You are on page 1of 24

SPE 59333

A Mechanistic Study of Single-Well Steam Assisted Gravity Drainage


Ashok K, Singhal, SPE, Swapan Das, SPE1, Jon Goldman and Alexandru T Turta, SPE

Copyright 2000, Society of Petroleum Engineers Inc.


This paper was prepared for presentation at the 2000 SPE/DOE Improved Oil Recovery
Symposium held in Tulsa, Oklahoma, 35 April 2000.
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
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300
words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

Abstract
Single Well Steam Assisted Gravity Drainage (SW-SAGD)
process was evaluated by closely examining performance of
field projects in the public domain, physical modeling and
numerical simulation.
It was seen that in order to be economically acceptable, field
implementation emphasizes the Single Well (SW) aspects
(near well bore heating), as opposed to the SAGD aspects
(creation of a large steam chamber).
The reported steam-oil ratio of about one for SW-SAGD
field projects includes contributions due to primary
production. Obviously, the performance will be attractive
where the primary production is strong. Presence of mobile
water or gas in the vicinity of a target location will have a
negative effect on performance.
The economic performance can be stronger than that for
primary and cyclic steam stimulation but perhaps not as strong
as for a dual well steam assisted gravity drainage. Therefore
targets should typically have a continuous pay thickness of 10
to 15 m.
Potential improvements can be obtained from optimizing
well bore configurations by way of reducing steam by-pass
and extending the effective well length utilized for production
during the process. Other optimizing ideas involve the use of
gases as steam additives and injection-production rate
scheduling.
The process is essentially a variation of cyclical steam
stimulation (CSS) rather than of SAGD. Using certain

elements of the SW-SAGD process, one should be able to


profitably modify the cyclic steam stimulation (CSS) process
while using horizontal wells.
Introduction
The Single Well Steam Assisted Gravity Drainage (SWSAGD) process is a steam injection based oil recovery process
in which the same horizontal well is used for steam injection,
as well as for oil production. Steam is injected through
insulated tubing, whereas hot oil and condensed water are
pumped to the surface through a separate tubing string. It was
initially thought of as a variation of the dual well Steam
Assisted Gravity Drainage process (SAGD). In SAGD, steam
is injected via a horizontal well and heated oil and condensate
produced via another horizontal well, placed directly
underneath the injector. The main driving mechanism being
gravity.
While considering steaming via horizontal wells, the idea of
heating the near well bore region by steam circulation and
draining the heated oil was evaluated by Huygen 1. Although
early field trials did not meet with success2,3, problems of
delivering high quality and evenly distributed steam over the
well length were identified. In 1997, Nzekwu et al.4 patented a
concept of using the same single well for injection, as well as
for production, by creating a steam chamber and promoting a
counter-current flow (steam rising and oil plus condensed
water draining to the same well). Appropriate insulated steam
injection tubing (Insulated Concentric Coiled Tubing or ICCT)
was developed5 and the concept was field tested by Elan
Energy (now Ranger Oil) at Cactus Lake and many other
places in Canada6,7. Mobil Oil Canada tested another version
of the concept at Celtic3. Both operators reported strong oil
rates (upwards of 40 m3/d) and oil steam ratios (about one)
during the first year or so of operations involving heavy oil
with viscosities of several thousands mPa.s (cp). However, the
reported oil-steam ratios were computed on a gross basis and
not on an incremental basis, i.e. oil production included the
primary component. Heat balances for the two projects
indicated that the performance of these projects was
dominated by near well bore heating as opposed to creation of
a large steam zone. On the other hand, the patent disclosure4
and results of physical model studies8 (and also the current
study) suggested that significant steam chambers could be
created during SW-SAGD operations. A puzzling question

ASHOK K. SINGHAL, SWAPAN DAS, JON GOLDMAN AND ALEXANDRU TURTA

was why the operators chose to settle for near well bore
heating instead. Encouraging field performance of the above
two projects (Cactus Lake and Celtic) also generated
additional curiosities. These included main mechanisms;
conditions under which SW-SAGD provides strong
performance and any improvement one could suggest on the
practices of these two operations. This paper reports the
results of analyses of field projects and physical modeling,
which attempted to address these questions. It may be pointed
out that when the same well is used for injection as well as
production, it involves a risk that a significant portion of steam
will travel though the well without entering the reservoir.
Design of well completion such that steam bypass is
minimized and high quality steam is delivered at the sand face,
are therefore very important to this kind of operation.
In addition to the above publications, several reports based
on simulation studies2,9,10 of the SW-SAGD process have
recently appeared in the public domain. Oballa9 concluded that
well bore behavior (frictional pressure drop and compositional
changes) dominated the drainage process. Although she saw
potential in the process, she pointed out the difficulties in
managing such projects. Shen2 studied the effect of capillary
threshold on steam injection and concluded that it may not
always be easy for the steam to overcome these thresholds and
enter the formation. However, any undulations in the well bore
trajectory can partially help overcome capillary thresholds. He
also pointed out that any dilation of the formation in the near
well bore region would reduce capillary thresholds and
significantly enhance the performance. Near well bore dilation
as well as undulating trajectories are common features related
to drilling of horizontal wells. Sawhney et al.10 pointed out
avoidance of excessive steam production/ by-pass as the
reason behind reducing steaming rates in field operations.
Important insights on stability of the process and the
importance of steam by-pass while using the same well for
injection and production were obtained from these papers. In
order to obtain additional insights, it was decided to conduct
an in-depth analysis of these two operations, followed by
physical and numerical modeling of different aspects of a SWSAGD project to obtain clues about the process.
Field Experience: Rangers Cactus Lake SW-SAGD
Operations. Figures 1-3 present some significant features of a
typical wells completion and performance at the Cactus Lake
Project (courtesy Elan Resources). Strong oil rate performance
(up to 100 m3/d) was reported at the well 7C16-16-36-28W3
(Figure 2). From their publications6,7, interviews with
personnel and simulation (input data made available to us,
courtesy Elan Resources) several additional clues were
obtained. Data from an observation well revealed that there
was significant pressure depletion during SW-SAGD
operation, partially due to production operations at offsetting
wells. Fiber optics probe at one SW-SAGD operation
indicated that the well length heated to steam temperature was
rather small (much less than 100 m, representing less than
10% of the horizontal section). From history match of

SPE 59333

performance it was seen that lateral heating was extremely


poor-no more than a few meters. However heating extended to
several meters immediately above and below the SW-SAGD
well.
Due to pressure drop (estimated at 50 kPa) from the toe to
the heel, distribution of heat injection into the reservoir was
not uniform. Consequently, the heated region around the well
is of variable extent (resulting steam chamber is illustrated in
Figure 4). In this region, pressure gradients along the direction
of the well cause a partial drive of oil towards the heel. This
pressure drop has an important influence on the amount of
steam entering the formation and the amount of oil and water
produced by the horizontal well. Some amount of steam
always bypasses the reservoir and return along with well
effluent. As a result, the actual steam injection into the
reservoir is smaller than the amount leaving the steam
generator. Well configuration and lateral pressure drops within
the well thus play important roles in determining thermal
efficiency of operations. Large lateral pressure drop within the
horizontal section would tend to drive more steam into the
formation. This could be achieved by reducing the annulus
clearance through which steam travels to the toe of the well.
The trajectories of the wells had the usual undulations. There
were also indications of multiple steam zones near various
highs of the undulating wells. This is consistent with results
of Shen 2. Because of these, increasing of steam injection rate
would have driven steam away from where it could be most
profitably employed.
At high rates of steaming, oil production rate was relatively
low and at the same time, relatively hot. The strategy of
optimizing steaming rate included throttling back when
production temperatures became too high. These efforts were
aided by the use of chlorides in the effluent as tracers. Salt
balances also helped identify excessive steam by-pass and
production of fluids from regions away from the wells. In
general, their steaming rate declined over time.
Since the reservoir pressure was gradually depleting due to
primary or thermal operations at different wells, SW-SAGD
performance seem to deteriorate at thermal operations that
were initiated when the pressure had already declined (Fig.3).
Long term success of SW-SAGD operations therefore would
require some pressure support to the reservoir.
In Rangers experience, wells located higher up within the
pay section seemed to provide a stronger SW-SAGD
performance than those completed near the base of the way,
again confirming that gravity assisted drainage played a
relatively minor role in their operations. Simulation confirmed
a significant heating of the reservoir and production of the
mobilized oil from underneath the SW-SAGD well.
Field Experience: Mobils Celtic SW-SAGD Operations.
SW-SAGD was applied at 5 wells at Celtic, operated by Mobil

SPE 59333

A MECHANISTIC STUDY OF SINGLE-WELL STEAM ASSISTED GRAVITY DRAINAGE

Oil Canada, during 1996-98. Injection/ production data on 4


SW-SAGD wells and a dual well SAGD project operated in
the same part of the pool are shown in Fig.5. Mobil has not
published results on these projects, other than a presentation
by Saltuklaroglu3. While Ranger used ICCT to deliver high
quality steam at the sand-face, Mobil used a nitrogen blanket
in the annulus (around injection/ production tubing) to provide
an extra insulation to the steam tubing. The use of nitrogen
also helped in reducing the bottom hole pressure. Three
observation wells were used for pressure monitoring purposes.
They increased and decreased their steaming rates in a cyclic
manner, each cycle lasting for several months. They achieved
a calendar day oil rate of 28 m3/d /well for the first 18 months
at a steam-oil ratio between 1.2 and 1.4 but there is a suspicion
that steam rates were under-estimated. Production was shut-in
during periods of high steaming rates whereas in the case of
Cactus Lake, it continued simultaneously with injection
operations.
It was concluded that SW-SAGD operations provided a
better oil rate performance than primary production or Cyclic
Steam Stimulation (CSS), but not as good as the dual well
SAGD. They state that inter-well communication via
fracturing should be avoided, and they found bottom hole
pressure monitoring to be extremely useful. A heat balance for
a typical Celtic well operation again revealed that SW-SAGD
operation was essentially dominated by a near well bore
heating. Simulation also revealed that steam zone was of an
insignificant size. Temperature profiles suggested occurrence
of some convection around this zone, which might be limiting
the size of the heated region The SW-SAGD wells were
placed in the middle, rather than at the base of the pay zone in
both these projects.
In both of the projects discussed here, the stabilized initial oil
rate was about twice of the primary oil rate. Due to continued
pressure depletion, rates could not be sustained. In case of the
Celtic project, cyclic steam injection partially maintained the
reservoir pressure. In both these projects, the operators
proactively monitored reservoir performance and used
simulation for identifying the need for any interventions to
optimize wells performance. One of the interventions was
injection/ production scheduling to achieve an efficient near
well bore heating/ oil drainage. The underlying motivation
being a promotion of heating by inducing convection and
rapid production of oil thus mobilized.
It appears that in both cases, the shift in emphasis from the
SAGD aspects (creation of a large steam chamber) to the SW
(single well) aspects was forced by the need to achieve and
sustain strong oil rates at low steam to oil ratios. One of the
unanswered questions is: can an economic operation be
devised such that it benefits both from creation of a sizable
steam chamber and near well bore heating beyond
conduction? For investigating these issues, it was decided to
conduct a physical and numerical modeling study of the SWSAGD process, especially in the well sections dominated by

injection and convective heating.


Physical and Numerical Modeling Study
In order to recreate conditions similar to those in field
implementation of SW-SAGD in the laboratory, a semi-scaled
physical model capable of operation at high pressure (4 MPa)
was constructed. This model consisted of a flanged box with
inside dimensions of 41.91 cm x 26.67 cm x 8.9 cm
(16.5x10.5x3.5) containing 55 thermocouples; its inside
was suitably insulated against heat losses. It was filled with
0.5 mm glass beads with a permeability in the range of 85-120
darcies. The model was initially saturated with water, which
was then displaced by a mineral oil (40 mPa.s viscosity), until
no more water was produced, in order to obtain uniform initial
oil saturation. In some experiments, this oil was blended with
4.3 weight percent propane to represent live oil with a
dissolved gas content of 20 m3/m3. A quarter inch stainless
steel tubing with an I.D. of 3.06 mm (0.12), placed along the
long edge of the model, represented the horizontal well (nonscaled). The well bore tubing had a number of drilled holes to
represent the slots. The well was wrapped in fine wire mesh to
exclude any particulate material from the effluent. In a few
experiments, a rod with a diameter of 2.6 mm (0.106) was
inserted within the well to represent well bore equipment
which would increase pressure drop along the well, thereby
pushing more steam into the formation and reducing steam bypass. A SW-SAGD (horizontal) well, placed in a 18 m thick
reservoir (4 darcy permeability) containing an oil with a live
oil viscosity of 4,000 mPa.s (North Cactus Lake Sparky pool)
was the prototype. The model represents an element of
symmetry around a 84 m section, adjacent to the steam exit
end.
The experimental set-up is shown in Fig. 6. It consisted of a
steam generator, the model, a back-pressure regulator, a heat
exchanger to cool the effluent, a water-oil-gas separator and a
data logger. The steam generator was custom built such that
the rate of steam injection could be varied to achieve the
desired injection rate profile, and the injected amount
accurately measured. Likewise, the back-pressure regulator
was modified such that the dome pressure could be adjusted to
obtain the desired production schedule. From the temperature
and flow rate data of the heat exchanger/ condenser at the
effluent, one could estimate its enthalpy.
The performance of some physical model experiments was
numerically simulated using the discretized well option of the
thermal simulator STARS.
Results
It may be recalled that the physical model represents only a 84
m section of the horizontal well near the steam entry point into
the formation. As compared to the performance of the entire
well, results from a physical model experiment that dealt only
with a small section of the horizontal well near the inlet end
will be more pessimistic in terms of steam-oil ratios. At the
same time they will be more optimistic from a point of view of

ASHOK K. SINGHAL, SWAPAN DAS, JON GOLDMAN AND ALEXANDRU TURTA

rate enhancement and oil recovery. A grossly simplified


estimation of oil rate during SW-SAGD operation is presented
next to illustrate this point. Suppose, the SW-SAGD well
produces at an initial (primary) rate of 20 m3/d in a heavy oil
pool. Steam is injected into this well at a rate of 50 m3/d. Next
to the steam inlet point, a 40 m section is utilized for injection.
Another section (40 m) experiences convection dominated
production (with a ten folds oil rate enhancement) and the rest
of the well length is heated by conduction (with a 50%
enhancement in oil rate). An oil rate of 28 m3/d from the
conduction heated section and 16 m3/d from the convection
heated section results, for a total oil rate of 44 m3/d. The
apparent steam-oil ratio is thus 1.14. If we consider only the
injection and the convection dominated sections, we may
conclude that the steam-oil and water-oil ratios are about 3.
However, the convection-dominated component of production
will decline faster than the conduction dominated production
due to oil saturation depletion. Also, both of these will decline
as the pressure depletes and gas/ water saturations rise. Thus,
oil rate and recoveries from a physical model test representing
only the injection plus convection dominated sections of the
horizontal would be highly optimistic. Since the main
emphasis in this study was on obtaining mechanistic insights
based on a comparison between different cases, the results
from the physical model are viewed in the context of this
illustration.
It was decided to follow a parametric approach whereby
effects of changes in a few selected parameters are
individually studied, keeping all other features unchanged (as
much as feasible). It was realized that the SW-SAGD process
involves a complex interaction between well bore
configuration, injection-production schedule and reservoir
performance. Different physical model results were compared
on the basis of total oil recovered before the steam-oil ratio
exceeded a value of 6. As mentioned above some of the results
were also history matched using a reservoir simulation
(thermal) model. A total of 25 experiments were conducted,
and some of them are discussed here.
Low Pressure Tests. The first 13 runs were conducted at low
pressures (440 kPa or 45 psig), mainly to identify appropriate
experimental procedures and explore improvement ideas. In
general, high steam injection rates led to creation of larger
steam zones and stronger oil rates. However, at very high
injection rates, oil recovery actually decreased as steam-oil
ratios rapidly rose. Temperature profiles and heat exchange
data suggested that the well length effectively heated (which
was accepting injection) increased with injection rate. At the
same time, proportion of steam travelling through the well and
not entering the reservoir (by-pass) decreased. Thus, an
optimal steam rate was indicated. This was not pursued any
further as these effects might be significantly influenced by
the design of physical model and its operation. At low rates,
insufficient heating and hence poor performance was noted.
For similar rates, stronger oil rates were noticed if the total
fluid production rate was increased. Thus the performance was

SPE 59333

seen to be extremely sensitive to variations of injection/


production rates.
The performance for live oil at low pressures was
considerably more encouraging than for dead oil. Also,
introducing an insert in the well (as described earlier to
increase the resistance to flow within the well) resulted in
reduced steam by-pass and showed significant increases in
reservoir heating, oil rates and recovery.
By conducing an experiment with the model upside down,
such that the well was placed near the top of the pay zone,
significantly more oil was recovered before reaching a steamoil ratio of 6, as compared to the case with this well at the base
of the pay zone. Oil rates were comparable but there was a
significant amount of convection underneath the well (hot
steam condensate sinking and mobilizing oil upwards).
Simulation results agreed with the laboratory results and
partially explain the observation of one of the operators7 that
wells located higher within the pay zone resulted in a stronger
performance.
The SW-SAGD performance was significantly improved
when small amounts of CO2 (0.43 to 5.2 mole %) were used as
additive to steam. It was also noticed that any mobile water/
gas close to the SW-SAGD well would be preferentially
produced (over oil) due to their higher mobility.
High Pressure Tests. These experiments were carried out at
930 kPa (120 psig) or 3600 kPa (505 psig). Specifically, roles
of reservoir pressure, variable (cyclically increasing and
decreasing) steam injection rate, and back-pressure were
studied. As mentioned earlier, well bore inserts and live oil
were also used in some of the experiments.
Experiment with a constant rate of steam injection at 40
gms/ min at 3415 kPa; dead oil; no well insert. Steam was
injected for 4 hours, during which a minor steam region was
created. Most of the steam injected was instantaneously
produced back. Oil rate varied between 14 and 3 ml/ minute
(Fig. 7). Steam-oil ratio after the first hour became
unattractive, i.e.>6 during which 600 ml of oil was recovered
(Fig.8), perhaps indicating that the effective steam injection
into the model had become inadequate. The next step was to
explore a more appropriate rate[modification of steam by-pass
was deferred to a later run]. Steaming rate was increased to 80
gms/ minute and a spurt in oil production was seen, as the
steam chamber grew relatively more rapidly. The effluent
contained even higher amounts of heat. Steam oil ratios
became acceptable for the next 40 minutes (Fig. 9), during
which 850 ml of additional oil was recovered.
A history match of the model run indicated that around the
entrance end of the well, steam override and slumping of the
steam condensate were occurring simultaneously (Fig. 10 and
11). Furthermore, contours for 50oC roughly coincided with
those for 90% oil saturation and those for 90o C with oil

SPE 59333

A MECHANISTIC STUDY OF SINGLE-WELL STEAM ASSISTED GRAVITY DRAINAGE

saturation contours of 70%. Thus, most of the oil drainage


from above the well is aided by a viscosity reduction due to
heating. The results further suggested that a) well bore
modeling/ steam by-pass needs to be improved, especially
since the efficiencies gradually deteriorate; and b) in order to
sustain acceptable steam-oil ratios, steam rates may have to be
gradually reduced, and occasionally increased in a cyclical
fashion.
The above experiment was repeated with the back-pressure
cycled between 3550 and 1485 kPa (500 and 200 psig) in 5
cycles over 5 hours, representing pressurizing and depressurizing (blow-down) around the SW-SAGD well. The
rates and recovery were less efficient. It was inferred that to be
beneficial, pressure manipulation must trigger some other
phenomenon such as flashing of the solution gas (dead oil was
used in this experiment).
Experiment with a constant rate of steam injection at 40
gms/ minute at 3415 kPa; dead oil; using a well insert.
During this 4 hour test, a small steam chamber formed and
1650 ml oil was recovered (as compared to 1223 ml during the
previous test). Oil rate rose to 16 ml/ minute and then
gradually declined (Fig. 12). In general, oil rate improved by a
third due to the well insert (blocking off 78% of the flow
area). Steam oil ratios were initially around 2.5 but gradually
rose to 6 over 90 minutes. From temperature plots and
numerical simulation, a limited effective well length
utilization was noted. This is due to slow propagation of heat
along the well length and in lateral direction to the well,
caused by the prevailing hydrodynamic/ heat flux conditions.
This poses a real risk of not draining all of the mobilized and
heated oil before the economic limit. Varying of injectionproduction rates may potentially overcome some of these
problems.
Experiments with a cyclically variable rate of steam
injection (average 40 gms/ minute at 3415 kPa). The
injection rate was increased and decreased in a cyclical
manner and was kept the same in all cycles (schedule # 1).
Later on, in some runs, the schedule was varied for different
cycles (steaming schedule #2, Fig.13). In schedule #1, steam
rates were decreased every 15 minutes from 60 gms/ minute to
50, 40, 30 and 20 gms/ minute and then increased to 40 and 60
gms/ minute in steps of 15 minutes each, with a full cycle
lasting over 90 minutes. Four such cycles were implemented.
As the steam rates were increased, oil rates and steam oil
ratios dramatically improved, but could not be sustained for
long. Upon varying injection rates, oil rates varied in tandem.
However, there was a significant time lag between rate
increases and corresponding steam-oil ratio response. The
overall performance deteriorated in subsequent cycles.
However, performance for the variable rate was a definite
improvement over a corresponding constant injection rate
case.

The above experiment was repeated, this time using the live
oil (with 4.3% propane content). The overall performance was
similar but swings in oil rates were somewhat smaller. The
steam oil ratios were slightly lower for the first two hours and
deteriorated thereafter. From effluent temperatures it appears
that upon lowering rates, steam bypass decreased and steam
utilization improved. It was speculated that a different
injection strategy might provide a stronger performance in
terms of oil rates and steam oil ratios. One such strategy
involved the use of shorter cycles (30 minutes), successively
increasing peak rates (to 90 gms/ minute in 4 cycles) and not
allowing steaming rate to fall below 30 gms/ minute. This
schedule (#2, Fig.13) was followed in the next set of
experiments.
Experiments with a cyclically variable rate of steam
injection (schedule #2 with an average of 60 gms/ minute).
Three experiments were performed with a back-pressure of
3495 kPa: using dead oil and no well insert; using dead oil and
a well insert; and using live oil and a well insert. Simulation
indicated that just like the other cases studied, extent of
heating along the well length and in the transverse direction
were limited (Fig. 14). Furthermore, oil drainage was
predominantly occurring from the heated regions, although
steam condensate slumping to the base of the pay was also
contributing to oil production.
The cumulative oil production, oil rates and steam-oil ratios
for these three experiments are depicted in Figures 15, 16 and
17, respectively. The overall performance is an improvement
over the previous steaming strategy. For example, upon four
kg of steam injection into the model containing live oil (and
no well insert) the earlier steaming schedule #1 yielded 650 ml
oil over the first 95 minutes. For the alternate schedule (#2,
with well insert this time) the corresponding oil recovery was
750 ml over 60 minutes.
The use of a well insert definitely improved the efficiency of
steam utilization. The live oil case again showed a slight
improvement over the dead oil case for the first two hours in
terms of steam-oil ratio (Fig.17). However, the cumulative oil
recovered during this period is slightly smaller (Fig.15). With
well bore inserts, highs in oil rates were sustained longer and
lows did not dip to very low values. However, these high oil
rates could not be sustained for long.
Thus, reduction of steam bypass and a judicious injectionproduction strategy are key to improving economics of these
projects.
Implications
From the foregoing, certain inferences about performance of
SW-SAGD projects in heavy oil pools could be derived in
response to some of the following basic questions:
Under what conditions SW-SAGD could provide strong
performance?

1.
2.
3.
4.
5.
6.

ASHOK K. SINGHAL, SWAPAN DAS, JON GOLDMAN AND ALEXANDRU TURTA

Delivery of a high quality steam at the sand face via an


insulated tubing.
Increasing lateral pressure gradient within the well in
order to force more steam into the reservoir.
Minimal interference or pressure depletion from other
wells operations.
Placing the wells in the middle of the pay, with the
trajectory being as straight as feasible.
Placing wells away from regions with mobile water or
gas.
Implementing a cyclic steam injection schedule and
proactive reservoir monitoring for identifying the need for
any interventions to optimize wells performance.

What are the main features during an economically


acceptable operation?
Enhancement of primary production mechanisms by
reservoir heating which includes: creation of a small
steam zone near the steam inlet end into the reservoir and
convection of heated oil adjacent to it; and near well bore
conduction heating of most of the well length.
Convection underneath the well whereby heated
condensate displaces the warmed oil upwards.
Steam bypass via the annulus helps to keep the produced
oil warm, near well bore region heated throughout the
horizontal section within the reservoir, and helps lift oil
from the well.
Pressure depletion which helps drive oil along the length
of the well and increases the volumetric flux for steam,
and provides a push to oil production via evolved
solution gas.
Near bore well dilation helps overcome capillary
thresholds and enable more steam entry into the reservoir.
What kind of improvements can one suggest?
1.Occasionally pulling back liner to serially heat different
sections of the SW-SAGD well and thus increase utilization
of the well length7.
2.Devices to increase lateral pressure drop within wells so that
more steam is pushed into the reservoir and steam bypass is
minimized.
3.Injection/ production strategy to evenly heat a large portion
of the reservoir and efficiently drain the heated/ mobilized
oil.
4.Use of small amounts of gas (CO2 or methane) or a diluent
as gas additives.
What are the key screening guidelines?
Since SW-SAGD is a variation of cyclic steam stimulation
(CSS) using horizontal well, all criteria for CSS and for
placing horizontal wells will also be applicable to SW-SAGD.
Also, since one of the mechanisms is an acceleration of
primary production, pools with strong primary production
performance of horizontal wells are likely to yield a strong
SW-SAGD performance. Thus, a live oil viscosity of 2000
mPa.s, a permeability of one darcy or greater and a continuous

SPE 59333

pay zone of 10 m or thicker are the obvious criteria for


selecting prospects. Another factor would be to avoid
situations where capillary threshold is excessive so that steam
easily enters the reservoir. With a few pull-backs of the liner, a
SW-SAGD well could be economic for a 10 m of continuous
pay zone. For zones of 15 m or thicker, a dual well SAGD is
likely to be more attractive. Therefore SW-SAGD is more
appropriate for prospects with a thickness of 10 to 15 m of
continuous heavy oil pay.
Conclusions
SW-SAGD, for an economical operation (strong early oil
rates, low steam-oil ratio) must focus on SW aspects (near
well bore heating) as opposed to the SAGD aspects
(creating of large steam chambers). Its operation is therefore
more of a variation of cyclic steam-stimulation using
horizontal wells whereby production is not necessarily shut off
when injection is occurring, and vice-versa.
In general, cycling of steam injection rates resulted in strong
oil rate performance during SW-SAGD operations.
The use of well inserts was seen to significantly enhance the
oil rate and recovery performance. Improvements were also
observed when a small quantity of gas was used as steam
additive.
Recommendation. It is recommended that a SW-SAGD mode
of operation be evaluated as an alternative to a steam huff and
puff (CSS) using horizontal wells, as it could provide an
attractive economics under certain conditions described in the
previous section.
Acknowledgements
The authors gratefully acknowledge the financial support of
participants in this Joint Industry Project (1997-99) of the
Petroleum Recovery Institute (PRI). These were: AOSTRA
(Alberta Department of Energy), Alberta Energy Co.,
AMOCO Canada, Canadian Occidental, Chinese Petroleum
Corporation (Taiwan), Fletcher Challenge, Gulf Canada,
Husky Oil Operations, INTEVEP (Venezuela), Japan National
Oil Corporation, Mobil Oil Canada, Petro-Canada, Petroleo
Brasileiro S. A. (Brazil), Petrom R.A. (Romania), Phillips
Petroleum and Ranger Oil Limited. Useful and stimulating
discussions and, encouragement from different industry
participants was the basis of many of the ideas presented here.
Special thanks go to Mike McCormack (Ranger), Mehmet
Saltuklaroglu (Mobil), Brad Ferguson (AMOCO), Don
Anderson (Husky) and Bill Good (AOSTRA). The authors
also acknowledge assistance of Darryl Darwent of PRI in
preparation of this manuscript.
References
1. Huygen, H. H. A. and Black, J. B., Steaming through Horizontal
Wells and Fractures a Scaled Model Test, Second European
Symposium on Enhanced Oil Recovery, Paris, November 1982.
2. Shen, C., Numerical Investigation of SAGD Process Using a

SPE 59333

A MECHANISTIC STUDY OF SINGLE-WELL STEAM ASSISTED GRAVITY DRAINAGE

Single Horizontal Well, SPE 50412, presented at the Third SPE/


PS International Conference on Horizontal Well Technology,
Calgary, Alberta, November 1-4, 1998.
3. Saltuklaroglu, M., Mobils Experience with Single Well SAGD,
presented at the Mini-Forum on Potential and Promise of SWSAGD, organised by the Horizontal Well Special Interest Group
of the Petroleum Society of CIM, March 24, 1999.
4. Nzekwu, B. I., Sametz, P. D. and Palensky, P. J., Single
Horizontal Wellbore Gravity Drainage Steam Flooding Process",
US Patent 5,626,193 dated May 6, 1997.
5. Luft, H. B., Pelensky, P. J. and George, A. E., Development and
Operation of a New Insulated Concentric Coiled Tubing String
for Continuous Steam Injection in Heavy Oil Production, SPE
30322, presented at the SPE International Heavy Oil Symposium,
Calgary, Alberta, June 19-21, 1995.
6. McCormack, M., Fitzgibbon, J. and Horbachewsky, N., Review
of Single Well SAGD Field Operating Experience, presented at
the Sixth SPE/ Petroleum Society One Day Conference on
Horizontal Well Technology, Calgary, Alberta, November 12,
1998.
7. McCormack, M., The Renaissance of Single Horizontal Well
Thermal Processes, presented at the Mini-Forum on Potential
and Promise of SW-SAGD, organised by the Horizontal Well
Special Interest Group of the Petroleum Society of CIM, March
24, 1999.
8. Nasr, T. N., Golbeck, H., Korpany, G. and Pearce, G., SAGD
Operating Strategies, SPE 50411, presented at the Third SPE/ PS
International Conference on Horizontal Well Technology,
Calgary, Alberta, November 1-4, 1998.
9. Oballa, V and Buchanan, W. L., Single Horizontal Well in
Thermal Recovery Processes, SPE 37115, presented at the
Second SPE/ PS International Conference on Horizontal Well
Technology, Calgary, Alberta, November 18-20, 1996.
10. Sawhney, G., Eddy, D. and Peters, E., Pressure Controlled
Gravity Drainage (PCGD): Method of Effective Utilization of
Steam with Horizontal Producers in Heavy Oil Pools, Paper 9797, presented at the 48th Annual Technical Meeting of the
Petroleum Society, Calgary, Alberta, Canada, June 8-11, 1997.

ASHOK K. SINGHAL, SWAPAN DAS, JON GOLDMAN AND ALEXANDRU TURTA

SPE 59333

SPE 59333

A MECHANISTIC STUDY OF SINGLE-WELL STEAM ASSISTED GRAVITY DRAINAGE

10

ASHOK K. SINGHAL, SWAPAN DAS, JON GOLDMAN AND ALEXANDRU TURTA

SPE 59333

SPE 59333

A MECHANISTIC STUDY OF SINGLE-WELL STEAM ASSISTED GRAVITY DRAINAGE

11

12

ASHOK K. SINGHAL, SWAPAN DAS, JON GOLDMAN AND ALEXANDRU TURTA

SPE 59333

SPE 59333

A MECHANISTIC STUDY OF SINGLE-WELL STEAM ASSISTED GRAVITY DRAINAGE

13

14

ASHOK K. SINGHAL, SWAPAN DAS, JON GOLDMAN AND ALEXANDRU TURTA

SPE 59333

SPE 59333

A MECHANISTIC STUDY OF SINGLE-WELL STEAM ASSISTED GRAVITY DRAINAGE

15

16

ASHOK K. SINGHAL, SWAPAN DAS, JON GOLDMAN AND ALEXANDRU TURTA

SPE 59333

SPE 59333

A MECHANISTIC STUDY OF SINGLE-WELL STEAM ASSISTED GRAVITY DRAINAGE

17

18

ASHOK K. SINGHAL, SWAPAN DAS, JON GOLDMAN AND ALEXANDRU TURTA

SPE 59333

SPE 59333

A MECHANISTIC STUDY OF SINGLE-WELL STEAM ASSISTED GRAVITY DRAINAGE

19

20

ASHOK K. SINGHAL, SWAPAN DAS, JON GOLDMAN AND ALEXANDRU TURTA

SPE 59333

SPE 59333

A MECHANISTIC STUDY OF SINGLE-WELL STEAM ASSISTED GRAVITY DRAINAGE

21

22

ASHOK K. SINGHAL, SWAPAN DAS, JON GOLDMAN AND ALEXANDRU TURTA

SPE 59333

SPE 59333

A MECHANISTIC STUDY OF SINGLE-WELL STEAM ASSISTED GRAVITY DRAINAGE

23

24

ASHOK K. SINGHAL, SWAPAN DAS, JON GOLDMAN AND ALEXANDRU TURTA

SPE 59333