Note: Descriptions are shown in the official language in which they were submitted.
1323~61
This invention relates to the recovery of highly viscous oil or bitumen from a
subterranean formation of tar sands. More particularly it relates to a method
of achieving effective thermal communication between injection and production
wells in a formation and also achieving increased production of the highly
viscous oil from tar sands formations of a weak uncemented type.
In many different areas of the world there are subterranean formations
containing high viscosity immobile oil or bitumen which cannot be recovered by
conventional methods, that is the tar-sand type of formation. Some tar sands
are described as granular materials which are close to mechanical incompetence
at in situ stress conditions. To produce the bitumen it is necessary to heat
it. There are many methods proposed in the prior art for stimulating
production from wells penetrating such formations as well as for establishing
thermal communication between wells in such formations to enhance the
production of the viscous oil or bitumen.
Canadian Patent 753.474 of Boberg relates to a process for thermal stimulation
of a producing well. The production of oil from the well is interrupted,
steam is injected into the formation in order to reverse the direction of flow
in the formation surrounding the well for a time sufficient to drive the steam
into the formation a certain distance. This is followed by backflowing and
producing steam condensate and oil from the well while maintaining a back
pressure on the well to prevent substantial revaporization of the condensate.
This cyclic process ~s commonly known as "cyclic steam stimulation" and is a
well-known process. This "cyclic steam stimulation" is intended for tar sands
or any other formation in which productivity can be enhanced by heating the
bitumen or tar to be produced.
-1-
.:., - :
., -
1323~61
U.S. Patent 3.280.909 of Closmann et al involves preheating and flooding a tarsand formation in order to recover oil therefrom. A zone of oil-bearing
formation between any pair of an injection and production well or adjacent
such wells is preheated in order to reduce the viscosity of the oil to be
moved out of the zone between the two wells. First, a horizontal fracture is
formed between injection and production wells to provide fluid flow
communication. Then a hot fluid is injected down, preferably, both wells and
through the fracture, or through one well at a time while the other well is
shut in. This is followed by injection of a drive or displacement fluid into
the injection well and the production of drive fluid and oil from the
production well.
USP 4,130,163 of Bombardieri relates to the recovery of hydrocarbons from a
viscous tar sand deposit utilizing at least two wells for injection and
production. A heated fluid, as for example steam, is injected simultaneously
down both wells until a substantial zone is heated in the formation into which
the wells penetrate and communication is established between the wells. Then
one well is shut in and hydrocarbons are produced from the other. When the
producing rates decline below a predetermined level, steam is introduced under
relatively low pressure at the shut-in well to continue production but to
prevent steam breakthrough at the production well. This cycle can be repeated.
USP 4.299.284 of Brown et al involves a method wherein production wells are
treated with sn aqueous fluid containing surfactants to reduce permeability of
the formation. The production wells are treated with this diverter" fluid to
promote sweep efficiency or horizontal conformance of the displacement process
for oil recovery. As a result of the treatment, there is alleged to be a
reduction in the tendency to develop an assymetrical sweep area and the
displacement and recovery process are allegedly enhanced.
,
~ :
'. . '
: : :
1323~1
USP 4.637.461 of Hight discloses a method relating to modified inverted nine
spot and thirteen spot well patterns where several wells extend at least
partly through the formation in a horizontal direction. The method involves
cyclic steam stimulation of production wells with a thermal fluid prior to
being placed on production and then the thermal fluid is injected at the
central injection well. Once breakthrough from the central injection well is
reached at any of the infill production wells, these are converted into
injection wells and injection continued at these interior wells of the pattern.
U.S. Patent 4.612.989 of Rakach et al relates to the recovery of oil from tar
sands and involves at least one injection and one production well in fluid
communication with each other. Steam is injected through the injection well
at less than a fracturing rate but at a sufficient pressure to displace the
heated oil to the production well. When breakthrough of steam occurs at the O
production well, the latter is shut in for a period of time. The injection of
steam through the injection well is then increased to form a fracture and the
injection well shut in and the production well opened to produce
hydrocarbons. The steps can be repeated in a cyclic matter.
Canadian Patent 1.122.113 of Britton et al relates to a process for
establishing a zone of increased heat and of increased fluid mobility between
an injection well and a production well vertically penetrating an oil
reservoir such as a tar sands reservoir. The method involves horizontal
hydraulic fracturing between the wells, followed by injecting steam at a very
high rate through the injection well while producing fluids at the production
well, the steam being injected at a sufficient pressure and for sufficient
time to maintain parting of the formation along the fracture system between
the wells and effect channel flow of fluids through the parted fracture system
between the wells and conduction heating of substantial reservoir volume. The
method is alleged to achieve high thermal efficiency. Although the method is
: .
: '
1323~1
not specified as being restricted to certain types of tar sand formations, it
appears that relatively thin reservoirs or shallow reservoirs having low
permeability are of particular interest insofar as the claimed process is
concerned. It is clear from this patent that the pressure of the steam
injection at the injection well at high rates must be greater than or equal to
the overburden stress in order to maintain parting of the fracture system as
required.
Canadian Patent 1.004.593 of Wang et al relates to recovery of viscous oil
from a tar sand formation in which a relatively water-impermeable zone
overlies a water-permeable zone. Steam is injected into the permeable zone to
establish a hot water-permeable channel between at least one production and
one injection well, the steam being injected at the highest practical rates to
give a tar mobility of at least 15 md/cp in as short a time as possible.
During this procedure the backpressure of the well from which fluids are
produced is kept as low as practical to promote a high rate of steam advance
through the permeable zone. The injection of steam is continued after fluid
communication is established at about the same rate or slightly less and at
the same time the production well is throttled back so that a hot,
water-permeable channel is established and expanded to the maximum possible
extent. Then the injection rate is cut back to the maintenance level of
temperature and pressure with a high back pressure being maintained on the
production wells. After maintenance of the pressure and temperature to
provide the required flow viscosity of the bitumen, the pressure in the
formation is reduced by increasing the production of fluids at the production
well. These steps are then carried out in a cyclic manner for increased
production of bituminous fluid.
All the foregoing processes are found to be unsatisfactory from the point of
view of achieving fluid communication and promoting areal conformance between
injection and production wells, as well as obtaining enhanced production of
. ~ . :
:. .: .
. ~ . . . .
: ~ .
.~: :. :
:, - : ~
:
1323~
bituminous fluid from tar sand formations of a type with which the present
invention is particularly concerned. In fact, processes such as that of
Canadian Patent 1,122,113 of Britton as such are not effective or even
operable in the tar sand formations of present concern. One reason is that
the patented process depends on a type of formation wherein tensile parting by
conventional fracturing methods is possible and the fracture so formed can be
maintained.
The particular formations with which the present invention is found to be
suitable are such as are found at Cold Lake Alberta, some in the Athabasca tar
sands, in California, Venezuela, and the Soviet Union. Such formations are
characterized as being nweak" formations which are subject to shear failure
rather than tensile failure. Adequate fluid communication between wells in
such formations cannot be achieved by conventional processes. The conditions
that characterize "weak" formations can best be determined by field
observations and laboratory testing.
Knowledge of the in situ stress state is essential to characterize whether or
not the formation is weak. The vertical in situ stress can be estimated by
integration of high quality density logs from the surface to the depth of
interest and the minimum in situ stress may be determined by low rate, low
volume, cyclic hydraullc fracturing as described by Gronseth, J.N. and Kry,
P.R., (1983). (Instantaneous shut in pressure and its relationship to the
minimum in situ stress. In Hydraulic Fracturing Stress Measurements, edited
by M.D. Zoback and B.C. Haimson, National Academy Press, ~ashington, D.C.)
The maximum in situ principal stress cannot easily or reliably be determined,
but what data exists in Western Canada, for example as described in Kry, P.R.
and Gronseth, J.N., 1983. (In situ stresses and hydraulic fracturing in the
Deep Basin. Journal of Canadian Petroleum Technology, Vol. 22, No. 6, pp.
31-35) suggests that the maximum horizontal stress exceeds the minimum stress
by 20~ to 100~. In summary, the minimum in situ stress and whether it is
vertical or horizontal can be reliably determined.
~3~3~1
The most important indicator of whether the formation is weak is the
observation of injection at pressures less than the in situ stress and at
rates that exceed those expected from the undisturbed formation injectivity.
Undisturbed formation injectivity is the limit of the ratio between injection
rate and the excess pressure above reservoir pressure which enables the
injection to take place as the excess pressure tends to zero. (At zero excess
pressure there is of course zero injection rate). In strong, cemented
formations, injectivity that exceeds undisturbed formation injectivity
indicates tensile parting, that is hydraulic fracturing. This is also the art
taught as central to Canadian Patent 1,122,113 where the formation is floated
by having the injection pressure exceed the minimum in situ stress which is
the weight of the overburden in the applications that are taught. Tensile
parting occurs only if the minimum in situ stress is exceeded by the injection
pressure. In a weak formation, a low rate stress test shows that undisturbed
formation injectivity is exceeded in oil sands for almost the lowest rates at
which injection can be conducted, e.g. several liters per minute. At Cold
Lake, for example, subsequent injection at several cubic meters per minute
occurred at pressures up to 2 MPa less than the minimum in situ stress. In
this situation, formation disturbance occurs leading to the increased
injectivity. This cannot occur due to tensile parting but rather is due to
material failure. The material undergoes localized shear failure due to
differences in the in situ stressses and the elevation of the pore pressure to
values within a few MPa of the minimum in situ stress. The material failure
increases porosity in the uncemented granular reservoir material, and
increased injectivity results: injection occurs at pressures less than the in
situ stress at rates that exceed those expected from the undisturbed formation
injectivity.
Local shear failure is the expression of material failure for an uncemented
granular material. Material failure occurs when the state of stress exerted
on a part of the material approaches the failure envelope for the material.
It is accompanied by dilation of the material, that is, increases in porosity
and corresponding increases in permeability. (Lambe, T.~., and ~hitman, R.V.,
1969. Soil Mechanics. John ~iley and Sons, Inc., New York, ISBN 0471
S1192-7.) It is the increase in permeability that gives rise to improved
in;ec~ivity.
. .
- : ~ . . :
1~23~61
Laboratory tests on the mechanical strength of reservoir material can show
whether or not it is a sufficiently weak material for the existing in situ
stress state, so that material failure will occur before tensile parting
occurs. In such tests as described by Dussealt, M.B. and Morgenstern, N.R.,
1978. (Shear strength of Athabasca Oil Sands. Canadian Geotechnical Journal,
Vol. 15, pp 216,238); Agar, J.G., 1984. (Geotechnical behaviour of oil sands
at elevated temperatures and pressures. Ph.D. Thesis, Department of Civil
Engineering, University of Alberta, Edmonton, Canada); Agar, J.G.,
Morgenstern, N.R. and Scott, J.D., 1983. (Geotechnical Testing of Alberta Oil
Sands at Elevated Temperatures and Pressures. Proceedings 24th U.S. Symposium
on Rock Mechanics, June 1983, pp 795-806.); Head, R.H., 1980. (Manual of Soil
Laboratory testing. Pentech Press, London.), a sample of reservoir material
usually recovered as core is subjected to confining stresses and pore
pressure. These are changed in prescribed manners corresponding to various
stress paths until failure is observed. By altering the stress paths and
using a number of samples, the envelope of stress conditions for which
material failure occurs can be defined. By comparison of this envelope with
the in situ stress state and its variations due to pore pressure increases
during the injection, conclusions as to whether or not material failure will
occur can be made.
A minor observation that helps identify formations that might be sufficiently
weak is that uncemented sands have grains that can more easily move relative
to each other as will occur during material failure. Cemented sands have
intergranular bonds that must be broken before such movement can easily
occur. Depending on the degree of cementation this can make cemented sands
considerably stronger, such that tensile parting will occur before material
failure. Not all uncemented granual materials have the same failure
envelope. An example illustrating this is the locked grain structure observed
by Dussealt, M.B. and Morgenstern, N.R., 1979. (Locked sands. Quarterly
Journal of Engineering Geology, Vol. 12, pp 117-131.)
' ' "' ~ ' '.~'~'
~:
~323~1
In the formation at Cold Lake, the injection of low viscosity fluid (water),
at pressures lower than the in situ stress state and rates higher than those
expected from the undisturbed formation injectivity has been observed.
Injection does start out above the minimum in situ stress but as that initial
fluid pressurizes some of the pore volume, that pore volume can fail, increase
in permeability, accept fluid and grow. The material failure zone grows but
at the pressure required for material failure, rather than the pressure to
maintain tensile parting.
Observations at Cold Lake have shown that when material failure occurs in a
virgin reservoir, the propagation of the failure proceeds in a single
preferential direction close to the orientation of the maximum in situ
stress. The width of the failed zone is set by the size of the initial
hydraulic fracture. The shape of the failed portion of the reservoir after
steam injection is like a cigar - long and narrow. Such regions are
unsuitable for communica~ion except between wells lying along the narrow cigar
shape. Even then, unless the channel is sufficiently large, the communication
path is not suitable for sufficient recovery of bitumen. The present
invention combines conventional technology for hydraulic fracturing in the
presence of the correct in situ stress state with an appropriate steam
injection strategy to Dvercome limitations of typical operations and achieve
desirable thermal communication and areal conformance.
Thus, the present invention was developed as a method applicable to theaforementioned type of formation for enhanced recovery of viscous oil from
such tar sand formations.
The present invention provides a process for achieving thermal communication
and areal conformance in a weak uncemented tar sand formation between at least
one injection well and one production well penetrating said formation, which
comprises
(1) steam stimulation of said at least one production well by injection of
steam for a time and at a volume and pressure sufficient to drive steam into
said oil-bearing formation in the vicinity of said well,
.
- . : .
~ '
1323~61
(2) terminating injection of steam to said at least one production well and
producing fluids from said at least one production well while maintaining back
pressure on said well to prevent substantial revaporization of steam
condensate forming part of said fluids, until the bituminous fluid production
rate is below an economically acceptable value,
(3) when the stress state favors horizontal fracturing, injecting a high
viscosity fracturing fluid at said injection well at a rate and for a
sufficient time to propagate horizontal fracturing from said injection well to
said at least one production well,
(4) immediately subsequently injecting steam through said injection well at a
rate and at a pressure necessary to promote material failure in said formation
and which will decrease to less than overburden pressure while still in a
state of net overinjection and producing steam condensate and bitumen at said
at least one production well for a time sufficient to establish fluid
communication and heated areal conformance between said injection well and
said at least one production well.
In a preferred embodiment the foregoing process also includes the additional
steps of (5) reducing the rate of injection of steam at said injection well
below that of said rate of production of fluids from said production well
until the proportion of bitumen in said fluids from said production well
decreases below a predetermined level, and (6) increasing the rate of
injection of steam at said injection well to increase the rate of production
of bitumen in the fluids produced from said production well. In the foregoing
process steps (1) and (2) may be carried out on a cyclic basis and also steps
(5) and (6) may be repeated sequentially on a cyclic basis.
In the drawings:
Figure la is schematic representation in plan view of an inverted seven spot
well pattern during steps (1) and (2) of the foregoing process at the level
where steam is in;ected in the formation.
~32356~
Figure lb is a sectional schematic view taken along the lines A-A' of figure
la.
Figure 2a is a schematic plan view of the inverted seven spot well pattern at
the level of injection at the end of step t3) of the process according to the
invention.
Figure 2b is a sectional schematic view along the lines A--A' of Figure 2a.
Figure 3a is a schematic plan view of the inverted seven spot well pattern at
the level of injection during step (4) of the process according to the
invention.
Figure 3b is a schematic sectional view along the line A--A' of Figure 3a.
Figure 4a is a diagrammatic representation in plan view of the seven spot well
pattern during step (5) of the process according to the invention at the level
where steam is injected into the formation.
Figure 4b is a cross-sectional representation along the lines A--A' of Figure
4a.
Figure Sa is a diagrammatic representation in plan view of the inverted seven
spot well pattern during step (6) of the aforementioned process according to
the invention at the level where the steam is injected into the formation.
Figure 5b is a sectional diagrammatic view along the level A--A' in Figure Sa.
Figure 6 is a schematic representation of the relative locations of the wells
at an experimental pad at Cold Lake Alberta wherein the method according to
the invention was carried out as described in the Example herein.
Figure 7 is a graph showing the injection pressure of steam vs. time in days
from the time of hydraulic fracturing of the formation for the method of the
invention as carried out and described in the Esample herein.
Figure 8 is a graph showing cumulative volumes of steam injected and fluid and
oil produced from the time of hydraulic fracturing of the formation for the
method according to the invention as described in the Example herein.
Figure 9 is a schematic diagram showing the results from simulation studies of
the hydraulic fracture treatment and the first 30 days of steam injection to
the central well in the method of the invention as carried out and described
in the Esample herein.
-10-
'
,, ~
~:
1~23~
Figure 10 is a graph showing bottomhole and wellhead temperatures as well
production rates of fluids for well 26 in Fig. 6 for the method of the
invention as carried out and described in the Example herein.
The process according to the present invention is applicable to any pattern of
injection and production wells with a minimum of one injection well and at
least one production well spaced therefrom but preferably with at least two
spaced production wells not colinear with the injection well. However, the
process is particularly suited to the use of an inverted pattern such as a
seven spot well pattern, which type of pattern is well-known to those skilled
in the art and involves a central injection well surrounded by six production
wells in the form of a hexagon.
The field in which the wells are drilled comprises an oil bearing reservoir
bounded above and below by impermeable formations. The injection and
production wells are provided in this field and all penetrate into the
reservoir. They are drilled by conventional drilling methods using drill bits
on a tool string which is removed after the hole is drilled to its final depth
to allow the installation of casing which is cemented in place. The casing
and cement are conventional for thermal recovery operations and each well is
provided with production tubing which places the wells in fluid communication
with the well heads in conventional manner. Each well head is provided with
suitable inlet and outlet connections, again in conventional manner. The
production wells and the injection wells are sealed at the lower end by the
cementing operation and use of conventional well cementing equipment.
Communication of each well with the surrounding reservoir is achieved by the
use of a perforating gun which gun creates holes in the casing when detonated
at the desired depth in the well, or one can use a method of notching wherein
a nozzle which e~ects a fine stream of fluid containing particulates is used
which wears the casing away as it is rotated. Both of these are conventional,
but the former method is more economical. The wells may be substantially
perpendicular to the planar surface of the terrain or they may be ~deviated~
'.
' ~ ,.
1323561
or "directionally~ drilled as is well-known in the art so that the surface
area over which the wells are distributed is relatively small compared to the
area over which the wells penetrate at the level of communication with the
reservoir.
The first two stages of the process according to the present invention are
preferably carried out in a cyclic manner and in themselves are conventional
steps known as "cyclic steam stimulation~. The purpose of this initial stage
in the process is to establish a heated portion of the reservoir in the
vicinity of each production well to lower the viscosity of the oil. Since
material failure will occur in a weak formation during this injection in the
vicinity of each of the production wells, reservoir dilation will occur
improving productivity of the wells. The expanded reservoir volume around
each of the producers will then also alter the stress state in the vicinity of
the central well to favour horizontal fracture over vertical fracture prior to
injection of highly viscous fracturing fluid at the injection well. First of
all, steam is injected into each production well and a total of about 5,000 to
15,000 cubic meters of steam is injected at each production or offset well at
a pressure sufficient to drive the steam into the formation and heat the
formation. This is generally carried out over a period of one or two months
or so depending on the particular formation and the well characteristics.
Suitable pressures of steam will depend on the depth and type of formation.
The rate of injection of the steam is suitably over a period of a month or two
at a rate between 100 and 300 m3/d. Steam must be injected to heat a
sufficiently large reservoir volume. Numerical studies and empirical
observations for the formation at Cold Lake Alberta have shown the total
volumes suitable to be as aforementioned.
~hen the predetermined volume of steam injection is achieved, steaming is
terminated and the production of condensate and bituminous fluid is begun from
these wells while maintainin8 a back pressure on each of the wells to prevent
-12-
' " ' ,
':
. ~ ,,' ~ ' ` ,.` ~
1323~
substantial revaporization of condensate within the formation or within the
wellbore. The production of fluids is continued until the bituminous fluid
stops flowing, that is, when it becomes too cold to continue the flow. Then,
steam is again injected at each of the production or offset wells and the
process repeated. The number of cycles will depend on how much heating is
required around each production well and generally for the purposes of the
present invention will comprise one to three cycles, whereas cyclic steam
stimulation in conventional processes usually involves of the order of eight
to fifteen cycles. The next step in the process can only proceed when a
determination of the stress pattern at the central injection well indicates
that horizontal rather than vertical fracturing is favoured and this test can
be carried out by the test described in KrY et al to which reference has
previously been made.
Figures la and lb show by diagrammatic representation the heated areas 10
around production wells 1, 2, 3, 4, 5 and 6 surrounding injection well 7.
Areas 11 in these figures show where material failure has occurred by virtue
of the injection of steam and increase of pore pressure around the production
wells 1-6.
In the overall process according to the invention it is desired to affect the
stress characteristics of the formation to promote horizontal fracturing
between the injection well and each production well as this allows
communication between a mulitiplicity of wells whereas vertical fracturing, if
favoured, can only result in limited well connectivity.
The next stage in the process according to the invention is the injection of a
high viscosity fracturing fluid through the central injection well when the
stress characteristics of the reservoir favour horizontal fracturing over
vertical fracturing. It has been found that the fracturing fluid to be used
at the time of entry into the formation through the injection well should have
a fluid loss coefficient of less than about 10 4 m s 1/2 and an effective
, ~
:
132~
viscosity at a shear rate of S00 s 1 of greater than 0.1 Pa s. These
characteristics are most economically achieved by the use of a cross-linked
guar gum as the fracturing fluid as for example MITEY-GEL LT-1300*
(Halliburton) or YFG* (Dowell) with Jlll* thickener which is a well-known such
fluid. These characteristics can be achieved with several fracturing fluids
from many different service companies. The characteristics are achieved quite
economically with water based fracturing fluids that are prepared by adding
guar gum to water to form a base gel with a moderate nonlinear viscosity. The
high viscosity properties are created with special proprietary chemical
formulations from hydraulic fracturing service companies that utilize metallic
ions such as titanium to act as complexing agents which cross link the guar
gum structure in the fluid and significantly increase its viscosity. Fluid
loss control is often also accomplished by addition of fine mesh size sand
(such as 100 mesh) added in low concentration to the fracturing fluid. In
conventional operations, which are quite satisfactory for the present
invention, the fracturing fluid is mixed to specifications by a service
company at the well site.
The fracturing fluid is preferably injected at the injection well at a rate of
about 5 to about 20 cubic meters per minute. It is found that rates lower
than about 5 cubic meters per minute are not particularly effective and higher
rates than about 20 cubic meters per minute can cause safety problems. The
injection of the fracturing fluid is carried out for a sufficient time to
propagate horizontal tensile fracturing between the injection well and each
production well or until such time as the fracturing fluid reaches the
production wells. In theory, the production wells can be monitored for an
increase in pressure which indicates the arrival of the fracturing fluid in
the formation at those wells. Alternately, this can be estimated from
calculations using a fracturing simulator (such as TAaFRAC* a commercial
*Trade Mark
:.
:~ :
132'~61
product from SIMTECH Consulting). Fracturing simulators use the input
conditions for the fracturing treatment such as the stress state, the
injection volumes and rates, the mechanical properties of the reservoir and
the fluid properties of the injected hydraulic fluid to then numerically solve
the governing equations for the propagation of a tensile parting fracture with
fluid leaking from the faces into the pore system and out of the fracture.
Such a simulator gives a prediction of fracture size as a function of volume
injected. Often assumptions need to be made as to the shape of the perimeter
of the hydraulic fracture, but for horizontal fractures a circular shape
assumption may be quite reasonable. To generate a fracture with a 100 m
radius typically requires the injection of 500 to 1500 m .
Figures 2a and 2b show in diagrammatic representation the situation at the
level of the injection of the fracturing fluid at the end of this stage of the
process wherein the fracturing fluid has propagated a tensile parting fracture
to each of the production wells. In Figures 2a and 2b, the previously heated
volumes surrounding the production wells are indicated at 10 while 12
indicates where fracturing fluid is present. During this stage of the
operation the production wells are shut in to prevent any uncontrolled flow of
fluid to the surface while the hydraulic fracture treatment is being
conducted. This is not considered to be a likely event, but the consequences
could be a severe disruption of the hydraulic fracturing treatment during its
execution.
The next stage in the process is that injection of the fracturing fluid at the
in;ection well is halted and injection of steam therethrough is begun at a
higher rate than is normally used for cyclic steam stimulation of wells. The
rate of in;ection of the steam is, according to the invention, preferably
greater than about 300 m3/d and at a pressure which decreases to less than
the overburden pressure as the material failure begins to dominate over
, ~
.
' .
1 3 2 ~
tensile parting promoted during the hydraulic fracturing phase.
Simultaneously, steam condensate and bitumen are produced from the production
wells at maximum rate. The rate of injection of the steam through the
injection well is limited mainly by safety factors and will probably not
exceed about 1,000 m3/d. A typical figure for this stage of the process is
about 500 m3/d of steam. This injection with simultaneous production at a
high rate of water condensate and bituminous fluid from the production wells
is continued until the reservoir between the injection well and each
production well has a significantly heated portion. This can be determined by
calculations with a numerical simulator similar in principle to the fracturing
simulator or more directly, observations of flowing temperatures can be made
at the production wells to determine when heat first arrives and hence
determine the length of time subsequent to that first arrival that hot fluids
must continue to flow in order to generate a sufficiently large heated portion
of the reservoir. If insufficient heating of the pathway between injector and
producer occurs, then through the narrowest part of that pathway subsequent
flow of bitumen may be cooled by conduction heat losses from the narrow
portion resulting in increased bitumen viscosity and plugging of the pathway.
The initial heat arrival time will depend on the specific injection rates and
how well material failure was suppressed during the hydraulic fracturing phase
and on the extent of material failure above and below the initial fracture.
The sooner that hot fluids arrive, the thinner is the material failure zone
and the longer must be the subsequent flow of hot fluids to sufficiently heat
the reservoir between injector and producer. In the experience at Cold Lake,
hot fluids began to reach the producers after about 100 days of injection
(total volume 43,000 m injected and 5,000 m3 of total fluids produced per
injector). High rate in3ection (>300 m3td) was continued for an additional
100 days to ensure thermal communication. In other words, the steam injection
at high rates is continued until there is thermal communication between the
injection and each production well and significant areal conformance has been
achieved. In practical terms the injection of steam at relatively high rates
-16-
: :. . :.. : ..
- .:: -:
~ -' ~ ; "
:
1323561
is continued with the total fluids produced at the production wells equalling
the fluids injected at the injection well until the water to bitumen ratio
from the production wells is excessive. At the end of this stage thermal
conformance of a heated area has been achieved over a significant portion of
the reservoir as shown in Figures 3a and 3b, the previously heated volumes
being shown at 10, the area or volume permeated by the fracturing fluid being
shown by 12 and the volume heated by steam injection being shown at 13.
As preferred steps according to the invention in order to achieve increased
production of bituminous fluids from the production wells, the next stage is
that the rate of injection of steam at the injector well is reduced to a rate
below the production rate from the offset wells and the pressure is reduced.
Thus, the heated water in the formation vaporizes and promotes exchange of
water and bitumen and hence flow of the latter. In fact, one can even cut off
the injection of steam at well 7 completely until one has reached the
preferred operating pressure. The production wells are operated within safety
constraints at maximum rate of fluid handling capability. The bitumen will
continue to flow and be produced at the production wells until such time as
there is a general cooling in the formation which adversely affects the
production rate. ~hen this step of the process is terminated will depend
strictly on economics but in general will be when the bitumen rate decreases
below that expected during high injection rate flooding. The injection rate
of steam at the injection well 7 is then increased to a rate higher than the
maximum production rate and higher than at least 300 m3/d in order to reheat
the formation with concurrent maximum production rates at the production
wells. During the stage when the rate of steam injection is decreased and the
flow of bitumen increased from the production wells, it has been found in
formation at Cold Lake, Alberta that the volume ratio of water condensate to
bitumen is about 3. Figures 4a and 4b show the condition of the reservoir
during the low pressure injection of steam stage of the process including
liquid (14) and vapour (15) zones superimposed on the zone (13) heated during
. :
. .. : ' '"
,
,
~ 323~
the prior high rate steam injection phase. Figures 5a and 5b show the
condition of the reservoir when high steam injection rates are resumed to
increase pressures and temperatures in the reservoir at 16.
The last two stages in the process can be repeated sequentially for as many
cycles as required and as long as bitumen is being economically produced from
the reservoir via the production wells.
EXAMPLE
The following specific example of the method carried out at a formation at
Cold Lake, Alberta illustrates the process according to the present
invention. Figure 6 illustrates the relative locations of the wells. To
conduct the experiment 7 wells were reserved from the corner of a 30 well
pad. The wells were drilled with an interwell distance of 105 m and the 7
wells reserved formed an inverted seven spot. The wells were all drilled to
the base of the oil sand (500 m), however the quality of the lower part of the
reservoir was considered to be marginal and thus the seven spot wells were
perforated near the middle of the reservoir. The top of the formation was at
460 m depth.
The six surrounding wells were carried through a first cycle of cyclic steam
stimulation. On average, 6500 m3 of steam was injected to each well and
production carried out over about a 5 month period. Each well on average
produced 3500 m3 of fluid including about 1400 m3 of bitumen. During this
time the stress state was monitored in the central well and it was observed
that by the end of the first cycle production, the minimum in situ stress had
decreased below the weight of the overburden. This means that vertical
fracturing would be favoured. To again change the stress state to favour
horizontal fracturing required the injection of more fluid to the producers.
It was thus necessary to undertake a second cycle of steam stimulation in the
six surrounding wells. This was accomplished with an average injection
-18-
;: :
:
. - ,
:' . ,
.
: : : ;. :.,
.
1323~6~
of 6200 m of steam which was varied between the wells to provide an
approximate balance with the net fluid injection over the life of the wells to
that date.
The stress state was assessed at the end of the second cycle of steam
injection and found to favour horizontal fracturing. The initial weight of
the overburden at the depths of the perforations was 9.9 MPa while the minimum
horizontal in situ stress was determined to be 11.5 MPa. The pore pressure
was at 10.3 MPa due to the second cycle of steam injection to the surrounding
6 wells suggesting that the vertical stress was somewhat elevated compared to
initial conditions by the injection to the surrounding producers. ~ith these
conditions the hydraulic fracture treatment was conducted using a guar gum
aqueous based crosslinked gel that during the fracture treatment had an
effective viscosity at a shear rate of 500 s 1 of 0.1 Pa s and a fluid loss
coefficient of 1.4 x 10 4 m s 1/ . 750 m3 of fluid were injected at a
bottom hole pressure of 12.9 MPa and fluid injection rate of 8 m3/min.
Steam injection was begun within 40 min of completion of the injection at a
bottom hole injection pressure of 11.5 MPa.
For the first 30 days the injection rate averaged 550 m /d at an injection
pressure of 11.7 MPa. Net overinjection was still occurring to the pattern
with more than twice as much condensed volume of steam injected as the total
volume of fluid produced. The rate was decreased to average 370 m3/d after
30 days of injection while the injection pressure dropped to 11.2 MPa. ~ith
no reduction in net injection rate after lO0 days the injection pressure had
decreased below lO MPa and by 200 days it had decreased to near 8 MPa. From
30 to 200 days there was essentially balance between the condensed volume of
steam injected and the volume of produced fluids. These results are shown in
Figures 7 and 8 which show the injection pressure history and the cumulative
volumes injected and produced from the pattern from the time of the hydraulic
fracture. The continuous support of flow from the producing wells
demonstrated that communication had been achieved between the central well and
each of the surrounding producers.
-19--
.
:'
1~23~
It was during this time interval that it was observed that the well to the
northeast of the pattern (24 in Fig. 6) was being strongly influenced by the
injection to the central well. Observation of a pressure response in the
closest well (23) in the pattern during the hydraulic fracture treatment
indicted some assymetry in that direction, suggesting that material failure
was playing some role during the hydraulic fracture treatment: that ~t was
not completely suppressed in favour of tensile parting. Consequently, that
well (24) was included in the pattern and operated as the bounding well to the
pattern in that direction. The intermediate well (23) was only intermittently
operated and served essentially as an observation well.
In this case, the steaming to the surrounding wells changed the stress state
significantly from the ~irgin condition. It is this change in stress state
that required such high pressures to initiate the material failure process.
The pore pressure must be very close to the in situ stress values since all
three stress components are close to each other. In virgin reservoirs where
higher differences between stress components can exist, pore pressures
significantly less than the minimum in situ stress value can result in
material failure. These results are clear from consideration of the equation
for the failure pressure with the assumption of a Mohr-Coulomb failure
envelope with typical friction angle of ~ - 35 .
Pr = 2 tah+ av- (ah- a~ in] (1)
~here ~ is the vertical in situ stress (minimum principal assumed) andc~h
is the marimum horizontal in situ stress (maximum principal stress). ~hen the
6tresses approach equality, the failure pressure approaches their value. ~hen
there i8 a contrast of 2 MPa between the minimum and maximum, Equation l shows
the failure pressure is 0.~ ~Pa less than the minimum in situ stress. As
overinjection continues, the size of the zone of failure must increase.
-20-
~ ' ' ''
.
1323~
Eventually the size of this zone becomes sufficiently large that the shear
support of the overburden which gives rise to the increase in vertical stress
can no longer support the area of pressurized reservoir and the contrast
between the vertical stress and horizontal stress must increase, reducing the
failure pressure while the reservoir stays in a state of net overinjection.
Simulation of the hydraulic fracture treatment indicated a radius for thé
fracture of slightly less than the well spacing (17 in Figure 9). Simulation
of the first 30 days of steam injection to the central well with a simulator
accounting for material failure (Settari, A., 1988. Modeling of Fracture and
Deformation Processes in Oil Sands: Paper presented at the 4th UNITAR/UNDP
Conference, Edmonton, Alberta. Settari, A., Kry, P.R., Yee, C.T., 1989.
Coupling of Fluid Flow and Soil Behaviour to model injection into uncemented
oil sands. Journal of Canadian Petroleum Technology, Vol. 28, No. 11, pp
81-92.) suggested that heat should be distributed out to about 85 m radius (18
in Figure 9). These results are shown in Pigure 9.
Thermal communication occurred about 100 days after the hydraulic fracture
treatment. The time of arrival of thermal communication is defined as the
time at which half the ultimate thermal response has been observed. This is
shown in Figure 10 which is a particularly clear example. The bottomhole
temperature was estimated from the well head temperature and the production
rates using an approYimate relationship for wellbore heat losses. As shown in
Figure 10 for well 26 in Figure 6, the estimated bottom hole temperature
decreases from the day of the hydraulic fracture treatment as the reservoir
near the producer c0018 due to production. After about 60 days for this well,
the temperature reaches its minimum value and begins to increase to the
maximum value estimated after about 140 days. It reaches the half way point
in the transition at about day 100. Similar timing was observed for the other
producers.
-
-21-
.
: .
1~23~
The high rate injection continued for another 100 days after this thermal
communication time. At that point the central injector was shut in for 40
days while production continued and the pressure fell to near 3 MPa.
Subsequently low rate injection was restarted to the central well. Production
continued for about 400 days until a repressurization cycle was started. This
latter interval is also shown in Figures 7 and 8 which are the injection
pressure history and the cumulative volumes injected and produced after the
hydraulic fracture. To the point of repressurization including the cyclic
steam stimulation cycles, 185,000 m3 of steam have been injected and 178,000
m of fluids have been produced, including, 40,000 m of bitumen.
-22-
: . : : ."
:
~:
. ~ ; ~- '
