Вы находитесь на странице: 1из 8

Experimental and Theoretical

Investigation of Methane-Gas-Hydrate
Dissociation in Porous Media
M.H. Yousif, * SPE, H.H. Abass, * * SPE, M.S. Sellm, and E.D. Sloan, SPE, Colorado School of Mines
SPE 1832.0
Summary. Vast quantities of natural gas deposits exist in the form of solid hydrates. Methane gas hydrate was formed and dissociated
for the first time in Berea core samples. A three-phase ID model was developed to simulate the process of gas production from Berea
sandstone samples containing methane hydrate by means of a depressurization mechanism. The model closely matched the experimental
data of gas and water production, the progress of the dissociation front, and the pressure and saturation profiles.

Natural-gas hydrates are crystalline, ice-like substances belonging pO, and temperature, TO. At time t=O, the pressure at the bound-
to a class of compounds called clathrates. 1 These chemical com- ary x=O is lowered to a new pressure that is less than the equilibri-
pounds have natural-gas molecules bound within almost spherical um pressure, Pe' and is held constant thereafter. Hydrate
water cages through physical rather than chemical bonds. Two dissociation begins; pressure and saturation distributions develop
natural-gas-hydrate crystal structures result from the combination throughout the system. Water and gas flow through the system be-
of these water cages. The crystal properties of both structures have cause of the resultant pressure gradient and are produced at the
been reviewed in detail. 2 When all cavities of either structure are boundary x=O.
occupied, there are about 15 gas molecules per 85 water molecules. For simplicity, we assume that the depressurization process is
Therefore, the enclathrated gas concentration is much higher than carried out under isothermal conditions. Such an assumption is valid
the gas solubility in water (usually on the order of one part per thou- only if the boundary pressure is not far enough below the equilbrium
sand). The effect of this concentration is that each volume of hydrate pressure to ensure slow dissociation. With this restriction, the trans-
can contain as many as 170 volumes of gas at standard conditions. port equations governing hydrate dissociation and the flow of gas
Kvenvolden 3 estimated that the total gas reserve in hydrates is and water through the porous medium are
1.87 x 10 17 std m 3 •
Many schemes have evolved for the recovery of such a large
resource of natural gas bound in the solid state. The three most
practical schemes are (1) thermal stimulation, in which an external
source of energy is used; (2) depressurization, in which the pres-
sure of an adjacent gas phase is lowered to cause decomposition;
and (3) inhibitor injection, in which methanol or some combina- ........................ (2)
tion of inhibitors is used to disequilibrate the system. Of these
schemes, only the first two have been addressed through mathe-
. a(4)P H SH)
matical modeling. and -mH= .............................. (3)
Thermal-stimulation models have emerged from four labora- at
tories. 4-7 Depressurization, on the other hand, has been the method
used in the single example of hydrate production from the Mes- The quantity -mH in Eq. 3 represents the local mass rate ofhy-
soyakha field in the U.S.S.R. Verigin et al. 8 proposed an isother- drate dissociated per unit volume as a result of depressurization.
mal depressurization model in which heat was assumed to flow The quantities mg and mw in Eqs. 1 and 2 represent the corre-
instantaneously from the surroundings to the hydrate body. Holder sponding local mass rate of gas and water produced per unit volume,
and Angert 9 proposed a methane-hydrate global depressurization respectively. These are related by
model in which the heat of dissociation came from the sensible heat
of the reservoir itself. Burshears et al. 10 extended this model 9 to mH=mg+mw ..................................... (4)
gas mixtures. Yousif et at. 11 proposed a model similar to the Rus-
sian moving-boundary model, but allowed for a temperature gra-
dient within the hydrated zone; this model, which has had some M
and mg =mH g . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (5)
limited experimental confirmation, shows the inverse relationship NHMw+Mg
of production with time, relative to the work of Verigin et al. 8
In all the models proposed to date, the flow equations in a hy- The local gas-generation rate caused by hydrate dissociation may
drate zone were not considered as a complement to the equations be obtained from the Kim-Bishnoi model 12 :
of mass and energy. One purpose of this work was to model the
hydrate-depressurization process considering equations of change mg =kdAs(Pe -p) . ................................. (6)
for both mass and momentum for each of the three phases present
(gas, water, and hydrate) in the porous medium. It was our pur- During the dissociation process, the volume occupied by the gas
pose to provide experimental measurements of slow, isothermal hy- and water continuously increases with time as a result of depletion
drate depressurization in porous media for confirmation or of the hydrate phase. Accordingly, the PV occupied gas and water
refinement of the model. and the permeability of the porous medium will be changing con-
tinuously with time. Let 4>wg denote the PV occupied by water and
Depressurization Model gas per unit volume of porous medium. Assuming that the hydrate
Consider a porous medium of uniform porosity occupying the region phase forms a continuous film coating the interior of the solid sur-
O<x<L. Initially, hydrate, water, and gas are at uniform pressure, face of the porous medium, we may use the parallel-cylinder
model 13 to obtain the following relationship for the specific sur-
face area occupied by water and gas:
'Now at BP Research.
, 'Now at Halliburton Services.
Copyright 1991 Society of Petroleum Engineers As=(4)~gl2k)lh, .................................. (7)

SPE Reservoir Engineering, February 1991 69



,!. 70

~ . 10. 5 60 u

~ ~::::E 50 ItJ
(I) ::;)
II:: (I)
0 ItJ (I)
II:: D.. 40 ItJ
0 II::

c.. D..
20 ~
10 10
0 10
b.l 10 100 1000

Fig. 1-Poroslty/permeablilty relationship. Fig. 2-Relatlve permeability and capillary pressure.

where cf>wg and k change with time. An empirical relationship be- Numerical Solution
tween k and cf>wg was obtained from field data of a Berea sandstone No analytical solution is available for the system given by Eqs. 8
reservoir, 13, * as shown in Fig. 1. through 12; therefore, a numerical solution was considered. Nu-
Introducing Darcy's law [vx = -(kkrIJL)op/ox] into Eqs. 1 merical approximation of the space and time derivatives was con-
through 3, we obtain sidered, which resulted in a set of simultaneous nonlinear
finite-difference equations. This set of equations was linearized;
the solution was obtained with an iterative method.
The space and time derivatives of the equations were approxi-
mated by a block-centered finite-difference scheme. The central-
difference approximation was used for the spatial derivatives; a for-
ward difference was used for the time derivatives. This method gives

In addition, the following auxiliary equations close the system:

Sg +Sw +SH= I .................................. (11)
and Pc(Sw)=Pg-Pw' ............................... (12)
The capillary pressure and the relative permeability relationships
for Berea sandstone shown in Fig. 2 were taken from Amyx et
at. 13 Note that these relationships depend on the porosity, perme~
ability, and pore-size distribution of the medium. Thus, as cf>wg and where Ox is the central-difference operator defined by oxf=f(x+
k change with time, we expect Pc(Sw) and kr(Sw) to change ac- !:J.x12)-f(x-!:J.x12) and Or is the forward-difference operator de-
cordingly. As discussed later, however, the change was found to fined by otf=f(t+tJ.t)-f(t).
have a marginal effect on the changes of porosity and permeability According to the block diagram in Fig. 3, Eqs. 15 through 17
in a porous medium containing hydrates during dissociation. may be discretized as
Eqs. 8 through 12 constitute a system of equations for the five
unknowns Pg, Pw' Sg, Sw' and SH' The initial and boundary con-
ditions are given below.
Initial conditions (t =0 and 0 sx sL).
p=po, ......................................... (13a)
SH=SH' ....................................... (13b)
Sw=S{f" ....................................... (13c)
and Sg =Sf . ...................................... (13d)
Boundary conditions.
p(O,t)=po at X=O for t~O ........................ (14a)
and op(L,t)/ox=O at x=L for t~O . .................. (14b)

• Personal communication with D.G. Davis, Colorado School of Mines, Golden, CO, 1988.

70 SPE Reservoir Engineering, February 1991

0.5 . . . . - - - - - - - - - - - - - - - - - - - ,

=111 x-O NO fLOW

, ,

.1 ± z 0.3
Fig. 3-Block diagram. ~
~ 0.2

3.5 r------------------, ~
:x: - 85 min

'0 3.0
.~( c
f 0.0


0 It)
6 It) GO

w R
Fig. 5-Model performance, hydrate-saturation distribution
;:, during dissociation.

g: 2.5 1 - - L - - - - - - - - - ' - - - - - - - 1

0.9 . . . . - - - - - - - - - - - - - - - - - - - ,


DISTANCE (m x 10
) iu 0.7

Fig. 4-Model performance, pressure distribution during hy-

drate dissociation. ~ 0.6

Inserting the auxiliary Eqs. II and 12 into Eq. 18 through 20 0.5
gives the following set of equations for Pg' Sw, and SH:
rg rg
- 1 (Pgkk
- -) (Pgi+l -Pgi ) - 1- (Pgkk
- -) (p gi -P gi - 1 ) 0.4
ax II-gax i+'h ax II-gax i-'h
+mg . =ilt- {[tPiPg.(l-Swi-SH)]n+l_[tPiPg (l-Swi-SH)]n}.
I I . I I I

.................................... (21) DISTANCE (m x 10 )

Fig. 6-Model performance, gas-saturation distribution during


Initial conditions (i= 1. 2 .. . n and t=O).

Pi=po. SHi =SfJ. SWi =S:t" and Sgi =Sg.
Boundary conditions.
P'h =Po for t~O and Pn+'h =Pn for t~O.
Eq. 23. which describes the hydrate dissociation into gas and
water. is solved first explicitly for the hydrate saturation SH.' Sub-
sequently. the finite-difference equations (Eqs. 21 and 22) a.ri solved
for the gas pressure. Pg .• and the water saturation. Sw .• at each
gridpoint with the Newton-Raphson method. I

The central-difference representation given by these equations

results in some coefficients that must be evaluated at the intercell Model Simulations
boundaries. Arithmetic mean weighting is used to approximate the Figs. 4 through 14 show typical results from the model. Table
intercell values for II- and P. harmonic weighting is used for k. and 1 lists the initial and the boundary conditions and the physical prop-
upstream weighting is used for kr- The initial and boundary con- erties used for the results in these figures. The equilibrium pres-
ditions in discretized form are given below. sure shown in Table I is that for methane hydrate.

SPE Reservoir Engineering. February 1991 71


0.4 0.20

i z 0.3
0 0.15

~ i
~ 0.2
0 0.10
£I: £I:

0.1 0.05

0.0k-~~-L~~~L-~~~~1~0~--L-~~~15 0.00 A--'-~-'---'---t-5--''--L-''''''''-'-..,J1\::-0-L--'----'--'L.......:J·

DISTANCE (m x 10"') DISTANCE (m x 10
Fig. 7-Model performance, water-saturation distribution dur-
Ing hydrate dissociation. Fig. 9-Poroslty variation during hydrate dissociation.

70 16

60 14
,.... 2- 12
"t:I 50
5 z
~ 10
5 40 (,)
I- 8
1&1 30 0
a.. IE 6
~ 4
III (,)

~ 0
10 III

15 25 50 75 1 0 150 1 5
DISTANCE (m x TIME (min)

Fig. 8-Absolute permeability variation during hydrate dis- Fig. 10-Model performance, location of the hydrate-
sociation. dissociation Interface.

Figs. 4 through 7 show the pressure and saturation distributions Fig. 6 shows the gas-saturation distribution with time. The plot
for t= 10, 50, and 85 minutes. It is clear from Fig. 4 that a sharp shows the gas evolution to be in a reverse S shape, initially under-
pressure front travels through the porous medium as dissociation going a period of exponential growth before slowing again. The
proceeds. Ahead of the front, the saturations of the hydrate, gas, initial gas-saturation buildup is caused by the initial low absolute
and water phases (shown in Figs. 5 through 7) are undisturbed at permeability when hydrates occupy a large portion of the pore space.
their initial values on the right of each figure. Behind the front (on Initially, the rate at which the gas is generated is much larger than
the left in Figs. 4 through 7), the pressure is essentially uniform can be accommodated by the available flow passage, causing the
at the outlet pressure Po. Within this region, hydrate dissociation exponential buildup in the gas saturation shown in all the curves
is distributed throughout a zone controlled by kinetics and pres- in Fig. 6. At later times, however, each gas-saturation curve shows
sure drop. an inflection point at the end of the exponential region. The inflec-
In Fig. 5, the hydrate-dissociation interface is widely distributed tion point results from the increase in both absolute permeability
over the core. The leading edge of the hydrate-dissociation inter- and relative permeability of gas as its saturation reaches about 57 %
face is defined as the point at which SH=SZ, At the time the trail- (a water saturation of 43% on the relative permeability plot, Fig.
ing edge of the dissociation interface reaches zero (between 50 and 2). In Fig. 6, the 85-minute curve achieves the maximum value
85 minutes), the leading edge is more than halfway through the of 80% gas saturation, coexisting with the immobile water, which
length of the core. is at 20% saturation.

72 SPE Reservoir Engineering, February 1991

2500 9

2250 8

2000 7
e 1750 ,....
0 0
w .2, 6
:::> 1500 0:::
0 W
0::: ~ 5
!l. 1250 ~
~ 1000
> 4
~:::> 750 :::::!l
:::::!l 2

50 75 100 125 150 175

TIME (min)
TIME (min)

Fig. 11-Model performance, cumulative gas produced. Fig. 12-Model performance, cumulative water produced.

Both hydrate- and gas-saturation curves of Figs. 5 and 6 can be terpreted as a magnified solution-gas drive. The large gas front
used to explain the water-saturation curves of Fig. 7. The gas sweeps the water from the porous medium so that only the immo-
evolves at the leading edge of the hydrate interface very rapidly. bile water remains, in accordance with the large water production
At the vicinity of the interface, the pressure drops sharply from (up to 80%) indicated by Makogon 14 in the Messoyakha field. This
the equilibrium pressure to the final pressure, causing substantial effect may indicate a significant production hindrance during the
gas expansion. The rapid gas evolution expands to occupy most recovery of natural gas from hydrates.
ofthe pore space, expelling the water produced during the dissoci-
Fig. 8 also indicates an increase in absolute permeability with
ation process and thereby creating a water front. As shown in the
water relative permeability curve of Fig. 2, at about 40% water time as hydrate dissociates. Fig. 9 shows the porosity increase with
saturation, the water front gains appreciable mobility and flows time as the hydrate dissociates tQ give more pore space. Both Figs.
toward the outlet. This effect creates a localized maximum value 8 and 9 are related to the hydrate-dissociation rate through the
of water saturation, as shown in Fig. 7. This behavior may be in- porosity/permeability relation shown in Fig. 1. While Patil et at. 15



r - -"1

m I
_ _ _ _I ~ BATH

Fig. 13-Experlmental apparatus.

SPE Reservoir Engineering, February 1991 73


III I Model Performance Testing

- L..L.L-LJL-LJU-J'- . Number of blocks 50
AIM! Ol234567G
Length of the model, cm 15
+ L..L.l-1.f-'tlt''tlt''-. Cross-sectional area, cm 2 11.4
-------'l~ Permeability, md
Temperature, K
1Ox10- 6
I) 2 Initial pressure, MPa 3.17
3 Porosity, % 18.8
Initial water saturation, % 17.0
I 4

. Initial hydrate saturation, % 42.76

~~ ~""'
Po, MPa 2.495
CORE . kd' kmol/(m 2 ·Pa·s) 4.4x10- 16
. Ilw, mPa·s
c w , Pa 6.9x10- 6

E 0.001
~t, seconds 0.0125
Fitting the Experimental Data'
Fig. 14-Electrlc circuitry of the resistance-measuring device.
Number of blocks 20
Initial water saturation, % 18.0
have made a start at quantifying the relationship of permeability Absolute permeability, md 100
to porosity in hydrate cores, much more remains to be done. Relative permeability as indicated in Fig. 2
Reaction rate constant, kmol/(m 'Pa's) 7.4x10- 16
Fig. 10 shows the linear position of the leading edge of the
hydrate-dissociation front with time. This linear relationship should • All properties the same as above with the exceptions listed.
not be confused with our previous work, 11 in which the dissocia-
tion front varies with the square root of time, as is characteristic
of a moving-boundary problem. Figs. 11 and 12 show the cumula- tubing, as shown in Fig. 14. Measurement of the electric resistances
tive gas and water production, respectively, as a function of time. provided an additional method to check for hydrate formation be-
sides the pressure drop in the core sample. A more detailed descrip-
Experimental Work tion of the apparatus is given in Ref. 11.
Methane hydrate was formed and dissociated in Berea sandstone
cores (15.2 cm long and 1.3 cm in diameter) with the apparatus Experimental Procedure. The core sample is initially evacuated
shown in Fig. 13. The core sample used in this experiment had and saturated with 1.5 wt% aqueous NaCI solution. The core is
an absolute permeability of 100 md and a porosity of 18.8%. A pressurized externally to 1034 kPa higher than the injection pres-
core was enclosed with heat-shrink plastic tubing and contained in sure (the pressure at which gas and water displacements are per-
a stainless-steel pressure bomb. A pressure-controlled pump was formed). Then, several PV's of the brine solution are circulated
used to inject gas and water into the core. Injection pressure was under high pressure throughout the core to ensure full saturation
controlled to within 173 kPa of the desired value. Two pressure and stability. Later, gas injection is commenced at a pressure of
transducers were used to measure inlet and outlet presures of the 7791 kPa and a temperature of273.7 K. Gas injection is continued
core. The system temperature was maintained at 273.7 K in a liq- until the desired water and gas saturations are achieved in the core.
uid bath containing a mixture of equal volumes of water and glycol. The produced water and gas volumes are closely monitored during
The electric resistance was measured with four pairs of electrodes this step. The outlet valve is then closed, and gas injection is main-
implanted at equal distances along the core under the heat-shrink tained during the hydrate formation. When no more gas uptake is

•• 400.0 r -=ELE-==CTR=OD=E:-:;"-=I------------,
...... ... •• 0 I

, 300.0
.• t :r
0- II 0

. ~
6.0 ~

ILl w
(/l .....
.! ....... ... ...
~ 200.0

.... \J.. \..

ILl 4.0 ~.~\ ·\0 ~
\ ~ to •

A1 A2 A3 A4

0.0 Ok'-'-...............,5;hO!;;!0................-t1-l;'00~0~.L.L.l+1~50>b0~.LU.*'20~0-b0!-'-'-1..J..L2W.500
0.0 ~..L.L.1.L.L';5HobI;Uou..u""1Hot,;0~0..L.L.1.L.L71~50~0~""""21d0t,;0~0..L..L..L.u..,!;2~50~0
TIME. t (min)
TIME, t (min)

Fig. 15-lnlet and outlet pressure behavior during hydrate for- Fig. 16-Electrlc resistance behavior during hydrate forma-
mation and the annealing process. tion and the annealing process.

74 SPE Reservoir Engineering, February 1991

2500 16
.-.. 2000 ::::E
u ~ 12
e z
\U 0

::::E 10
0 0
I- 8
~ 1000
u 6
::> z
0 0
n: ~
ll. 500 U

50 10 150 20 250
TIME (min) 60 120 180 2
TIME (min)

Fig. 17-Model match of experimental results, cumulative Fig. 18-Model match of experimental results, location of
gas produced. hydrate-dissociation Interface.

possible, the inlet valve is closed. At this stage, the bath tempera- ble to water. The relative permeability curve for gas (k rg vs. Sw)
ture is maintained at 273.7 K to allow hydrate formation to con- was shifted upward 3 % to make the system slightly more permea-
tinue for a period of 5 to 34 hours. When no change in the pressure ble to gas.
and the electric resistance with time is observed, the hydrate for- The dissociation reaction rate constant, kd' was decreased from
mation automatically ceases. 10 -II kmoll(m 2 . Pa' s), as reported by Kim et af. 12 for pure hy-
In many instances after the hydrate is formed, especially with drate dissociation, to 10 -16 kmol/(m 2 . Pa' s). The value reported
the lower-permeability cores, a pressure drop of up to 2.8 MPa by Kim et af. was obtained for conditions where mass-transfer
across the core is detected. One or more cycles of annealing proc- resistance was negligible. For hydrate dissociation in porous me-
ess are performed to eliminate this pressure drop and to ensure more dia, however, we expect the mass-transfer resistance to be substan-
uniform hydrate distribution along the core. In the annealing proc- tial, resulting in a lower value of k d • Future work will incorporate
ess, the hydrate is dissociated and reformed by subsequent heating the theory of dispersion in porous media to estimate the mass-
and cooling ofthe core. The effect of the annealing process on the transfer resistance.
behavior of the pressure and electric resistance during a typical hy-
drate formation run is presented in Figs. 15 and 16, respectively. Conclusions
The resistance changes in Fig. 16 show that hydrate formation is 1. A three-phase (gas/water/hydrate), ID numerical model was
not uniform along the core length, causing the permeability to vary. developed to simulate mathematically the process of isothermal hy-
Thus, the permeability is lower at the inlet of the core, where more drate dissociation in Berea sandstone samples. Depressurization was
hydrate is formed. considered the drive mechanism of hydrate dissociation in porous
After hydrate formation, dissociation is commenced at constant media.
outlet pressure and bath temperature. The rate of gas produced dur- 2. The model differs from previous ones in implementing the ki-
ing this step is measured by water displacement into a graduated netics of hydrate dissociation in porous media; the flow of water
cylinder over I-minute intervals. The inlet and outlet pressures and associated with hydrate dissociation was considered.
the electric resistance along the core are recorded at the same time 3. The experimental and numerical model results showed that
intervals. The change in the electric resistance during this stage an appreciable amount of water is produced simultaneously with
is used to track the dissociation front. At the end of the dissocia- gas production as a result of hydrate dissociation in porous media.
tion process, the water produced during this step is determined. 4. The mathematical model provided a satisfactory match to the
experimental data.
Experimental Results and Model Matching
Figs. 17 and 18 represent the experimental data for the cumulative Nomenclature
gas produced and the dissociation-front location, respectively. In A = cross-sectional area of the model, m 2
both figures, the data are shown as circles and the model fit as a As = specific surface area, m - I
solid line. Total cumulative water produced as predicted by the C w = isothermal water compressibility, Pa - I
model was 9.7 cm 3 , as opposed to 12.8 cm 3 measured at the end
f = arbitrary function
of the dissociation process. Overall, the agreement between the
k = absolute permeability, md
model predictions and the experimental data is quite satisfactory.
To obtain a satisfactory match of the model to the data, the abso- kd = dissociation reaction rate constant, kg/(m2. Pa' s)
lute permeability, relative permeability, and reaction rate constant kre = relative permeability of Phase £
were varied. The initial absolute permeability varied linearly from L = total length of the core, m
25 x 10 -6 md at the outlet to 2 x 10 -6 nid in the undissociated- me= mass rate per unit volume of Phase £, kg/(m3 's)
hydrate region. These data have not been measured; however, our Me = molecular weight of Component £, kg/kmol
experimental observations qualitatively suggest this type of variation. n = total number of blocks
The relative-permeability/water-saturation curve (krw vs. Sw) N H = hydrate number, dimensionless
was shifted upward about 20% to make the system more permea- Pc = capillary pressure, Pa

SPE Reservoir Engineering. February 1991 75

Authors Superscripts
I n = time level, n=O, 1, 2, 3 ...
o = boundary conditions

Ox = difference operator for space derivatives
0, = difference operator for time derivatives
fl = difference operator

1. Powell, H.J.M.: "The Structure of Molecular Compounds: Part IV-
Clathrate Compounds, J. Chern. Soc. (1948) 61, 194.
2. Jeffrey, G.A.: "Hydrate Inclusion Compounds," Inclusion Compounds,
Majeed H. Yousif is an engineering J.L. Atwood, J.E.D. Davis, D.D. MacNichol (eds.), Academic Press,
specialist at BP Research's Warrensville New York City, 1 (1984) 135.
Research Center In Cleveland, OH. He 3. Kvenvolden, K.A.: "Methane Hydrate-A Major Reservoir of Car-
holds a PhD degree from the Colorado bon in the Shallow Geosphere?" Chemical Geology (1988) 71, 41.
School of Mines In petroleum engineer- 4. McGuire, P.L.: "Methane Hydrate Gas Production by Thermal Stimu-
Ing. His research Includes hydrocarbon lation," Proc., Fourth Canadian Permafrost Conference, H.M. French
phase behavior and natural-gas hy- (ed.), Calgary (1982).
drates as a potential energy resource 5. Bayles, G.A., Sawyer, W.K., and Malone, R.D.: "A Steam Cycling
and as a problem to the 011 and gas In- Model for Gas Production From a Hydrate Reservoir," Chern. Eng.
dustry since 1986. Hazlm H. Abassis Comm. (1986) 47, 225.
a development engineer in the Stimula- 6. Kamath, V.A. and Godbole, S.P.: "Evaluation of Hot-Brine Simula-
Abass tion Research & Engineering Dept. of tion Technique for Gas Production From Natural Gas Hydrates," JPT
Halliburton Services Research Center in (Nov. 1987) 1379-88.
Duncan, OK. He holds a BS degree from the U. of Baghdad 7. Selim, M.S. and Sloan, E.D.: "Hydrate Dissociation in Sediment,"
in petroleum and mining engineering and MS and PhD degrees SPERE (May 1990) 245-51; Trans., AIME, 289.
from the Colorado School of Mines, both In petroleum engi- 8. Verigin, N.N., Khabibullin, I.L., and Khalikov, G.A.: "Linear Prob-
neering. His research Interests Include unconventional gas lem of the Dissociation of the Hydrates of a Gas in a Porous Medi-
reservoirs (tight sand, coal bed methane, Devonian shale, and urn," Izvest. Akad. Nauk. SSR, Mekhanika Gaza No. I (Jan.lFeb. 1980)
gas hydrates), reservoir engineering and stimulation, and rock 174.
mechanics. M.S. Sellm, professor of chemical engineering 9. Holder, G.D. and Angert, P.F.: "Simulation of Gas Production From
and petroleum refining at the Colorado School of Mines In a Reservoir Containing Both Gas Hydrates and Free Natural Gas," paper
Golden, has primary research Interests In modeling transport SPE III 05 presented at 1982 SPE Annual Technical Conference and
phenomena. He holds a PhD degree In chemical engineer- Exhibition, New Orleans, Sept. 26-29.
Ing from Iowa State U. E.D. Sloan Is a professor of chemical 10. Burshears, M., O'Brien, T.J., and Malone, R.D.: "A Multi-Phase,
engineering and petroleum refining at the Colorado School Multi-Dimensional, Variable Composition Simulation of Gas Production
of Mines. His primary research covers physical properties of From a Conventional Gas Reservoir in Contact With Hydrates," paper
natural gas hydrates. He holds a PhD degree in chemical en- SPE 15246 presented at the 1986 SPE Unconventional Gas Technology
gineering from Clemson U.
Symposium, Louisville, May 18-21.
11. Yousif, MJ. et al.: "Depressurization of Natural Gas Hydrate in Berea
P e = equilibrium pressure, Pa Sandstone Cores," J. Inclusion Phenomena & Molecular Recognition
Pe = pressure of Phase Pae, Chern. 1990, 8, 77-88.
Po = pressure at boundary x=O, Pa 12. Kim, H. C. et al.: "Kinetics of Methane Hydrate Dissociation, " Chern.
Eng. Sci. (1987) 42, No.7, 1645.
Se = saturation of Phase e 13. Amyx, J.W., Bass, D.M., and Whiting, R.L.: Petroleum Reservoir
t = time, seconds Engineering-Physical Propenies, McGraw-Hili Book Co., New York
flt = time increment, seconds City (1960) 98, 194.
T = temperature, K 14. Makogon, P.L.: "Les hydrates de gas: del'energie congelee," La
ve = velocity of Phase e, mls Recherche (1987) 18, 1192.
V = grid volume, m 3 15. Patil, S.L. et al.: "Measurement of Gas Permeability in Hydrate Satu-
x = distance, m rated Unconsolidated Cores," paper presented at the 1988 ACS Chemical
Congress of North America, Toronto, June 5-10.
flx = distance increment, m
€ = convergence tolerance
SI Metric Conversion Factors
Jl.e = viscosity of Phase e, Pa' s
Pe = density of Phase e, kg/m3 cp x 1.0* E-03 Pa's
ft x 3.048* E-Ol m
q, = porosity of medium, dimensionless ft2 x 9.290304*
q,wg = PV occupied by gas and water per unit volume, E-02 m2
OF (OF +459.67)/1.8 K
in. x 2.54* E+OO cm
in. 2 x 6.451 6* E+OO cm 2
Ibm mol x 4.535924 E-Ol kmol
Ih = boundary condition (x=O)
md x 9.869233 E-04 Jl.m 2
n+ Ih = boundary condition (x=L) psi x 6.894757 E+OO kPa
i = Gridpoint i OR °R/1.8 K
i± 1 = Gridpoint i± 1
i ± Ih = boundaries of a block containing Point i • Conversion factor is exact. SPERE
g = gas
H = hydrate Original SPE manuscript received for review Oct. 2, 1988. Paper accepted for publication
Oct. 9, 1990. Revised manuscript received Sept. 17, 1990. Paper (SPE 18320) first presented
w = water at the 1988 SPE Annual Technical Conference and Exhibition held in Houston, Oct. 2-5.

76 SPE Reservoir Engineering, February 1991