Fuel 115 (2014) 706–713

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Energy efficiency and emissions intensity of SAGD

Ian D. Gates a,⇑, Stephen R. Larter b
a
Department of Chemical and Petroleum Engineering, University of Calgary, Canada
b
PRG, Department of Geosciences, University of Calgary, Canada

h i g h l i g h t s

 Steam-Assisted Gravity Drainage is process of choice for extra heavy oil recovery.
 Average recoveries from heavy oil and oil sand reservoirs are typically low.
 Some SAGD operations are actually not net energy generating.
 New oil sands processes needed to improve energy and carbon dioxide intensities.

a r t i c l e i n f o a b s t r a c t

Article history: Currently, to mobilize and produce bitumen from Athabasca oil sands reservoirs, Steam-Assisted Gravity
Received 5 February 2013 Drainage (SAGD) is the method of choice. SAGD, requires large amounts of energy and emits significant
Received in revised form 18 July 2013 volumes of greenhouse gases to the environment. Here, we discuss the thermal efficiencies, energy bal-
Accepted 18 July 2013
ances, and emissions of SAGD. While the world’s heavy oil and oil sand resource is large, average recov-
Available online 9 August 2013
eries from heavy oil and oil sand reservoirs are typically low, ranging from 5% to 15% for cold heavy oil
production and from 25% to 60% percent for steam-based in situ processes. This is for two reasons: firstly,
Keywords:
geological heterogeneity and secondly, ubiquitous large scale fluid property heterogeneities are common
Steam-Assisted Gravity Drainage
SAGD
on a range of spatial scales. Thus, there is a strong motivation to develop better recovery processes with
Bitumen lower energy and emission intensities. The thermal efficiencies, energy balances, and emissions of SAGD
Thermal process show a very wide range of field performance for the current thermal recovery projects in Alberta, with
Energetics earlier pilots being more successful. The data suggests that at the extreme, some operations are actually
not net energy generating with injected energy via steam, exceeding recovered chemical energy in recov-
ered oil. Differential pricing of oil and natural gas, the main steam generating fuel, still permits these
extreme cases to be economically profitable due to low natural gas prices. In all cases, carbon dioxide
intensity is high.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction [5]. For heavy oil (between 10° and 20°API), the dead oil viscosity
can range up to the thousands, or tens of thousands of cP. For bitu-
With global estimates of over 6 trillion barrels, the majority of men (less than 10°API), the viscosity typically ranges from the tens
the world’s petroleum is stored in the form of biodegraded heavy of thousands to over 10 million cP at reservoir conditions. One use-
oil and oil sand (bitumen) reservoirs [1,2]. In Western Canada ful property of heavy oil and bitumen is that its viscosity drops by
alone, there are over 1.7 trillion barrels of bitumen trapped in oil several orders of magnitudes when the oil temperature is raised
sand reservoirs [3]. This volume in place is second only to the con- from original reservoir conditions, typically between 5 and 15 °C,
ventional oil reserves of Saudi Arabia [4]. With current surface to steam temperatures, typically over 180 °C. For example, for a
mining and in situ recovery technology, it is estimated that roughly typical Athabasca bitumen, the dead oil viscosity at 100 °C is equal
10% of oil sands resource in Western Canada is recoverable. The to roughly 220 cP [6]. For a successful in situ oil sands bitumen
key challenge for producing heavy oil and bitumen is complex, recovery process, two requirements must be met: first, it is neces-
compartmentalized reservoir geology and its high and variable in sary to raise the oil mobility (often done by lowering its viscosity
situ viscosity related to variable levels of crude oil biodegradation which results from raising its temperature) until it can be moved
by natural forces such as gravity, and secondly, it is necessary to
⇑ Corresponding author. Tel.: +1 (403) 220 5752; fax: +1 (403) 284 4852. move the mobilized oil to a production wellbore so it can be pro-
E-mail address: [email protected] (I.D. Gates). duced to the surface.

0016-2361/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.fuel.2013.07.073
 I.D. Gates, S.R. Larter / Fuel 115 (2014) 706–713 707

2. In situ bitumen recovery processes 8. The maintenance of an effective density difference between flu-
ids in the depletion chamber (vapor and steam condensate) and
Currently, commercial steam-based in situ processes used to re- fluids in the oil sands (bitumen and water) at the chamber edge
cover bitumen from oil sands reservoirs are either one of Cyclic to enable gravity drainage.
Steam Stimulation (CSS), or Steam-Assisted Gravity Drainage
(SAGD) [7]. In this work, we will focus on SAGD although the anal- If any one or more of these factors are adverse or the process
ysis and results, conceptually, also apply to CSS. CSS is a cyclic high controls fail e.g. the steam chamber is not charged with steam as
pressure steam injection process that uses a single well where the there is a complete loss of the injected steam to a thief zone (such
formation is fractured by a steam injection pulse followed by a per- as a water saturated reservoir zone) rather than the growing steam
iod of soaking and then oil and water co-production [7]. SAGD, dis- chamber, then SAGD productivity and thermal efficiency suffers
played in cross-section in Fig. 1, is a continuous steam injection and the process may be rendered in viable and not commercially
process that uses two horizontal wells. The top well injects steam, successful. Given that the driving force in SAGD is gravity drainage
at near reservoir pressure, into a depletion chamber that forms in of heated oil and condensed water, process productivity is directly
the oil sands reservoir. The steam releases its remaining latent heat tied to the underlying fluid mobility field in the reservoir con-
through condensation at the edge of the chamber, heating the oil trolled by geological and fluid compositional heterogeneity [8].
sands (oil, sand, and water) there. The viscosity of the heated oil The oil sand reservoirs are geologically very variable and complex
drops and under the action of gravity drains through the steam reservoirs with many interbedded sandstones and shale [9,10] that
chamber, or down its edges, to the production well at the base of together generate a complex variably permeable medium that
the chamber [7]. The key controls on the productivity and effi- steam, oil and water must migrate through. Actual oil sands reser-
ciency of the SAGD process are: voirs are completely different from the homogeneous sandstones
with uniform fluids envisaged by the reservoir engineers that
1. The rate of reduction of oil viscosity with increased temperature developed the early SAGD process [7]. Geological heterogeneity
i.e. the extent of viscosity reduction given an amount of latent impacts the recovery process through permeability changes of
heat added; the reservoir sandstones within the oil column and the shale or
2. The rate of heat transfer at the edge of the steam chamber (set mudstone barriers and baffles that prevent or retard fluid flow,
by the thermal conductivity of the oil sand and any convective respectively [10,11]. The more laterally extensive the barrier, the
contributions due to steam condensate or hot oil flow into the longer it takes steam or production fluids to go around it and the
oil sand); longer it takes for mobilized oil to get to the production well. Also,
3. The relative amount of injected steam heat lost to the cap rock non-productive reservoir within the oil column represents a heat
or thief zones e.g. zones of mobile gas or water, compared to sink which erodes the thermal efficiency of the process. The main
delivering heat to the bitumen; impact of fluid compositional heterogeneity is the due to effect of
4. Any geological barriers such as shales or siltstones in the reser- vertically and laterally varying oil phase viscosity.
voir that would interfere with steam rise and gravity drainage The oil phase viscosity at the bottom of the oil column, due to
of oil to the production well; biodegradation occurring over geological time scales, can be sev-
5. Natural oil compositional variations that yield relatively low oil eral times to several tens of times the oil phase viscosity at the
viscosities at the top of the oil column and relatively high vis- top of the oil column over a scale of as little as a few tens of meters
cosities at the bottom of the oil column with often, lateral vis- [5]. These viscosity variations also occur horizontally with 3–5
cosity gradients which perturb uniform steam chamber times variations along lateral length scales of about 1000 m [8].
development and can slow the production of oil; Furthermore, these spatial oil phase viscosity variations persist,
6. The relative permeability of oil in the reservoir at the edge of to a lesser degree, at the steam temperatures used in SAGD. During
the steam chamber and its effect on oil phase mobility there; SAGD startup, the variability of the oil phase mobility is thus often
7. The ability of the recovery processes operator to prevent live largely controlled by oil viscosity and water saturation rather than
steam production by maintaining a liquid barrier between the by variations in the absolute reservoir permeability. Further, at the
injection and production wells to enforce steam trap control, edge of the steam chamber, the majority of the flowing oil is at
and; temperatures below the steam temperature (typically about 75%
of the steam temperature) and at this temperature the oil phase
viscosity can vary four to six times over the height of the oil col-
umn which is of similar order of magnitude to the variations in
absolute permeability [7]. Thus, during SAGD production, both per-
meability and oil viscosity variations are important and adversely
impact theoretical SAGD productivity.
Native
A key limitation on the SAGD process is set by deployment ap-
bitumen
Steam Chamber proaches. During large scale deployment of SAGD, to scale produc-
tion, rather than tailor individual well settings to highly variable
local geological realities, large scale SAGD operations are com-
monly deployed in a rather uniform cookie cutter style with uni-
form process characteristics. Thus, despite very variable local
Injection Well reservoir and fluid property combinations, SAGD injection and pro-
duction wells are placed at the base of the oil column, typically 2 m
above the base of the oil zone for the production well with the
injection well 5 m above that. The lowest viscosity oils are typically
Bitumen in about the upper portion of the oil column and thus placing the
flow zone Production Well
wells higher in the formation than directly at the bottom will en-
hance production rates in real reservoirs, not only at start-up but
Fig. 1. Cross-section schematic of the Steam-Assisted Gravity Drainage (SAGD) also beyond [8], but practice is to place essentially all wells at
process. the base of the oil column as this is where the theoretically
708 I.D. Gates, S.R. Larter / Fuel 115 (2014) 706–713

maximum oil recovery would be attained in an ideal SAGD process from the returned hot water is often used to pre-heat boiler feed
working in an ideal reservoir. water via heat exchange prior to its entry into the steam generator.
For SAGD, average oil recovery factors are between 40% and 60% This produced energy must be accounted for in the energy balance
[7]. Although optimized SAGD can, in the best reservoirs, yield rea- as follows:
sonably high recovery factors, economic and environmental costs
can be large given the amount of steam required to extract oil from Q produced ¼ 0:90b/So qo cpo ðT s  40  T r Þ þ /Sw qw cpw ðT s  40  T r ÞcBV
the reservoir. In this paper, the energy efficiency and emissions ð4Þ
intensity of the SAGD process is assessed both theoretically and
as deployed at scale, from publically field data available from the and used in the tEOR and tSOR calculations. In Eq. (4), the produced
ERCB in Alberta. fluids have been taken to be at a temperature equal to 40 °C below
that of the injected steam.
Fig. 2a displays the tEOR versus steam injection pressure (set by
3. Theoretical energy efficiency for an oil sands recovery process
the steam saturation temperature, Ts) and Fig. 2b shows the tSOR
versus steam injection pressure both for 60% and 100% steam qual-
Based on established bitumen heat capacities [7], the theoreti-
ity taking energy from the produced fluids into account as de-
cal minimum amount of energy required to heat Athabasca bitu-
scribed by Eq. (4). The results show that the tEOR and tSOR rise
men from reservoir temperature (10 °C-viscosity equal to
as the steam pressure and steam temperature increases and that
millions of cP) to steam temperature (200 °C and a dead oil viscos-
under ideal heating conditions, for the majority of SAGD operations
ity of 10 cP) is equal to about 1.75 GJ/m3 oil. This is simply the sen-
which operate in the range of from about 1000–5000 kPa steam
sible heat required to raise the temperature of a cubic meter of
injection pressure, the tEOR lies between about 1.0 and 2.5 GJ/m3
bitumen from 10 to 200 °C. For heating oil within the reservoir,
for 60% steam quality delivered at the edge of the chamber. Over
not only is the oil volume heated but also any trapped water and
the same pressure range, the tSOR ranges from about 0.38–
the mass dominant solid mineral phases in the rock volume con-
0.9 m3 CWE steam/m3, for 60% steam quality at the chamber edge.
taining the oil. Thus, ideally, the amount of energy required to raise
The tSOR curve gives the theoretical minimum steam-to-oil ratio if
the temperature of the oil also includes the minimum sensible en-
the steam’s latent heat was ideally transferred to the oil sand with
ergy, Q, required to heat the water and rock of the oil sand to the
produced energy as given by Eq. (4) which implies that average
temperature at whichoil will flow:
steam-to-oil ratios below this value are not possible unless some
Q ¼ b/So qo cpo ðT s  T r Þ þ /Sw qw cpw ðT s  T r Þ þ ð1  /Þqr cpr ðT s  T r ÞcBV other energy or material vector such as solvent reduces the viscos-
ð1Þ ity of the bitumen. Given the difficulty of distributing steam
uniformly within an oil sands reservoir due to its geological and
where BV is the bulk volume of the reservoir rock, / is the porosity,
Sx is the fluid saturation of phase x, x and cpx are the average density
and specific heat capacity of phase x over the temperature range (a) 3
from Ts to Tr, and the subscripts o, w, and r refer to the oil, water,
and rock phases, respectively. By re-arranging Eq. (1), the amount 2.5
of energy required to raise the temperature from Tr to Ts per unit
oil volume, tEOR, becomes: 2
tEOR, GJ/m3

 
Q Sw 1/
tEOR ¼ ¼ qo cpo þ qw cpw þ qr C pr ðT s  T r Þ ð2Þ 1.5
/So BV So /So
The energy required can be converted to an equivalent amount of 1
steam, expressed as cold water equivalents (CWE), at steam quality
xs. This gives the theoretical minimum possible steam-to-oil ratio,
0.5
tSOR:
 
ms S 1/
tSOR ¼ qo cpo þ w qw cpw þ qr cpr ðT s  T r Þ ð3Þ 0
0 1000 2000 3000 4000 5000 6000 7000
hL þ xs km So /o
Steam Pressure, kPa
where hL and kv are the liquid water phase enthalpy and steam la-
tent heat, respectively. To express the steam as a cold water equiv-
(b) 1
alent, vs is set equal to the specific volume of water at standard 0.9
conditions (0.001 m3/kg). The density, heat capacities, liquid enthal-
py, and latent heat of water depend on the temperature. The steam 0.8
quality, xs, is that at the edge of the steam chamber. From typical 0.7
tSOR, m3/m3

steam quality gradients within the saturated steam chamber, at 0.6

the edge of the chamber, the steam quality ranges between 0.1
0.5
and 0.6 [12]. Eq. (3) gives the ideal amount of steam, at temperature
Ts, required to mobilize the oil to temperatureTs. That is, in other 0.4
words, given the porosity and fluid saturations, the tSOR is the min- 0.3
imum achievable, to mobilize the oil to its flow viscosity at temper-
0.2
ature Ts. In steam-based recovery processes, the produced fluids are
typically about 40 °C below the injection temperature and typically 0.1
about 90% on average of the water injected is produced and if oil is 0
0 1000 2000 3000 4000 5000 6000 7000
mobilized by steam heating, typically about 90% of the oil is pro-
Steam Pressure, kPa
duced from the reservoir zone where oil can flow (the low recovery
factors of steam-based recovery processes arise from regions of the Fig. 2. (a) Theoretical energy-to-oil ratio, tEOR, and (b) steam-to-oil ratio, tSOR,
reservoir that are not contacted by steam). This produced energy versus steam chamber pressure for SAGD.
 I.D. Gates, S.R. Larter / Fuel 115 (2014) 706–713 709

fluid compositional heterogeneity, the actual steam-to-oil ratio of a for methane combustion (0.000802 GJ/mol). Fig. 3 shows the CO2
block of oil sand would be some value greater than the tSOR. Also, emitted per unit bitumen recovered, versus steam pressure under
since the oil saturation and porosity varies in the reservoir, the ideal heat transfer to the oil sand. The results show that the amount
tSOR will vary in the reservoir. The lower the porosity, the higher of carbon dioxide emitted per unit volume oil produced is large. For
the tSOR since the rock volume increases and oil volume drops example, for a steam recovery process operating at about 3000 kPa,
per unit bulk volume. Similarly, the lower the oil saturation, the under ideal heating conditions, just over 0.2 tonnes of carbon diox-
higher the tSOR. Given the variability of bitumen composition in ide are emitted per cubic meter of bitumen recovered for 0.6 steam
oil sands reservoirs, there will also be some small variability of quality provided to the edge of the chamber. In reality, boiler effi-
the tSOR curve since each composition implies a different heat ciencies will be less than 100%.
capacity but this will be a minor secondary factor. The tSOR is a
spatially-distributed variable parameter within an oil sands 4. Field data analysis for actual SAGD operations
reservoir.
The theoretical steam-to-oil ratio tSOR, derived above, includes Fig. 4 displays cumulative steam-to-oil ratios versus time for all
the heat losses to the formation water and mineral grains of the major commercial SAGD operations in Alberta reported by project/
bulk reservoir rock volume which also holds the bitumen. There field. The data was obtained from public databases [13] and con-
are also unavoidable heat losses associated with SAGD, including sists of over 200 SAGD well-pairs. The results show that although
surface line and wellbore heat losses, losses of heat to formation there has been a reduction in the cSOR with time, the cSOR is lev-
water and sand grains, and losses to the overburden and under- eling off for most operations, at values above 2 m3/m3. Many of the
strata. This implies that the actual steam-to-oil ratio achieved in operations display a decline of cSOR caused by oil production vol-
the field will always be larger than the tSOR. Another factor that umes increasing after the initial circulation stage of SAGD. How-
leads to higher steam-to-oil ratio than the tSOR is the oil recovery ever, a projection of these cSOR declines with time shows that
factor which is substantially controlled by geological heterogene- the cSORs will still persist at levels above 2 m3/m3. The results
ity. Even if the oil in a volume of reservoir rock is heated so that show that the range of cSOR values narrows with time and that
it is mobile, some of it will not gravity drain and move from the it appears to be converging to within the range of 2–4 m3/m3 cSOR,
reservoir rock due to high capillary entry pressures at fine grained with the largest variability in cSOR typically occurring in the first
lithology boundaries with, in extreme cases, local shales acting as four project years. Some of the project cSOR profiles increase
baffles, barriers and underseals to oil and water flow. From Eq. suddenly and then decline. This is the result of additional SAGD
(3), the thermal efficiency of the steam-based field operation is well-pairs coming on production through a project. The variation
then given by: in energy content of the steam used among the different opera-
tSOR tions is not assessed here as the steam will have different enthal-
gTH ¼ ð5Þ pies at the different injection pressures of each SAGD project.
cSOR
Recalling the cSOR results of the ideal SAGD heating case shown
where the cSOR is the cumulative steam-to-oil ratio of the field in Fig. 2b with cSOR values below 1 m3/m3, the majority of the real
operation. Since the cSOR from field operations does not include field well-pair cSORs are many times higher than that of the ideal
the energy recovered from produced fluids, this is not included in case.
the tSOR calculation in Eq. (5). Figs. 5–10 display cSOR values versus average steam injection
Once the steam requirements are known, as described by the rate for all SAGD well-pairs in Alberta after 1, 2, . . ., 6 years of oper-
tSOR, then the theoretical amount of carbon dioxide emitted per ation, respectively. As shown in Fig. 5, after 1 year of operation, the
unit volume recovered bitumen, tCOR, can be determined by: range of project cSOR is large, being from about 1.64 to over 40 m3/
tSORqCWE ðhL þ xs km Þsg m3 with an average value after year 1 equal to about 34.9 m3/m3.
tCOR ¼ MW CO2 ð6Þ The large cSOR values are due to steam circulation that occurs
DHr
for several months to warm the region between the injection and
where qCWE is the density of water (1000 kg/m3), the subscript ‘sg’ production wells prior to oil production. The data reveals that up
denotes steam conditions in the steam generator, MW CO2 , is the to steam injection rates equal to about 400 m3/day, variability in
molecular mass of carbon dioxide, and DHr is the heat of reaction cSORs is largest. Beyond about 400 m3/day, the variability appears
reduced. The field data also reveals that in some cases it is possible

250

16
200
14
tCOR, kg CO2/m3

12
150
cSOR, m3/m3

10

100 8

6
50
4

2
0
0 1000 2000 3000 4000 5000 6000 7000
0
Steam Pressure, kPa 0 12 24 36 48 60 72 84 96 108 120 132 144
Time, Months
Fig. 3. Theoretical emitted carbon dioxide-to-bitumen ratio, tCOR, at chamber edge
for SAGD with a steam quality equal to 0.6 and 1 (ideal) versus steam chamber Fig. 4. Cumulative steam-to-oil ratio, cSOR, versus time for all SAGD well-pairs
pressure. The efficiency of the steam generator is equal to 0.75. operating in Alberta, Canada. Values over 16 m3/m3 are not plotted.
710 I.D. Gates, S.R. Larter / Fuel 115 (2014) 706–713

50 10
45 9
40 8
35 7
cSOR, m3/m3

cSOR, m3/m3
30 6
25 5
20 4
15 3
10 2
5 1
0 0
0 100 200 300 400 500 600 700 800 900 1000 0 100 200 300 400 500 600 700 800 900 1000
Steam Rate, CWE m3/day Steam Rate, CWE m3/day
Fig. 5. cSOR of SAGD well-pairs after 12 months of operation. Well-pairs with cSOR Fig. 7. cSOR of SAGD well-pairs after 36 months of operation.
greater than 50 m3/m3 are not plotted.

50 10
45 9
40 8
35 7
cSOR, m3/m3

cSOR, m3/m3
30 6
25 5
20 4
15 3
10 2
5 1
0 0
0 100 200 300 400 500 600 700 800 900 1000 0 100 200 300 400 500 600 700 800 900 1000
Steam Rate, CWE m3/day Steam Rate, CWE m3/day

Fig. 6. cSOR of SAGD well-pairs after 24 months of operation. Fig. 8. cSOR of SAGD well-pairs after 48 months of operation.

to obtain relatively low cSORs across the entire range of steam 10

injection rates. As displayed in Fig. 6, after two years of operation,
9
the cSORs have dropped significantly with averages equal to about
6.05 m3/m3 and a range between 1.53 and 37.2 m3/m3. The cSOR is 8
dropping now since the SAGD steam chambers are maturing and 7
becoming more productive per unit volume of steam injected.
cSOR, m3/m3

6
More of the data is clustered at cSORs under 5 m3/m3. After three
years of operation, as shown in Fig. 7, the cSORs have fallen further, 5
with a range between 1.82 and 8.47 m3/m3 with an average equal 4
to about 3.27 m3/m3. The minimum cSOR has risen since the sec-
3
ond year which suggests that the low values observed in the sec-
ond year may be a transient effect in many cases resulting from 2
oil stimulated during start-up steaming. As shown in Fig. 8, after 1
four years of operation, the cSORs now range from 1.65 to 0
7.83 m3/m3 with an average equal to about 3.12 m3/m3. Several 0 100 200 300 400 500 600 700 800 900 1000
of the well-pairs are below 2 m3/m3 and are demonstrating good Steam Rate, CWE m3/day
performance. The variability of the data has reduced most likely
Fig. 9. cSOR of SAGD well-pairs after 60 months of operation.
because the steam chambers are now mature and well-established
within the reservoir. After 5 years of operation, as plotted in Fig. 9,
the average cSOR is equal to about 3.13 m3/m3 with range from
1.56 to 8.39 m3/m3. The average cSOR is starting to level off above 2.5 m3/m3. The data at 6 years is similar to that at 5 years
although the range is not narrowing. The variability is even further suggesting that the performance of the well-pairs is leveling off.
reduced and it appears that the lowest cSORs result at low steam In general the best performing well pairs are from regions with
injection rate. As shown in Fig. 10, after 6 years of operation, the better quality reservoir which have thick, highly oil saturated accu-
range of cSOR is between 1.66 and 9.14 m3/m3 with average value mulations with few shale barriers and high vertical permeability
equal to about 3.21 m3/m3. For those well-pairs under 2 m3/m3, throughout the reservoir. Operator experience is clearly an
their performance is good however the majority of well-pairs are important factor but reservoir geology is king!
 I.D. Gates, S.R. Larter / Fuel 115 (2014) 706–713 711

10 values do not include the additional energy required for field oper-
9 ations e.g. electricity for water and oil pumping, water treatment
and separation or fluid storage and pipelining/transportation.
8
The heating value of bitumen is equal to about 43 GJ/m3 [14].
7 This implies that at tSOR = 0.7 m3/m3, the chemical energy output
cSOR, m3/m3

6 within the bitumen from combustion for a steam-based process is

roughly about 21 times the energy supplied per unit volume bitu-
5
men produced under ideal heating conditions. The solid curve plot-
4 ted in Fig. 12 displays a plot of cSOR versus the ratio of energy
3 produced, in the form of chemical energy contained in the bitumen
2
if combusted, to the energy injected in the form of steam taking
into account 75% efficient steam generation. The results show that
1
the larger the cSOR, the lower the net energy return and figure also
0 shows that the decline in energy return is very rapid at lower cSOR
0 100 200 300 400 500 600 700 800 900 1000
values. The breakeven point, where the energy output equals the
Steam Rate, CWE m3/day
energy input in the process, is at cSOR values equal to around
Fig. 10. cSOR of SAGD well-pairs after 72 months of operation. 11.5 m3/m3. This means that above this cSOR, the SAGD process
is a net energy consumer and thus is not an energy generation pro-
cess. The economic viability of the process under current fiscal re-
gimes is less related to cSOR and the energy balance, as in
situations where natural gas is used to generate steam the price
differential between low gas and high oil prices means that eco-
nomic return does not depend on an energy efficient recovery pro-
cess. The majority of extracted bitumen is upgraded to synthetic
crude oil (SCO), which in turn is refined into transportation fuels.
On average, a further 6 GJ per m3 of bitumen processed is con-
sumed to upgrade the bitumen to synthetic crude oil [15]. This va-
lue depends on whether a coking (about 5–6 GJ/m3) or ebullated
bed (6–7 GJ/m3) upgrader configuration is used. SCO has on aver-
age a heating value similar to crude oil which is equal to about
39 GJ/m3 [14,15]. However, during upgrading, approximately 25%
of the bitumen volume is lost [15,16]. This implies that the heating
value content per unit volume bitumen recovered from the reser-
voir is equal to about 31 GJ/m3. For refining and conversion of
SCO to transportation fuels (average heating value equal to about
46 GJ/m3), the average energy intensity for refining SCO is equal
Fig. 11. Thermal efficiency, defined by Eq. (5), of field SAGD operations.
to about 2.7 GJ/m3 SCO [15]. Transportation fuels, mainly consist-
ing of gasoline, diesel, and jet fuel, have an average heating value
Fig. 11 displays the thermal efficiency, as calculated by Eq. (4), equal to about 46 GJ/m3 [15]. Given that typical volume losses
for all SAGD well-pairs plotted versus steam injection pressure. for refining crude oil are typically between 5% and 8% (we assume
From a heat engine point of view, if the reservoir temperature is 5% here), on a bitumen volumetric basis, the heating value of the
about 10 °C, for steam temperatures ranging from 180 to 250 °C, transportation fuel is equal to about 35 GJ/m3 bitumen recovered
the thermal efficiency will remain limited to values less than 38% from the reservoir. Also plotted in Fig. 12 (dashed curve) is the
and 46%, respectively. The field SAGD data reveal that the majority net energy return for in situ recovery with both upgrading and
of well-pairs are operating with a thermal efficiency less than 40%. refining taken into account. The results reveal that the overall
For those few well-pairs operating above this limit, this is probably
an artifact of interaction between proximal well pairs and implies
that some well-pairs are not operating independently and thus one 24
may appear highly efficient whereas its neighbor does not. Well- 22
pairs operating between 30% and 40% efficiency are probably oper-
Chemical Energy Out / Steam

20
ating as thermally efficiently as can be expected, given the factors 18
that detract from ideal SAGD behavior as listed above. Real SAGD is 16
Energy In, GJ/GJ

about 30% as efficient as ideal SAGD with many well pairs less effi- 14
cient than this!
12
10 Energy Return for In Situ Recovery Only

5. Energy analysis 8
6 Energy Return for In Situ Recovery
Plus Upgrading Plus Refining
Eq. (2) gives the theoretical energy required to heat a volume of 4
oil sand reservoir containing a unit volume of bitumen to raise the 2
temperature from Tr to Ts. At a steam injection pressure equal to 0
2.5 MPa (the approximate average value used in many Alberta 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
SAGD operations), the tSOR is equal to 0.71 m3/m3 and the amount cSOR, m3/m3
of energy required is equal to about 2 GJ/m3 oil produced. At cSORs Fig. 12. Energy out in form of chemical energy of product per unit energy generated
equal to 3.5 and 4.5 m3/m3, the amount of energy required is equal in the form of injected steam for steam-based recovery and case taking upgrading
to about 9.5 and 12.2 GJ/m3 produced oil, respectively. These and refining into account.
712 I.D. Gates, S.R. Larter / Fuel 115 (2014) 706–713

net energy breakeven point is now equal to a production cSOR of 10000

CO2 Per Net Energy Produced, g/MJ

about 6.5 m3/m3. Based on the cSOR field data displayed in Fig. 4, CO2 Emitted for In Situ Recovery
Plus Upgrading Plus Refining
many operations exceed this value and thus are not net energy
generation processes yet may be ‘‘economic’’! With disconnected 1000
price markets for natural gas and bitumen, it is possible for bitu-
men recovery under these conditions to be economically viable to-
day even though it makes no sense to pursue such an energy
100
inefficient process when cSOR values are high.
CO2 Emitted for In Situ Recovery Only

6. Carbon dioxide emissions analysis 10

Given the amount of steam generated, by using the enthalpy of

steam and combustion stoichiometry, the amount of carbon diox-
1
ide generated can be determined. As shown in Fig. 3, for steam 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
injection at 2100 kPa, with 60% steam delivery to the chamber cSOR, m3/m3
edge, the tSOR is equal to about 0.7 m3/m3 and consequently about
190 kg of carbon dioxide is generated per m3 bitumen produced. Fig. 14. CO2 emitted per unit net energy generated (100% steam quality generated
with efficiency of the steam generator equal to 0.75) from steam-based recovery
This assumes that the steam generator has an efficiency of 75%. process operating at 2100 kPa.
Fig. 13 displays the CO2 intensity, measured in kg CO2 generated
per unit volume bitumen produced versus the cSOR. At a cSOR
equal to 3.5 m3/m3, roughly 716 kg of carbon dioxide is released 7. Conclusions
to the atmosphere per m3 bitumen generated. On a carbon dioxide
per unit energy basis, at a tSOR equal to about 0.7 m3/m3, approx- A thermodynamic limit on the thermal efficiency for steam-
imately 2.2 m3 of carbon dioxide is generated per GJ of energy in based bitumen recovery processes has been developed in the form
the form of chemical energy contained in the bitumen. At a cSOR of a theoretical steam-to-oil ratio(tSOR). The analysis shows that
equal to 3.5 m3/m3, about 9.1 m3 of carbon dioxide is emitted per although some SAGD operations are achieving good steam-to-oil
GJ chemical energy contained in the produced bitumen. Fig. 14 dis- ratios, many are not achieving thermally efficient operation, with
plays the mass of CO2 emitted per unit energy generated (in the cumulative steam-to-oil ratios many times the theoretical value.
form of chemical energy in the bitumen) versus the cSOR of the This results from combinations of geological realities, operator
bitumen recovery process for the case with no cogeneration of decisions and the limitations of the SAGD process. The results dem-
electricity from waste heat. The results show that the CO2 emitted onstrate that on an energy and carbon dioxide emissions basis,
grows significantly as the cSOR increases. If the CO2 generated bitumen or bitumen-based energy recovery processes need to step
from bitumen upgrading to SCO and SCO refining to transportation well beyond the capabilities of current steam-based bitumen
fuels is taken into account, the overall net production of CO2 per recovery processes, such as SAGD, if practical and sustainable en-
unit energy realized in the transportation fuel is raised dramati- ergy balance and emissions scenarios are to be achieved from the
cally. Here, for bitumen upgrading to SCO, the amount of CO2 in situ oil sands operations.
equivalent emitted per m3 of bitumen processed ranges from about
233 kgCO2/m3 for a coking configuration to 399 kgCO2/m3 for an
Acknowledgements
ebullated bed configuration [15]. For this study, the average value
is used (316 kgCO2/m3 bitumen processed). For refining of SCO, the
Financial support was provided by the Natural Sciences and
amount of CO2 released is taken to be 360 kgCO2/m3 bitumen
Engineering Research Council of Canada (NSERC), Carbon Manage-
recovered from the reservoir and the amount generated for bitu-
ment Canada and the Canada Research Chairs program.
men transportation is taken to be 20 kgCO2/m3 bitumen recovered
[15].
References

[1] Roadifer RE. Size distributions of the world’s largest known oil and tar
accumulations. In: Meyer RF, editor. Exploration for heavy crude oil and
CO2 Per Volume Bitumen Produced, kg/m 3

5000
natural bitumen: American Association of Petroleum Geologists Studies in
4500 Geology, vol. 25. Tulsa: American Association of Petroleum Geologists; 1987. p.
3–23.
4000 [2] Head IM, Jones DM, Larter SR. Biological activity in the deep subsurface and the
origin of heavy oil. Nature 2003;426:344–52.
3500
[3] AED (Alberta Economic Development). Oil Sands Industry Update. Alberta
3000 Department of Energy; 2008. <http://www.energy.gov.ab.ca/com/default.
htm>.
2500 [4] CIA (Central Intelligence Agency). The World Factbook; 2009. Available at CIA
<https://www.cia.gov/library/publications/the-world-factbook>
2000 [5] Larter SR, Adams JJ, Gates ID, Bennett B, Huang H. The origin, prediction and
1500 impact of oil viscosity heterogeneity on the production characteristics of tar
sand and heavy oil reservoirs. J Can Petrol Technol 2008;47(1):52–61.
1000 [6] Mehrotra AK, Svrcek WY. Viscosity of compressed Athabasca bitumen. Can J
Chem Eng 1986;64(5):844–7.
500 [7] Butler RM. Thermal recovery of oil and bitumen. Calgary (Alberta,
Canada): GravDrain Inc.; 1997.
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 [8] Gates ID, Adams JJ, Larter SR. The impact of oil viscosity heterogeneity on the
production characteristics of tar sand and heavy oil reservoirs. Part II:
cSOR, m3/m3 intelligent, geotailored recovery processes in compositionally graded
reservoirs. J Can Petrol Technol 2008;47(9):40–9.
Fig. 13. Mass of emitted carbon dioxide-to-oil ratio versus cSOR (100% steam [9] Ranger MJ, Gingras MK. Geology of the Athabasca oil sands – field guide and
quality generated with efficiency of the steam generator equal to 0.75) from steam- overview. Alberta (Canada): Department of Earth and Atmospheric Science,
based recovery process operating at 2100 kPa. University of Alberta; 2006. 119 p.
 I.D. Gates, S.R. Larter / Fuel 115 (2014) 706–713 713

[10] Fustic M, Bennett B, Adams J, Huang H, MacFarlane B, Leckie D, et al. Bitumen [13] IHS Energy Inc. AccuMap oil and gas database for Alberta, Canada; 2009.
and heavy oil geochemistry: a tool for distinguishing barriers from baffles in [14] Rosenfeld J, Pont J, Law K, Hirshfeld D, Kolb J. Comparison of North American
oil sands reservoirs. Bull Can Petrol Geol 2011;59(4):295–316. and imported crude oil lifecycle GHG emissions. Final report prepared for
[11] Strobl RS, Muwais WK, Wightman DM, Cotterill DK, LiPing Y. Geological Alberta Energy Research Institute, TIAX LLC, Cupertino, California, USA, and
modeling of McMurray formation reservoirs based on outcrop and subsurface MathPro Inc., West Bethesda, Maryland, USA; 2009.
analogues. In: Pemberton GS, James DP, editors. Can Soc Petroleum Geologists [15] Keesom W, Unnasch S, Moretta J. Life cycle assessment comparison of North
Mem, vol. 18. Western Canada: Petroleum Geology of the Cretaceous Lower American and imported crudes. Final report prepared for Alberta Energy
Manville Group; 1997. p. 292–312. Research Institute, Jacobs Consultancy, and Life Cycle Associates, Chicago,
[12] Gates ID, Kenny J, Hernandez-Hdez IL, Bunio GL. Steam injection strategy and Illinois, USA; 2009.
energetics of Steam-Assisted Gravity Drainage. SPE Reserv Eval Eng [16] Furimsky E. Emissions of carbon dioxide from tar sands plants in Canada.
2007;10(1):19–34. Energy Fuels 2003;17(6):1541–8.