You are on page 1of 7

SPE-180394-MS

Life After SAGD – 20 Years Later


S. M. Farouq Ali, University of Calgary

Copyright 2016, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Western Regional Meeting held in Anchorage, Alaska, USA, 23–26 May 2016.

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 have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper 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 SPE copyright.

Abstract
I presented a paper ⬙Is There Life After SAGD?⬙ at the 1996 Western Regional Meeting1. The paper was
published and elicited some 20 discussions. Today, SAGD - Steam-Assisted Gravity Drainage - is a
commercially successful recovery method in Canada, if the price of oil is right. The subject of this paper
is what we have learned in the past 20 years, and what is the future of SAGD in oil (tar) sands and heavy
oil exploitation.
At last count there were 18 SAGD projects in Alberta, and none anywhere. Bitumen production by
SAGD is about 800,000 B/D. There have been failures also - in Alberta, California, Venezuela, etc.
The oil recovery factor can be as high as 60%, depending on how one defines it, and as low as 10-20%.
The steam-oil ratio varies from 1.6 Bbl steam/Bbl oil, to 8 Bbl/Bbl. (Note that these are based on 95%
downhole steam quality and are to be multiplied by ~1.3 to compare with typical steam injection SOR’s).
There are large variations even within the same project. The heat production can be as much as 50% of
that injected. The desired downhole steam quality is 100%. Thus SAGD is a unique recovery method, but
its application requires great caution.
Some of the assertions of 20 years ago still apply. A few new aspects of SAGD have come to the fore.
Many variations of the basic process have been proposed and tested. Many have failed. In some of the less
successful projects, solvents and other additives are being injected with steam to improve performance.
The attractive features of SAGD are described, and application guidelines are offered, based on
failures. SAGD is here to stay, but what are the limits to application? This is the focus of the paper.

Introduction
Steam-Assisted Gravity Drainage, or SAGD, is now a well-established oil recovery method for bitumen
and extremely viscous oils. At the moment there are ten major commercial operations, and many smaller
ones, consisting of about 1500 well pairs, and producing about 800,000 B/D of bitumen. Considering that
30 years ago bitumen production by SAGD was negligible, this is an incredible achievement. In the 1996
Western Regional Meeting, I presented a paper1 with a similar title that opened up much discussion,
published in JCPT over a period of two years. The purpose of the original paper was to point out the major
deficiencies of the theory and practice of SAGD. This paper revisits those concerns, and posits new ones,
in the light of the experience gained in the past 20 years.
2 SPE-180394-MS

Some 200 papers have been published on SAGD since the original article was written. This paper is
not a literature review. A listing of the published papers can be found in Ref. 2. These can be divided into
four groups: 1. A vast number of papers on numerical simulation studies; 2. Many on SAGD modifica-
tions; 3. Many on the use of additives, and 4. A small number on the theoretical aspects of SAGD. In the
following, only the papers relevant to the present discussion will be mentioned.
SAGD Features
SAGD was introduced by Butler, McNab, and Lo in 19793. In the well-known scheme, two horizontal
wells, each ~1 km long are placed on top of each other in the same vertical plane, with a spacing of 5 m
in most projects in the Alberta oil sands (see inset, Fig. 1). Well pairs are spaced at 100 m laterally.
Initially, steam is circulated in the annulus of each well for three months, or so, to heat the oil sand in
between to a temperature of ~90 C. The actual temperature attained has to be estimated, based on
conduction heating, as there is no simple way to measure it over the 1 km distance along the well pair.
The lower well is then put on production, while injecting steam into the upper well. The numbers given
above are for most projects in Alberta. The well spacing of 5 m is taken from the graph Butler4 gave on
the basis of steamflood theory, shown in Fig. 1. The inset shows the well scheme described above.

Figure 1—Relation between the vertical separation of the wells and oil viscosity (After Butler4).

The commercially successful SAGD projects are all in bitumen formations with viscosity around one
million cp. SAGD has been tested in less viscous oils, in particular in California, in the range of a few
1000 cp, without success, as the steam migrated to the top of the formation directly, given the high
mobility of in-place oil, high vertical permeability, and also the well separation of 25-30 m, as given by
Fig. 1 (in some projects this was reduced to the midpoint of the formation thickness).
The SAGD process operates essentially at a single temperature (or pressure) - the steam temperature.
The pressure difference between the wells is kept at a low value, less than 500 kPa, to minimize the chance
of the injected steam short-circuiting into the producer. This implies that only the latent heat component
of the injected steam is utilized to expand the steam chamber. Liquid enthalpy is produced back. Chamber
growth is necessary for oil production. Production rate control is very important: if the rate is too low, the
chamber would fill with condensate, if the rate is too high, steam will be pulled into the producer. The
SPE-180394-MS 3

attempt is to maintain a water layer (steam trap) between the injector and the producer, by maintaining
a temperature in the producer that is a certain amount (called ⬙subcool⬙ - about 10 C) below the steam
saturation temperature at the producer pressure. The complexity of this can be appreciated by the fact that
this condition has to be met over the 1 km length of the well.
Conduction heat transfer was postulated as the only heating mechanism in the original theory. Later
investigators5 added convection and two-phase flow, and demonstrated the importance of the same.
SAGD is driven by gravity only. So it is important to ensure that steam enters the formation at a very
high quality approaching 100%. This implies that the entrained liquid must be knocked out of the injected
steam at the wellhead. Gravity would be most effective if the in situ steam quality is high, volume is large,
and density is low; that would require a low operation pressure - not always feasible.
Current Status
At the moment there are many SAGD operations. The more important ones are listed in Table 1. The
current production rates are lower as a result of shut-downs and steam injection rate reductions in view
of the low oil price. Total production is about 800,000 B/D from approximately 1500 well pairs. Total
steam injection is estimated at 400,000 m3/day (2.4⫻106 B/D CWE).

Table 1—Selected commercial SAGD projects


Company - Field Production, B/D Steam-oil ratio, Bbl/Bbl

EnCana - Foster Creek 118,000 1.7


EnCana - Christina Lake 83,000 2.4
Devon - Jackfish 50,000 2.7
Suncor - MacKay River 25,000 3.0
Suncor - Firebag 175,000 2.7
Nexen - Long Lake 20,000 5-7
Conoco-Phillips - Surmont 23,000 3.0
MEG Energy - Christina Lake 32,000 2.4
Husky - Tucker Lake 9,000 4.0
JACOS - Hangingstone 6,000 3.5
CNRL - Kirby South 25,000 2.8

Unforgiving Geology
Among the many factors - technical and human - determining the profitability of a SAGD project, the
most important is formation geology. Gravity potential gradient - of the order of 10 kPa/m - is small and
is the only driving force. The flow rate depends strongly on the effective vertical permeability to oil In
a layered formation, the effective vertical permeability is close to the lowest permeability. Thus if a zero
permeability shale layer is present, the effective permeability is zero. In SAGD projects, wells are placed
in the zone of highest permeability - as a result, a portion of the resource has to be left unrecovered. Thus,
oil recoveries are often given in terms of the ⬙SAGDable⬙ oil. This applies also in terms of the well
completion in the horizontal plane. Thus caution is needed when comparing the recovery factors of SAGD
with those of CSS (Cyclic Steam Stimulation), where the basis is the entire resource.
Additional geological degradation of a SAGD prospect is caused by the presence of IHS (Inclined
Heterolithic Stratification), bottom/top/middle water, gas cap, and any fractures - natural or induced. The
presence of high water saturation ⬙thief zones⬙ leads to the distortion of a steam chamber, and a loss of
oil recovery. Top water is particularly undesirable - communication of the steam chamber with top water
can lead to a sharp increase in SOR. The large variations in SOR in Table 1 are primarily due to geological
variations. This is shown more vividly by Fig. 2, where plots of SOR are seen for the operating projects.
4 SPE-180394-MS

Figure 2—Steam-oil ratio variation in SAGD projects (Oilsands Review, May 2013).

The original UTF (Underground Test Facility), where the wells were drilled from tunnels below the oil
sand (200 m deep) was a success because the operating pressure of about 2200-2400 kPa minimized the
effect of thief zones in the oil sand. In contrast, the operating pressure in Peace River was of the order of
8000 kPa. More on that later.
Theory - Limitations and Extensions
SAGD theory is based on the well-known equation reported in the original reference3, which is as follows:

where qo is oil production rate in m3/s per m well length, ko is effective oil permeability in m2, g is
9.80665 m/s2, ␣ is thermal diffusivity, in m2/s, ␸ is porosity, as fraction, ⌬So is initial oil saturation minus
the steamflood residual oil saturation, as fraction, h is the fully developed chamber thickness, m is a
constant based on the viscosity equation, and vs is oil viscosity at steam pressure, in m2/s. This steady state
equation was derived on the basis of a heat front at steam temperature advancing into bitumen, that is
mobilized at the front and flows by gravity to the formation base. The flow lines derived from the above
equation move farther and farther away from the lower producing well. Besides, the equation gave much
higher rates than those observed in the field. A correction was made by drawing tangents to the flow lines
through the centre of the producer, resulting in a multiplier of 1.5 in place of 2 under the square root sign.
The rates calculated were still too high, so in the next modification the tangents were extended to the top
of the formation, resulting in a multiplier of 1.3 in place of 2 under the square root sign. Details are in
Ref. 4.
The main problem with the above equation is that it has nothing to do with the steam chamber
formation and growth. Thus, according to the equation, if oil viscosity is reduced by increasing steam
temperature, oil rate will increase without limit. However, as the temperature (viz. pressure) increases, the
latent heat content of steam decreases, and the steam chamber size decreases. In the original theory, the
steam injection rate and the steam-oil ratio are calculated after calculating the oil rate. Butler4 gives a
methodology for calculating the oil rate with time, partly based on experimental observations.
Zargar and Farouq Ali6 developed two unsteady state analytical models which calculate the oil rate
coupled with the steam chamber. One of these considers a constant steam injection rate, and the other one
SPE-180394-MS 5

a constant volumetric rate. In the second model the steam injection rate has to be increased with time to
obtain a constant oil production rate. Figure 3(a) shows the oil rate with time for two different pressures.
As expected, the rate decreases with increasing operating pressure, all other variables remaining un-
changed. Figure 3(b) shows the cumulative SOR with time - it is seen that the SOR is much higher at the
higher operating pressure. The original theory gives an oil rate that increases with an increase in pressure.

Figure 3(a)—Oil production rate at two different operating pressures.

Figure 3(b)—Cumulative steam-oil ratio at different operating pressures.

Mutations of the Original Process for Performance Improvement, and for


Making It Work
Many modifications of the original SAGD process have been proposed for various reasons. One of these
is the high energy consumption of the process, especially in a mature project when the oil rate is declining.
Butler4 proposed gas injection with steam to maintain chamber pressure, with a reduced steam injection
rate. A secondary objective was to ⬙insulate⬙ the steam chamber ceiling for reducing heat loss. In the field,
this process has had mixed results, as much of the gas tends to accumulate in the highest temperature
region, on the sides of the chamber, thus reducing bitumen relative permeability, and hence the flow rate.
6 SPE-180394-MS

A variation of SAGD using solvents only was proposed by Butler in 1991, called VAPEX. The well
scheme was the same as in SAGD, however no steam is injected. Solvent diffusion-dispersion into
bitumen is a slow process. Two well pairs were used for a VAPEX field test sponsored by several industry
partners. In one case, the process was started with steam, and then converted to solvent only.
In the other case, solvent was used from the beginning. Both tests were failures, with bitumen
production rates of 1-10 m3/day compared to rates up to 170 m3/day predicted by simulations.
Many variations of SAGD using solvents have been proposed, discussed in detail by Bayestehparvin
and Farouq Ali7. The objectives are (1) to improve the performance of an already successful process, and
as a result, reduce the steam injection rate, increase well pair spacing, thus increasing the bitumen in place
per pair, etc.; (2) make SAGD successful in a heterogeneous reservoir, hoping that the solvent would help
spread steam better; and (3) in general, make SAGD applicable to unsuitable reservoirs. Some of the
processes have been successful in increasing the bitumen production rate, but no data on the cost- benefit
analysis of the solvents has been published. The choice of the solvent is based in one case, on the
proximity of the steam saturation and the solvent saturation temperatures, in another, on the extent of
instability caused by the solvent, leading to increased mixing with bitumen. In general, the mass transfer
of solvent to bitumen is a slow process, governed by diffusion and mechanical mixing. Bayestehparvin
and Farouq Ali8 showed that bitumen mobilization by a solvent is at least 3 times slower than by heat.
Another variation of SAGD, with the use of solvents combined with electrical heating is purported to
reduce greenhouse gas (GHG) emissions as well as water usage, so that the need for steam may be
eliminated. There are a number of problems with these concepts. Field success is yet to be demonstrated;
besides both the manufacture of solvents and production of electricity entail GHG emissions in excess of
those produced by steam generation. As for water, currently over 90% of the produced water is recycled.

Looking Ahead
SAGD will continue to make an impact on bitumen recovery as it is only one of the two recovery methods
(the other one is CSS, operated at parting pressures), which have been successful in the oil sands. In the
current (March, 2016) low oil price (bitumen price is $10-15 lower than WTI oil price) environment,
SAGD is not economic in many projects, even though the fuel cost is at an all time low, with the steam
costing less than $1 per barrel. Many of the SAGD variations currently being tested would be abandoned
for various reasons, one of which is the limited availability of suitable solvents at a low price. The field
results have made it quite clear that in some reservoirs, SAGD is the wrong choice. Use of solvents,
additives, foam, etc., being currently tested, will not improve the process, because the success depends
primarily on geology. However, other variations will be developed, one of which is doing in situ
combustion in a mature SAGD chamber, which makes little technical or economic sense.
In some projects, steam injection rate has been reduced, or even shut down, because of the current
economic situation. The author showed in a recent publication that shutting down a SAGD operation leads
to a permanent loss of oil recovery.

References
1. Farouq Ali, S.M.: ⬙Is There Life After SAGD?⬙ JCPT, June 1997.
2. Al Bahlani, A.M., and Babadagli, T.: ⬙A Critical Review of the Status of SAGD: Where Are We and What Is Next?⬙.
SPE 113283, presented at SPE Western Regional and Pacific Section AAPG Joint Meeting, 29 March-4 April, 2008,
Bakersfield.
3. Butler, R.M., McNab, G., and Lo, H.: ⬙Steam-Assisted Gravity Drainage,⬙ presented at the Oil Sands Conference,
Jasper, 1979.
4. Butler, R.M.: Thermal Recovery of Heavy Oil and Bitumen, GravDrain, Calgary (1988).
5. Sharma, Jyotsna, and Gates, I.D. ⬙Convection at the Edge of a Steam-Assisted-Gravity-Drainage Steam Chamber,⬙
SPE 142432, SPEJ, Sep 2011.
SPE-180394-MS 7

6. Zargar, Zeinab, and Farouq Ali, S.M.: ⬙Analytical Treatment of SAGD - Old and New⬙, SPE 180748, presented at the
SPE Canada Heavy Oil Technical Conference, 7-8 June, 2016.
7. Bayestehparvin, Bita, Farouq Ali, S.M., and Abedi, J.: ⬙Dissolution and Mobilization of Bitumen at Pore Scale,⬙
SPEl74482, preseented at SPE Canada Heavy Oil Technical Conference, 9-11 June, Calgary.
8. Bayestehparvin, Bita, Abedi, J., and Farouq Ali, S.M.: ⬙Use of Solvents With Steam - State-of-the-Art and
Limitations,⬙ SPE 179829, SPE EOR Conference at Oil and Gas West Asia, 21-23 March, Muscat.

You might also like