Professional Documents
Culture Documents
SPE-110738-Enhanced Heavy Oil Rec by As Flooding
SPE-110738-Enhanced Heavy Oil Rec by As Flooding
Abstract
This study presents the results of laboratory core studies
investigating the recovery mechanisms of alkali-surfactant
flooding in heavy oil reservoirs. Specifically, mixtures of
water and alkali-surfactant systems have been injected into
cores containing heavy oil (11 000 mPas and 15 000 mPas).
Salinity is varied in order to generate oil-in-water vs. water-inoil emulsion systems, and the effects of generating different
emulsions are compared.
The application of this work is for the many heavy oil
reservoirs in countries such as Canada and Venezuela
containing viscous oil that still has some limited mobility
under reservoir conditions. Alkali-surfactant (AS) flooding
has considerable potential for non-thermal oil recovery after
primary production.
Experiments were performed on cores with varying
permeability, at different AS injection rates. All tests were
performed on gas-free oil systems. The response from direct
injection of AS systems is compared to AS injection after
waterflooding. Pressure and oil recovery information is
obtained from core floods, and these results are interpreted
based on a semi-theoretical framework obtained from phase
behavior and bulk liquid studies. It is demonstrated that both
oil-in-water and water-in-oil emulsions can lead to the
recovery of additional oil.
Alkali-surfactant flooding is already an established
technique in conventional oil reservoirs, whereby enhanced oil
recovery is a result of reduced trapping of oil due to the
lowered oil/water interfacial tension. In addition, the injection
of these chemicals may lead to the formation of emulsions, as
has been documented by previous researchers. In our work,
we demonstrate that in heavy oil systems, emulsion formation
is a necessary requirement for the production of heavy oil.
When these emulsions form, AS injection can lead to
considerable improvements in the flooding response, even
without the addition of polymers to stabilize the flood.
Introduction
Several countries in the world, notably Canada and Venezuela,
contain massive resources of unconventional heavy oil and
and bitumen. With issues of resource stability and rising oil
prices, international interest is now shifting rapidly towards
Canadas oil sands. The oil sands are characterized as
unconsolidated, high porosity and high permeability
reservoirs. While ease of flow is therefore not a significant
concern, the single biggest obstacle to successful recovery
from the oil sands is the high oil viscosity.
Heavy oil reservoirs are a special subset of the oil sands,
whereby the oil viscosity at reservoir temperature and pressure
varies on the order of 50 50 000 mPas (cP). While this oil
is still very viscous, it does have some limited mobility at
resevoir conditions. As much as 20% of the oil may be
recovered by solution gas drive1, but in many cases the
recovery is much lower. At the end of primary production,
significant oil still remains in the reservoir, but the reservoir
energy has now been depleted. This is the target for enhanced
heavy oil recovery.
In order to recover additional heavy oil after primary
production, a fluid usually has to be injected in order to
displace oil to the production wells. However, mobility ratio
concerns dominate displacement of viscous oil, and most EOR
processes focus on reduction of the oil viscosity or
improvement in the mobility ratio. Unfortunately, many of the
heavy oil reservoirs in Canada are relatively small and thin,
making them poor candidates for expensive thermal processes.
Ideally, the displacement mobility ratio should be improved in
an inexpensive (i.e. non-thermal) fashion.
This work
investigates the potential for alkali-surfactant flooding to be
used for enhanced heavy oil recovery.
The injection of alkalis and/or surfactants into oil
reservoirs is not a new technology. As early as in the 1920s,
Nutting proposed the injection of alkaline solution into
reservoirs for oil recovery2. The injection of a combination of
alkali and surfactant was discussed in the 1950s by Reisberg
and Doscher3. Since then, chemical injection (alkali and/or
surfactant) has become an accepted enhanced oil recovery
methodology in many conventional oil applications.
Surfactants are a special class of molecule that are both
hydrophobic and hydrophilic, thus the most stable
configuration for these molecules is at the oil-water
interface4,5. In surfactant flooding, these molecules are
generally injected along with water in order to reduce the oilwater interfacial tension. This leads to a reduction in capillary
forces, that may trap oil in rock pores6,7. Alkali flooding is a
special subset of chemical flooding, whereby the surfactant is
SPE 110738
SPE 110738
recovery and water cut information from the larger core size.
In cores M and N, overburden pressure of 2070 kPa (300 psi)
was also applied, which is responsible for the lower
permeability compared to cores A to I. Finally, core O is the
larger version of cores J to L, using the same sieved sand.
In the Brn AS floods studied in Table 2, Cores Q to X are
the same core as A to I in Table 1. Likewise, cores Y to AB
are the same core as J to L in Table 1, and core AC is the
larger version (cores M and N in Table 1). In between
experiments, cores were flooded with toluene to remove any
oil from the rock. The cores were then flushed using DI water
in order to remove any surfactant that may have adsorbed onto
the sand grains. Finally, the cores were flushed with methanol
and were then dried using air. The core porosity and
permeability was re-measured before starting a new
experiment, in order to ensure that the core properties
remained unchanged and results from different experimental
runs could be compared against one another. Since the cores
were packed using unconsolidated material and in many cases
overburden pressure was not present, the measured
permeability to water changed between experiments. This was
especially pronounced in the higher permeability systems, as
evidenced in Tables 1 and 2.
After the cores were saturated with either DI water or brine
and their permeability had been measured, heavy oil (viscosity
shown in Tables 1 and 2) was injected at a constant rate using
a digital ISCO pump until the cores were at their irreducible
water saturation. At this point, injection of either water or AS
solution was performed at the constant rates shown in Tables 1
and 2. It should be noted that the velocities listed in these
tables are Darcy velocities, calculated by dividing the injection
flow rate (set by the ISCO pump) by the cross-sectional area
of the core. In core floods with only one test, this means that
either a waterflood or a primary AS flood was performed. In
these floods, the experiment was terminated after around 5 8
pore volumes (PV) of fluid had been injected. In systems with
both waterfloods and AS floods, a waterflood consisting of
approximately 5 PV of fluid was injected, followed by around
2 PV of AS solution.
Produced fluid volumes were measured using low field
NMR, since the heavy oil is sticky and may adhere to the
walls of the glass collection containers, making visual
determination of produced oil and water cut difficult to
measure accurately. After the NMR spectra had been
analyzed, the oil recovery was verified by separation with
toluene. Pressures were recorded using transducers connected
to an automatic data acquisition system.
The results of the experiments are divided into findings
from the bulk liquid study and the core floods. The data
collected from the core floods consist of produced fluid
volumes and pressure, thus the actual recovery mechanisms
that are occurring at the pore scale can only be inferred based
on an understanding of how the systems behave in the absence
of porous media.
Bulk Liquid Study
Oil and water are immiscible fluids, therefore when they are
mixed energy has to be put into the system in order to
emulsify them. The energy required for emulsification is
proportional to the interfacial tension between the two
fluids5,26. For fluids that are dis-similar, such as oil and water,
the energy requirement can be significant. Shear or agitation
is required to overcome the Gibbs free energy of formation
and, more importantly, the Laplace equation of capillary
pressure in a droplet26. In order to form an O/W or W/O
emulsion, therefore, two requirements must be met: the oilwater IFT must be reduced, and shear must be added to the
system. The addition of alkali and surfactant is used to reduce
the oil-water IFT such that emulsification can occur with a
minimum of applied shear.
Figure 1 shows the measured oil-water IFT values as a
function of alkali content, with the surfactant fixed at 0.1 wt%
(1000 ppm). As mentioned previously, due to the similar
densities of oil and water, when interfacial tension is higher
than 1 mN/m the apparatus cannot measure the IFT even at its
maximum rotational speed. Several comments can be made
regarding Figure 1. First, considering the DI water data: in the
absence of any alkali the surfactant alone cannot reduce the
IFT down to measurable values. Even with 0.1 % alkali
(1000 ppm) added, the IFT is too high to be measured. After
0.2% alkali has been added to the aqueous phase, however, the
IFT values plateau at around 0.02 mN/m for the entire range
of alkali contents tested. For the brine system, even in the
absence of alkali it appears that the surfactant alone can
reduce IFT down to measurable values. In the presence of
salt, the IFT has a minimum at 5000 ppm alkali, and additional
alkali leads to increases in the IFT.
The salinity effect observed in Figure 1 has also been
previously noted by other researchers8,25. Salt tends to shift
the salinity minimum to lower alkali concentrations, and the
presence of additional alkali then acts like increased salinity.
Therefore, if injecting brine, the IFT in the system seems less
stable than for DI water. In either case, however, it can be
seen that with the addition of alkali and surfactant to the
system, the interfacial tension between oil and water is
reduced by two to three orders of magnitude.
While low IFT is certainly a requirement for
emulsification to occur, the results from Figure 1 do not
indicate which type of emulsion will be generated. When oil
and water are agitated at low IFT conditions, both O/W and
W/O emulsions are formed simultaneously. The emulsions
are thought to break exponentially, at a rate governed by the
frequency of droplet collisions and by the energy barriers to
droplets coalescing upon contact27. The emulsion that
separates more slowly is the emulsion that is macroscopically
said to have formed.
Due to the high viscosity of heavy oil, when the oil phase
is continuous, water droplets will collide at a much lower
frequency than oil droplets in a continuous, low viscosity
water phase. For this reason, W/O emulsions are much more
common than O/W emulsions in heavy oil systems5,27.
However, oil droplets in water can also be stabilized by
electrostatic repulsion of similarly-charged surfactant
heads26,28. Several other rules of thumb can also be used to
identify which emulsion type should form in bulk liquid
systems5. First, if the volume of one fluid is much larger than
the other fluid, it is more likely that the fluid with the larger
volume will be the continuous phase. Second, if the surfactant
is more soluble in one phase, that phase will also more easily
be continuous. Finally, surfactants made from monovalent
SPE 110738
SPE 110738
(3)
Pemul = ow + cos
R r 3
k
(5)
M eff = w = w o
o w ko
The effective mobility ratio is calculated for waterflooding
and DI AS flooding in Figure 11. In this figure, the injected
pore volumes have been normalized in order to reduce the
scatter in the floods. Waterfloods are carried out to a
normalized value of one PV injected, and secondary AS floods
then go to two PV injected. For the waterfloods, the effective
mobility ratio is low until water breakthrough, as expected.
After breakthrough, Meff is on the order of 40 50. Upon
injection of DI AS solution, a remarkable reduction in Meff is
observed. The effective mobility ratio decreases down to the
order of unity, meaning that the flood actually becomes stable
at early times. This corresponds to the region in Figures 9 and
10 where the pressure gradient increases in the core. In all of
the floods, no produced O/W emulsion was observed, thus the
proposed recovery mechanism is that of emulsification &
entrapment.
Effective mobility ratio for Brn AS floods are presented in
Figure 12.
Once again, during waterflooding Meff is
approximately 50 after one normalized PV of fluid has been
injected, and with Brn AS addition the mobility ratio improves
significantly. In both DI AS and Brn AS flooding, therefore,
the mobility ratio response is very similar. Even after the
mobility ratio increases once again, the final value of Meff is
around 20, indicating that even after pressure declines again in
the system the injected fluid is still less mobile than pure
water.
The recoveries from secondary DI AS flooding are shown
in Table 3, and those for secondary Brn AS flooding are in
Table 4. The average recovery efficiencies for both types of
SPE 110738
SPE 110738
water cut. It is the pressure and water cut response that is key
to understanding what is happening: emulsion droplets are
forming due to the low IFT and shear from flow, but these
droplets are not being entrained in the flowing aqueous phase.
Rather, they must be plugging off the water channels, and this
is leading in turn to improved sweep efficiency and a reduced
effective mobility ratio.
In systems with high shear, O/W emulsions were actually
produced under DI AS flooding. In these same systems, if
flooded with Brn AS solution after DI AS flooding, again the
pressure in the core built up but now the produced fluid
appeared to be a W/O emulsion. Therefore, although in many
cases actual emulsions were not observed in the produced
fluids, the bulk liquid experiments are still useful in order to
determine which type of emulsion should be forming in the
porous medium. It is possible that even under DI AS
conditions multiple emulsion types may be forming in the
core38, but the results from low permeability experiments
show that the bulk liquid emulsion systems can be used at
least as an indication of what should be happening in reservoir
applications.
Figure 15 plots the total recovery efficiency against the
minimum value of Meff for the secondary DI and Brn AS
floods, the entrainment secondary DI AS floods and the
tertiary Brn AS floods. In all cases, the total recovery
efficiency is directly correlated to the minimum value of Meff.
This shows that improved recovery is due to an improvement
in the mobility ratio of the system.
Figure 15 shows that the smallest values for the minimum
Meff occur for secondary DI and Brn AS floods. In Table 5,
these systems also have the highest recovery efficiency values.
During emulsification & entrainment or tertiary Brn AS
flooding, Meff is not reduced to such low values, and the
recovery efficiencies in Table 5 are likewise not as high.
The recovery efficiency values in Table 5 are the average
from all of the core floods performed in this study. The
largest standard deviation in repeated measurements is 0.014
RF/PV inj, so this was used as the error bar to determine if
floods had higher or lower recoveries when compared to one
another. If two systems were different by > 0.014, they were
said to be different in their recovery efficiency. Table 6
compares the results from waterflooding in DI vs Brn systems.
The results differ by less than this error bar, so it can be said
that waterflooding at the same rate is not affected by the
salinity of the injected fluid. When comparing waterflooding
to primary AS flooding, however, the results differ by a larger
amount. This indicates that primary AS floods are more
effective than waterfloods for both DI AS and Brn AS
systems.
Obviously, Table 5 shows that secondary AS floods are
much more efficient than primary AS floods. This can be
understood through a simple Jamin pressure drop
calculation30. If an emulsion droplet is surrounded on both
sides by a low IFT fluid, then the pressure drop to force this
droplet through a pore constriction is small. This is the state
of a primary AS flood after breakthrough of the AS solution.
Alternatively, if the emulsified droplet is contacted on the
back end with a low IFT fluid but on the front end with a high
IFT fluid, as is the case for AS flooding after waterflooding,
the Jamin pressure drop is significantly higher in this situation.
4 k P
L o
(7)
(8)
dim = actual
app
Figure 16 plots the total recovery efficiency of different
floods vs. dimensionless shear rate. For waterflooding and
primary DI AS flooding, the most effective oil recovery region
occurs before the breakthrough of the water or AS solution. If
pressure in the system builds up further, this leads to improved
flow rates by Darcys Law. This explains the weak trend
showing that at higher dimensionless shear rate, the recovery
efficiency is slightly higher for these systems. For secondary
AS flooding, however, the trend is opposite. The best
recovery is achieved under conditions of lowest dimensionless
shear rate. This indicates that entrapment is in fact the main
recovery mechanism for these DI AS systems. At lower shear
rate, emulsified oil droplets are formed but the shear is not
high enough to force these droplets through the pore throats.
At elevated shear, the trapping of the oil droplets is reduced,
SPE 110738
actual
app
dim
P
x
ow
os
ws
=
=
=
=
=
=
=
=
=
=
=
References
1.
2.
3.
4.
5.
6.
7.
8.
SPE 110738
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
10
SPE 110738
Core ID D (m)
A
0.011
B
0.011
C
0.011
D
0.011
E
0.011
F
0.011
G
0.011
H
0.011
I
0.011
J
0.011
K
0.011
L
0.011
M
0.036
N
0.036
O
0.026
L (m)
0.103
0.103
0.103
0.103
0.103
0.103
0.103
0.103
0.103
0.105
0.105
0.105
0.211
0.211
0.532
0.45
0.45
0.45
0.45
0.45
0.45
0.45
0.45
0.45
0.49
0.49
0.49
0.45
0.45
0.41
WF rate AS rate
o
(mPas) (m/d) (m/d)
11000
0.26
11000
0.51
15000
0.51
11000 0.51
11000 0.51
0.51
15000 0.51
0.26
15000 0.51
1.02
15000 0.51
2.05
15000 0.51
2.05
11000
0.51
15000 0.51
0.51
15000 0.51
0.26
11000 0.42
0.42
11000 0.023 0.42
11000 0.51
0.51
k (D)
19.22
16.25
14.98
19.98
15.84
12.71
12.72
12.05
11.44
1.19
0.91
0.97
8.7
8.7
0.79
Core ID
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
WF total AS
total
RF/PV inj RF/PV inj
0.098
0.064
0.046
0.037
0.035
0.127
0.043
0.133
0.038
0.093
0.037
0.095
0.031
0.094
0.066
0.050
0.066
0.047
0.055
0.050
0.512
0.051
0.208
0.048
0.055
Core ID D (m)
Q
0.011
R
0.011
S
0.011
T
0.011
U
0.011
V
0.011
W
0.011
X
0.011
Y
0.011
Z
0.011
AA
0.011
AB
0.011
AC
0.036
L (m)
0.103
0.103
0.103
0.103
0.103
0.103
0.103
0.103
0.105
0.105
0.105
0.105
0.211
0.45
0.45
0.45
0.45
0.45
0.45
0.45
0.45
0.49
0.49
0.49
0.49
0.45
k (D)
15.98
11.04
15.9
11.74
17.94
11.55
11.81
10.24
1.48
1.19
0.92
0.97
9.05
WF rate AS rate
o
(mPas) (m/d) (m/d)
11000
0.51
11000
0.51
11000
0.51
11000
0.51
11000 0.51
0.51
11000 0.51
0.51
15000 0.51
0.26
15000 0.51
1.02
11000
0.51
15000
0.51
15000 0.51
0.51
15000 0.51
0.26
11000 0.42
0.42
Core ID
Q
R
S
T
U
V
W
X
Y
Z
AA
AB
AC
WF total AS
total
RF/PV inj RF/PV inj
0.089
0.109
0.076
0.086
0.029
0.109
0.030
0.121
0.041
0.200
0.034
0.188
0.065
0.070
0.035
0.123
0.026
0.098
0.045
0.237
System
DI waterflood
Prim DI AS
Sec DI AS
Entrainment
Brn waterflood
Prim Brn AS
Sec Brn AS
Tert Brn AS
Total
RF/PV inj
0.042
0.069
0.180
0.062
0.034
0.083
0.148
0.061
SPE 110738
11
System
DI WF vs Brn WF
DI WF vs Prim DI AS
Brn WF vs Prim Brn AS
Prim DI AS vs Prim Brn AS
Sec DI AS vs Sec Brn AS
Difference
No
Yes
Yes
Yes?
Yes
0 ppm NaCl
20 000 ppm NaCl
0.1
0.01
0
5000
10000
15000
20000
12
SPE 110738
0.25
1400
1200
500
Pressure
Oil RF
300
0.1
200
0.05
100
0
0
0
1000
800
0.4
600
0.3
400
0.2
200
0.1
0
0
PV Injected
500
0.7
Pressure
0.6
Water cut
0.5
300
0.4
200
0.3
0.2
100
0
4
0.9
0.8
1000
0.7
Pressure
800
0.6
Water cut
0.5
600
0.4
400
0.3
0.2
200
0.1
1200
Pressure (kPa)
Pressure (kPa)
0.8
600
0.1
0
0
PV Injected
PV Injected
700
1000
Effective Mobility Ratio
600
Pressure (kPa)
1400
0.9
700
4
PV Injected
400
0.5
Oil RF
400
0.15
0.6
Pressure
Pressure (kPa)
0.2
600
Pressure (kPa)
0.7
700
500
Waterflood
400
AS Flood
300
200
100
10
1
0.1
0.01
100
0.5
1.5
2.5
PV Injected
Core M AS
Core N AS
Core E AS
Core G AS
Core H AS
Core I AS
Core F AS
13
1000
0.3
100
0.25
SPE 110738
10
1
0.1
0.5
1.5
0.15
0.1
0.05
0
0.01
0.01
0
0.2
2.5
Core W
Core AB
Core AC
Core AA
10
100
Secondary AS Floods
Entrainment
Secondary Brn AS
Tertiary Brn AS
100
10
1
0.1
0.01
0.001
0
0.5
1.5
2.5
1000
Waterflood
Primary AS
Secondary AS
Entrainment
0.1
0.01
Normalized PV Injected
1
Core K AS
Figure 13.
ratio
0.1
Core L AS
10
100
Core O AS
0.4
Pressure
0.35
Pressure (kPa)
Oil RF
0.3
10000
0.25
8000
0.2
6000
0.15
4000
0.1
2000
0.05
12000
14000
0.1
Secondary DI AS
Secondary Brn AS
0.01
0
0
10
10
100
PV Injected