are molar density of free gas and aqueous phases. The
parameters Nm and Nw in the equations are number of methane and water molecules in
67
methane hydrates, respectively. Parameters Nm and Nw are the constants that are known
for a given hydrate structure. The value of Nw is 46 whereas the value of Nm is normally
between 6 to 8 for the case of methane hydrate. Parameters xm and ym are mole fractions
of methane in aqueous and free gas phases, respectively. The saturated value of
dissolved methane gas in aqueous phase can be calculated from Henry’s law because of
⎧ Pg
⎪ if Pg < Pe
⎪H
x sat
m = ⎨ m
⎪ Pe if Pg ≥ Pe
⎪⎩ H m 5.3
The value of Henry constant for methane gas can be evaluated from[10]:
The saturated molar fraction of water in gas phase is determined using partial
Pυ
y gsat =
Pg 5.5
Pg is the gas phase pressure and Pυ is the water vapor pressure which can be
68
⎧ ⎛ 7258 −6 2 ⎞
⎪exp ⎜ 73.649 − T − 7.3037 ln T + 4.1653×10 T ⎟ , T ≥ 273.15 K
⎪ ⎝ ⎠
Pυ = ⎨
⎪exp ⎛⎜15.6217 − 6415.37 ⎞⎟ ×10 6 , T < 273.15 K
⎪⎩ ⎝ T + 5.5171 ⎠ 5.6
In this study, it is assumed that there is no fluid flow across the reservoir
boundaries (no flow boundaries). The heat transfers in the x and y directions are
negligible comparing to heat transfer in the z-direction because the temperature gradients
in the z direction are much higher than the temperature gradients in the x- and y-
directions. The heat transfer between reservoir and its surrounding can be modeled using
the energy balance equation used in this study, which takes into account heat transfer due
∂
{φ ( ρ g S g hg ) + φ ( ρ a S a ha ) + φ ( ρ H S H hH ) + (1 − φ ) ( ρ r hr )}
∂t
(QE − Qdis + Q fus )
[
= ∇ ⋅ (K h ∇T ) − ∇ ⋅ ρ gν g hg + ρ aν a ha + ] Vb 5.7
In equation (5,7), the variable Kh is the heat conductivity of formation. QE, Qdis,
and Qfus are external heat source, heat of hydrate dissociation, and heat of fusion of ice,
[28]
respectively. The value of heat of fusion of ice is 80 cal/g or 144 BTU/lb . The
variable vl, and hl are velocity and molar enthalpy of phase l, respectively.
69
The amount of methane-hydrate dissociation energy (based on 1 mole of released
d (ln P)
Qdis = ZRT 2 5.8
dT
hydrate phase equilibrium relationship. The unit of Qdis is J/kg-released methane gas and
T is temperature in K.
hydrate is[22]:
T = temperature in K
The molar enthalpy of water and gas phase in equation (5.7) at standard pressure
hi (T ) = ∫c pi dT = c piT 5.10
In this work, the molar enthalpy of gas phase at pressure P and temperature T is
70
where hg = gas enthalpy at pressure Pg
In this study, the ideal liquid behavior was assumed for aqueous phase.
Accordingly, the molar enthalpy of water phase at high pressure is assumed to be equal to
In this work, the Peng-Robinson Equation of State is used for estimating the gas
properties. Then, the molar gas enthalpy in equation (5.11) can be expressed as:
where A = aP/R2 T2
B = bP/RT
a = a c [1 + κ (1 − Tr ) ] 2
ac = 0.45724 R 2Tc2 / P
b = 0.07780 RTc / P
ω = acentric factor
−a a
F (Tr ) = − c [ 1 + κ (1 − Tr ) ] κ Tr
bRT bRT
71
∂
∂t
{
φ [ ρ g S g (c pg T + hgres RT ) ] + φ [ ρ w S w c pwT ] + φ [ ρ H S H hH ] + (1 − φ ) [ ρ r c pr T ]}
∇ ⋅ ( ρυ c p T ) = ρυ ⋅ ∇(c p T ) + c p T ⋅ ∇( ρυ ) 5.14
⎛ ∂c p ⎞ ⎛ ∂c ⎞
∇ ⋅ (c p T ) = c p ∇T + T ∇c p = c p ∇T + T ⎜⎜ ⎟⎟ ∇T = ⎜⎜ c p + T p ⎟⎟ ∇T
⎝ ∂T ⎠ ⎝ ∂T ⎠ 5.15
and the mass balance of mobile phase “l” can be expressed as:
Ql ∂
∇ ⋅ ( ρ lυ l ) = − (φ ρ l S l )
Vb ∂t 5.16
Substitution of equation (5.14), (5.15), and (5.16) into equation (5.13) yields:
⎛ ∂c pg ⎞ ⎛ ∂c ⎞
− ρ gν g ⎜⎜ c pg + T
∂T
( )
⎟⎟ ∇T − ρ gν g ∇ ⋅ hgres RT − ρ wν w ⎜⎜ c pw + T pw ⎟⎟ ∇T
∂T ⎠
⎝ ⎠ ⎝
(Q E − Qdis + Qicefus ) ⎛ Q g c pg T Qw c pwT Q g hgres RT ⎞
+ − ⎜ + + ⎟ + ∇ ⋅ ( K h ∇T )
Vb ⎜ Vb V V ⎟
⎝ b b ⎠
⎛ ∂c pg ⎞ ∂T ⎛ ∂c pw ⎞ ∂T ∂T
= φρ g S g ⎜⎜ c pg + T ⎟⎟ + φρ w S w ⎜⎜ c pw + T ⎟⎟ + (1 − φ ) ρ r c pr
⎝ ∂T ⎠ ∂t ⎝ ∂T ⎠ ∂t ∂t
⎛ ∂c pr ⎞ ∂T ∂ (φρ H S H hH ) ∂ (φρ g S g )
+ (1 − φ ) ρ r T ⎜⎜ ⎟⎟ + − hgres RT
⎝ ∂T ⎠ ∂t ∂t ∂t
∂ (1 − φ ) ⎛ ∂Pg ⎞
+ ρ r c pr T ⎜ ⎟
∂Pg ⎜⎝ ∂t ⎟⎠ 5.17
equations with ten unknowns which are Pg, Pa, Sg, Sa, SH, xm, xw, ym, yw, and T. The
primary variables are defined as the unknowns that are directly solved from the
72
discretized mass and energy balance equations (finite difference approximation) and the
remaining unknowns are determined from auxiliary equations such as equations (5.3) and
(5.5). Typically, primary variables will be selected from the natural variables such as
pressure, temperature, phase saturation, and concentrations (mole fractions). They will
be selected according to the phases appeared in each grid block. The sets of primary
variables for each phase appearance case are summarized in Table 5-1.
Table 5-1: Sets of primary variables for each phase appearance case
Case Phases Primary Variables
1 H+A T, P a , S H
2 G+A T, P g , S a
3 H+A+G T, S a , S g
For two-phase case, one phase pressure and temperature are used as the
independent variables. They are solved simultaneously from the discretized governing
equations. Only one phase saturation needs to be solved. Mole fractions of each
component are not considered unknowns because they can be determined from equations
For the three-phase case, free gas phase pressure and temperature are not both
independent variables but they can be calculated as a function of the other using the
equilibrium P-T correlation. In this work, the temperature was selected as the primary
variable and PG is calculated from the equilibrium correlation. The aqueous phase
pressure is then calculated using capillary pressure relationship. For this case, aqueous
73
and free gas phase saturations are selected as primary unknowns and hydrate phase
tracking its saturation value. If a saturation of any phase become less than or equal to
zero, that phase disappears. The identification of phase appearance is more complicated
and is determined using the location of gridblock condition which reflect to the P-T
diagram of hydrate phase equilibrium. In the phase appearance identification, during two
consecutive iterations, the hydrate phase stabilization boundary in the P-T diagram
cannot be directly passed. An example of the protocol followed for phase appearance
Let us assume that hydrate and aqueous phases (H+A) are present in a previous
iteration such as point “A” in Figure 5-2. The gas phase pressure at current iteration level
is compared with the equilibrium pressure at the block temperature. Three possibilities
a) If PG > Pe, such as point “B” in the figure, there should not be appearance of any
new phase. Thus, the primary variables for this case are not changed.
b) If PG = Pe, such as point “C” in the figure, the three phases (hydrate, aqueous, and
free gas) should coexist at this condition. Therefore, a free gas phase should
appear. Then the primary variables are switched to T, Sa, and Sg.
c) If PG < Pe, such as point “D” in the figure, there should be only gas and aqueous
Thus, the phases presented in this block at current iteration level should be
74
hydrate, free gas, and aqueous phases (H+G+A). Then, the primary variables are
Pressure
Hydrate stabilization
boundary
Hydrate Stabilization Zone
A A
BB
C
increases due to the decrease of the pore spaces. The capillary pressure between free-gas
and aqueous phase in the formation also change when the saturations of aqueous and
hydrate phases change. Additionally, the change of aqueous and hydrate saturations also
change the heat conductivity of the system because they have different conductivity
values. All of these factors are required to be taken into account for solving the problem
75
5.2 Absolute and Relative Permeability, Capillary Pressure, and Heat conductivity
In this study, solid phase (hydrate and ice) in porous medium was considered as
immobile phase and only aqueous and free gas phases can flow in porous medium. Thus,
the presence of solid phase changes porosity and absolute permeability of porous rock.
In the case that the data of rock permeability at various hydrate and ice saturations are not
available, based on “tube-in-series” model of pore space, the change of rock permeability
n
k ⎛ φ − φc ⎞
= ⎜⎜ ⎟⎟
k0 ⎝ φ0 − φc ⎠ 5.18
k 0 = permeability at φ 0
n = correlation parameter
k = the permeability at φ
According to the experimental data, the value of parameter “n” can be as high as 10[11].
k
The values of at various hydrate saturations (for the case k0 = 44 md, φ 0 = 0.3, φc =
k0
0.07, and n = 3) calculated from equation (5.18) are shown in Figure 5-3.
76
Permeability at various hydrate saturations
1.0
0.8
0.6
k/k0
0.4
0.2
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Hydrate saturation
For relative permeability calculation purposes, the immobile solid phases (hydrate
and ice phases) were treated as parts of porous rock, and the normalized aqueous and free
gas saturations were used in relative permeability calculation. The normalized aqueous
and free gas saturations can be calculated from the following equations:
Sa
S a* =
Sa + S g 5.19
Sg
S g* =
Sa + S g 5.20
77
Sa = aqueous phase saturation
The calculated normalized water and free gas saturations are then used to
determine relative permeability data to each phase from relative permeability data using
linear interpolation technique. Note that the relative permeability data can be either
provided to the simulator by users or generated from the default correlations of the
simulator. Equation (5.21) shows the default relative permeability correlations of the
na
⎡ S * − S ira ⎤
k ra = ⎢ a ⎥
⎣ 1 − S ira ⎦
ng
⎡ S g* − S irg ⎤
k rg = ⎢ ⎥
⎢⎣ 1 − S ira ⎥⎦ 5.21
Here na and ng are the exponential parameters for aqueous and free-gas phases,
respectively.
The change of the capillary pressure between aqueous and free-gas phases due to
78
Pcap ( S H , S aA ) = H ( S H ) ⋅ F ⋅ Pe ⋅ ( S aA ) c 5.22
SA = water saturation
H (SH) = 1 + wBx(a,b,SS)
SH = hydrate saturation
discontinuity in the derivative which may cause the unstable of the numerical solution[10].
The capillary pressure at various hydrate saturations (SirA = 0.25, SirG = 0.02, w =
9.28, a = 2.1, b = 2.2, c = -0.65, and PGE = 1.55x104 Pa) calculated from equation (5.22)
79
Capillary Pressure at Various Hydrate Saturations
100
Pc (psi)
10
0.1
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Sh
The appearance of hydrate phase can change the heat conductivity property of
porous medium because rock, hydrate, aqueous and free gas have different conductivity
[25]
properties. It has been recently studied and found that the commonly used linear
model based on the saturation-weighted contributions of phases and of the solid matrix is
not adequate[11]. According to the study [25], a more appropriate estimate of the composite
thermal conductivity of the porous media can be calculated from the extension of the
80
θ = θ dry + ( S a0.5 + S H0.5 )(θ wet − θ dry ) + φ S I θ I 5.23
The following correlations are used to calculate density and viscosity of aqueous
In this study, a linear inerpolation technique is used to calculate the aqueous phase
density at temperature T and 1 atm from the density data of water. Then the
and pressure p. Note that the effect of the amount of methane dissolved in aqueous phase
ρ a (T , p ) = ρ a (T , 1atm ) ⋅ [1 + c w ( p − p sc ) ] 5.24
81
The density of free-gas phase can be calculated from the following real gas law:
ρ g (T , p) = ( p ⋅ MW ) / ( z ⋅ R ⋅ T ) 5.25
Equation of State is used for calculating the compressibility factor of free-gas phase in
this study.
In this study, the linear interpolation technique is used to calculate the aqueous
phase viscosity at temperature T from the viscosity data of water. It is assumed that the
⎛ ⎛ ρ ⎞ yv
⎞
μ g = 1x10 − 4 k v EXP ⎜ xv ⎜⎜ g ⎟⎟ ⎟
⎜ ⎝ 62.4 ⎠ ⎟ 5.26
⎝ ⎠
where kv =
[ 9.4 + 0.02 MW ] T
g
1.5
209 + 19MW g + T
yv = 2.4 − 0.2 xv
986
xv = 3.5 + + 0.01MWg
T
The units of T, ρg, and μ g in this correlation are R, lb/ft3, and cP, respectively.
82
5.4 Well Models for Reservoir Simulation
In this work, the Peaceman’s wellbore model is used. The well-bore model is used
to determine well flow rates and/or flowing bottomhole pressures. The detail of the
wellbore model used in this work is discussed below. Note that the positive value of flow
rate represents injection rate whereas the negative value of flow rate represents
production rate.
In this study, the explicit method was used for calculating water flow rate from
multi-layer wells. A constant pressure gradient (i.e., the difference between well-block
pressure and bottomhole pressure) is assumed for all completion layers in this method.
Thus, water flow rate from layer k of a vertical well can be calculated from total water
J wk
q wk = q wSP ⋅ nk
∑J
k =1
wk 5.27
The variable Jwk is productivity index of aqueous phase for well block layer k (in
⎡ ⎤
⎢ ⎥
⎢ 2π k k rw h ⎥
J wk =
⎢ ⎛ rw ⎞⎥
⎢ μ w Bw ln⎜⎜ + S ⎟⎟ ⎥
⎢⎣ ⎝ re ⎠ ⎥⎦ k 5.28
83
where rw = wellbore radius
μ w = aqueous viscosity
equations:
k = kx ⋅ky
5.29
(k y / k x )1 / 2 (Δx) 2 + (k x / k y )1 / 2 (Δy ) 2
re = 0.28
( k y / k x )1 / 4 + (k x / k y )1 / 4 5.30
The values of kx and ky are the rock permeability in x- and y-directions. After
calculating water flow rate from each layer in vertical wells, gas flow rate from each
layer is then calculated from the mobility ratio of gas to water phases as shown in the
following expression:
⎧ k rg / μ g B g ⎫
q gk = q wk ⎨ ⎬
⎩ k rw / μ w Bw ⎭ k 5.31
84
μ g , μ w = gas and aqueous viscosities
Similar to the water flow rate specification case, gas flow rate from layer k of a
vertical well can be calculated from total gas flow rate specification shown below:
J gk
q gk = q g SP ⋅ nk
∑J
k =1
gk 5.32
Again, Jgk is the productivity index of gas phase of well block layer k which can
be calculated from:
⎡ ⎤
⎢ ⎥
⎢ 2π k k rg h ⎥
J gk =
⎢ ⎛ rw ⎞⎥
⎢ μ g B g ln⎜⎜ + S ⎟⎟ ⎥
⎢⎣ ⎝ re ⎠ ⎥⎦ k 5.33
μ g = gas viscosity
Again, k and re can be calculated from equations (5.26) and (5.27), respectively.
Similarly, water flow rate from each layer can be calculated from the mobility
85
⎧⎪ k / μ B ⎫⎪
q wk = q gk ⎨ rw w w ⎬
⎪⎩ k rg / μ g Bg ⎪⎭ k 5.34
In this work, it is assumed that the bottomhole pressures for all phases are the
same. The Peaceman’s wellbore model for this case can be written as:
⎛ ⎞
⎜ ⎟
⎜ 2π k k rf h ( Pf − Pwf , k ) ⎟
q kf = ⎜ ⎟
⎜ μ B ln⎛⎜ rw + S ⎞⎟ ⎟
⎜ f f ⎜r ⎟ ⎟ 5.35
⎝ ⎝ e ⎠ ⎠k
μ f = fluid viscosity
86
Pwf,k = sandface pressure of the well block layer k
Again k and re can be calculated from equations (5.26) and (5.27), respectively.
Since a constant pressure gradient for every layer for the flow rate specification
cases is assumed, the value of sandface pressure of well block layer k can be calculated
Again, it is assumed the constant pressure gradient along the horizontal well (the
difference between well-block pressure and bottomhole pressure). Thus, water flow rate
from section l of a horizontal well can be calculated from total water flow rate
J wl
q wl = q wSP ⋅ nl
∑J
l =1
wl 5.37
87
The variable Jwl is productivity index of aqueous phase for well block l (in the
⎡ ⎤
⎢ ⎥
⎢ 2π k k rw Ll ⎥
J wl =
⎢ ⎛ rw ⎞⎥
⎢ μ w Bw ln⎜⎜ + S ⎟⎟ ⎥
⎢⎣ ⎝ re ⎠ ⎥⎦ l 5.38
μ w = aqueous viscosity
k = km ⋅ kn 5.39
Subscripts m and n represent the m-n plane which is perpendicular to the well direction.
For example, m and n refer to y and z directions for horizontal wells placed along the x-
(k m / k n )1 / 2 (Δn) 2 + (k n / k m )1 / 2 (Δm) 2
re = 0.28
(k m / k n )1 / 4 + (k n / k m )1 / 4 5.40
88
Again, subscript m and n represent the m-n plane which is perpendicular to the
well direction.
Similar to the vertical well case, and after calculating water flow rate of each
section of a horizontal well, gas flow rate of well block section l can be then calculated
from the mobility ratio of gas to water phases as indicated in the following equation:
⎧ k rg / μ g B g ⎫
q gl = q wl ⎨ ⎬
⎩ k rw / μ w Bw ⎭ l 5.41
Similar to the water flow rate specification case, gas flow rate from section l of a
horizontal well can be calculated from total gas flow rate specification using the
following equation:
J gl
q gl = q g SP ⋅ nl
∑J
l =1
gl 5.42
Again, Jgl is the productivity index of gas phase of well block section l which can
be calculated from:
89
⎡ ⎤
⎢ ⎥
⎢ 2π k k rg Ll ⎥
Jgl =
⎢ ⎛r ⎞⎥
⎢ μ g B g ln⎜⎜ w + S ⎟⎟ ⎥
⎣⎢ ⎝ re ⎠ ⎦⎥ l 5.43
μ w = aqueous viscosity
The values of k and re can be calculated using equations (5.36) and (5.37).
Similarly, and after calculating gas flow rate of each section of a horizontal well,
water flow rate of well block section l can be then calculated from the mobility ratio of
⎧⎪ k / μ B ⎫⎪
q wl = q gl ⎨ rw w w ⎬
⎪⎩ k rg / μ g B g ⎪⎭ l 5.44
90
c) Flowing Bottomhole Pressure Specification
well are equal to specified bottomhole pressure because there are no hydrostatic gradients
for this case. Accordingly, the flow rate of fluid f from section l of a horizontal well can
be calculated from:
⎛ ⎞
⎜ ⎟
⎜ 2π k k rf Ll ( Pf − Pwf , l ) ⎟
qf = ⎜
l
⎟
⎜ μ B ln⎛⎜ rw + S ⎞⎟ ⎟
⎜ f f ⎜r ⎟ ⎟ 5.45
⎝ ⎝ e ⎠ ⎠l
μ f = fluid viscosity
Again k and re can be calculated from equations (5.36) and (5.37), respectively.
91
5.5 Material and Energy Balance Checks
In order to check the accuracy of the simulation results, a material balance check
∑∑∑ ⎜
i =1 j =1 k =1 ⎝ 5.615 ⎠i , j , k i =1 i =1
IMBi = n
≅ 1.0
⎛ Vb
(φ yi Sg ρg + φ xi Sa ρa + φ ci SH ρH )⎞⎟
nx ny nz
∑∑∑ ⎜
i =1 j =1 k =1 ⎝ 5.615 ⎠i, j , k 5.46
i g g + ∑ ⎜ ∑ yi qgwell (l ) Δt l + ∑ xi qawell (l ) Δt l ⎟
i =1 j =1 k =1 ⎠i , j , k l =0 ⎝ i =1 i =1 ⎠
CMBi = n =0
≅ 1.0
⎛ Vb
(φ yi Sg ρg + φ xi Sa ρa + φ ci SH ρH )⎞⎟
nx ny nz
∑∑∑ ⎜
i =1 j =1 k =1 ⎝ 5.615 ⎠i , j , k 5.47
The correct temperature distribution should produce a good energy balance check.
The finite difference approximation should be consistent with the energy conservation if
the finite difference representation approximates the solution of the problem with an
Final Heat Content = Original Heat Content + Heat Added − Heat Loss 5.48
92
Each term in the above equation can be calculated from the following expressions:
∑∑∑V {[ (1 − φ ) ρ c ]
nx ny nz
n +1
Final Heat Content = b r pr + φ ( ρ g S g c pg + ρ w S w c pw ) i , j ,k
Ti ,nj+,1k
i =1 j =1 k =1
+ [φρ H S H hH (T )]i , j ,k
n +1
}
∑∑∑V {[ (1 − φ ) ρ c ]
nx ny nz
Original Heat Content = + φ ( ρ g S g c pg + ρ w S w c pw ) i , j ,k Ti ,nj ,k
n
b r pr
i =1 j =1 k =1
+ [φρ H S H hH (T )]i , j ,k
n
}
∑∑∑ {[ q ]
nx ny nz
n +1
Heat Loss = well
g Δt ρ gsc c pg + 5.615q wwell Δt ρ wsc c pw ) i , j ,k Ti ,nj+,1k
i =1 j =1 k =1
(
+ q gdis Δt Qdis )n +1
i , j ,k
}
∑∑∑ [ 5.615q ]
nx ny nz
Δt ρ wsc c pw )T
injected conditions
Heat Added = inject
w i , j ,k
i =1 j =1 k =1
93
CHAPTER 6
data from actual methane-hydrate reservoirs in the literature are not yet available.
previous studies can be found in the literature[6,11]. Therefore, at this point in time, the
validation of the developed simulator in this work can only be achieved by comparing the
simulation results of this work with the results from previous studies. In this work, the
simulation results from Burshears’ and Moridis’ works[6,11] are used for the validation
purposes. In Moridis’ work, the cylindrical-coordinate system was used whereas the
rectangular-coordinate system is used in this work. The two reservoirs have the same
volume. Gas is produced through a production well located at the center of the reservoir
at a constant gas flow rate at 2.5 MMSCF/day. The production well is completed along
The class 1 methane hydrate reservoir structure discussed in Holder’s work [5] was
used in Moridis’ study and it is also used in this work for the examination of production
characteristics from class 1 gas hydrate deposit. From the literature review, most of the
gas production rate scheme were examined. In this work, the study of gas production
94
than a constant gas flow rate production scheme to gain a better insight for the activation
The reservoir structure in Holder’s work consists of two zones: hydrate zone in
the upper part of the system and free gas zone in the lower part of the reservoir. The
Initially, methane hydrate and water saturations in the hydrate zone are 70 and 30
percents, respectively, whereas the average free gas and water saturations in the free gas
zone are 70 and 30 percents, respectively. There is a well located at the center of the
system and it is completed only in the free gas zone. The overall structure of the reservoir
50 ft
50 ft
95
6.1 Reservoir Rock and Fluid Properties
In this work, the reservoir rock is assumed to be sandstone with uniform porosity
and homogeneous and isotropic permeability distribution. The reservoir rock properties
used in this model validation are summarized in Table 6-1. Note that the input data for
the developed simulator must be in the oil-field unit, but since the SI unit was used in
Moridis’ work, the rock and fluid properties used in this validation are therefore shown in
b) Methane-Hydrate Properties
In this work, it is assumed that the change of the density of methane-hydrate due
96
Table 6-2: Hydrate Properties[11]
Hydrate Properties Value
Density* 3
57.43 lb/ft (920 kg/m3) [3]
Specific heat 0.5016 BTU/lb-oF (2.10 kJ/kg-K)
Heat conductivity 0.2600 BTU/hr-ft-oF (0.45 W/m-K)
* This value was not provided in Moridis’ study
The change of the density of aqueous phase can occur when either pressure or
temperature (or both) change. The densities of aqueous phase at various temperatures and
14.70 psia used in this work are shown in Table 6-3. The change of aqueous phase
density due to the change of pressure can be calculated using the compressibility of
aqueous phase. In this work, the compositional effect on aqueous phase density is
97
In this work, it is assumed that the change of aqueous phase viscosity due to the
change of pressure is negligible. However, the change of aqueous phase viscosity due to
the temperature change is taken into account. Table 6-4 shows the values of aqueous
phase viscosity used in this study at various temperatures. Again, the compositional effect
In this study, the change of rock permeability due to solid saturation is calculated
using equation (5.17). Equation (5.20) is used for calculating capillary pressure between
aqueous and free gas phases. Table 6-5 summarizes all the parameter values of these two
equations used in this study. Note that the information about the equation used for
calculating the change of formation permeability due to the change of hydrate saturation
was not provided in the Moridis’ study. In this study, equation (5.18) was used and the
values of the parameters in the equation were selected in such a way that the permeability
of the hydrate zone (SH = 0.70) is closed to zero in order to prevent the drainage of
aqueous phase in hydrate zone to the lower free-gas zone. The plot of this equation was
Table 6-6 shows the data of the relative permeabilities to aqueous and free-gas
99
6.2 Grid Structure of the Reservoir
discretization of this study. The additional 45-feet thick intervals of impermeable rock
are added to the top and the bottom of the reservoir. No fluid flow takes place in these
two impermeable zones but heat transfer across the upper and lower boundaries is
allowed. It is further assumed that this thickness is sufficient to allow accurate heat
[11]
exchange calculation with hydrate deposit over a 30-year long production period .
Figure 6-2 shows the grid structure of the reservoir in this study.
45 ft
50 ft
50 ft
Impermeable rock
45 ft
Methane-hydrate zone
Free-gas zone
Impermeable rock
Figure 6-2: Grid structure of the reservoir in this work
In this study, the total number of layers of free-gas and hydrate zones is 10 layers
(5 layers for each zone). Each layer is 10 feet thick. There are three layers for each
impermeable zone. Therefore, the grid structure is totally consisted of 16 layers in the z-
direction. The layer thickness and the rock properties for each zone are summarized in
Table 6-7. The cartesian coordinate system is used in this work whereas the cylindrical
100
coordinate was used in Moridis’ work. Thus, the grid-block dimensions used in this work
are not the same as the block dimensions used in the Moridis’ work. Note that the block
dimensions are selected as small as possible with due cognizance is given to the
The initialization for the reservoir simulation of Class 1 hydrate deposits is a challenging
task. It is found that initializing the entire system at once is not an easy task to obtain the
101
initial conditions close to the desired initial conditions. This is because the changes of
pressure and temperature in the free-gas zone will affect the conditions in hydrate zone.
For example, if the heat flux to the dissociation front from the lower and upper layers are
not equal, it causes the change of temperature at the dissociation front. Consequently, the
temperature of the entire system. Accordingly, it will be easier to initialize the entire
system when the change of the conditions of both zones is minimal. In order to achieve
In the initialization process, the reservoir is divided into two sections, one
corresponding to the hydrate zone (upper section) and the other one corresponding to the
two-phase zone (lower section). The bottom layer of the upper (hydrate) zone and the top
layer of the lower (free-gas) zone are treated as the boundaries representing the hydrate
dissociation front where free gas phase, aqueous phase, and methane hydrate coexist in
equilibrium.
Starting with the lower section, initial pressure, temperature, and saturation of
each phase are determined by running the simulator (without production or injection) to
achieve an equilibrium condition. Note that pressure, temperature, and phase saturations
of the first layer (hydrate dissociation front) of this zone do not change during this step.
The initial values used for this initialization are shown in Table 6-8. The temperature
gradient used for this zone is 0.0170 oF/ft and the temperature of the dissociation front is
56.30 oF [11]
. The temperature gradient of the system was not provided in this Moridis’
[30]
work. This value comes from a different work of Alp and Moridis when a similar
102
Table 6-8: Initial values for the initialization of the lower section
Layer no. Pg (psia) Sa SH T (oF) Remarks
9 1550.00 0.2950 0.7000 56.300 Dissociation front (G+H+A)
10 1550.00 0.3000 0.0000 56.465
11 1550.00 0.3000 0.0000 56.629
12 1550.00 0.3000 0.0000 56.794 Free gas zone (G+A)
13 1550.00 0.3000 0.0000 56.959
14 1550.00 0.3000 0.0000 57.123
15 0.00 0.0000 0.0000 57.288
16 0.00 0.0000 0.0000 57.452 Impermeable layer
17 0.00 0.0000 0.0000 57.699
18 0.00 0.0000 0.0000 58.029 Inactive layer with constant T
Note that the program automatically sets the gas pressure of any block where
three phases (gas, aqueous, and methane hydrate) are present to be equal to the
dissociation pressure corresponding to the block temperature. Therefore, the gas pressure
of the dissociation front layer is automatically set to be 1548.06 psia corresponding to the
temperature (56.30 oF) of this layer even though the pressure of this layer from the input
file was 1,550 psia. In this initialization, the small gas saturation (0.005) was introduced
in the dissociation front layer so that three phases appear in the dissociation front layer.
The results from the initialization of this section are shown in Table 6-9.
103
Results in Table 6-9 show that aqueous phase saturation in the lower part of the
The next step is to initialize the upper section. The initial values for the
initialization of this section are shown in Table 6-10. The results from the initialization
of this section are shown in Table 6-11. Note that the program reads gas phase pressure
and aqueous phase saturation from the input data and then uses the capillary pressure
correlation to calculate aqueous phase pressure. In the case that there is no gas phase in a
grid block, the gas phase pressures of all layers in the upper section are set to be a little
bit higher (1,550 psia) than the gas pressure at the dissociation front layer (1,548.06 psia)
to ensure that the conditions of all the layers above the dissociation front are inside the
Table 6-10: Initial values for the initialization of the upper section
o
Layer no. Pg (psia) Sa SH T ( F) Remarks
1 0.00 0.0000 0.0000 54.571 Inactive layer with constant T
2 0.00 0.0000 0.0000 54.901
3 0.00 0.0000 0.0000 55.230 Impermeable layer
4 0.00 0.0000 0.0000 55.477
5 1550.00 0.3000 0.7000 55.641
6 1550.00 0.3000 0.7000 55.806
Hydrate zone (H+A)
7 1550.00 0.3000 0.7000 55.971
8 1550.00 0.3000 0.7000 56.135
9 1548.06 0.2950 0.7000 56.300 Dissociation front (G+H+A)
104
In the third step, the two sections are combined together and still keep the
pressure, temperature, and saturations of the dissociation front constant. Then the model
is run until achieving a new equilibrium condition. Table 6-12 shows the initial values
for this step. The next step is to set layer no.9 as a regular layer (T, P and S can change)
and run the simulation to achieve a new equilibrium condition. The results of this step are
Table 6-12: Initial values for the initialization of the entire reservoir
Layer no. Pg (psia) Sa SH T (oF) Remarks
1 0.00 0.0000 0.0000 54.571 Inactive layer with constant T
2 0.00 0.0000 0.0000 54.909
3 0.00 0.0000 0.0000 55.275 Impermeable layer
4 0.00 0.0000 0.0000 55.438
5 1530.66 0.3000 0.7000 55.584
6 1535.01 0.3000 0.7000 55.812
Hydrate zone (H+A)
7 1539.36 0.3000 0.7000 55.970
8 1543.71 0.3000 0.7000 56.136
9 1548.06 0.2950 0.7000 56.300 Dissociation front (G+H+A)
10 1537.05 0.2591 0.0000 56.344
11 1537.44 0.2650 0.0000 56.415
12 1537.83 0.2747 0.0000 56.486 Free gas zone (G+A)
13 1538.22 0.2926 0.0000 56.555
14 1538.60 0.3334 0.0000 56.621
15 0.00 0.0000 0.0000 56.726
16 0.00 0.0000 0.0000 57.051 Impermeable layer
17 0.00 0.0000 0.0000 57.507
18 0.00 0.0000 0.0000 58.029 Inactive layer with constant T
105
Table 6-13: Results from the initialization of the entire reservoir
Layer no. Pg (psia) Sa SH T (oF) Remarks
1 0.00 0.0000 0.0000 54.571 Inactive layer with constant T
2 0.00 0.0000 0.0000 55.040
3 0.00 0.0000 0.0000 55.532 Impermeable layer
4 0.00 0.0000 0.0000 55.971
5 1534.75 0.3000 0.7000 56.060
6 1539.10 0.3000 0.7000 56.117
Hydrate zone (H+A)
7 1543.45 0.3000 0.7000 56.183
8 1547.81 0.3000 0.7000 56.257
9 1552.16 0.2949 0.6965 56.341 Dissociation front (G+H+A)
10 1538.87 0.2619 0.0000 56.594
11 1539.26 0.2651 0.0000 56.653
12 1539.65 0.2747 0.0000 56.609 Free gas zone (G+A)
13 1540.03 0.2926 0.0000 56.638
14 1540.42 0.3334 0.0000 56.614
15 0.00 0.0000 0.0000 56.734
16 0.00 0.0000 0.0000 57.054 Impermeable layer
17 0.00 0.0000 0.0000 57.508
18 0.00 0.0000 0.0000 58.029 Inactive layer with constant T
The result shows a slight decrease of hydrate saturation at the dissociation front
It was found that the hydrate saturations at the dissociation front of the reservoir
can be slightly changed to the desired values (0.70) without disturbing the obtained initial
condition because this change does not significantly affect the heat transfer properties of
this layer. Accordingly, the computed initial pressure and temperature of the entire
system can still be used. Therefore, the saturations of aqueous and hydrate phases in the
upper section are changed to the desired values. Table 6-14 shows the initial conditions
106
Table 6-14: Initial conditions of the reservoir used in this study
Layer no. Pg (psia) Sa SH T (oF) Remarks
1 0.00 0.0000 0.0000 54.571 Inactive layer with constant T
2 0.00 0.0000 0.0000 55.040
3 0.00 0.0000 0.0000 55.532 Impermeable layer
4 0.00 0.0000 0.0000 55.971
5 1534.75 0.3000 0.7000 56.060
6 1539.10 0.3000 0.7000 56.117
Hydrate zone (H+A)
7 1543.45 0.3000 0.7000 56.183
8 1547.81 0.3000 0.7000 56.257
9 1552.16 0.2950 0.7000 56.341 Dissociation front (G+H+A)
10 1538.87 0.2619 0.0000 56.594
11 1539.26 0.2651 0.0000 56.653
12 1539.65 0.2747 0.0000 56.609 Free gas zone (G+A)
13 1540.03 0.2926 0.0000 56.638
14 1540.42 0.3334 0.0000 56.614
15 0.00 0.0000 0.0000 56.734
16 0.00 0.0000 0.0000 57.054 Impermeable layer
17 0.00 0.0000 0.0000 57.508
18 0.00 0.0000 0.0000 58.029 Inactive layer with constant T
An uniform block dimension is used in this work. The effects of block dimensions
(Δx and Δy) on simulation results are shown in Figure 6-3. Table 6-15 shows the average
percent deviation of the results against the results from the smallest grid block
dimensions (Δx and Δy =50 ft) and the computational time it can reduced.
Table 6-15: The effects of grid block dimensions on the results and computational time
Dimensions (feet) Differences (%) Reduction Time (%)
50* 0.00 0.00
75 2.92 12.00
100 6.99 26.00
150 13.07 34.00
* reference case
The grid block dimensions in x and y directions (Δx and Δy) used in this study is
selected as 100 feet as the model could execute about 26 percent faster while maintaining
an average derviation of 7 percent with respect to the reference case of block dimensions
of 50 feet.
107
Gas Production Rate
20
15
10
0
0 10 20 30 40 50
Time (days)
0.25
0.20
0.15
0.10
0.05
0.00
0 20 40 60 80 100
Time (days)
Figure 6-3: The effects of grid block dimensions (Δx and Δy) on the simulation results
108
6.4 Comparison of Results
the center of the reservoir at a constant gas flow rate at 2.5 MMSCF/day[11]. The input
data of reservoir structure, rock and fluid properties, and initial conditions of this
validation case are shown in the Table 6-1 to Table 6-7, and Table 6-14, respectively. In
this case, a production well is completed only in the free-gas zone. Figure 6-4 shows the
comparison between the cumulative produced gas and cumulative dissociated gas from
this work and that from Moridis’ work [11]. Note that this comparison is shown in SI unit
because the simulation results in Moridis’ work were presented in SI unit system. The
saturation and temperature distributions along the x-z plane (at about 56 meters from the
well) are shown in Figure 6-5, Figure 6-6 shows the distribution plots from the Moridis’
work[11].
109
Cumulative Methane Gas
1.50E+08
Cumulative Produced Gas
Cumulative Dissociated Gas (this work)
Cumulative Dissociated Gas (Moridis' work)
Cumulative Methane gas (Std. m )
3
1.00E+08
5.00E+07
0.00E+00
0 200 400 600 800 1000 1200 1400 1600 1800 2000
time (days)
Figure 6-4: Cumulative produced and dissociated gas from Moridis’s[11] and this work
P (Mpa) T (C) SH Sg Sa
-30.00 -30.00 -30.00 -30.00 -30.00
0 year
1 year
-35.00 -35.00 -35.00 -35.00 -35.00
2 years
3 years
5 years
Z (m)
-45.00 -45.00 -45.00 -45.00 6 years
-45.00
Figure 6-5: Temperature and saturation distributions along the vertical direction
(at r = 56 m) of this work
110
Figure 6-6: Temperature and saturation distribution along the vertical direction
(at r = 50 m) of Moridis’s work[11]
The simulation results from the two models are somewhat different. The
cumulative dissociated gas predicted from Moridis’ work is higher than the predictions
achieved from this work. The differences can be attributed to the use of different
correlations for calculating some parameters and fluid physical properties (such as the
amount of methane gas that can dissolve in the aqueous phase, aqueous phase density,
aqueous phase viscosity) and the change of rock permeability due to hydrate saturation of
the two simulators. Another difference between the two simulators is that the wellbore
model was not incorporated in the TOUGH FX simulator which was used in the Moridis’
study. For TOUGH FX simulator, the gas production rate from each individual artificial
grid layer was calculated by dividing the total gas production rate (2.5 MMSCF/day) by
the number of artificial grid layers of the production zone in order to obtain the gas
production rate for each layer, and then these calculated values were specified for each
layer and they did not change for the entire simulation period. On the other hand, a
111
wellbore model as suggested by Peaceman is incorporated in this model. Therefore,
production rate from each artificial grid layer can change according to the mobility of
each mobile phase whereas the total gas production rate is still equal to 2.5 MMSCF/day.
This could cause the difference between the results from the two models. The lack of
wellbore model in the TOUGH FX simulator limits the capability of the model to
simulate the systems that use a specified bottom-hole pressure production scheme. The
difference of the coordinate systems that used in the two models might also contribute to
The hydrate saturation profiles shown in Figure 6-5 and Figure 6-6 indicate that
hydrate dissociation propagates from the initial dissociation front upward to the top of the
hydrate zone. This is because the pressure reduction moves upward from the production
region (free-gas zone) to the non-production region (hydrate zone). The nearer to the
production region, one sees the higher pressure drops (which cause the higher
dissociation rate).
Similar to the results from Moridis’ study, the emergence of a second dissociation
front that forms at the top of the hydrate zone and advances downward is observed. This
is because production and dissociation occurred in the system causes the decrease of
reservoir temperature. Consequently, the rate of heat transfer from the surrounding above
the hydrate zone to the top of hydrate zone increases, and this triggers the hydrate
Figure 6-7 shows the incremental material balance checks of methane and water
components of the obtained simulation results which indicate very good material
112
Incremental Material Balance
IMBw IMBm
1.0000004
1.0000002
1.0000000
0.9999998
IMB
0.9999996
0.9999994
0.9999992
0.9999990
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Time (days)
Figure 6-7: Incremental material balance checks for methane and water components
on the x-z plane (cross section of the system) at various times are shown in Figure 6-8 to
According to the hydrate saturation distribution plots, the hydrate saturation just
above the production zone is increasing with time. This is because the temperature in this
region drops faster than the other region resulting in the hydrate formation in this region.
The temperature distribution plots show the drop of temperatures in the dissociation
regions because hydrate dissociation is endothermic and it might be able to slow down
the dissociation process if the rate of temperature reduction is high enough to make the
dissociation pressure below the reservoir pressure (hydrate phase become stable).
113
Aqueous Saturation at t = 0 year Aqueous Saturation at t = 1 year
-30 -30
-35 -35
-40 -40
-45 -45
-50 -50
-55 -55
-60 -60
-500 -335 -170 0 170 335 500 -500 -335 -170 0 170 335 500
-35 -35
-40 -40
-45 -45
-50 -50
-55 -55
-60 -60
-500 -335 -170 0 170 335 500 -500 -335 -170 0 170 335 500
-35 -35
-40 -40
-45 -45
-50 -50
-55 -55
-60 -60
-500 -335 -170 0 170 335 500
-500 -335 -170 0 170 335 500
114
Hydrate Saturation at t = 0 year Hydrate Saturation at t = 1 year
-30 -30
-35 -35
-40 -40
-45 -45
-50 -50
-55 -55
-60 -60
-500 -335 -170 0 170 335 500 -500 -335 -170 0 170 335 500
-35 -35
-40 -40
-45 -45
-50 -50
-55 -55
-60 -60
-500 -335 -170 0 170 335 500 -500 -335 -170 0 170 335 500
-35 -35
-40 -40
-45 -45
-50 -50
-55 -55
-60 -60
-500 -335 -170 0 170 335 500 -500 -335 -170 0 170 335 500
115
Gas Saturation at t = 0 year Gas Saturation at t = 1 year
-30 -30
-35 -35
-40 -40
-45 -45
-50 -50
-55 -55
-60 -60
-500 -335 -170 0 170 335 500 -500 -335 -170 0 170 335 500
-35 -35
-40 -40
-45 -45
-50 -50
-55 -55
-60 -60
-500 -335 -170 0 170 335 500 -500 -335 -170 0 170 335 500
-35 -35
-40 -40
-45 -45
-50 -50
-55 -55
-60
-60
-500 -335 -170 0 170 335 500
-500 -335 -170 0 170 335 500
116
Temperature at t = 0 year Temperature at t = 1 year
-30 -30
-35 -35
-40 -40
-45 -45
-50 -50
-55 -55
-60 -60
-500 -335 -170 0 170 335 500 -500 -335 -170 0 170 335 500
-35
-35
-40
-40
-45
-45
-50
-50
-55
-55
-60
-60
-500 -335 -170 0 170 335 500
-500 -335 -170 0 170 335 500
-35 -35
-40 -40
-45 -45
-50 -50
-55 -55
-60 -60
-500 -335 -170 0 170 335 500 -500 -335 -170 0 170 335 500
117
Another validation with a simplified gas-hydrate reservoir simulation is also
performed in this work. The simulation results from Burshears’ study[6] are used in this
validation. The reservoir structure used in his study is similar to the reservoir structure
shown in Figure 6-1. The characteristics and properties of this reservoir are summarized
in Table 6-16. Note that the values of parameters and properties used in the previous
validation case are used if they are not provided in the Burshears’ work. In this
simulation, the production well is operated with a constant gas flow rate at 10
MMSCF/day.
Table 6-16: Characteristics and properties of the reservoir in Burshears’ work [6]
Reservoir thickness 100 ft
Gas zone thickness 50 ft
Hydrate zone thickness 50 ft
Reservoir porosity 30%
Reservoir permeability 44 mD
Initial pressure 3,000 psia
Initial temperature 65.8 oF
Gas composition 100% CH4
Thermal conductivity 1.56 BTU/ft-Hr-oF
Production rate (gas) 10 MMSCF/day
The reservoir grid structure (number of layers and their thickness) shown in Table
6-7 is also used in this case. Following the same initialization procedure previously
discussed, the initial conditions of this reservoir shown in Table 6-17 are realized.
118
Table 6-17: Reservoir initial conditions used in the validation
o
Layer no. Pg (psia) Sa SH T ( F) Remarks
1 0.00 0.0000 0.0000 64.239 Inactive layer with constant T
2 0.00 0.0000 0.0000 64.845
3 0.00 0.0000 0.0000 65.376 Impermeable layer
4 0.00 0.0000 0.0000 65.754
5 2980.29 0.3000 0.7000 65.820
6 2984.66 0.3000 0.7000 65.857
Hydrate zone (H+A)
7 2989.02 0.3000 0.7000 65.893
8 2993.39 0.3000 0.7000 65.930
9 2997.75 0.2950 0.7000 65.966 Dissociation front (G+H+A)
10 2984.01 0.2637 0.0000 66.015
11 2984.75 0.2709 0.0000 66.048
12 2985.50 0.2832 0.0000 66.081 Free gas zone (G+A)
13 2986.24 0.3071 0.0000 66.115
14 2986.98 0.3650 0.0000 66.150
15 0.00 0.0000 0.0000 66.262
16 0.00 0.0000 0.0000 66.620 Impermeable layer
17 0.00 0.0000 0.0000 67.123
18 0.00 0.0000 0.0000 67.697 Inactive layer with constant T
Figure 6-12 (a) shows a very good agreement between the cumulative dissociated
gas predictions from Burshears’ work and this study. Figure 6-12 (b) shows the change of
the simulation results when different heat transfer mechanisms are taken into account in
the calculation.
119
Cumulative Methane Gas (MMSCF)
10000
Cumulative dissociated gas (Burshears' study)
9000
6000
5000
4000
3000
2000
1000
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
(a) Comparison between the results from Burshears’ work and this study
2000
1500
1000
500
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
120
The dark blue line in Figure 6-12 (b) represents the cumulative dissociated gas
when heat conduction, heat convection, and heat loss from the production are taken into
account. In the same figure, the red line represents the cumulative dissociated gas when
only heat conduction is taken into account and the green line represents the cumulative
dissociated gas when heat conduction and heat loss by taking fluids out of the system are
taken into account. The red and the green lines are not significantly different indicating
that the effect of heat loss by taking the fluids out of the system does not significantly
affect the simulation results. However, it shows that the simulation results change when
heat convection is taken into account indicating the needs of incorporation of heat
convection into the model for more accurate results. The result indicates higher
dissociation when heat of convection is taken into account because the model predicts
more effective heat transfer between hot and cold regions for this case. Consequently, the
model predicts higher heat support for the dissociation resulting in the higher dissociation
rate in this case. Note that, according to the simulation results generated, the difference of
the predicted cumulative dissociated gases of the two cases was up to 9.5 percents (for
The two comparisons show the consistency of the simulation results of the model
developed in this work and those from the models developed in previous studies
121
6.5 Study of Production Characteristics
In this section, the gas production characteristics from a class 1 methane hydrate
reservoir are studied using the simulator developed in this work. Note that the initial
condition and rock and fluid properties from the validation (with Burshears’ work)
section are used as the input data for all of the further simulation studies in this work.
(methane) reservoirs. The two systems have the same initial conditions for the free-gas
zone. The thickness of free-gas zone is 50 feet. The conventional gas reservoir has no
hydrate zone and it is also treated as a non-isothermal system. For each system, there is
only one production well located at the center of the system and it is completed in the
middle of the initial free-gas zone. The production well is operated with constant bottom-
hole pressure at 14.7 psia. The reservoir properties of the two systems are summarized in
Table 6-18 and Figure 6-13 shows the structure of the two reservoirs.
Figures 6-14 and 6-15 show the gas and water production characteristics of the
two systems. Note that the simulator developed in this study was also validated by
comparing the obtained simulation results for a conventional methane gas reservoir with
the simulation results (for the same reservoir) from a commercial reservoir simulator
(CMG). The details of the system used in this comparison and the compared results are
shown in Appendix B. The slight differences of the results are attributed to the use of
122
different correlations for calculating fluid properties (such as density and viscosity) in the
two simulators.
Table 6-18: Reservoir properties of the conventional and methane hydrate reservoirs
Methane-hydrate Conventional gas
Properties reservoir reservoir
Reservoir thickness 100 ft 50 ft
Gas zone thickness 50 ft 50 ft
Hydrate zone thickness 50 ft 0 ft
Reservoir porosity 30% 30%
Reservoir permeability 44 mD 44 mD
Initial pressure 3,000 psia 3,000 psia
Intial temperature 65.8 oF 65.8 oF
Gas composition 100% CH4 100% CH4
Thermal conductivity 1.56 BTU/ft-Hr-oF 1.56 BTU/ft-Hr-oF
Hydrate
50 ft zone
Free-gas Free-gas
50 ft zone
50 ft zone
123
Gas Production Rate (MMSCF/day)
Conventional gas reservoir Methane-hydrate reservoir
110
100
90
Gas Production Rate (MMSCF/day)
80
70
60
50
40
30
20
10
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
3000
2000
1000
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
Figure 6-14: Gas Production characteristics of hydrate and conventional gas systems
124
Water Production Rate
Conventional Gas Reservoir Methane-Hydrate Reservoir
30
20
15
10
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
7500
5000
2500
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
Figure 6-15: Water Production characteristics of hydrate and conventional gas systems
125
According to the simulation results, gas production rate from the gas-hydrate
reservoir is higher than that from the conventional gas reservoir because gas hydrate
phase is capable of providing additional amount of gas to the system. From this
simulation exercise, the cumulative gas production at the end of 1,000 days from the
methane hydrate reservoir (3,678.4 MMSCF) was higher than the cumulative gas
production from the conventional gas reservoir (2,651.4 MMSCF) by about 38.74 percent
and the gas production rate from the methane hydrate reservoir is still higher than that
It can be seen that both gas and water production rates exponentially decline with
time for a conventional gas reservoir. However, for a methane-hydrate reservoir, gas
production rate decreases with time whereas water production rate increases with time.
This is because the dissociation of methane hydrate releases both free gas and water to
the system. The released water flows down from the hydrate zone to the lower part of the
reservoir due to the gravitational forces resulting in the increase of aqueous saturation in
the lower part of the free-gas zone. Figure 6-16 shows the significant increase of aqueous
saturation of the well block (for the gas-hydrate system) resulting in the increase of the
mobility of aqueous phase. On the other hand, aqueous phase saturation for the
conventional gas reservoir increased in the early period of production and gradually
decreased after approximately 300 days of production. One can see that, for the gas
hydrate reservoir, the shapes of water production rate and aqueous phase saturation plots
are consistent.
126
Water Saturation at Well Block
0.2850
0.2840
0.2835
0.2830
0 200 400 600 800 1000
Time (days)
0.70
Water saturation (fraction)
0.60
0.50
0.40
0.30
0.20
0 200 400 600 800 1000
Time (days)
Figure 6-16: Water saturation of well block of conventional gas and hydrate systems
For the gas hydrate reservoir, aqueous phase saturation at the well block decreases
during the early period of production because some of aqueous phase is taken out from
the system through the production well. Aqueous phase saturation starts increasing at the
127
end of 60 days because the released water from the dissociation in the hydrate zone flows
downward and reaches the well block. The amount of water entering to the well block is
higher than the amount of water leaving out from the block (some water flows down to
the lower portion and some water is taken out through the production well) resulting in
the increase of aqueous phase saturation in this block. The increase of aqueous phase
mobility yields the higher water production rate resulting in the decrease of the difference
between the amount of water entering and leaving the well block. Consequently, the
aqueous phase saturation increases with the lower rate and becomes more or less
constant. This approximately happens during the first 250 to 700 days of production.
Meanwhile the aqueous phase saturations as well as pressures of the blocks below the
well block keep increasing causing aqueous phase in the well block more difficult to flow
downward whereas the amount of water (from the dissociation) entering to the well block
does not decrease at the same rate. Consequently, the aqueous phase in the well block
starts accumulating. As a result, aqueous phase saturation in the well block starts
increasing which can be seen in the late production period in Figure 6-16.
Figure 6-17 shows the well block pressure at various times of the conventional
gas reservoir more rapidly decreases than the block pressure of the methane-hydrate
reservoir resulting in the more rapid decrease of gas production rate in the conventional
gas reservoir. This is because the hydrate decomposition releases methane gas and water
to the system. Accordingly, the well-block pressure of the methane hydrate system is
higher than the well-block pressure of the conventional gas system at any particular time.
From this simulation study, the gas pressure of the well block at the end of 1,000 days in
128
the methane hydrate reservoir is about 378 psia whereas the well block pressure in the
2500
2000
Pg (psia)
1500
1000
500
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
Figure 6-17: Well-block pressure of the conventional gas and methane-hydrate reservoirs
higher because there is more amount of gas available (free gas in the free-gas zone plus
methane gas in methane hydrates). For this particular example, if the thickness of the
conventional gas reservoir becomes 70 feet, the cumulative gas production at the end of
1,000 days of the new conventional gas reservoir and the methane-hydrate reservoirs will
be approximately the same. Figure 6-18 shows the cumulative gas productions of the two
systems. The cumulative gas production curves of the two systems are different even
though their cumulative gas productions at the end of 1,000 days are approximately the
same. This is because the initial thickness of free gas zone of the gas hydrate reservoir is
50 feet and the thickness of this zone increases with time due to the dissociation of gas
129
hydrate in the hydrate zone. On the other hand, the initial thickness of free gas zone of
the conventional gas reservoir is 70 feet and it never changes with time. Therefore, there
is more gas available to be produced for the conventional gas reservoir resulting in the
higher gas production rate during the early period of production. Note that the same
3000
2000
1000
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
Figure 6-18: Gas production of the conventional gas and methane-hydrate reservoirs
One can conclude from this simulation study that the presence of gas hydrate on
the top of free-gas reservoir can significantly improve the productivity of the underlying
gas reservoirs (using a constant bottom-hole pressure production scheme) are different.
For a conventional gas reservoir, both gas and water production rates decrease with time,
130
whereas, for a gas-hydrate reservoir, gas production rate decreases with time meanwhile
The depressurization paths at three different locations in the reservoir on the P-T
25 ft
1 Hydrate
50 ft zone
Free-gas
50 ft 3
zone
25 ft
Depressurization Path
Stabilization boundary Point (1)
3250
2750
Pressure (psia)
2250
1750
1250
750
45 50 55 60 65 70
Temperature (K)o
Temperature ( C)
131
Depressurization Path
Stabilization boundary Point (2)
3250
2750
Pressure (psia)
2250
1750
1250
45 50 55 60 65 70
Temperature (K)
Temperature ( oC)
Depressurization Path
2750
Pressure (psia)
2250
1750
1250
750
250
45 50 55 60 65 70
Temperature((K)
Temperature oC)
132
Figure 6-19 (b) represents the depressurization path of point 1 for 1,000 days of
production. For this point, the depressurization path starts from the point inside the
hydrate stabilization zone. Once the dissociation of hydrate phase at this point takes
place, the T-P condition at that point is on the 3-phase equilibrium line and after that the
change of P-T conditions at that point follows the equilibrium line. The P-T condition at
this point will go outside the hydrate stabilization zone when all the hydrate phase at that
point dissociates. At point 2 (Figure 6-19 (c)), the similar path as point 1 is obtained but
the starting point of this path is on the equilibrium line as the three phases coexist at the
initial condition. Similarly, at point 3 (Figure 6-19 (d)), the dissociation path starts from
the point outside the hydrate stabilization zone and it does not move into the hydrate
stabilization zone because no hydrate formation takes place at this point for the entire
simulation time.
performance of the production. In this section, the effects of four different well-
are examined. The height of the completion zone in every case is 10 feet. In this study,
three well-completion locations are in the free-gas zone and one well-completion location
is in the dissociation front layer. The penetrating zone of the production wells for Cases
1, 2, and 3 are located at the middle, top, and bottom of the initial free-gas zone,
respectively. For Case 4, the well is completed in the initial dissociation layer (the bottom
133
of the initial hydrate zone). Note that the well-completion locations in each case do not
change for the entire simulation. The production well is operated with a constant bottom-
hole pressure at 14.7 psia for every case. Figures 6-21 and 6-22 show the characteristics
Case 1 Case 2
Case 3 Case 4
134
Gas Production Rate
Location 1 Location 2 Location 3 Location 4
125
75
50
25
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
3500
3000
2500
2000
1500
1000
500
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
135
Water Production Rate
Location 1 Location 2 Location 3 Location 4
100
80
Production Rate (STB/day)
60
40
20
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
50000
40000
30000
20000
10000
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
136
The simulation results shows that gas production rates exponentially decrease
with time for every case, meanwhile the water production rates for each case have
different characteristics because the change of aqueous saturation at the well block with
time for each case are different. The water production characteristic for Case 1 has been
discussed in the previous section (6.5.1). For Case 2, the increase of aqueous saturation of
the well block happened earlier than Case 1 because the well block is just below the
dissociation front. Hence, the released water from the dissociation at the front requires
less time to flow to the well block. The increase of the aqueous phase saturation improves
the mobility of aqueous phase resulting in the increase of water production rate.
The change of aqueous phase saturation of the well block is controlled by the
amount of released water (from the dissociation) entering to the well block, the amount of
water leaving the well block through the production well, and the amount of water
flowing down to the lower part of the free-gas zone. This difference controls the change
of aqueous phase saturation and, consequently, water production rate. At about the end of
500 days of production, aqueous phase saturation started decreasing because the amount
of aqueous phase leaving the well block was higher than the amount of aqueous phase
entering the well block. Consequently, water production rate declined after 500 days of
production in this case. For Case 3, it is similar to Case 2 but the increase of aqueous
phase saturation required more time because the well block location of this case is lower
than the well block location in Case 2. Again, the change of aqueous phase saturation of
the well block is controlled by the difference of the amount of water entering and leaving
the well block. It can be seen that the water production rate at late time in Case 3 is
higher than that in Cases 1 and 2 because the well-completion location in Case 3 is lower
137
than the locations in the other two cases, which means that aqueous phase saturation of
the well block in Case 3 is higher than the aqueous saturations in the other two cases.
Consequently, the water production rate in Case 3 is the highest. Similarly, the water
production rate of Case 1 is higher than the water production rate of Case 2 because the
For Case 4, the water and gas production rates are less than the rates in the other
three cases because the permeability of the well block in this case is significantly smaller
(causing by the presence of methane hydrate) than the permeabilities in the other cases.
The water production rate increased with time because of the increase of the rock
Therefore, it is suggested that one should not complete a well (either producer or injector
The gas production rates in Cases 2 and 3 decrease more rapidly than the gas
production rate in Case 1 because of the effects of low permeability region above the
completion zone (for Case 2) and the effects of no flow boundary below the completion
zone (for Case 3). The gas production rate in Case 4 is significantly lower than the gas
production rates in the other cases due to the permeability issue discussed before.
The gas production rate in Case 2 drops more rapidly than that of Case 1 during
the early period of production. Hence, the cumulative gas production in Case 1 is higher
than that of Case 2 during 1,000 days of production (see Figure 6-21). However, the
cumulative gas production in Case 2 was, eventually, a bit higher than Case 1 (see Figure
6-23). This is because the completion zone of the well in Case 2 is higher than the
completion zone in Case 1 providing that the gas saturation of the well block in Case 2 is
138
higher than the gas saturation of the well block in Case 1. Accordingly, more gas is
3500
3000
2500
2000
1500
1000
500
0
0 250 500 750 1000 1250 1500
Time (days)
Figure 6-24 shows percent hydrate recovery and cumulative dissociated gas. Case
1 can dissociate methane-hydrate faster than the other cases during the first 500 days of
operation. However, at the end, Case 2 can dissociate all the methane-hydrate a bit sooner
than Case 1. Case 4 requires much more time than the other cases to dissociate all
methane-hydrate in the reservoir because it is harder to take the fluids out of the reservoir
(through the production well) due to low permeability issue previously mentioned.
139
Hydrate Recovery
90
80
Hydrate Recovery (%)
70
60
50
40
30
20
10
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
1500
1200
900
600
300
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
Figure 6-24: Hydrate recovery and cumulative dissociated gas for different
well-completion locations
140
For the previous four cases, the well-completion locations do not move with
respect to the location of dissociation front which moves upward during the dissociation
process. The following additional cases in which the well-completion locations are
moved with respect to the dissociation front are examined. In Case A, a production well
is initially completed in the middle of the initial free gas zone. The thickness of free-gas
zone increases with time because the dissociation front moves upward during the
dissociation process. Therefore, the completion location must be moved upward with
respect to the dissociation front location so that the completion location is always in the
middle of free-gas zone. In Case B, a production well is initially completed at the top of
the initial free-gas zone and just below the initial dissociation front. Similarly, the
completion location must be moved with respect to the location of the dissociation front
so that the completion location is always at the top of free-gas zone. Figure 6-25 shows
the cumulative gas and water productions of these two cases as compared to the
production plots of Case 1. Figure 6-25 shows that Cases A and B yield higher
cumulative gas production than Case 1 (by 6.75 and 3.36 percents, respectively)
indicating that moving the completion location improves gas production performance.
This is because the completion locations in cases A and B are moved to higher locations
which have larger gas saturation resulting in higher gas production rate and cumulative
gas production and lower water production (Figure 6-25). Case B yields lower gas
production than Case A (by 3.28 percent) because the completion location in Case B is
just below the dissociation front which releases water from the dissociation process at the
time, resulting in the high water saturation just below the dissociation front.
141
Cumulative Gas Production
Location 1(non-moving) (A) Location 1 (moving) (B) Location 2 (moving)
4000
3500
3000
2500
2000
1500
1000
500
0
0 250 500 750 1000 1250 1500
Time (days)
40000
30000
20000
10000
0
0 250 500 750 1000 1250 1500
Time (days)
142
As a result, Case B yields higher water production and lower gas production than
It can be concluded from this examination that one should not complete a well in
the hydrate zone because permeability in this zone is significantly lower than the
permeability in free-gas zone. The simulation results show that a moving well-completion
location strategy yields better gas production performance than a fixed well-completion
location strategy. Completing a production well in the middle of free-gas zone with a
moving completion location strategy yields the best gas production performance.
In this section, the effect of well spacing on the production performances from a
class 1 methane-hydrate reservoir shown in Figure 6-26 is examined. There are a number
of production wells in this system. Considering a well located in a gray square in Figure
6-26, the interferences from the production wells surrounding this well causes the no flow
boundary around the center well (dash line in Figure 6-26). Due to the symmetry of the
reservoir, each portion (square area) has the same pressure, temperature, and saturation
distributions at any time. Therefore, the simulation study of the entire system can be
143
Figure 6-26: System configuration of the production characteristics study
14.7 psia. The production characteristics of a reservoir using three different well-drainage
areas are examined. The gas production rate and cumulative produced gas from one well
for different well-spacing systems are shown in Figure 6-27. Figure 6-28 shows the
cumulative dissociated gas and percent hydrate recovery for the three different systems.
144
Gas Production Rate
22.95 acres 45.00 acres 74.38 acres
125
75
50
25
0
0 200 400 600 800 1000 1200 1400 1600
Time (days)
10000
9000
8000
7000
6000
5000
4000
3000
2000
1000
0
0 200 400 600 800 1000 1200 1400 1600
Time (days)
Figure 6-27: Gas production from one well for different well-spacing systems
145
Cumulative Dissociated Gas
22.95 acres 45.00 acres 74.38 acres
5000
3000
2000
1000
0
0 200 400 600 800 1000 1200 1400 1600
Time (days)
Hydrate Recovery
22.95 acres 45.00 acres 74.38 acres
100
90
80
Hydrate recovery (%)
70
60
50
40
30
20
10
0
0 200 400 600 800 1000 1200 1400 1600
Time (days)
146
During the very early production period, the cumulative gas production lines from
the three systems are the same because the pressure transient has not reached the
boundary of the system. The cumulative gas production line of the smallest well spacing
system (22.95 acre) deviates from the cumulative gas production lines of the other two
systems indicating the effect of the system boundaries on the gas production. The
deviation of the cumulative gas production lines of the other two systems happen later
because the pressure transient takes more time to reach the boundary of the 45-acre
system. The plots of cumulative dissociated gas show the same characteristics.
The gas production rate (from one well) of the larger well-drainage area is higher
because the amount of gas and methane hydrate available is higher. Thus, the cumulative
produced gas (from one well) of the larger well-spacing system is higher. The results
indicate that the smaller well-spacing system requires less time to dissociate 100 percent
and 1,540 days to dissociate 100 percent of methane hydrate for the 22.95, 45.00, and
74.38 acres of well spacing systems, respectively. Figure 6-29 shows percent hydrate
In order to compare the production efficiency of these three systems, the system
must have the same drainage area. In this comparison, the cumulative gas production and
hydrate recovery from a 450-acre reservoir using these three different well spacings are
investigated. In this case, the number of production wells for the 22.95, 45.00, and 74.83
acre well-spacing systems are approximately 20, 10, and 6, respectively. The cumulative
produced gas and hydrate recovery of the three cases for this reservoir are shown in
Figure 6-30.
147
Hydrate recovery (%)
100
80
40
20
0
1 2 3
22.95 45.00 74.38
Well spacing (acres)
100
80
60 100.00 88.89
40
20
0
1 45.00 2 74.38
Figure 6-29: Hydrate recovery of different well-spacing at 750 and 1,080 days
148
Cumulative Gas Production
22.95 acres 45.00 acres 74.38 acres
80000
Cumulative Gas Production (MMSCF)
70000
60000
50000
40000
30000
20000
10000
0
0 200 400 600 800 1000 1200 1400 1600
Time (days)
Hydrate Recovery
22.95 acres 45.00 acres 74.38 acres
100
90
80
Hydrate recovery (%)
70
60
50
40
30
20
10
0
0 200 400 600 800 1000 1200 1400 1600
Time (days)
Figure 6-30: Cumulative gas production and hydrate recovery of the three
different well spacings for a 450 acre reservoir
149
As expected, the smaller well spacing system yields higher cumulative produced
gas and faster hydrate dissociation. This indicates that the gas production performance
and hydrate recovery efficiency can be improved by increasing the number of wells in the
system (or reduce well spacing). However, increasing the number of wells require higher
investment. The cumulative produced gases (at the end of 1,600 days of operation) of the
three cases are shown in Table 6-19. Note that in this simulation study, the gas
production rates of all cases become very small at the end of 1,600 days of production
indicating that the ultimate cumulative gas productions of the three cases do not
Table 6-19: Cumulative produced gases for three different well spacings
Well spacing Number of well Gas Production Improvement
(acres) (MMSCF) (%)
22.95 20 74365 15.34
45.00 10 69891 8.40
74.38* 6 64474 0.00
* Reference case
Note that the values shown in Table 6-19 may change from one system to the
other system. Therefore, these values shall not provide the accurate optimum well
spacing for systems different from the one studied in this work.
150
6.5.4 Implementation of a Different Production Schedule
wells as shown in Figure 6-31 using four different production schedules are examined in
this section. All production wells are operated at 14.7 psia (constant bottom-hole pressure
scheme) but they are put on production at different times. Four different production
500 ft 500 ft
= Production well
500 ft
1 2
1,000 ft
3 4
1,000 ft
500 ft
1,000 ft 1,000 ft
Figure 6-31: Well structure in the methane-hydrate reservoir used in this study
151
Figure 6-32 shows the percent methane-hydrate recovery and cumulative
dissociated gas for each of these four production schedules and the plots of gas and water
productions for these cases are shown in Figures 6-33 and 6-34, respectively.
Hydrate Recovery
Case (a) Case (b) Case (c) Case (d)
100
90
80
70
Hydrate Recovery (%)
60
50
40
30
20
10
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
5000
4000
3000
2000
1000
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
Figure 6-32: Hydrate recovery and cumulative dissociated gas for each case
152
Total Gas Production Rate
Case (a) Case (b) Case (c) Case (d)
450
400
350
Production Rate (MMSCF/day)
300
250
200
150
100
50
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
15000
Cumulative Produced Gas (MMSCF)
12500
10000
7500
5000
2500
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
153
Total Water Production Rate
Case (a) Case (b) Case (c) Case (d)
200
175
150
Production Rate (STB/day)
125
100
75
50
25
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
70000
Cumulative Water Production (STB)
60000
50000
40000
30000
20000
10000
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
154
The results show that Case (a) can dissociate 100 percent of methane-hydrate
earlier than the other four cases providing the higher cumulative dissociated gas in Case
(a). On the other hand, the cumulative gas production of Case (a) is lower than the
cumulative gas productions of Cases (b), (c), and (d). As shown in Figure 6-34, the
cumulative production and production rate of water in Case (a) is higher than that of the
other cases.
The plots of water saturation of the well blocks for these four cases are shown in
Figure 6-35. The increase of water saturation at the well blocks in Case (a) is higher than
that in the other cases resulting in the higher increase of water production rate in Case (a).
The change of aqueous phase saturation for these cases can be explained by the
discussion of the saturation change in section 6.5.1 because the shapes of the water
saturation plots shown in Figure 6-35 are similar to the shape of the saturation plot in
For Case (a), the changes of aqueous phase saturations at each well block are the
same because of the system’s symmetry (the system can be symmetrically divided into 4
parts). According to Figures 6-32 and 6-35, methane hydrate can be dissociated more
rapidly in Case (a) because all the production wells are put on production at the
dissociation) into the system in this case resulting in the higher (and sooner) increase of
aqueous phase saturations as shown in Figure 6-35. Thus, water production rate in Case
(a) is the highest. On the other hand, higher increase of aqueous phase saturation implies
higher decrease for free-gas phase saturation. Hence, the mobility of free-gas phase in
Case (a) more rapidly decreases than the other cases. Therefore, the gas production rate in
155
Case (a) drops more rapidly than the production rate in the other cases resulting in a
0.8
0.7
Sw (fraction)
0.6
0.5
0.4
0.3
0.2
0 200 400 600 800 1000
Time (days)
0.7
Sw (fraction)
0.6
0.5
0.4
0.3
0.2
0 200 400 600 800 1000
Time (days)
156
Water saturation at well block of Case (c)
well no. 1 well no. 2 well no. 3 well no. 4 Case (a)
0.8
0.7
Sw (fraction)
0.6
0.5
0.4
0.3
0.2
0 200 400 600 800 1000
Time (days)
0.7
Sw (fraction)
0.6
0.5
0.4
0.3
0.2
0 200 400 600 800 1000
Time (days)
Among the remaining cases (Cases b, c, and d), the hydrate dissociation rate in
Case (d) is higher than that in the other two cases. As a result, the water production in this
case is higher whereas the gas production is lower than those in the other two cases.
157
Hydrate dissociation rate in cases (b) and (c) are slightly different. Therefore, gas and
water production from these two cases are not significantly different. The plots of water
saturation distribution on the well block layer (on the x-y plane) at various times for these
158
Aqueous Saturation at t = 0 day Aqueous Saturation at t = 60 days
-1000 -1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
Figure 6-36: Aqueous phase saturation of well block layer for Case (a)
159
Aqueous Saturation at t = 0 day Aqueous Saturation at t = 90 days
-1000 -1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
Figure 6-37: Aqueous phase saturation of well block layer for Case (b)
160
Aqueous Saturation at t = 0 day Aqueous Saturation at t = 60 days
-1000 -1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
Figure 6-38: Aqueous phase saturation of well block layer for Case (c)
161
Aqueous Saturation at t = 0 day Aqueous Saturation at t = 90 days
-1000 -1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500
500
1000
1000
-1000 -500 0 500 1000
-1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
Figure 6-39: Aqueous phase saturation of well block layer for Case (d)
162
Figure 6-40 to Figure 6-43 show hydrate saturation distribution of the initial
dissociation front layer (layer no. 9 in Table 6-17) at various times for the four production
schedules. As expected, Case (a) requires less time (160 days) than the other cases for
dissociating all the hydrate phase in the initial dissociation front layer because all the
wells are operated at the same time. Case (d) requires less time (270 days) than the other
two remaining cases because two wells are put on production at t = 0 and Cases (b) and
(c) need approximately the same amount of time (300 days) for dissociating all the
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
Figure 6-40: Hydrate saturation of the initial dissociation front layer for Case (a)
163
Hydrate Saturation at t = 0 day Hydrate Saturation at t = 90 days
-1000 -1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
Figure 6-41: Hydrate saturation of the initial dissociation front layer for Case (b)
164
Hydrate Saturation at t = 0 day Hydrate Saturation at t = 60 days
-1000 -1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
Figure 6-42: Hydrate saturation of the initial dissociation front layer for Case (c)
165
Hydrate Saturation at t = 0 day Hydrate Saturation at t = 90 days
-1000 -1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
-500 -500
0 0
500 500
1000 1000
-1000 -500 0 500 1000 -1000 -500 0 500 1000
Figure 6-43: Hydrate saturation of the initial dissociation front layer for Case (d)
166
Figure 6-44 shows the gas to water ratio and the cumulative gas to water ratio for
these four cases. The gas to water ratio is calculated by dividing total gas production rate
by total water production rate at any particular time. Similarly, the cumulative gas to
water ratio is the ratio of the total cumulative gas production to the total cumulative water
production. They represent the amount of produced gas per unit volume of produced
Both gas to water and cumulative gas to water ratios for Case (a) decrease more
rapidly than the ratios for the other cases because of the higher increase of water
production rate and decrease of gas production rate as previously discussed, meanwhile
the ratios for Cases (b) and (c) are higher than the other two cases, and the ratios for Case
(b) is a bit higher than the ratio for Case (c). The cumulative gas to water ratio
(MMSCF/STB) and cumulative produced gas (MMSCF) at the end of 1,000 days for the
The results indicate that the ratio for Case (b) is the highest among the four cases
and it shows that the ratio could improve 119 percent by delaying the operation of
production wells. However, the cumulative gas production in Case (b) is higher than that
of Case (a) just about 10.8 percent, not 119 percent. This is because the cumulative water
production in Case (a) is higher than that of Case (b). When considering about the
cumulative gas production, Case (c) represents the best production schedule among the
167
Gas-Water Ratio
case (a) case (b) case (c) case (d)
80
70
60
GWR (MMSCF/SCF)
50
40
30
20
10
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
70
60
CGWR (MMSCF/SCF)
50
40
30
20
10
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
168
Table 6-21: Cumulative Gas to Water Ratio at 1,000 days for each case
Ratio Amount
Case
CGWR % Improvement Qg % Improvement
a* 0.208 0.0 14714 0.0
b 0.456 119.0 16297 10.8
c 0.449 115.6 16383 11.3
d 0.367 75.9 16040 9.0
* Reference case
The percent hydrate recovery, cumulative gas production, and cumulative gas to
water ratio obtained from this simulation study indicate that Cases (b) and (c) are the best
production schemes (the production efficiencies of these two cases are not significantly
different) among the four cases. They yield higher gas production even though less
hydrate is dissociated in the system and as thereby leaving still more hydrate to be
recovered.
The results shows that lowering hydrate dissociation rate by putting production
shown in section 6.5.3, lowering the hydrate dissociation rate by reducing number of
production wells in the system should not necessarily improve production efficiency. The
following simulation exercise has been performed to verify this observation. The two
different production strategies for the system shown in Figure 6-31 are investigated. In
Case 1, all wells are put on production at the same time whereas only wells no. 1 and 4
are put on production (at the same time) in Case 2. The plots of percent hydrate recovery
and cumulative produced gas of these two cases are shown in Figure 6-45.
169
Hydrate Recovery
Case 1 Case 2
100
90
80
60
50
40
30
20
10
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
15000
Cumulative Produced Gas (MMSCF)
12500
10000
7500
5000
2500
0
0 100 200 300 400 500 600 700 800 900 1000
Time (days)
Figure 6-45: Hydrate recovery and cumulative produced gas of Cases 1 and 2.
than Case 1, and the cumulative gas production at the end of 1,000 days of Case 1
(14,714 MMSCF) is higher than the cumulative gas production at the end of 1,000 days
170
From this simulation study, it could be concluded that in a multiple-well system,
putting every well on production at the same time yields the higher dissociation rate but it
would result in lower gas production. Putting the wells on production at different times
reduces hydrate dissociation rate but it seems to improve the recovery efficiency.
Decreasing hydrate dissociation rate by reducing the number of production wells does not
171
CHAPTER 7
system has been developed to examine the production characteristics from methane-
hydrate reservoirs. The model accounts for heat transfer due to conduction and
capillary pressure between free-gas and aqueous phases are also taken into account. The
technique) from Class 1 methane-hydrate reservoirs. In this study, the constant bottom-
characteristics of Class 1 methane-hydrate reservoirs have been derived from this work:
• Unlike conventional gas reservoirs in which both gas and water production rates
exponentially decline with time when the well is operated at a constant bottom-
hole pressure, gas production rate exponentially decreases with time meanwhile
water production rate increases with time for gas-hydrate reservoirs. Because gas-
hydrate dissociation releases both free-gas and aqueous phases to the system, the
released aqueous phase flows down to the free-gas zone (due to the gravitational
force) resulting in the increase of aqueous phase saturation in the free-gas zone.
172
• The effect of well-completion location on gas production performance has been
better gas production performance than all the cases of fixed well-completion
(using a moving completion location strategy) yields the best gas production
performance (it provides the highest gas production and lowest water production).
• The effect of well spacing on the production efficiency was also investigated. The
larger well spacing provides higher gas production per well because more gas and
gas hydrate are available. However, the total gas production rate and cumulative
gas production (for a specific reservoir) of the larger well-spacing system are
lower than the total gas production rate and cumulative gas production of a
smaller well-spacing system. This implies that, for a specific reservoir, drilling
more well yields higher gas production but it requires higher investment. Based
on the simulation results in this study, the cumulative gas production decreased
just about 8.4 percent when the well spacing increases from 45.0 acre to 74.38
acre. This value could change depending upon the rock and fluid properties and
reservoir structure.
performance has also been examined. The simulation results show that starting all
production wells at the same time provides faster hydrate dissociation. However,
the more effective gas production is observed when starting production wells at
different times. This is because the faster hydrate dissociation yields higher rate of
water released from gas hydrate phase to the system. The released water flows
173
down to the free-gas zone due to the gravitational effect resulting in the increase
phase increases meanwhile the mobility of free gas decreases. Therefore, the
faster dissociation rate does not necessarily improve gas production efficiency.
The simulation results show the improvement of gas production efficiency when
the production wells are put on production at different time (which yields lower
dissociation rate than the case when all wells are put on production at the same
174
BIBLIOGRAPHY
1. Makogon Y. and Holditch S., “Gas Hydrates as a Resource and a Mechanism for
(2002)
SPE Paper 75691 presented in the SPE Gas Technology Symposium held in
3. Sloan E.D. JR, “Clathrate Hydrates of Natural Gases”, 2nd Marcel Dekker, Inc,
4. Moridis G.J. and Collett T.S., “Strategies for Gas Production From Hydrate
Containing both Gas and Free Natural Gas”, SPE paper 11105 presented at the
57th SPE Annual Technical Conference and Exhibition in New Orleans, LA,
175
7. Yousif M.H., Abass H.H., Selim M.S., and Sloan E.D, “Experimental and
8. Chuang Ji, Goodarz Ahmadi, and Duane H. Smith, “Natural Gas Production
9. Sun X., Nanchary N., and Mohanty K.K., “1-D Modeling of Hydrate
11. Moridis G.J., Kowalsky M.B., and Pruess K., “Depressurization-Induced Gas
14. Ozisic M.N., “Heat Conduction”, 2nd ed., John Wiley&Sons, Inc., (1993)
176
15. Phale H.A., Zhu T., White M.D., and McGrail B.P., “Simulation Study on
16. White M.D. and Oostrom M., “STOMP-Subsurface Transport Over Multiple
17. Udding M., Alberta Research Council Inc., Coombe, D.A., Computer Modeling
Group Ltd., Law, D.A., Schlumberger Reservoir Fluid Center, and Gunter, W.D.,
18. http://www.pet.hw.ac.uk/research/hydrate/hydrates_what.html
19. Van der Waals J.H. and Platteeuw J.C., “Clathrate Solutions”, Adv. In Chem.
20. Holder G.D., Zetts S.P., Pradham N., “Phase Behavior in Systems Containing
22. Kamath V.A., “Study of heat transfer characteristics during dissociation of gas
Pittsburgh, PA (1984)
23. Makogon Y.F., “Hydrates of Hydrocarbons”, Penn Well, Tulsa, OK. (1997)
177
24. Moridis G.J., “Numerical Studies of Gas Production From Methane Hydrates”,
75691 (2002)
25. Verma A., and Pruess K., Tsang C.F., and Witerspoon P.A., “A Study of Two-
26. Xu T., Ontoy Y., Molling P., Spycher N., Parini M., and Pruess K., “Reactive
27. Somerton W.H., et al., “Thermal Behavior of Unconsolidated Oil Sands”, SPE
28. Lee A. L., Gonzalez M. H., and Eakin B. E., “The viscosity of Natural Gases”,
JPT, (1996)
30. Alp Doruk, “Gas Production from Hydrate Reservoirs”, Master Thesis, Middle
(1993)
178
APPENDIX A
system shown in Figure A-1. The control volume has a volume ΔxΔyΔz, porosity φ, and
permeabilities kx, ky, kz in the x, y, and z directions, respectively. This control volume
hosts up to two mobile phases (aqueous and free gas) and two immobile phases (gas
hydrate and ice) at saturations Sa, Sg, Sh and Si, respectively. Figure A-1 shows the molar
flow rates for component i. Variable Qi* represents the external sink or source term of
component i.
x
N ia z + Δz
N ig
z + Δz
* N ig
y z
Q i y
N ia y
N ig N ig
x x + Δx
N ia x
N ia x + Δx
N ig
y + Δy
N ia y + Δy
N ia z
N ig
z
179
The change in the amount of component i within the elementary volume is a
results of fluid leaving and entering the elementary volume. Over a period of time Δt, one
The units of each of the terms used in this derivation are lbmole/day. In this
derivation, a positive sign is assigned to flow entering the elementary volume whereas a
negative sign to flow leaving the elementary volume. Accordingly, the molar balance of
⎡ ⎤ ⎡ ⎤
− ⎢ ∑ ( N ilx ) x+Δx + ∑ ( N ily ) y +Δy + ∑ ( N ilz ) z +Δz ⎥ + ⎢ ∑ ( N ilx ) x + ∑ ( N ily ) y + ∑ ( N ilz ) z ⎥
⎣l = a , g l = a, g l = a, g ⎦ ⎣l = a , g l = a, g l = a, g ⎦
⎡ ⎤ ⎡ ⎤
⎢φ ∑ ( S l Cil )⎥ − ⎢φ ∑ ( Sl Cil )⎥
⎣ l = a , g ,h ⎦ t +Δt ⎣ l = a , g ,h ⎦t
+ Qi* = Δx ⋅ Δy ⋅ Δz A.2
Δt
Here, (Nils)s is the number of mole of component “i” in “l” phase entering the
elementary volume in the s direction at position “s” and (Nils)s+Δs is the number of mole of
component “i” in “l” phase leaving the elementary volume in the s direction at position
“s+Δs”.
The molar flow of the ith component is a function of velocity of the phase, the
concentration of the component in the phase, and the area perpendicular to the flow as the
following expression:
N ils s= x, y ,z
= Cl xil vls s= x, y,z
As s= x, y ,z
N ils s = x + Δx , y + Δy , z + Δz
= Cl xil vls s = x + Δx , y + Δy , z + Δz
As s = x + Δx , y + Δy , z + Δz
A.3
180
where: vls is superficial velocities of the “l” phase in the “s” direction
xil is the mole fraction of the ith component in the “l” phase
Ax = Δy ⋅ Δz ; Ay = Δx ⋅ Δz ; and Az = Δx ⋅ Δy A.4
⎡ ⎤ Δx
− ⎢( ∑ (Cl xil vlx Ax ) x+Δx − ∑ (C xil vlx Ax ) x ⎥
⎦ Δx
l
⎣ l = a, g l = a, g
⎡ ⎤ Δy
− ⎢( ∑ (Cl xil vly Ay ) y +Δy − ∑ (C xil vly Ay ) y ⎥
⎦ Δy
l
⎣ l = a, g l = a, g
⎡ ⎤ Δz
− ⎢( ∑ (Cl xil vlz Az ) z +Δz − ∑ (Cl xil vlz Az ) z ⎥
⎣ l = a, g l = a, g ⎦ Δz
⎡ ⎤ ⎡ ⎤
⎢φ ∑ ( Sl xil Cl )⎥ − ⎢φ ∑ ( Sl xil Cl )⎥
⎣ l = a , g ,h ⎦ t +Δt ⎣ l = a , g ,h ⎦t
+ Qi* = Vb A.5
Δt
Diving equation (A.5) by ΔxΔyΔz and then taking the limits of Δx, Δy, Δz, and Δt
∂ ⎡ ⎤ ∂ ⎡ ⎤ ∂ ⎡ ⎤
− ⎢ ∑ (C l xil vlx Ax )⎥ Δx − ⎢ ∑ (C l xil vly Ay )⎥ Δy − ⎢ ∑ (C l xil vlz Az )⎥ Δz
∂x ⎣l = a , g ⎦ ∂y ⎣l = a , g ⎦ ∂z ⎣l = a , g ⎦
∂⎛ ⎞
+ Qi* = Vb ⎜⎜ φ ∑ ( S l xil C l ) ⎟⎟
∂t ⎝ l = a , g , h ⎠ A.6
The fluid flow mechanic in porous media can be described by Darcy’s law as
k s k rl ∂Φ l
vls = − 5.615 A.7
μl ∂s
181
where ks is the absolute permeability in the “s” direction
ρl
Cl = = ρl A.8
MWl
∂ ⎡ ⎛ xil ρ l Ax k x k rl ∂Φ l ⎞⎤ ∂ ⎡ ⎛ xil ρ l Ay k y k rl ∂Φ l ⎞⎤
⎢ ∑ ⎜⎜ ⎟⎟⎥ Δx + ⎢ ∑⎜ ⎟⎥ Δy
∂x ⎣⎢l = a , g⎝ μl ∂x ⎠⎦⎥ ∂y ⎢⎣l = a , g⎜⎝ μl ∂y ⎟⎠⎦⎥
∂ ⎡ xil ρ l Az k z k rl ∂Φ l ⎤ Qi* Vb ∂ ⎛ ⎞
− ⎢ ⎥ Δz + = ⎜ φ ∑ ( S l xil ρ l ) ⎟
∂z ⎣ μl ∂z ⎦ 5.615 5.615 ∂t ⎜⎝ l = a , g , h ⎟
⎠
A.9
Equation (A.9) is the general compositional mole balance equation. The mole balance
Methane:
∂ ⎡ Ax k x k rg ⎛ ∂Φ g ⎞ Ak k⎞⎤ ⎛ ∂Φ a
⎢ ym ρ g ⎜⎜ ⎟⎟ + x m ρ a x x ra
⎟⎥ Δx ⎜
∂x ⎢⎣ μg ⎝ ∂x ⎠ μa⎠⎥⎦ ⎝ ∂x
∂ ⎡ Ay k y k rg ⎛ ∂Φ g ⎞ A k k ⎛ ∂Φ a ⎞⎤
+ ⎢ ym ρ g ⎜⎜ ⎟⎟ + x m ρ a y y ra ⎜⎜ ⎟⎥ Δy
∂y ⎣⎢ μ g ⎝ ∂y ⎠ μ a ⎝ ∂y ⎟⎠⎦⎥
∂ ⎡ Az k z k rg ⎛ ∂Φ g ⎞ Ak k ⎛ ∂Φ a ⎞⎤ Qm
+ ⎢ ym ρ g ⎜⎜ ⎟⎟ + x m ρ a z z ra ⎜ ⎟⎥ Δz +
∂z ⎢⎣ μg ⎝ ∂z ⎠ μa ⎝ ∂z ⎠⎥⎦ 5.615
⎛ V ⎞ ∂
= ⎜ b ⎟ ( y mφ S g ρ g + x m φ S a ρ a + N mφ S H ρ H ) A.10
⎝ 5.615 ⎠ ∂t
182
Water:
∂ ⎡ Ax k x k rg ⎛ ∂Φ g ⎞ A k k ⎛ ∂Φ a ⎞⎤
⎢(1 − y m ) ρ g ⎜⎜ ⎟⎟ + (1 − x m ) ρ a x x ra ⎜ ⎟⎥ Δx
∂x ⎣⎢ μg ⎝ ∂x ⎠ μ a ⎝ ∂x ⎠⎦⎥
∂ ⎡ Ay k y k rg ⎛ ∂Φ g ⎞ Ak k⎞⎤ ⎛ ∂Φ a
+ ⎢(1 − y m ) ρ g ⎜⎜ ⎟⎟ + (1 − x m ) ρ a y y ra
⎟⎟⎥ Δy⎜⎜
∂y ⎢⎣ μg ⎝ ∂y ⎠ μa ⎠⎥⎦ ⎝ ∂y
∂ ⎡ Az k z k rg ⎛ ∂Φ g ⎞ A k k ⎛ ∂Φ a ⎞⎤ Qw
+ ⎢(1 − y m ) ρ g ⎜⎜ ⎟⎟ + (1 − x m ) ρ a z z ra ⎜ ⎟⎥ Δz +
∂z ⎢⎣ μ g ⎝ ∂z ⎠ μ a ⎝ ∂z ⎠⎥⎦ 5.615
⎛ V ⎞ ∂
= ⎜ b ⎟ ((1 − y m ) φ S g ρ g + (1 − x m ) φ S a ρ a + φ S I ρ I + N wφ S H ρ H ) A.11
⎝ 5.615 ⎠ ∂t
183
APPENDIX B
In this section, the simulation results (for a conventional gas system) from the
model developed in this work are compared with the results from a commercial reservoir
The developed model is used for simulating the gas production from a two-
center of the reservoir as shown in Figure B-1 and the reservoir properties are shown in
Table B-1.
184
Table B-1: Reservoir and grid block properties for model validation
Reservoir porosity 0.2
Top depth (ft) 1,000
Reservoir thickness (ft) 100
dx, dy (ft) 500
kx,ky (md) 100
Initial Pressure (psi) 800
Initial water saturation 0.6
Reservoir temperature (oF) 40
Well bore radius (ft) 0.25
The comparison for the isothermal system between the results from the model and
commercial simulator of various well specification values are shown in Figure B-2 to
Figure B-4.
The comparisons show the very good matches of the results from the developed
model and the commercial simulator. It indicates the robustness of the simulator
185
CUMULATIVE GAS PRODUCTION CUMULATIVE WATER PRODUCTION
(Qw is specified @ 50 STB/day) (Qw specified @ 50 STB/day)
17500
1500.0
15000
1250.0
12500
Qg (MMSCF)
Qw (STB)
1000.0
10000
750.0
7500
500.0
5000
250.0 2500
0
0.0
0 50 100 150 200 250 300 350 400
0 50 100 150 200 250 300 350 400
6.50 750
6.00 700
5.50 650
Qg (MMSCF/day)
5.00 600
Psf (Psi)
4.50 550
4.00 500
3.50 450
3.00 400
2.50 350
2.00 300
0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400
Time (days)
Time (days)
Figure B-2: The comparison when water flow rate was specified at 50 STB/day
17500.0
1250
15000.0
1000
12500.0
Qg (MMSCF)
Qw (STB)
750 10000.0
7500.0
500
5000.0
250
2500.0
0
0.0
0 50 100 150 200 250 300 350 400
0 50 100 150 200 250 300 350 400
Time (days) Time (days)
750
60.0
700
650
55.0
600
Qw (STB/day)
Psf (Psi)
50.0
550
500
45.0
450
40.0 400
350
35.0
300
30.0 250
0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400
Figure B-3: The comparison when gas flow rate was specified at 4 MMSCF/day
186
CUMULATIVE GAS PRODUCTION CUMULATIVE WATER PRODUCTION
(Psf is specified @ 500 psia) (Psf is specified @ 500 psia)
1250 12500.0
1000 10000.0
Qg (MMSCF)
Qw (STB)
750 7500.0
500 5000.0
250
2500.0
0
0.0
0 50 100 150 200 250 300 350 400
0 50 100 150 200 250 300 350 400
16
125.0
14
12 100.0
Qg (MMSCF/day)
Qw (STB/day)
10
75.0
8
6 50.0
4
25.0
2
0 0.0
0 50 100 150 200 250 300 350 400 0 50 100 150 200 250 300 350 400
Figure B-4: The comparison when sand face pressure was specified at 500 psia
187
VITA
SUNTICHAI SILPNGARMLERT
Education
Ph.D., Petroleum and Natural Gas Engineering
2003–2007
The Pennsylvania State University, University Park, PA, U.S.A.
Thesis: “Numerical Modeling of Gas Recovery from Methane Hydrate Reservoirs ”
G.P.A. 3.81/4.0
Honors &
• Research Assistantship at the Pennsylvania State University, PA, USA. (2003-
Activities present)
• Member of the Society of Petroleum Engineers (2003 – present)
• Recipient of the National Science and Technology Development Agency
Scholarship (1997 – 1999)