Steam Flooding of Naturally

Fractured Reservoirs: Basic Concepts

and Recovery Mechanisms
A. Mollaei, IOR Research Institute
B. Maini, University of Calgary

to overcome reservoir heterogeneity. Many simulation studies and

Abstract laboratory corefloods have shown that steam fingers do not develop
A review of important issues in steam injection in naturally in high permeability streaks.
fractured reservoirs (NFRs) is presented. The effect of tempera- CO2 liberation during steam injection in carbonate reservoirs
ture on physical properties of crude oils and rocks and the thermo- can provide additional enhanced oil recovery, as has been noted
chemical alteration of crude oil are discussed. in many experimental and pilot projects(3-5). Here we will briefly
The recovery of oil from NFRs can be modelled as a two step review the mechanism and role of CO2 generation during steam
process: first the oil is expelled from the matrix blocks through injection.
mechanisms that impose a pressure gradient within each ma-
trix block and then it is swept through the fracture network to a
production well by mechanisms that impose a pressure gradient
within the fracture network. The recovery mechanisms associ-
Basic Concepts
ated with steam injection in NFRs and their characteristic times
are presented. The most important recovery mechanism in matrix
The Effects of Temperature on Physical
blocks is differential thermal expansion between oil and the ma- Properties of Rock and Fluids
trix pore volume and the strongest mechanism in fracture network
For modelling steam injection processes, the physical properties
is the reduction of viscosity ratio (μo/μw). The matrix oil recovery
of both reservoir fluids and rock matrix are needed over a range of
mechanisms are relatively independent of oil gravity, making
temperatures covering the initial reservoir temperature to the steam
steam an equally attractive recovery process in fractured light and
temperature. The effect of temperature on physical properties of
heavy oil reservoirs.
the reservoir that are subject to change during thermal processes is
The mechanism and impact of CO2 generation during steam in-
summarized below:
jection in carbonate reservoirs are discussed. The rate of CO2 gen-
eration is controlled by the rate of heat conduction from fracture Viscosity
into the matrix. For a specific reservoir the rate of heat conduction
is a function of temperature and injection rate of steam and these The viscosity of heavy oil decreases very rapidly as the tem-
can be optimized to make use of the in situ generated CO2. perature increases. The correlation used to establish the ASTM
viscosity/temperature chart can be used to model the effect of tem-
perature, as follows:
Introduction
Heavy oil in naturally fractured carbonate reservoirs is an impor- ( )
log log ν + f  = c − b log T
. ................................................................. (1)
tant resource, which accounts for one-third of total heavy oil world-
wide. Many fractured reservoirs in the Middle East, former Soviet where ν is the kinematic viscosity, T the absolute temperature and
Union and Canada are candidates for thermal heavy oil recovery. b, c and f are coefficients.
Steam injection processes, which have been used extensively to re-
cover heavy oil from non-fractured reservoirs, were not applied to Density
fractured reservoirs until recently. This was primarily based on the
The density (ρ) and its inverse the specific volume (v), depend
belief that the injected steam would bypass the oil through the frac-
on pressure and temperature. Their variation with temperature and
tures and would be ineffective in recovering the oil. However, the
pressure is described by two coefficients:
results of experimental, theoretical and pilot tests which have ap-
1) The coefficient of thermal expansion, defined as
peared in the literature since early 1980s, show the feasibility of
heavy oil recovery from fractured reservoirs using steam injection.
Fractured carbonate reservoirs represent a unique target for ap- ( )(
β = 1/v dv/dT ) p = − (1/ ρ)( dρ /dT ) p .................................................... (2)
plication of enhanced oil recovery technology. High divalent ion
concentrations in reservoir waters and extensive fracture networks
appear to preclude use of chemical processes and gas injection tech- 2) The compressibility, defined as
niques with the possible exception of miscible CO2. Schulte and
Vries(1) did an experiment to show the feasibility of in situ com- ( )(
k = − 1/v dv/dP )T = (1/ ρ)( dρ /dρ)T ............................................... (3)
bustion in densely fractured reservoirs, such as those that occur in
the Middle East in Iran and Oman. However, because of the chan-
neling of injected air, in situ combustion probably cannot be sus- The thermal expansion of rocks is relatively small. According
tained in fractured formations(2). Because of these reasons, more to measurements for carbonate rocks over a wide range of tem-
consideration has been given to steam injection. The key to suc- perature, the average value of β is between 9 × 10−5 − 1.8 × 10−4
cess of steam processes is the way in which heat conduction acts K−1. The density of liquids at pressure P and temperature T can be
January 2010, Volume 49, No. 1 65
estimated from ρo, βo, ko at reference conditions Po, To using a first absolute permeability increases, is constant or decreases when tem-
order approximation: perature increases. It should be noted that the various studies have
been carried out using very different materials and fluids. There-
( )
ρ / ρo = 1− βo T − To + K o P − Po ( ) .................................................. (4)
fore, each case is a specific one. The phenomena observed may be
explained as follows: (a) the thermal expansion of solids can cause
a change in the size of pore throats in low permeability media; (b)
The coefficient of thermal expansion β is approximately 10−3 interactions between the injected fluids and the matrix can be tem-
K−1 for hydrocarbons at 20°C. For crude oils, β increases when ρ perature sensitive and (c) chemical changes in the constituents
decreases. The coefficient of thermal expansion of water is about of heated rocks can modify the porosity and permeability of the
2 × 10−4 K−1. medium(6–11).

Heat Capacity
Thermo-Chemical Alteration of Reservoir
When there is no chemical reaction and no change in state, the
variation in enthalpy H with temperature at constant pressure can
Rock and Crude Oil
be expressed by: Generally, heating of the reservoir can lead to two kinds of
chemical reactions: (1) pyrolysis and (2) oxidation. The first group
dH = C p dT of reactions occurs with heating in the presence of an inert gas with
........................................................................................... (5) or without the presence of water phase. The second group occurs
in the presence of oxygen in gaseous phase, in the liquid phase and
The heat capacity per unit mass of solids, liquids and gases gen- on surface of the solids. Since there is no oxygen in steam injec-
erally increases with temperature and the effect of temperature can tion process, the pyrolysis reactions are more likely to occur in the
be expressed as: reservoir during steam injection. These reactions can be classified
into three types:
C p =b+cT+ƒT 2 1) Cracking: The cracking involves rupture of C-C bonds and
.................................................................................... (6) the formation of molecules with lower molecular weights. A
typical reaction is as follows:
where cp is the specific heat at constant pressure and b, c and ƒ are
coefficients. Cn+mH 2( n+m)+2 → Cn H 2 n + Cm H 2 m+2
At room temperature, the heat capacity for natural dry rocks lies ................................................ (8)
between 0.75 and 0.88 Kj.kg−1.K−1. Its value for a crude oil is ap-
proximately 1.8 kJ.kg−1.K−1 and it is equal to 4.18 kJ kg−1 K−1 for 2) Dehydrogenation: in these kinds of reactions, the number of
water. The heat capacity of a complex mixture is equal to the sum carbon atoms remains unchanged; only C-H bonds are de-
of the heat capacities of its constituent elements. stroyed, and unsaturated hydrocarbons are formed:
Thermal Conductivity (λ)
Cn H 2 n+2 → Cn H 2 n + H 2
Heat transfer in a porous medium often occurs by thermal con- ....................................................................... (9)
duction. As long as the flow rate in the porous medium is not too
high, the fluids and the reservoir rock are at the same temperature. 3) Condensation: Condensation between two hydrocarbons
Under this condition, it is possible to consider the real medium as leads to the formation of a molecule with a higher molecular
a continuum having an equivalent thermal conductivity. The effec- weight. When the reactants are alkanes and alkenes, conden-
tive thermal conductivity depends on the thermal conductivity of sation often leads to the formation of aromatic compounds.
the matrix, on the conductivity of the saturating phases and on their Temperature has a strong effect on the type of reaction and the
relative distribution in the medium and it is preferable to use experi- amount of gas and solid in products. Reactions that occur at low
mentally determined values. temperature (<350°C) are referred to as visbreaking while high
The thermal conductivity of most natural rocks changes slowly temperature reactions are called cracking. The term visbreaking ex-
with temperature. The thermal conductivity λ of a liquid decreases presses the reduction in oil viscosity. These reactions are character-
with increasing temperature, according to: ized by the formation and retention of products in the liquid phase.
It implies that production of gases and solids is minimum. Cracking
(
λ / λ o =l − b T − To ) ............................................................................. (7)
is known to begin at temperatures above 260°C and production of
solids and gases becomes significant above 350°C.

where λo is the thermal conductivity at reference temperature To

and b is a constant. At room temperature, the thermal conductivity
Oil Recovery Mechanisms in Fractured
lies between 0.13 and 0.15 W.m−l.K−1 for liquid hydrocarbons and Reservoirs During Steam Injection
is approximately 0.6 W.m-1.K-1 for water. NFRs differ from non-fractured reservoirs in that the fractures
Interfacial Tension, Wettability and Capillary Pressure provide flow paths with permeabilities that can be orders of mag-
nitude higher than the remainder of the formation. Generally, the
The interfacial tension between hydrocarbons and water de- storage capacity of the fracture network is significantly lower
creases with increasing temperature. The water/solid surface con- than that of the matrix blocks. Thus, fractures can control fluid
tact angle decreases as temperature increases, which indicates that flow within the reservoir without contributing much to its storage
the rock surface becomes more water-wet. The polar compounds capacity.
affecting the way a rock surface is wetted may be desorbed at high For oil to be recovered from such reservoirs, a pressure gradient
temperatures, which makes the surface more water-wet. must be established within matrix blocks on a pore level. This pres-
sure gradient then displaces the oil from one pore to the next and
Porosity and Permeability of the Rock ultimately, to the production well by means of the fracture network.
Rocks have very low values of the thermal expansion coeffi- If a high permeability fracture network exists within a reservoir, it
cient. If their permeability depends only on the geometry of the may not be possible to impose a pressure gradient across a matrix
pores, its change with temperature would be a function of the small block simply by injecting fluid into a well and pressurizing the frac-
thermal expansion coefficient, so it should be only a minor ef- ture network(3).
fect. Reported results concerning the influence of temperature are The recovery of oil from NFRs can be modelled as a two step
contradictory. Depending on test conditions, it has been reported that process: oil is expelled from the matrix blocks through mechanisms
66 Journal of Canadian Petroleum Technology
that can impose a pressure gradient within each matrix block and temperatures, the molecular structure of some oils can be altered
then it is swept through the fracture network to a production well through breaking of chemical bonds. For a heavy oil heated to
by mechanisms that impose a pressure gradient within the fracture 575°F (300°C) permanent decrease in oil density of approximately
network. 1% has been observed, which could result in an increase in the pro-
duced oil volume of 1%.
Mechanisms of Oil Expulsion From Matrix to Gravity Drainage
Fracture During Steam Injection in NFRs Gravity drainage can occur in reservoir containing vertical frac-
Steam injection can be a viable recovery process for many frac- tures saturated with either water or gas. The differential hydrostatic
tured reservoirs. A number of mechanisms operate during steam head between the fluid in the fractures and the oil In the matrix
injection to establish the required pressure gradient within matrix blocks will establish a vertical pressure gradient and tend to force
blocks to drive the oil into the fracture network. As will be shown, oil out of the matrix blocks. For a 35°API oil in the matrix blocks
these mechanisms are largely independent of oil gravity, making and water in the fracture, the resulting differential pressure gradient
steam injection in NFRs equally attractive to light and heavy oil would be 0.065 psi/ft (1.47 kPa/m). This low differential gradient
reservoirs. is not likely to recover a significant amount of oil, particularly in
reservoirs with low matrix permeability. Gravity drainage is, there-
Thermal Expansion fore, not thought to be a significant factor during steam injection
with liquid filled fractures.
During steam injection processes, both the matrix minerals For gas filled fractures, a differential pressure gradient on the
and pore saturating fluids are heated and will expand. Values for order of 0.3 psi/ft (6.8 kPa/m) could be generated. This differential
the thermal expansion coefficient (β) for crude oil vary, but 7.2 gradient would be sufficient to displace oil, particularly if the ma-
× 10−4 m3/(m3.K) is a reasonable average. Matrix minerals ex- trix block permeability is not too low. However, before any oil can
pand into the pore volume upon heating and reduce the porosity. be displaced, there must be a sufficient gravity head between the
These effects combine to yield a differential thermal expansion matrix oil and fracture gas to overcome the capillary entry pressure
coefficient to expel fluid from the matrix block of approximately of the matrix block. The ability of steam to enter the matrix blocks
9 × 10−4 m3/(m3K). will also be restricted by condensation of the steam vapour at the
For an increase in matrix block temperature of 200°C, the re- fracture face as the matrix block heats.
sulting differential thermal expansion of fluid and pore volume is Reis(12) concluded that gravity drainage is not a significant re-
approximately 18% of the pore volume. This volume of fluid will covery mechanism in NFRs, except possibly in thick formations
be expelled from the matrix block upon heating. having a high matrix permeability, a low matrix capillary pressure
and continuous, vertical, steam filled fractures.
Capillary Imbibition
At equivalent saturations, matrix capillary pressure is much
higher than fracture capillary pressure (Pcm>>Pcf) and this differ- Cyclic Steam Injection
ence represents the imbibition driving force, which is an impor- During cyclic steam injection, the total reservoir pressure in-
tant recovery mechanism in fractured carbonates. Since the rock creases during injection and then decreases during production. Be-
becomes more water-wet at higher temperatures, the matrix cap- cause low permeability can inhibit the flow of fluids within matrix
illary pressure (Pcm) increases with temperature and this leads to blocks, the pressure changes within the matrix blocks may lag be-
higher capillary imbibition force, which has been called “thermally hind that of the fracture network, establishing temporary pressure
augmented imbibitions.” The amount of oil expelled depends on the gradients within the matrix blocks.
interfacial tension, wettability and pore geometry of the fluid rock During cyclic steam injection, it may be possible to drop the
system. Oil expulsion from capillary imbibition varies with recov- pressure more quickly than the temperature can decline by thermal
eries ranging from a few percent to as high as 70%. conduction. Since the matrix blocks will be at or near the saturation
Also Dreher et al.(7) showed that isothermal imbibition followed temperature of steam at the injection pressure, dropping the pres-
by steam injection can lead to a noticeable oil rate response, but sure will cause the water in the matrix block to flash to steam. This
that the water/oil ratio may cease to be economic if the isothermal will expel oil by gas drive. This mechanism has recovered oil from
imbibition has already produced a significant oil response and if matrix blocks in laboratory tests(8). It is similar to gas generation,
the relative permeability is unchanged because of heating. For ex- but can act at lower steam temperatures.
ample, if a reservoir has been waterflooded for some years but Sor
is not too small at the end of the waterflooding and relative perme- Alteration of the Rock Matrix
abilities improve during heating the fractured reservoir by steam
In addition to gases being generated from reactions in fluids at
injection, thermally augmented imbibition would provide a signifi-
high temperature, many reservoir minerals also undergo chemical
cant oil rate.
alterations. Studies have indicated that mineral alterations during
Gas Generation steam injection tend to consolidate oil bearing rock by changing
the structure of the pore filling material. Chemical alterations of the
Significant amounts of gas can be generated in a reservoir during matrix generally decrease the permeability of the reservoir, but the
steam injection from various water/oil or water/matrix chemical re- effects on oil recovery are difficult to predict and will depend on the
actions(4). The volume of gas generated can exceed the pore volume rock/fluid chemistry.
of the affected reservoir. These gases include CO2, H2S and light
hydrocarbons. While both the temperatures required for gas gen- Distillation (Oil Vapourization)
eration and the gas composition depend on the composition of the The distillation of light hydrocarbon components is a proven re-
fluids and matrix minerals, significant generation can occur at tem- covery mechanism in steamflooding of non-fractured reservoirs. As
peratures as low as 450°F (232°C). This gas can displace oil from the steam flows past the oil, thermodynamic forces evaporate some
the matrix as its volume increases. of the hydrocarbon components into the steam. Because the amount
The volume of oil recovered from gas generation is uncertain of distillation is limited by thermodynamic equilibrium, significant
but may be similar to recoveries in non-fractured reservoirs by so- volumes of oil are distilled only when many pore volumes of va-
lution gas drive. It could contribute recovery factors in the range of pour pass through the pore network. Since the steam vapour in a
15% to 25%. fractured reservoir flows primarily through the fracture network,
very little steam flows through the matrix blocks to allow distilla-
Chemical Reactions
tion. Even when gas generation occurs, the number of pore volumes
It has been discussed that chemical reactions such as visbreaking of vapour flowing through the matrix blocks will be relatively small
and cracking occur and can improve the oil properties. At high and the amount of distilled hydrocarbons carried out with the gas
January 2010, Volume 49, No. 1 67
will be minor. Distillation, therefore, is not thought to be significant the crude oil. When there is steam in fractures the thermodynamic
in matrix, but it may be effective in fractures. conditions would be conducive for oil vapourization. The conse-
quences of such presence of vapourized light hydrocarbons are:
Solution Gas Drive 1) The hydrocarbon phase downstream from the steam front is
A reactivation of a solution gas drive has been postulated as a enriched in light components. The viscosity of the crude en-
mechanism for increasing oil recovery during steam injection proj- countered by the advancing steam front is reduced (this effect
ects. However, a solution gas drive can only exist when the reser- is more pronounced with heavy oils). The result would be a
voir pressure is at or below the bubblepoint and there is a sufficient decrease in the residual oil saturation and an increase in oil
amount of gas dissolved in the oi1. In reservoirs with an initial gas recovery.
saturation, this gas would be expanded upon heating and would be- 2) The residual oil in the steam zone has an increasing content
have similar to a reservoir with an active solution gas drive. Oil re- of heavy fractions that are less and less volatile, and the oil
covery from this mechanism, if it occurs, would be similar to that saturation decreases with time.
of gas generation. 3) Sometimes, because of the increase in light constituents,
heavy components such as asphaltenes may precipitate in-
Rock Compaction side the porous medium. Changes in permeabilities and wet-
As the reservoir pressure decreases in unconsolidated sand for- tability can occur.
mations, the sand can compact as the grain contacts bear more of 4) In other cases, the lighter oil produced makes possible
the overburden weight, reducing the effective porosity of the res- the elimination of residues deposited near the wellbore
ervoir. Substantial quantities of oil in the Venezuelan tar sands are during cold production, again affecting permeabilities and
recovered by this mechanism. This mechanism relies on the rear- wettability.
rangement of matrix grains in poorly cemented formations, and Other mechanisms such as thermal expansion, chemical altera-
will not be effective in competent formations. Because virtually all tion of oil, solution gas drive can also sweep oil within fractures but
NFRs are in highly competent lithologies (e.g., dolomites, cherts, they are not as strong(1, 6, 9–11).
shales, and sandstones), compaction is not expected to be important
in NFRs(2, 7, 8, 12).
Oil Expulsion Rates
Mechanisms of Oil Displacement Through Our discussion has shown that recovery mechanisms exist within
the Fracture Network to the Production Well NFRs that can expel significant quantities of oil from matrix blocks
into the fracture network. Each of these mechanisms has their own
During Steam Injection in NFRs characteristic oil expulsion rate that can be used to evaluate oil pro-
The second stage of recovery process starts after the oil has been duction rates. Also the characteristic oil expulsion times can be cal-
expelled from matrix into the fracture. In this stage the oil must be culated, which give the time needed for expelling the mobilized
swept through the fracture network by imposing a pressure gradient oil by the identified recovery mechanisms. Calculations for pres-
within the fracture network. Fracture has much higher transmissi- sure depletion mechanism are shown below and the calculations for
bility than matrix and the oil moves much faster within the frac- other mechanisms are similar.
tures. In densely fractured reservoirs, it may be difficult to generate The characteristic time for oil to be expelled by pressure deple-
a high pressure gradient in the fractures but even a small pressure tion can be found through the unsteady-state flow equation for a
gradient is enough to sweep the oil. The most important mecha- slightly compressible liquid. For linear flow this equation is:
nisms that contribute to oil recovery from fractures during steam
injection are: ∂2 P cϕµ ∂P
=
∂x 2 k ∂t ..................................................................................... (9-2)
Reduction of Viscosity Ratio (μo/μw)
The oil viscosity reduction with increasing temperature is an im-
portant factor in all thermal recovery processes. Generally, the oil The following solution to this equation can be found through
viscosity is reduced faster than water viscosity as temperature in- separation of variables.
creases and the greatest viscosity reduction occurs with the most
viscous oils. The viscosity ratio reduction is effective in decreasing
the fractional flow of hot water, consequently increasing the hot ( )
P x , t − Pf ∞   2
nπ
= ∑ Bn exp −  

k   nπ 
t sin  x 
water displacement efficiency. The situation is different for light Pi − Pf   L  cφµ   L 
n=1  
oils because in some cases the viscosity ratio (μo/μw) increases with .............................. (10)
temperature.
with
Effect of Temperature on Relative Permeabilities (kro and krw)
It is generally agreed that residual oil saturation Sor decreases and 2 L  nπ 
irreducible water saturation Swi increases with increasing tempera- Bn =
L
∫ sin  L x  dx
ture. Besides the viscosity ratio, the important temperature depen-
0 ................................................................. (11)
dent factors that affect relative permeabilities are wettability and
interfacial tension. When the rock is water-wet, the interfacial prop- The characteristic time for the pressure decline from fluid expul-
erties affected by temperature are water/oil interfacial tension and sion is given by the primary decay mode (n = 1) as:
the thickness of the wetting film. When the rock is preferentially
oil-wet, the adsorption equilibrium of polar compounds changes
L2 cϕµ
with increasing temperature towards less adsorption making the τ=
surface less oil wet. π 2 k ............................................................................................ (12)

Distillation (Oil Vapourization) Table 1 lists the characteristic expulsion times for other oil re-
An important phenomenon during steam injection is the steam covery mechanisms that drive the oil from matrix into fractures. As
distillation of the light components of crude oil. If the pressure is seen in Table 1, these characteristic times are largely independent
lower than the sum of the vapour pressures of water and oil, the of oil gravity, making steam injection in NFRs equally attractive to
liquid mixture will boil and give off a vapour phase composed of light and heavy oil reservoirs. To illustrate the time scales for oil
steam and organic compounds. Like steam, some hydrocarbons will expulsion for the recovery mechanisms outlined above, Reis(12) cal-
condense ahead of the steam front and then become mixed with culated the characteristic oil expulsion times for a 10 ft (3 m) wide
68 Journal of Canadian Petroleum Technology
Table 1: Characteristic time formula for different Table 2: Matrix block properties for characteristic
recovery mechanisms(12). times calculations(12).

Characteristic Matrix Block Property Value

Mechanism Expulsion Time (τ) Width 20 ft (6 m)
Height 10 ft (3 m)
L2 cϕµ Thermal diffusivity 0.029 ft2/hr (0.75 mm2/s)
Pressure depletion τ=
π2 k Porosity 15%
Viscosity 3 cp (3 × 10-3 Pa.s)
L2 Interfacial tension 3 dyne/cm
Thermal expansion τ=
απ 2 Permeability 0.1 md (9.87 × 10-17m2)
2 Gravity 30 °API (0.876 g/cm3)
Lµ ϕ Compressibility 4 × 10-6 psi-1 (5.8 × 10.7 kPa-1)
Capillary imbibition τ=
σ k Kerogen pre-exponential factor 1016 sec-1
 − E  Kerogen activation energy 52.5 kcal/mole
Chemical changes within the matrix τ = A exp  i  CO2 pre-exponential factor 0.1 sec-1
 RTa  CO2 activation energy 15 kcal/mole

Gravity drainage hµϕ

τ=
k ∆ρg controlled by the steam injection rate and the temperature. There-
fore, there are optimum steam injection rate and temperature for a
matrix block with a permeability of 0.1 md (9.87 × 10−5 µm2) and given situation. When the injection rate (or temperature) is higher
a 3 cp (3 × 10−3 Pa∙s) viscosity oil. The matrix block properties for than this level, there will not be effective balance between CO2 gen-
this example are summarized in Table 2 and the resulting character- eration and use. A large amount of CO2 will be generated in a short
istic expulsion times are shown in Table 3. time after steam injection but most of it will migrate to fractures
In this example, the characteristic oil expulsion time from pres- without dissolving in the oil or mobilizing substantial amount of
sure depletion is only a few hours. Oil recovery by thermal expan- oil.
sion could occur within a few weeks. Oil recovery from chemical If the saturation of free CO2 in fractures becomes high, it can by-
reactions depends on the temperature, but for temperatures above pass the oil without contributing to oil recovery. On the other hand,
450°F (232°C) could result in significant chemical reactions within since there is only a limited amount of CO2 that can be produced,
a one year, while for temperatures near 550°F (288°C) could have it would be a waste of CO2 liberation if we do not optimize the in-
significant reactions within days. The expulsion of oil by capillary jection rate and temperature to synchronize CO2 liberation and the
imbibition is much slower and is estimated to take approximately oil mobilization mechanisms. It should be possible to alter these
one year. Gravity drainage is too slow to be an effective recovery mechanisms such that there would be minimum amount of free CO2
mechanism(12). in the reservoir. For example, we can optimize the steam injection
conditions so that prolonged CO2 generation from the matrix is ac-
companied with lowering of the oil viscosity by swelling.
Co2 Liberation in Fractured Carbonate
Reservoirs During Steam Injection
Processes Conclusions
The decomposition of mineral carbonates existing in carbonate 1. The oil recovery from NFRs by steam injection involves two
rocks, mainly dolomite CaMg(CO3)2, upon heating yields a great distinct processes, one responsible for expelling the oil from
amount of CO2 and begins considerably sooner in the presence of matrix blocks into the fractures and the other driving the oil
steam. The CO2 is generated from the matrix blocks after heat and into the production well through the fracture network.
steam penetrate into the block and cause the decomposition of do- 2. Differential thermal expansion between the oil and the pore
lomite. Higher the rate of heat conduction from fracture into the volume is the most important incremental recovery mecha-
matrix, the faster would be the generation CO2. CO2 migrates to- nism in the matrix blocks of NFRs during steam injection.
ward fracture after generation and then flows through the fractures
toward the production wells. CO2 also dissolves in the oil, which in- 3. Reduction of viscosity ratio (μo/μw) during steam injection
creases oil’s volume and decreases its viscosity. CO2 also displaces is the most important oil recovery mechanism in the frac-
some oil by pushing it forward before dissolving in the oil. tures. The viscosity ratio reduction is effective in decreasing
For a single matrix block the amount of generated CO2 in a given the fractional flow of hot water, consequently increasing hot
time depends on the percentage of matrix volume that was heated water displacement efficiency.
in that time period. Therefore, for a given reservoir, the amount of 4. By estimating the characteristic times for different mecha-
generated CO2 is controlled by the rate of heat conduction from nisms it is possible to estimate the time it will take to expel
fracture into the matrix. the mobilized oil from matrix to the fracture.
Dreher et al.(7) used a thermal simulator to show the effects of 5. Decomposition of carbonates (mainly dolomite) upon heating
CO2 liberation by rock dissolution during steam injection in a reser- occurs during steam injection and liberates CO2, which en-
voir which had been water flooded for 6 years. Their results show hances oil recovery. The rate of CO2 liberation is controlled
that the liberation of CO2 from the matrix appears to create a rapid by of the temperature and rate of steam injection.
but short-lived oil rate response followed by a brief decline in oil
rate, followed by a second, more sustained oil rate response. The Table 3: Characteristic oil expulsion time (days)(12).
early oil response was thought to be caused by CO2 liberated in
the “shell” of the matrix cells where oil saturation is already low Recovery Mechanism Value
by prolonged waterflood. This response dies off fairly rapidly and Thermal heating time 15
before the “core” has been heated to a significant degree. The Mattax and Ky1e capillary imbibition 360
second, more sustained part of the oil response is due to a com- Gravity drainage 250,000
bination of CO2 liberation and thermally augmented imbibition in Pressure depletion 0.3
the “core.” Kerogen maturation (450oF) 61
CO2 generation (450oF) 360
It is apparent that in a specific reservoir the rate of CO2 libera- Kerogen maturation (550oF) 0.33
tion during steam injection is controlled by the rate of heat conduc- CO2 generation (550oF) 8.1
tion from fracture into the matrix and the rate of heat transfer is
January 2010, Volume 49, No. 1 69
 6. The recovery performance can be improved by using an op- 7. Dreher, K.D., Kenyon, D.E., and Iwere, F.O. 1986. Heat Flow
timum temperature and rate of steam injection that would During Steam Injection Into a Fractured Carbonate Reservoir. Paper
maximize the use of liberated CO2 for oil mobilization. SPE 14902 presented at the 1986 SPE/DOE fifth Symposium on En-
hanced Oil Recovery, Tulsa, 20–23 April. doi: 10.2118/14902-MS.
8. Hyne, J.B., Greidanus, J.W., Tyrer, J.D., Verona, D., Rizek, C.,
Clark, P.D., Clarke, R.A., and Woo, J. 1982. Aquathermolysis of
Acknowledgements Heavy Oils. Presented at the Second International Conference on the
Future of Heavy Crude and Tar Sands, Caracas, 7–17 February.
The authors would like to thank Roberto Aguilera (University of 9. Shahin, G.T., Moosa, R., Kharusi, B., and Chilek, G. 2006. The
Calgary) and S.M. Farouq Ali (University of Calgary) for their help Physics of Steam Injection in Fractured Carbonate Reservoirs: Engi-
and guidance in preparation of this paper. neering Development Options That Minimize Risk. Paper SPE 102186
presented at the SPE Annual Technical Conference and Exhibition,
San Antonio, Texas, USA, 24–27 September.doi: 10.2118/102186-
NOMENCLATURE MS.
A = pre-exponential factor 10. Sumnu, M.D., Brigham, W.E., Aziz, K., and Castanier, L.M. 1996. An
Experimental and Numerical Study on Steam Injection in Fractured
C = compressibility Systems. Paper SPE 35459 presented at the SPE/DOE Improved Oil
Cp = constant pressure heat capacity Recovery Symposium, Tulsa, 21–24 April. doi: 10.2118/35459-MS.
Ei = activation energy 11. Green, D.W. and Willhite, G.P. 1998. Enhanced Oil Recovery. Text-
g = acceleration of gravity book Series, SPE, Richardson, Texas 6.
12. Reis, J.C. 1990. Oil Recovery Mechanisms in Fractured Reservoirs
h = height of matrix block
During Steam Injection. Paper SPE 20204 presented at the SPE/
k = permeability DOE Enhanced Oil Recovery Symposium, Tulsa, 22–25 April. doi:
h = thickness of matrix block 10.2118/20204-MS.
M = mobility ratio
P = pressure Provenance—Original Petroleum Society manuscript, Steam Flooding of
Pf = fracture pressure Naturally Fractured Reservoirs: Basic Concepts and Recovery Mech-
Pi = initial pressure anisms (Paper 2007-128; SPE Paper 132485), first presented at the 8th
Canadian International Petroleum Conference (the 58th Annual Technical
R = gas constant Meeting of the of the Petroleum Society), June 12-14, 2007, in Calgary, Al-
t = time berta. Abstract submitted for review March 13, 2007; editorial comments
T = temperature sent to the author(s) February 21, 2009; revised manuscript received No-
vember 13, 2009; paper approved for pre-press December 5, 2009; final
Ta = temperature (absolute units) approval December 15, 2009.
TR = initial reservoir temperature
Ts = steam temperature
V = velocity Authors’ Biographies
x = length
Alireza Mollaei is currently a petroleum en-
α = thermal diffusivity
gineering PhD candidate at the University of
β = coefficient of thermal expansion Texas at Austin. He served as reservoir sim
ΔVo = recovered oil volume ulation and EOR researcher at the Improved
Δρ = density difference Oil Recovery Research Institute (IORRI) of
φ = porosity the R&D Directorate of the National Iranian
Oil Company (NIOC) from 2006 to 2008.
λ = thermal conductivity
His re search interests include forecasting of
µ = viscosity EOR processes using analytical modeling,
σ = interfacial tension developing EOR predictive models, sensi-
τ = characteristic time tivity and uncertainty analysis, experimental
inves tigation and numerical simulation of gravity drainage in frac-
tured and non-fractured porous media. Other interests include de-
REFERENCES signing a compositional simulation of gas condensate reservoirs
1. Schulte, W.M. and de Vries, A.S. 1985. In-Situ Combustion in considering relative permeability rate effect and different numer
Naturally Fractured Heavy Oil Reservoirs. SPE J. 25 (1): 67–77. ical simulation methods. He holds a B.Sc. degree from Petroleum
SPE-10723-PA. doi: 10.2118/10723-PA. University of Technology, Iran (PUT), an M.Eng. degree from the
2. Nolan. J.B., Erlich, R., and Crookston, R.B. 1980. Applicability of Uni versity of Calgary, Canada, and an M.Sc. degree from PUT, all
Steamflooding for Carbonate Reservoirs. Paper SPE 8821 presented in petroleum engineering.
at the SPE/DOE Enhanced Oil Recovery Symposium, Tulsa, 20–23
April. doi: 10.2118/8821-MS. Brij Maini is a professor of chemical and
3. Dehghani, K. and Ehrlich, R. 1998. Evaluation of Steam Injection petroleum engineering at the University of
Process in Light Oil Reservoirs. Paper SPE 49016 presented at the Calgary. He holds a BTech degree from the
SPE Annual Technical Conference and Exhibition, New Orleans, Indian Institute of Technology, Kanpur and
27–30 September. doi: 10.2118/49016-MS. a PhD degree from University of Wash-
4. Holladay, C.H. Jr. 1966. The Basic Effects of Steam on a Reservoir. ington, both in chemical engineering. Maini
Paper SPE 1666 presented at the SPE Eastern Regional Meeting, Co- served as senior staff research engineer and
lumbus, Ohio, USA, 10–11 November. doi: 10.2118/1666-MS.
group leader for heavy oil research at the Pe-
5. Penney, R., Moosa, R., Sahin, G., Hadhrami, F., Kok, A., Engen, troleum Recovery Institute for more than 20
G., van Ravesteijn, O., Rawnsley, K., and Kharusi, B. 2005. Steam
Injection in Fractured Carbonate Reservoirs: Starting a New Trend
years before taking up his current academic
in EOR. Paper IPTC 10727 presented at the International Petro- position in 1999. He has been studying the
leum Technology Conference, Doha, Qatar, 21–23 November. doi: unusual behaviour of primary production in several Canadian heavy
10.2523/10727-MS. oil reservoirs for nearly 20 years and has authored several papers
6. Baviere, M. ed. 1991. Basic Concepts in Enhanced Oil Recovery Pro- on this topic. Maini’s other research interests include bitumen re-
cesses, Vol. 33. London: Critical Reports on Applied Chemistry, SCI/ covery with Vapex and steam assisted gravity drainage processes
Elsevier Applied Science. and multiphase flow through porous media.
70 Journal of Canadian Petroleum Technology