Lai 1991
Lai 1991
Lai 1991
SPE 21788
This paper was prepared for presentation at the Western Regional Meeting held in Long Beach, California, March 20-22, 1991.
This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper,
as pres~~ted, have not been reviewed by th.e Society of 'petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily r.eflect
any position of th.e SoCiety of 'petroleum En~lneer~, ItS officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society
of Petroleum Engineers. PermISSion to copy IS restricted to an abstract of not more than 300 words. Illustrations may not be copied. The abstract should contain conspicuous acknowledgment
of where and by whom the paper is presented. Write Publications Manager, SPE, P.O. Box 833836, Richardson, TX 75083-3836. Telex, 730989 SPEDAL.
361
NUMERICAL STUDIES OF COLD WATER INJECTION
2 INTO VAPOR-DOMINATED GEOTIIERMAL SYSTEMS SPE 21788
.of injection on reservoir perfonnance and well productivity, it more than 20 km with an average width of about 3 km.
is still difficult to forecast the gain in production that might
result from the boiling of injected water under field condi- The Geysers field lies within Jurassic-Cretaceous Fran-
tions. Numerical studies of the effects of cold water injection ciscan rocks arranged in a series of thrust fault-bounded
on production and reservoir perfonnance of vapor-dominated slabs I2.13• The main steam reservoir is located in graywacke
systems have been perfonned by several workers including horizons being deeper towards the northwest. A felsite base-
Schroeder et 81.7; Calore et 81.8; and Pruess et 81.9. Schroeder ment is present throughout the field and its depth can be
et al. used a porous medium model and demonstrated the correlated to that of the main graywacke reservoir. At the
importance of gravity effects during injection into vapor- end of 1987 about 8xlO 11 kg of steam had been produced
dominated systems. Calore et al. perfonned numerical studies from the resource with about 22 % of this mass injectedlO.
of injection into superheated reservoirs and illustrated the
boiling process associated with such injection as well as the Pressure transient (primarily pressure build-up) data suggest
effects of physical parameters on reservoir perfonnance. More that the transmissivity of the fractured reservoir at The
recently, Pruess et al. developed an analytical solution to Geysers ranges from IxlO- l1 to IxHr lo m3• Core data indicate
study the boiling process resulting from cold water injection that the average matrix porosity is in the range of 1 to 4 %14.
into superheated zones using a one-dimensional radial, infinite Analysis of flow rate decline data, assuming that most of the
porous medium model, and showed that the boiling fraction of reserves are located in the matrix blocks, yielded values for
the injected water depends on the ratio of the injection rate to the so called recharge factor between 1 and lOIS. The
the penneability-thickness product. They also pointed out that recharge factor R is defined as l6
vertical flow effects may be important in cold water injection
k",
into vapor-dominated reservoirs. There still remains much R=C (I)
rY
uncertainty about cold water injection at The Geysers, " as the
thennodynamic effects of injection are not well known where k", represents the matrix penneability, D is the average
because most of the liquid moves under gravity towards the fracture spacing, and C is a constant with a value of IxloD.
bottom of the reservoir below drilled depth" 10. Steam entry data suggest a fracture spacing on the order of
100 m, yielding matrix penneability in the range of IxlO-17 to
IxHrl8 m2. These matrix penneability values are similar to
The purpose of the present study is threefold. First, the
those obtained from other geothennal fields such as Los
effects of water injection on production rates are investigated
Azufers, Mexico I7.18, and Nesjavellir, Iceland l9, and are con-
and the sensitivity of the results to reservoir properties as well
sistent with results of large scale models of The Geysers lO,l1.
as the rate of injection is addressed. Second, comparison stu-
dies of the effectiveness of injection into superheated zones
and two-phase vapor dominated zones are perfonned. Third, APPROACH
grid effects on the simulation results are investigated, espe-
cially with regard to spatial discretizations in the vertical A study of "cold water" injection into fractured vapor-
direction. Because three-dimensional field-scale models of dominated reservoirs, taking into account of the gravitational
The Geysers field have been and are being developedIO.l1, it is effects, is performed using a two-dimensional, cylindrical (ie.,
important to investigate the grid block resolution required near r-z coordinate) fractured-porous medium model. Figure 2
the injectors. shows a schematic representation of physical problem con-
sidered in the study. The model has a radius of 1,500 m and
THE GEYSERS FIELD a thickness of 2,500 m. The injection well is located at the
center of the model and the production wells are assumed to
be uniformly distributed in the region with radii extending
Over five hundred wells have been drilled at The
from 500 to 700 m. The open intervals of the injection and
Geysers, providing large amounts of data for characterization
production wells are assumed to be located in the top 1,000 m
of this resource. This information includes lithologic logs,
of the reservoir. This configuration is similar to actual field
directional surveys, steam entry locations, static and flowing
conditions at The Geysers, where single injectors are typically
temperature and pressure surveys, production and injection
surrounded by a number of producers. The domain is divided
histories, and geochemical data. These data have been col-
into 24 grid blocks in the radial direction with two grid blocks
lected by the field operators and most of the infonnation is
used for the production wells. Because changes in the thenno-
open file and available. Figure 1 shows a base map of the
dynamic conditions near the injection well are expected to be
field and the approximate reservoir boundaries inferred from
"dry" (non-production) wells. The productive area is large, a fine mesh is used with the first element having a
elongated in the northwest-southeast direction extending for radius of one meter. Within a radial distance of 100 m from
362
SPE 21788 C. H. LAI AND G. S. BODVARSSON 3
the injection well most of the elements are 10 m wide. The injection rate (25 to 50 kg/s).
grid size is increased logarithmically from lOO.to 500 m, and
from 700 to 1,500 m. The 1,000 m thickness of the produc- Some of the injectors at The Geysers are located in
tion zone is discretized vertically into two grid blocks, result- reservoir zones which are strongly depleted and contain
ing in two and four grid blocks for injection and production superheated steam (superheated zones). In other newly
wells, respectively. For the basic grid used in most of the developed areas where cold water injection may be per-
study, the lower 1,500 m of the reservoir are subdivided into fonned, the reservoir pressure is still reasonably high and both
9 layers using thicknesses of 4 x 300 m, 1 x 150 m, 2 x 50 steam and liquid water are present (two-phase zones). Thus,
m, and 2 x 25 m. This grid is determined to be sufficiently the effects of cold water injection into both superheated and
fine based upon a grid sensitivity study that will be described two-phase zones are investigated, as described below. It
in a later section. should be noted that in the simulations, we neglect effects of
adsorption and capillarity, which are poorly understood for the
Because the coupled heat and mass transfer processes system considered. Also we should note that although we
involved in the problem are too complex to be addressed have tried to design the problem for conditions at The
analytically, the numerical code MULKOM developed by Geysers, many of the important parameters are poorly known
Pruess20 is used in this study. Description of the code, its vali- including the reservoir thickness and heterogeneity, penneabil-
dation, and applications to various problems involving mass, ity, and liquid saturation in the matrix blocks ll .
heat and chemical transport in geologic media can be found in
the literature21 • Based on the available data, the permeability of INJECTION INTO SUPERHEATED ZONES
the fractures may be two to three orders of magnitude higher
than that of the rock matrix, so that the fractures may be con-
In strongly depleted areas, the pressure and temperature
sidered as major conduits for transport of fluid and heat in the
at the top of the reservoir may be close to 20 bars and 240
system, while the rock matrix acts as a source and sink for
DC, respectively, and the vertical pressure gradient is vapor-
mass and energy. In the numerical simulations, we use the
static. This information, along with material property data,
"Multiple Interacting Continua" method22 for modeling the
allows for approximate evaluation of the pressure and tem-
fractured characteristics of the reselVoir, assuming three sets
perature distribution throughout the reservoir. Before starting
of orthogonal fractures. The fracture porosity and spacing
the injection I production simulations, the system is run to
assumed in the simulations are 1 %, and 100 m, respectively.
steady state conditions by imposing a constant pressure and
Both the fracture and rock matrix are discretized into volume
temperature of 20 bars and 240 DC, respectively, in the top
elements; fluid, heat, and chemical transport in the fracture
elements, and a no flux boundary condition at the bottom of
network is represented by a two-dimensional r-z grid. To
the reservoir. The resultant pressure and temperature distribu-
include the interactions between the fractures and rock matrix
tions are then used as initial conditions for the exploitation
blocks, each fracture element is connected to a string of rock
simulations. Other boundary conditions considered in the
elements; in this study the rock matrix string contains three
simulations include a constant injection rate of 80 DC water
volume elements for each fracture element.
and deliverability specified for the production wells (see Eq.
2). In order to investigate the effects of physical parameters
The parameters that are held constant in all of the simu-
such as reselVoir properties and injection rate, five different
lations perfonned in the study are given in Table 1. Most of
cases are considered (see Table 2). For purposes of com-
these parameter values are considered to be appropriate for
parison, Case A which considers an injection rate of 25 kg/s
conditions and practice at The Geysers. The injection and permeabilities of lxlO- 13 and lxHr16m2 for the fractures
enthalpy of 334,930 J/kg corresponds to a water temperature and rock matrix blocks, respectively, is considered to be the
of 80 DC, which is reasonable given the fact that the injected base case.
water is a mixture of "hot" water from the condensers and
"cold" surface water. The productivity index (PI) is defined
Base Case Results
from
llt = L ~ P~ PI (P~-PwtJ (2) The general behavior obselVed from the simulations is
~ ~ liquid, steam I.I.j\ similar to that found using a porous medium modelS and is as
where llt is the total flow rate, ~' P~' 1.I.j\, and P~ are the rela- follows. When injection starts, the liquid plume migrates
tive permeability, density, viscosity, and pressure for phase ~, away from the injection well in both the horizontal and verti-
respectively, and Pwb is the assumed bottomhole pressure. cal directions. The horizontal and vertical movements are pri-
The productivity index is generally unknown and is chosen so marily controlled by the pressure gradient and gravity forces,
as to give a reasonable initial production rate for the assigned respectively. The gravity forces are extremely strong for this
363
NUMERICAL STUDIES OF COLD WATER INJECTION
4 INTO VAPOR-DOMINATED GEOTIIERMAL SYSTEMS SPE 21788
problem because of the large density difference between Figure 5 shows the temperature distribution in the fractures
. liquid water and steam at prevailing conditions. For instance, after five years of simulation. Comparison with Figure 3
the ratio of the liquid water and steam densities under reveals that the 240 OC isotherm closely follows the locations
saturated conditions at 240 °C is approximately 50. During of the outer surface of the plume, indicating that the
injection this enhances the downward movement of the plume temperature in the outer surface of the plume is relatively
and reduces the horizontal movement of the upper portion of large compared to that of the liquid core of the plume.
the plume. With time the horizontal movement of the upper
portion of the plume appears to terminate, and only downward Grid Effects
movement of the plume in the vicinity of the injection well
occurs. When the plume reaches the bottom of the reservoir,
In general, grid effects on the results of numerical prob-
it continues to migrate radially outward along the bottom of
lems, resulting from the coarseness and the orientation of the
the reservoir. The boiling process associated with the plume
grid relative to the flow field, may be large. In this study, the
migration along the reservoir bottom is similar to that effects of the grid block size are investigated, especially the
predicted in one-dimensional infinite systems9• The steam
gridding in the vertical direction, as our studies indicate that
derived from boiling flows upward and laterally towards both
the gridding in the horizontal direction is relatively unimpor-
the production wells and the outer surface of the shallow tant compared to that in the vertical direction. In particular,
plume, where it condenses and releases the latent heatS. we find that the grid used for the region near the reservoir
bottom has a large impact on the simulation results. We have
Figures 3 through 5 show some of the results for the not investigated grid orientation effects for this problem,
base case (Case A). Figure 3 shows the phase distribution in which may be significant.
the fractures after 5 years of simulation. At this stage, the
lateral expansion of the upper portion of the plume has almost The results of the numerical simulations show that the
terminated. With the exception of the bottom layer, the coarser grid the bottom layer, the slower the liquid plume
injected liquid plume, in general, is confined to a region moves away from the injection well. The reason for this
extending radially approximately 40 m from the injection artificial phenomenon is that in the numerical simulations, the
well. The outer surface of the liquid plume is surrounded by a heat contained in the elements of the coarse grid is higher
two-phase zone where the boiling of the injected water and than that in the fine one, resulting in a larger boiling fraction
the condensation of the flashed steam occurs. The two-phase of the injected water, thus slowing down the movement of the
zone extends further from the plume at greater depths, espe- liquid plume. To evaluate the grid effects on the simulation
cially at the bottom layer. This shows that the movement of results, numerical experiments are conducted by systemati-
the injected water is downward, due to the strong gravity cally refining the grid size with all parameters remaining the
effects, and then laterally outward in the bottom layers, where
same. Total of five different grid discretizations (ie., 7, 8, 9,
most of the boiling of the liquid occurs. These large 10, and 11 layers) in the vertical direction is considered in the
predicted vertical movements of the injected water are sug- study (Table 3).
gested by microearthquake studies at The Geysers23,24. The
microearthquakes are found to concentrate around the injec-
Figure 6 shows the variation of the total steam produc-
tors and often indicate a vertical liquid water movement to
depths exceeding 3,000 m below the injection zone. tion rate with time for different grids; all parameters
correspond to the base case. The results show that for this
problem, the production rates predicted for the cases with 7,
In the simulations, the injected water was tracked with a
8, and 9 layers and grid spacing for the bottom layer ranging
conservative tracer in order to investigate how much of the
from 100 to 300 m, are very similar at all times. For these
injected water boils and is produced as steam by the wells. It
cases, the production rate stabilizes after five years as the
is assumed that the original reservoir fluid is pure water and
boiling rate of the injected water is approximately equal to the
that when the injected water with a certain concentration of
total production rate. However, when the simulation domain
the tracer boils, both liquid and the evolving steam contain in the vicinity of the reservoir bottom is discretized into a
the same tracer concentration. This allows to quantify the finer grid (ie., 10 and 11 layers) with grid spacing of 50 or 25
mass fraction of the fluid derived from the injected water for m for the bottom layer, after approximately 20 years of simu-
any grid blocks. The tracer concentration distribution in the lation the predicted production rate begins to deviate from
fractures after five years is given in Figure 4, and shows that those calculated for the cases with coarse grids. This is
higher tracer concentrations (96 % by weight) occur at greater because for the fine grids, the bottom layer fills with single
depths, resulting from the displacement of the native fluid by phase liquid, thus reducing the boiling fraction of the injected
the injected fluid. At this time, a tracer concentration of
water. Figure 7 presents the cumulative production versus
approximately 70 % is observed in the production well area.
364
SPE 21788 C. H. LAI AND G. S. BODVARSSON 5
time for cases with different grids, and shows a difference of case (lxlO-16 m2) and in Case E the fracture permeability is
approximately 20 % between the cases after 30 years of simu- ten times lower that in the base case (1xIO- IS m2). These
lation. ranges of values are comparable with those obtained from
site-scale modelslO,ll As shown in Figures 10 and 11, changes
To further test the grid effects, the injection rate is in these material properties have little effects on the steam
changed from 25 to 50 kg/s with other conditions remaining production rate and the cumulative production. It is plausible
unchanged (ie., Case B; Table 2). Figure 8 shows the varia- that the matrix permeability will not greatly affect the results
tions of the production rate with time for various discretiza- for superheated conditions. as the mass of fluid contained in
tions of the reservoir thickness. It shows that the production the matrix blocks is small. The fracture permeabilities tested
rate predicted from the cases with 7 and 8 layers is very simi- are apparently high enough to support the rather small steam
lar, and that the predicted rate tends to stabilize after two flow through the system modeled.
years of simulation. After an initial transient, the production
rate derived for the cases with 9, 10, and 11 layers is nearly INJECTION INTO TWO·PHASE ZONES
constant during the 2 to 15 year period, but declines rapidly
in the next five years. Examination of the simulation results Cold liquid injection into two-phase vapor-dominated
reveals that after 20 years of simulation the bottom layer is zones is being carried out in newly-developed areas of The
filled with the injected liquid for the 9, 10, and 11 layer cases, Geysers. mostly in regions near the flanks of the field. The
thus limiting the boiling fraction of the injected water and thermodynamic conditions of two-phase zones are believed to
reducing the production rate. Figure 9 shows the cumulative be maintained through counterflow of liquid and steam that is
production versus time. For the cases with coarse grids they sustained by a vertical heat fiux 2S ,26,27. In order to obtain these
are 70 % higher than those for the fine girds. thermodynamic conditions that will be used as initial condi-
tions for the injection / production studies, it is easiest to per-
Sensitivity Studies form a numerical simulation on a single column with only
fracture elements. The boundary· conditions used for this
Sensitivity studies are performed to investigate the simulation include two-phase conditions in the deepest frac-
effects of reservoir properties and injection rate on the total ture elements and a heat loss of 0.5 W/m2 in the uppermost
productivity. Figure 10 shows the variation of the production fracture elements10; it results in a stable vapor-dominated heat
rate with time for the various cases listed in Table 2. When pipe28. If needed the lower boundary conditions are adjusted
no injection (Case C) is assumed, the production rate declines until the desired conditions in the shallow reservoir are
very rapidly and approaches zero after only 10 years of pro- reached. We obtain a vapor-dominated heat pipe with a near-
duction. This rapid decline is caused by the low mass capa- uniform liquid saturation of 5 % in the fractures, and a pres-
city of the system because only steam is present initially. In sure and temperature of 33.4 bars and 240 °C in the upper-
most fracture elements. These stable initial conditions are then
this case after about 10 years almost all of the initial reserves
prescribed for all of the fracture element columns, and the
(3xI(f kg) are produced (Figure 11).
two-phase boundary conditions at the bottom are replaced
with a constant (0.5 W/m 2) heat flux boundary condition.
In the case of 25 kg/s injection with the base case (Case
Finally, the rock matrix elements are added and given the
A) parameters, the production rate gradually declines (Figure
same pressure and temperature as the adjacent fracture ele-
10) and a total of 1.6xI010 kg of steam are produced during
ments. In this study. we assume a uniform initial matrix
the 30 year period (Figure 11). This mass recovery is more
liquid saturation of 25 %. We also conduct simulations with
than three times the initial reserves in-place, and approxi-
50 % liquid saturations and the results are similar.
mately 60 % of the injected fluid is vaporized and produced
by the wells. Using the base case parameters but doubling
The same cases as in injection into the superheated zone
the injection rate (Case B). the total mass recovery is 2.7x1010
are considered. In general, the predicted migration paths of
kg of steam, or nine times the initial reserves. In this case,
the injected water are similar to those obtained in the
about 50 % of the injected fluid vaporizes and is produced.
superheated zone problem. Figure 12 shows the phase distri-
Thus. injection into superheated zones significantly increases
the reservoir productivity. This conclusion is also supported bution for the base case (Case A) after five years of simula-
tions. It shows the vertical extension of the plume downward
by data from field injection operations6•
from the injection zone and a lateral extension along the
reservoir bottom. A two-phase zone of considerable thickness
The sensitivity of the results on fracture and rock matrix
still exists in the fractures in the bottom 700 m of the system,
permeability is also investigated. In Case D the matrix per-
whereas the fractures in the upper part have dried out due to
meability is an order of magnitude lower than that in the base
365
NUMERICAL STUDIES OF COLD WATER INJECTION
6 INTO VAPOR-DOMINATED GEOTHERMAL SYSTEMS SPE 21788
366
SPE 21788 C. H. LAI AND G. S. BODVARSSON 7
block thickness near that bottom should not exceed 25 to 4. Enedy, S, Grande, M., and J. L. R Smith, A Case His-
50 m, in order to accurately model the rate of boiling of tory of Steam Field Development, Reservoir Evaluation,
the injected water. and Power Generation in the Southeast Geysers, Geoth-
eraml Resources Council, Bulletin, Vol. 19, No.9, Oct.
6. Given the large sensitivity of the results to grid discreti- 1990.
zation and perhaps also to grid orientation, in site-scale
modeling one must carefully design the grid around 5. Bertrami, R., Calore, e., Celati, R, and D'Amore, F., A
injectors. This certainly makes conventional, unifonn Three-Year Recharge Test by Reinjection in the General
finite-difference gridding impractical for modeling The Area of Larderello Field: Analysis of Production Data.
Geysers as local grid refinements will be necessary. 1985 International Symp. on Geothennal Energy,
Kailua-Kona, Hawaii, Vol. 9, P. 293-298, 1985.
NOMENCLATURE
6. Gulati, M.S., Lipman, S.C., and Strobel, e. J., Tritium
Tracer Survey at The Geysers, Geothennal Resources
D = fracture spacing, L
Council, Transactions, Vol. 2, P. 237-239, 1978.
ku, = matrix penneability, L2
krj3 = relative penneability in phase Ii
7. Schroeder, R.C., O'sullivan, M.J., Pruess, K., and Rufilli,
Pwb = bottom hole pressure, F/l}
C., Reinjection Studies of Vapor-dominated Systems,
P~ = pressure for phase Ii, F/L2 Geothennics, Vol. 11, No.2, P. 93-118, 1982.
R = recharge factor given in Eq. 1
Srw = irreducible water saturation 8. Calore, C., Pruess, K., and Celati, R., Modeling Studies
of Cold Water Injection into Fluid-Depleted, Vapor-
Srg = irreducible steam saturation
Dominated Geothermal Reservoirs, Proceedings Eleventh
J..L~ = viscosity in phase Ii, F-T/l} Workshop on Geothennal Reservoir Engineering, Stan-
ford University, P. 161-168, 1986.
ACKNOWLEGEMENT
9. Pruess, K., Calore, C., Celati, R., and Wu, Y.S., An
The authors thank M.J. Lippmann and C. Doughty for Analytical Solution for Heat Transfer at a Boiling Front
critical review of this paper. This paper was supported by the Moving through a Porous Medium, Int. J. Heat Mass
California State Lands Commission and by the Assistant Transfer, Vol. 30, No. 12, P. 2595-2601, 1987.
Secretary of Conservation and Renewable Energy, Geothermal
Division of the U. S. DOE under Contract No. DEAC03- 10. Williamson, K.H., Reservoir Simulation of The Geysers
76SF00098. Geothennal Field, Proceedings Fifteen Workshop on
Geothermal Reservoir Engineering, Stanford University,
REFERENCES 1990 (in press).
1. Ramey, H.J., Jr., A Reservoir Engineering Study of The 11. Bodvarsson, G.S., Gaulke, S., and Ripperda, M., Some
Geysers Geothermal Field, Submitted as Evidence, Reich Considerations on Resource Evaluation of The Geysers,
and Reich, Petitioners vs. Commissioner of Internal Geothennal Resources Council, Transactions, Vol. 13, P.
Revenue, Tax Court of the United States, 52, T.e. No. 367-375, October 1989.
74, 1969.
12. McLaughlin, RJ., Tectonic Setting of Pre-Tertiary Rocks
2. Barker, RJ., Gulati, M.S., Bryan, M.A., and Riedel, and Its Relation to Geothennal Resources in The
K.L., Geysers Reservoir Perfonnance, Geotheraml Geysers-Clear Lake Area: In Research in The Geysers-
Resources Council, Transactions, Vol. 13, P. 349, 1989. Clear Geothennal Area, Northern California: Geol. Surv.
Prof. Paper, 1141, P. 3-25, 1981.
3. Haizlip, J.R and Truesdell, A.H., The Correlation of
Noncondensible Gas and Chloride in Steam at The 13. Thompson, RC., Structural Stratigraphy and Intrusive
Geysers, Geothennal Resources Council, Transactions, Rocks at The Geysers Geothermal Field, Geothennal
Vol. 13, P. 455-460, Oct. 1989. Resources Council, Transactions, Vol. 13, P. 481-485,
1989.
367
NUMERICAL STUDIES OF COLD WATER INJECfION
8 INTO VAPOR-DOMINATED GEOTHERMAL SYSTEMS SPE 21788
14. Gunderson, R.P., Reservoir Matrix Porosity at The 24. Zucca, Personal Communication, 1990.
Geysers From Core Measurements, Geothennal
Resources Council, Transactions, Vol. 14, Part II, P. 25. White, D.E., Muffler, L.lP., and Truesdell, A.H., Vapor
1661-1665, 1990. Dominated Hydrothennal Systems Compared with Hot-
Water Systems, Economic Geology, Vol. 66, No.1,
15. Ripperda, M., Gaulke, S., and Bodvarsson, G.S., The P.75-97, 1971.
Geysers Geothennal Field - Analysis of Production Data,
Paper in Preparation, 1991. 26. Martin, lC, Wegner, R.E., and Kelsey, FJ., One-
Dimensional Convective and Conductive Geothennal
16. Bodvarsson, G.S. and Witherspoon, P.A., Flow Rate Heat Flow, Proceedings Second Workshop on Geother-
Decline of Steam wells in Fractured Geothennal Reser- mal Reservoir Engineering, Stanford University, P. 251-
voirs, Proceedings Tenth Workshop on Geothennal 262, 1976.
Reservoir Engineering, Stanford University, P. 105-112,
1985. 27. Pruess, K., Celati, R., Calore, C., and Cappetti, G., On
Fluid and Heat Transfer in Deep Zones of Vapor-
Dominated Geothennal Reservoirs, Proceedings Twelfth
17. Contreras, E., Iglesias, E., and Razo, A., Initial Measure-
ments of Petrophysical Properties of Rocks from the Los Workshop on Geothennal Reservoir Engineering, Stan-
Azufres, Mexico, Geothennal Field, Proc. Eleventh ford University, P. 89-96, 1987.
Workshop on Geothennal Reservoir Engineering, Stan-
ford University, P. 51-57, 1986. 28. McGuinness M.J. and Pruess, K., Unstable Heat Pipes,
Proceedings of the Ninth New Zealand Geothennal
18. Iglesias, E., Contreras, E., Garcia, G., and Dominguez, Workshop, Aukland University, P. 147-151, Nov. 1987.
B., Petrophysical Properties of Twenty Drill Cores from
the Los Azufres, Mexico, Geothennal Field, Proc.
Twelfth Workshop on Geothennal Reservoir Engineer-
ing, Stanford University, P. 195-202, 1987.
368
1 seE 2178 Ii H
20
\ , 23
I
I
22 r 2< 19
,, I
,
I I
I
I
I
ca--------,--------
lL.4l'o I
4=1I
.......
I
I
I
28
I
I
I
I
I
I
O""'F.'
2 '4/1L1' 30
,
---~~Q---!-----O l--------L-------~------
: 32 J 33 I 34: I
r I : :
I lIN W
I
I
: \ ..
_1 L--O --~----_ . . . . . . _~
I :
• ,i· { I
I
10 J
I
I
I
--J----~---1--:::;;i-------1---
I
.. ,
APPROXIMATE .
3--~'--__I
I
RESERVOIR
-- I / BOUNDARJES-------
I I •
22 2< 19
I
r
I
I
I
r
---t=r---:
I
30
I
25 I
r
I
,
.
3'
r
\
o
I I
\ I ,
d':6'~O
\
\
: 2 : 1
r ,
370
I IPE 2178&
,.)
I
0_ I
g 1000
)
.c
I g-
"8. O 1500
c3
)-
2000
20 )1- 2500
o 300 600 900 1200 1500
Radial distance from injection well (m)
I I I I
2 1 Figure 3 Case A (injection into superheated zones): Phase distribu-
o 300 600 900 1200 1500
tion in the fractures after 5 years of simulation.
Radial distance (m)
Co>
::l
Figure 2 Schematic representation of physical problem.
0 0
\
\'"
500 ---------- 500
" " --------------
' ....
E
1000 ............ _-------------------- I
1000
.c
96
-5
0..
Q)
0-
Q)
0 1500 0 1500
2000
2500
0 300 600 900
Radial distance from injection well (m)
1200 1500
2000
2500"
o
gr ~oo·c
I
\
240°C
I
300
I
600
i
900
Radial distance from injection well (m)
I
1200
I
1500
Figure 4 Case A (injection into superheated zones): Tracer concen- Figure 5 Case A (injection into superheated zones): Temperature
tration (wt %) distribution in the fractures after 5 years of distribution in the fractures after 5 years of simulation.
simulation.
,,'.
i
tl.a.
leE 21781
60 I I
Injection Rate 25 kg/s
......
Cl
2.0 I .if I
.Y. Injection Rate 25 kgls
.........
III
.........
Cl
-
o
w
_ 1-
.::ti
......
I.IJ z
I- o
<I: i=
0:: 8 Layers u
Z 9 Layers ~ 1-
o o
a:
I- c..
U
=> w
C
o :X, >
i= 0.5
./' -"'~-'--':;'
0::
a.. 10 Layers --r- j
::::l
~
11 Layers ::::l
U
o 5 10 to 15 25 30
0.0 r
o
i i i
5 10 15
i i i
20 25 30
TIME (YEARS) TIME (YEARS)
. . ,. 7L>,... -•
0
w 7 Layers
~ ~=c::~~.,."".;....
';;; , ./" I ....
-
......
'-:~:;..-/ ~,"/?
Z 8 Layers /--. .
0 3
I.IJ •• .,..-.:'"'::....." ' " 8 Layers 1 Layers i=
'\'.
I- '........ 1 u .........••..···ii·Layers..
,~
:. ',' ; -.J..._ .... ::::l
. >.>-::;;;:~7-::-?:::~;1-
0
0
% , / Lay.:~
••• _.,. 10 - _ a: 2
c..
g ,
\- ,v - w ...~:':/
"ca..
o
=>
0::
"--'
'----, "7
9~
>
i=
<
~ 10 Layers
9 Layers
...J ~~~
::::l
~
::::l
U
o 5 10 15 20 25 30
o 5 10 15 20 25 30
TIME (YEARS)
TIME (YEARS)
Figure 8 Case B (injection into superheated zones): Production rate Figure 9 Case B (injection into superheated zones): Cumulative pro-
versus time for various grid sizes. duction versus time for various grid sizes.
IPE 21788
60 1 I
- t»
.::I.
3.0"1,-----------------_-,
-
"-01
I II
o
or-
L.J
......
or-
Ca\E" /./..,//
~
7"
....... Z 2.
o .' ••••••• Case A
W .'
~
....•••••....~aseD
I-
«
0:
·v \ _· , u
::::> .... / "
/.,.
z
o
I-
o 2
-
~ ~.;; _ -. __
~::~.h
o
o
I:t:
a..
L.J .....
. . . . . . '-"r"'/'"
/"'~
::::>
o
o
g:1 Case C
--
m_.~
-. >
~
<
:5 o.
••.•.••
./ / ••,
/'" Case E
Case C
Case E ./ /,,'" J ----
\',./ Case A ::::::E
::::>
#,#,,;
-.'1'1
J::'.- ..- ..- -- .- .- ..- -- -- -- -- -- -_..
"- -- "'- -- .. - .. - -- -- u
30 0.0 fC 1 I i I ( ,
o 5 10 15 20 25 25 30
TIME (YEARS)
0 0
2000
2000~kr 200°C
240°C
2500 2500+-
0 300 600 900 1200 1500 0 300 600 900 1200 1500
Radial distance from injection well (m) Radial distance from injection well (m)
Figure 12 Case A (injection into two-phase zones): Phase distribution
Figure 13 Case A (injection into two-phase zones): Temperature Dis.
in the fractures after 5 years of simulation.
tribution in the fractures after 5 years of simulation.
300
.......
III
250
.........
Cl
.::tt.
'-" 200
L.J
I-
«
et: 150
Z
0
I- 100
U
::>
0
0
et: 50
Q..
0
0 5 10 15 20 25 30
TIME (YEARS)
,.... 2
Cl
~
-- i."'!
Z
0
Case A
I-
U
::>
-_ ... - -- ---
0
0
et:
a.
....
>
t-
«( ,,"
",,-,,"-"\ Case E
...I
::>
::E
::>
U
10 15 20 25 30
TIME (YEARS)
374