Combined Cyclic Solvent Injection and Waterflooding in the Post-Cold

Dec 12, 2016 - ... Systems Engineering, Faculty of Engineering and Applied Science, University of Regina, Regina, Saskatchewan S4S 0A2, Canada...
0 downloads 0 Views 3MB Size
Article pubs.acs.org/EF

Combined Cyclic Solvent Injection and Waterflooding in the PostCold Heavy Oil Production with Sand Reservoirs Hongze Ma,† Desheng Huang,† Gaoming Yu,‡ Yuehui She,§ and Yongan Gu*,† †

Petroleum Technology Research Centre (PTRC), Petroleum Systems Engineering, Faculty of Engineering and Applied Science, University of Regina, Regina, Saskatchewan S4S 0A2, Canada ‡ College of Petroleum Engineering, Yangtze University, Jingzhou, Hubei 434023, P. R. China § College of Chemical and Environmental Engineering, Yangtze University, Jingzhou, Hubei 434023, P. R. China ABSTRACT: In this paper, cyclic solvent injection (CSI) and waterflooding (WF) were combined and studied to maximize their technical synergy and optimize the enhanced heavy oil recovery (EHOR) in the post-cold heavy oil production with sand (CHOPS) reservoirs. The original heavy oil sample was collected from the Colony formation in western Canada. The PVT data and viscosities of CH4/CO2/C3H8-saturated heavy oil were measured at different equilibrium pressures and Tres = 21 °C. A total of eight sandpacked laboratory tests were conducted to study and compare four different EHOR processes after the primary production: CSI, CSI + WF, simultaneous CSI + WF, and WF + CSI. In the last three processes, WF was applied after, simultaneously with, and prior to the CSI production, respectively. Three different pressure drawdown rates (6.8, 12.5, and 25.0 kPa/min) and two different solvents (CO2 and C3H8) were used to determine their specific effects on CSI + WF. The experimental results showed that CSI + WF had the highest heavy oil recovery factor (RF) of 30.1% in comparison with 28.9, 25.9, and 24.3% for WF + CSI, simultaneous CSI + WF, and CSI, respectively, when CO2 was used in CSI. The intermediate pressure drawdown rate of 12.5 kPa/min resulted in the highest heavy oil RF in CO2-CSI + WF. In addition, C3H8 was found to be a more effective extracting solvent than CO2 due to its more favorable PVT properties and larger heavy oil viscosity reduction.

1. INTRODUCTION The global heavy oil and bitumen reserves are estimated to be six trillion barrels, approximately 50% of which are located in western Canada.1 As a widely applied primary production process, cold heavy oil production with sand (CHOPS) contributes 20% of the crude oil production in Canada.2 The foamy oil flow and the high-permeability wormholes as a result of the sand production jointly lead to a high heavy oil production rate in the CHOPS.3 Overall, approximately 5−15% of the original oil-in-place (OOIP) in the heavy oil reservoirs in western Canada can be recovered during the primary pressuredepletion process, whereas the majority of the OOIP remains in the post-CHOPS reservoirs.4,5 More than 200 heavy oil waterflooding (WF) processes have been implemented in western Canada in the past 60 years as an economical secondary improved oil recovery (IOR) method.6,7 Nevertheless, a typical waterflood in the post-CHOPS reservoirs recovers on average only 2−7% additional heavy oil because of severe water channeling.8 Thus, some effective enhanced heavy oil recovery (EHOR) methods are required to further develop many post-CHOPS reservoirs. Thermal-based EHOR methods such as steam-assisted gravity drainage (SAGD) and cyclic steam stimulation (CSS) have been successfully applied in many oilfields. However, they are not suitable for a post-CHOPS reservoir with a thin pay zone, bottom water, gas cap, and/or low rock thermal diffusivity. This is because there are excessive heat losses into the overburden and underburden, and tremendous amounts of greenhouse gases are generated.9 Solvent-based EHOR methods such as vapor extraction (VAPEX) and cyclic solvent injection (CSI) are excellent alternatives and have some distinct © XXXX American Chemical Society

advantages in terms of the energy efficiency, produced oil quality, and environmental benefits.10 The major problem of these EHOR methods is the low heavy oil production rate, especially in the traditional CSI pilot tests or field-scale applications. For example, CO2 was cyclically injected into two wells in the Camurlu heavy oil field for three cycles in 1984.11 The oil production rate of 11.5 BOPD in CO2-CSI was almost the same as 11.0 BOPD in the primary production if the solvent injection and soaking periods were taken into account. A CO2CSI field pilot was implemented in the Halfmoon field (United States) in 1992.12 The incremental oil production rates were in the range of 1.7−2.4 BOPD, which were far lower than those predicted from the laboratory data. The heavy oil production rate was relatively high in the CSI production period. At the end, the cyclic CO2 injection was terminated due to the lack of the economic incentive. In summary, it is a key technical challenge to find EHOR processes better than the traditional CSI in the post-CHOPS reservoirs and considerably increase the heavy oil production rate and recovery factor (RF). In the past, some useful CSI variations have been explored to solve three major technical limitations of the traditional CSI. The first drawback of the traditional CSI is the heavy oil viscosity regainment due to pressure depletion in the solvent release process. The enhanced cyclic solvent process (ECSP) was proposed to slow the heavy oil viscosity regainment.13 A more volatile solvent (e.g., methane) was injected first to reinforce the solution-gas drive. Then, a more soluble solvent Received: October 7, 2016 Revised: December 5, 2016 Published: December 12, 2016 A

DOI: 10.1021/acs.energyfuels.6b02596 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

method and is summarized in Table 1. It can be seen from this table that there are no hydrocarbons under C9 and that there is a large lump

(e.g., ethane or propane) was injected later to maintain the reduced heavy oil viscosity. The experimental results of the four sandpacked tests with different solvent injection sequences showed that a higher heavy oil production rate and RF and a lower gas−oil ratio (GOR) were achieved in the ECSP than those in the traditional CSI. The second main pitfall of the traditional CSI is that the remaining foamy heavy oil at the end of the CSI production period was pushed back from the producer by the subsequently injected solvent. In this regard, the gasflooding-assisted cyclic solvent injection (GA-CSI) was investigated to prevent the back-and-forth movement of the foamy oil.14 A gas flood was applied immediately after the CSI production period to push the remaining foamy oil toward the producer. It was found from nine sandpacked tests that the heavy oil production rate was increased by 2−3 times and that the heavy oil RF was increased by 10−20%. Nevertheless, an initial communication between the CSI producer and the gas injector had to be established prior to each GA-CSI test, which took too much time and consumed some solvent. Third, an extremely high mobility ratio between the injected gas and the heavy oil in the traditional CSI resulted in an early gas breakthrough. To inject a foaming agent prior to the foamy oilassisted methane huff-n-puff (FOAM H-n-P) was proposed to enhance the sweep or displacement efficiency and strengthen the foamy oil flow.15 Eight sandpacked tests showed that an average of 43.29% incremental heavy oil was recovered in the FOAM H-n-P. Although several effective CSI variations have been developed to overcome the technical limitations of the traditional CSI, there have been fewer studies to explore the technical synergy of combining CSI and WF together in a twowell configuration in the post-CHOPS reservoirs. In this paper, the original heavy oil and brine samples were collected from the Colony formation in western Canada. Then, CH4/CO2/C3H8 solubilities in the Colony heavy oil, oil-swelling factors, densities, and viscosities of CH4/CO2/C3H8-saturated Colony heavy oil at different equilibrium pressures and the actual reservoir temperature of Tres = 21 °C were measured. A total of eight laboratory-scale tests in a sandpacked physical model were conducted to measure the heavy oil RFs and production rates, GORs and water−oil ratios (WORs), and reservoir pressures of different CSI + WF processes. These hybrid EOR + IOR processes take advantages of the enhanced microscopic displacement efficiency of the traditional CSI and the improved volumetric sweep efficiency of the secondary WF. Moreover, WF can effectively prevent the quick solvent liberation and continue to displace and produce the remaining foamy heavy oil.

Table 1. Compositional Analysis Result of the Colony Heavy Oila carbon no.

mol %

carbon no.

mol %

C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31

0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.79 2.61 2.21 3.80 4.20 4.51 4.49 4.67 4.43 4.18 3.87 3.52 3.16 2.99 2.54 2.47 2.21 2.22 2.24 1.89 1.97 1.68 1.43

C32 C33 C34 C35 C36 C37 C38 C39 C40 C41 C42 C43 C44 C45 C46 C47 C48 C49 C50 C51 C52 C53 C54 C55 C56 C57 C58 C59 C60 C61+ total

1.46 1.32 1.28 1.15 1.05 1.16 1.00 0.99 1.05 0.93 0.98 1.06 0.96 0.89 0.83 0.77 0.67 0.55 0.64 0.63 0.51 0.40 0.53 0.47 0.53 0.40 0.44 0.39 0.35 8.52 100.00

Well no.: 16A-3-59-7; ρo = 0.992 g/cm3; μo = 33 876 cP at Pa = 1 atm and Tres = 21 °C; MWo = 547.7 g/mol; asphaltene content of wasp = 18.3 wt.% (n-pentane insoluble).

a

sum of extremely heavy hydrocarbons (C61+ = 8.52 mol %) in the Colony heavy oil. The physicochemical properties of the Colony reservoir brine collected from the Bonnyville area at Pa = 1 atm and Tres = 21 °C are listed in Table 2. Its total dissolved solids (TDS) were measured to be 37 619 mg/L, which is in a typical range of 20 000− 60 000 mg/L for a reservoir brine from the post-CHOPS reservoirs in western Canada.16 The purities of CH4, CO2, and C3H8 (Praxair, Canada) used in this study were equal to 99.97 mol %, 99.998 mol %, and 99.5 wt %, respectively. 2.2. Viscosity Measurements and PVT Studies. A schematic diagram of the experimental setup for measuring the viscosity and PVT data of the solvent-saturated heavy oil at a prespecified equilibrium pressure and Tres = 21 °C is shown in Figure 1. CH4/CO2/C3H8saturated heavy oil was prepared by mixing the Colony heavy oil and pure CH4/CO2/C3H8 in two stainless steel cylinders (500-10-P-316-2, DBR, Canada) at Tres = 21 °C. Each heavy oil−solvent system was considered to be at an equilibrium state once the daily pressure change of the heavy oil−solvent mixture was less than 10 kPa, which was the accuracy of the pressure gauge used (PRI-PRO, Martel Electronics, United States). After the heavy oil−solvent system reached the equilibrium state, the viscosity (μmix), density (ρmix), solvent solubility (χ) in the heavy oil, and oil-swelling factor (SF) of the solventsaturated heavy oil were measured using the following experimental steps:

2. EXPERIMENTAL SECTION 2.1. Materials. In this study, the original heavy oil sample was collected from the Colony formation in the Bonnyville area, Alberta, Canada. The density and viscosity of the Colony heavy oil were measured to be ρo = 0.992 g/cm3 using a densitometer (DMA 512P, Anton Paar, United States) and μo = 33 876 cP using a viscometer (DV-II+, Brookfield, United States) at Pa = 1 atm and Tres = 21 °C, respectively. The molecular weight of the Colony heavy oil was measured to be MWo = 547.7 g/mol using an automatic highsensitivity wide-range cryoscope (Model 5009, Precision Systems Inc., United States). The asphaltene content of the Colony heavy oil was measured to be wasp = 18.3 wt % (n-pentane insoluble) using the standard ASTM D2007-3 method and filter paper (Whatman No. 5, England) with a pore size of 2.5 μm. The compositional analysis result of the Colony heavy oil was obtained using the standard ASTM D86 B

DOI: 10.1021/acs.energyfuels.6b02596 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

• Three different pumping rates of the solvent-saturated heavy oil were used to measure the viscosity and PVT data at each equilibrium pressure and Tres = 21 °C. The averages of the three measured results are reported as the measured viscosity and PVT data in this study. • The pressure of the heavy oil−solvent system in the heavy oil cylinder was then decreased to a lower prespecified equilibrium pressure using the syringe pump. • The viscosity and PVT data measurements were repeated at a lower equilibrium pressure. In this study, the characterization of CH4/CO2-saturated heavy oil was conducted at 5 relatively high equilibrium pressures of 5.0, 4.0, 3.0, 2.0, and 1.0 MPa, whereas the characterization of C3H8-saturated heavy oil was performed at 5 relatively low equilibrium pressures of 0.8, 0.7, 0.6, 0.4, and 0.2 MPa. Prior to the solvent-saturated heavy oil viscosity measurements, a standard-viscosity silicone liquid of S8000 (Cannon Instrument Company, United States) with μ = 33 583 cP at Tres = 21 °C was injected through the capillary tubing at different constant volume flow rates of 0.1−0.5 cm3/min to calibrate the capillary viscometer and determine the so-called “effective radius” of the capillary tubing. Then, the Poiseuille equation was applied to determine the solvent-saturated heavy oil viscosity:

Table 2. Physical and Chemical Properties of the Colony Reservoir Brinea temperature (°C) density (g/cm3) viscosity (mPa·s) pH at 25.0 °C specific conductivity (μS·cm−1) refractive index at 28 °C chloride (mg/L) sulfate (mg/L) total dissolved solids (mg/L) potassium (mg/L) sodium (mg/L) calcium (mg/L) magnesium (mg/L) a

21 1.03 1.2 7 57 143 1.3390 22 999 2 37 619 50 13 410 766 349

Well no.: 16A-3-59-7 at Pa = 1 atm. • The solvent-saturated heavy oil was pumped through several meters in series at a constant volume flow rate using a syringe pump (500 DX, ISCO Inc., United States). A back-pressure regulator (BPR-50, Temco, United States) was used to maintain the exit pressure above the prespecified equilibrium pressure so that the heavy oil−solvent mixture was in a liquid phase at all times. • The solvent-saturated heavy oil viscosity was measured using a house-made capillary viscometer which consisted of a 300 cmlong stainless-steel tubing (SS-T2-S-028-20, Swagelok, Canada) with an inner diameter of 0.175 cm and a wall thickness of 0.071 cm. • The solvent-saturated heavy oil density was measured using a densitometer (DMA 512P, Anton Paar, United States). • The flashing method was used to determine the solvent solubility in the heavy oil. Once the solvent-saturated heavy oil passed through the BPR, a quick separation of the heavy oil− solvent mixture occurred. The separated dead heavy oil was collected in a flask and weighed to be m0. The flashed gas was collected in a gas bubbler, and its volume was measured to be Vg. • At the same time, the pumped solvent-saturated heavy oil volume Vmix was read from the syringe pump and recorded to determine the oil-swelling factor.

μmix (Peq , Tres) =

π(reff )4 ΔP 8qmix L

(1)

where μmix(Peq,Tres) is the viscosity of the solvent-saturated heavy oil at each prespecified equilibrium pressure and Tres = 21 °C, ΔP and qmix are the measured pressure drop between the two ends of the capillary tubing and the preset constant heavy oil−solvent mixture volume flow rate of 0.1−5.0 cm3/min, respectively, and reff and L are the “effective radius” and actual length of the capillary tubing. In addition, the densitometer was calibrated using two standard liquids of S2000 and N7.5 (Cannon Instrument Company, United States) before the solvent-saturated heavy oil density was measured. The densities of S2000 and N7.5 are equal to 0.8805 and 0.7996 g/cm3, respectively, at Tres = 21 °C. The solvent-saturated heavy oil density was calculated using the calibrated correlation:

ρmix (Peq , Tres) = Aτ 2 + Bτ

(2)

where τ is the oscillation period read from the densitometer and A and B are two coefficients determined in the calibration process.

Figure 1. Schematic diagram of the experimental setup for measuring the viscosities and PVT data of the solvent-saturated heavy oil at different equilibrium pressures and Tres = 21 °C. C

DOI: 10.1021/acs.energyfuels.6b02596 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 2. Schematic diagram of the experimental setup for conducting the primary production and subsequent CSI/CSI + WF/simultaneous CSI + WF/WF + CSI. The solvent solubility in the solvent-saturated heavy oil at a prespecified equilibrium pressure and Tres = 21 °C was measured using the above-mentioned flashing method:

χ (Peq , Tres) =

Vg(1atm, Tres)ρg (1atm, Tres) mo

× 100%

location, whereas the WF injector was located at the center on the right-hand side of the physical model. The fluid injection unit included a solvent cylinder (Praxair, Canada) with a gas regulator (KCY Series, Swagelok, United States) and a gas flow meter (XFM17S, Aalborg, United States) for CSI. The gas regulator was set at a prespecified injection pressure to regulate the continuous solvent supply from the solvent cylinder during the solvent injection period. In CSI + WF, simultaneous CSI + WF, and WF + CSI, the syringe pump was used to inject the reservoir brine from a brine cylinder into the sandpacked physical model during the WF. The fluid production unit was comprised of a BPR (LBS4 Series, Swagelok, United States), a syringe pump (100DX, ISCO Inc., United States), a produced oil and water collector, a pair of gas bubblers for the produced gas, and a vacuum pump. The BPR was used to control the production pressure at a prespecified pressure drawdown rate. It should be noted that the injection and production pressures were measured at the same time using a pressure transducer (PPM-2, Heise, United States) and recorded in a personal computer automatically. The procedure for preparing each sandpacked physical model is briefly described as follows. The physical model was packed with the Ottawa sands (Bell & Mackenzie, Canada) of mesh sizes of 60−80 and repeatedly hammered to ensure that the sands were uniformly packed and distributed. It was tested to be leakage-free up to 3.0 MPa by using methane for 24 h. Afterward, its porosity was measured using the imbibition method and found to be in the range of 37.5−38.3%. Next, its permeability was measured using the Darcy’s law. Three different pressure drops of 4−10 kPa were applied at the two ends of the physical model, and the corresponding water volume flow rates of 4.8−15.0 cm3/min were recorded. The measured permeabilities of the sandpacked physical model were in the range of 3.9−5.0 D. After the permeability was measured, the wet sands were dried with highpressure air for at least 24 h. Then, the sandpacked physical model was saturated with the Colony brine through the imbibition process. Finally, CH4-saturated Colony heavy oil was used to displace the brine at respective constant injection and production pressures of 3.5 and

(3)

where ρg(1 atm, Tres) is the gaseous solvent density at Pa = 1 atm and Tres = 21 °C, which was calculated using the CMG WinProp module (version 2014.10, Computer Modelling Group Limited, Canada). The swelling factor of the solvent-saturated heavy oil is defined as the ratio of the volume of the solvent-saturated heavy oil at a prespecified equilibrium pressure and Tres = 21 °C to the volume of the dead heavy oil at Pa = 1 atm and Tres = 21 °C: SF(Peq , Tres) =

Vmix(Peq , Tres) mo /ρo (1atm, Tres)

(4)

2.3. Combined CSI and WF Tests. A schematic diagram of the experimental setup for conducting the primary production and CSI/ CSI + WF/simultaneous CSI + WF/WF + CSI is shown in Figure 2. This experimental setup was composed of three major operational units: a sandpacked physical model, a fluid injection unit, and a fluid production unit. The visual rectangular sandpacked physical model consisted of three plates. The rear plate was a stainless steel block with a rectangular cavity (L × W × H = 40.0 cm × 10.0 cm × 2.0 cm) for sandpacking. The front plate was a transparent acrylic glass plate so that the experimental process could be visualized. A thin transparent polycarbonate plate was placed in between to prevent the front plate from being scratched and corroded. In this study, the physical model was placed horizontally with the height of 2.0 cm to study a horizontal pay zone or slice of a heavy oil reservoir, which is perpendicular to the vertical well (injector or producer). For the primary production, the producer was located at the center on the left-hand side of the physical model. The CSI injector and producer were positioned at the same D

DOI: 10.1021/acs.energyfuels.6b02596 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels 3.0 MPa until the initial oil saturation (Soi) and the irreducible water saturation (Swi) were achieved. Table 3 summarizes the extracting solvents and reservoir characteristics of the 2D sandpacked physical model used in tests 1−8.

residual heavy oil saturations at different locations of the sandpacked physical model were measured at the ends of test 1 (CSI), test 3 (CO2-CSI + WF), and test 5 (C3H8-CSI + WF). Each sand sample was taken from a different location of the physical model, heated in an oven (OMH 750, Fisher Scientific, Canada) at T = 110 °C and 1 atm for at least 10 h, and weighed when the sand sample was dried completely. It is worthwhile to mention that no hydrocarbons were evaporated because the initial boiling point (IBP) of the Colony heavy oil was measured to be 156 °C. Then, the dry sand sample was rinsed using toluene as an extracting solvent. It was assumed that the residual heavy oil was completely removed when the color of the used toluene did not change. Finally, the cleaned sand sample was heated in the oven at T = 110 °C and 1 atm for at least 4 h to ensure that no toluene was left in the sand sample, which was weighed again at the end. The residual heavy oil saturation was determined from the measured weight change (i.e., the oil weight), the weight and density of the dry cleaned sand sample, the heavy oil density, and the porosity of the sandpacked physical model.

Table 3. Extracting Solvents and Reservoir Characteristics of the 2D Sandpacked Physical Model Used in Tests 1−8 with Colony Heavy Oil test no.

solvent

k (D)

ϕ (%)

Soi (%)

1 2 3 4 5 6 7 8

CO2 CO2 CO2 CO2 C3H8 C3H8 CO2 CO2

4.0 4.9 4.7 5.0 3.9 4.9 4.1 4.0

38.0 38.0 38.1 37.6 37.5 37.5 38.2 38.3

98.7 99.0 99.1 99.0 99.0 98.9 98.5 98.6

3. RESULTS AND DISCUSSION 3.1. Characterization of Three Heavy Oil−Solvent Systems. The measured viscosities and PVT data of CH4/ CO2/C3H8-saturated Colony heavy oil as a function of the reduced pressure at Tres = 21 °C are plotted in Figures 3a−d. In this study, the reduced pressure is defined as the ratio of the equilibrium pressure to the saturation pressure (Psat) for CO2/ C3H8 at Tres = 21 °C or the critical pressure (Pc) for CH4. Figure 3a depicts that C3H8-saturated Colony heavy oil has the lowest viscosity in comparison with that of the CH4/CO2saturated Colony heavy oil at Tres = 21 °C. The measured C3H8-saturated Colony heavy oil viscosity of 22 cP at the equilibrium pressure of 0.802 MPa and Tres = 21 °C is close to that of a typical dead medium crude oil at 1 atm and Tres = 21 °C. The viscosity reduction of the solvent-saturated heavy oil is a major EOR mechanism in the solvent-based EHOR process. It is anticipated that C3H8-saturated Colony heavy oil can be easily recovered due to its lowest viscosity and highest mobility by comparison. Accordingly, C3H8-CSI is more effective than CO2-CSI. It is found from Figure 3b that C3H8-saturated Colony heavy oil has a much reduced density, whereas CH4/CO2-saturated Colony heavy oil density is slightly decreased with the equilibrium pressure. The solvent-saturated heavy oil density reduction is the net effect of the solvent dissolution and the oil swelling. As shown in Figure 3c, CO2/C3H8 solubility in the Colony heavy oil is much higher than CH4 solubility in the Colony heavy oil at Tres = 21 °C. A high solvent solubility can appreciably enhance the microscopic displacement efficiency of CSI. This figure also indicates that C3H8 solubility is increased

The four enhanced heavy oil recovery processes were investigated in tests 1−8 after the primary production. The technical details are given in Table 4 and will be described below. In each test, the primary production was initiated at the actual initial reservoir pressure of Pi = 3.0 MPa and terminated at the final pressure of Pf = 0.2 MPa. A constant pressure drawdown rate of dP/dt = 5.0 kPa/min was applied in the primary pressure-depletion process. In test 1, CSI consisted of three periods: the injection, soaking, and production periods. In the CSI injection period, either CO2 or C3H8 was continuously injected into the sandpacked physical model as an extracting solvent at Pinj = 3.0 or 0.8 MPa and Tres = 21 °C for 40 min when no more solvent could be injected. Then, the injected solvent was soaked into the heavy oil until the reservoir pressure remained almost unchanged. In the CSI production period, the solvent injector was converted into the fluid producer. The production pressure was linearly reduced with time at a preset pressure drawdown rate from the ending pressure (Ps) of the CSI-soaking period to the prespecified ending pressure (Pe = 0.2 MPa) of the CSI-production period. In each cycle of tests 2−6 (CSI + WF), the WF was conducted after the CSI production period was completed. It should be noted that each cycle of the CSI + WF contains both CSI and WF. The reservoir brine was injected continuously from the WF injector to displace the remaining foamy heavy oil which was generated during the previous CSI production period. In test 7 (simultaneous CSI + WF), the reservoir brine was injected in the WF while the CSI production was underway. In test 8 (WF + CSI), the WF was commenced after the primary production was terminated. The subsequent CSI was implemented after the WF, which had an opposite production sequence of CSI + WF. The water volume injection rate was equal to qw = 0.5 cm3/min in CSI + WF, simultaneous CSI + WF, and WF + CSI. A total of 0.33 PV brine was injected in each WF cycle of the above-mentioned three different production processes. Moreover, the

Table 4. Enhanced Heavy Oil Recovery Processes of Tests 1−8 after the Primary Production at Tres = 21°C primary production

CSI injection

test no.

EHOR process

1 2 3 4 5 6 7 8

CSI CSI + WF CSI + WF CSI + WF CSI + WF CSI + WF simul. CSI + WF WF + CSI

Pi (MPa)

dP/dt (kPa/min)

tpro (min)

Pf (MPa)

tinj (min)

Pinj (MPa)

soaking ts (min)

3.0 3.0

5.0

560

0.2

production

Ps (MPa)

2.70 24 × 60

40 0.80

0.80 0.75

3.0

2.70

E

WF

dP/dt (kPa/min)

tpro (min)

Pe (MPa)

12.5 25.0 12.5 6.8 3.0 1.5 12.5 12.5

200 100 200 366 200 366 200 200

0.2

qw (cm3/min)

injected PV

0.5

0.33

DOI: 10.1021/acs.energyfuels.6b02596 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 3. Measured (a) viscosities, (b) densities, (c) solubilities, and (d) oil-swelling factors of CH4/CO2/C3H8-saturated Colony heavy oil at different reduced pressures P/Pc or P/Psat (Pc = 4.599 MPa for CH4; Psat = 5.868 MPa for CO2; and Psat = 0.858 MPa for C3H8) and Tres = 21 °C.

Table 5. Enhanced Heavy Oil Recovery Factors of Tests 1−8 Including the Primary Production at Tres = 21°C primary production test no.

RF (%)

1 2 3 4 5 6 7

11.4 11.3 11.2 11.5 10.5 10.8 10.1 primary production

cycle 1

cycle 2

cycle 3

RFCSI (%)

RFWF (%)

RFCSI (%)

RFWF (%)

RFCSI (%)

RFWF (%)

6.5 3.5 6.1 5.3 6.2 6.7 8.0

4.8 4.2 2.5 5.6 2.6

5.7 2.3 3.8 3.1 4.0 5.4 6.9

3.8 3.2 1.7 5.4 2.1

0.7 0.8 0.8 0.4 2.2 3.3 0.9

1.8 0.8 0.2 5.3 0.8

cycle 1

cycle 2

cycle 3

test no.

RF (%)

RFWF (%)

RFCSI (%)

RFWF (%)

RFCSI (%)

RFWF (%)

RFCSI (%)

8

10.9

6.2

6.0

3.5

0.7

0.9

0.7

cycle 4 RFCSI (%)

cycle 5

RFWF (%)

1.8 3.1

0.9 0.3

cycle 4 RFWF (%)

RFCSI (%)

RFCSI (%)

cycle 6

RFWF (%)

2.2

0.2

cycle 5 RFWF (%)

RFCSI (%)

RFCSI (%)

2.1

RFWF (%)

0.2

total RFCSI (%)

total RFWF (%)

total RF (%)

12.9 6.6 10.7 8.8 14.2 22.8 15.8

10.4 8.2 4.4 17.2 6.2

24.3 28.3 30.1 24.7 41.9 39.8 25.9

total RFWF (%)

total RFCSI (%)

total RF (%)

10.6

7.4

28.9

cycle 6 RFWF (%)

RFCSI (%)

smaller than that of the heavy oil−C3H8 mixture, the latter of which can swell to about 50% of the dead heavy oil volume at the equilibrium pressure of 0.802 MPa. The oil-swelling factor increases with the solvent solubility in the heavy oil at a given equilibrium pressure.

much more quickly at a higher equilibrium pressure, whereas CH4/CO2 solubility in the Colony heavy oil is increased almost linearly with the equilibrium pressure. Similar findings have also been reported in the literature.17,18 Figure 3d shows that the oil-swelling factor of the heavy oil−CH4/CO2 mixture is much F

DOI: 10.1021/acs.energyfuels.6b02596 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 4. (a) Measured heavy oil production rate, instantaneous GOR, pressure drop, and production pressure during the CSI production period of cycle 1 in test 3 (CSI + WF). (b) Measured heavy oil production rate, instantaneous WOR, pressure drop, and production pressure during the WF period of cycle 1 in test 3 (CSI + WF).

the oil producer at the end of the primary or CSI production.14 In the CSI + WF or WF + CSI, however, the foamy oil could be continuously and effectively displaced by the injected brine. Second, the reservoir pressure is maintained in the WF so that the heavy oil viscosity remains lower and the heavy oil mobility is higher in the combined CSI and WF. On the other hand, the combined CSI and WF can also have some shortcomings. The injected reservoir brine prevents the subsequently injected solvent from further contacting the residual heavy oil, which is referred to as the waterblocking effect.19 Hence, the heavy oil RF of CSI in CSI + WF or WF + CSI is slightly lower than that in CSI alone. In the simultaneous CSI + WF, the wellmaintained reservoir pressure substantially hinders the foamy oil formation, which is one of the major EOR mechanisms in CSI. This is why this EHOR process has a heavy oil RF much lower than that of CSI + WF or WF + CSI.

3.2. Different Production Processes. In this study, four different production processes after the primary production were conducted to examine the effect of a different EHOR process on the heavy oil RF in the post-CHOPS reservoir: test 1 (CSI), test 3 (CSI + WF), test 7 (simultaneous CSI + WF), and test 8 (WF + CSI). The heavy oil RFs (in terms of the OOIP) of the primary production, CSI and WF in Tests 1−8 are listed in Table 5. Test 3 (CSI + WF), test 7 (simultaneous CSI + WF), and test 8 (WF + CSI) give additional heavy oil RFs of 5.8, 1.6, and 4.6%, respectively, in comparison with test 1 (CSI). It should be noted that the WF alone as the subsequent IOR method after the primary pressure-depletion process in test 8 has a heavy oil RF of 6.2%, which is about half of 12.9% in test 1 (CSI). These results show that the combined CSI and WF processes are effective EHOR methods because they have two distinct technical advantages over the traditional CSI. First, it is known that not all the foamy heavy oil can reach G

DOI: 10.1021/acs.energyfuels.6b02596 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels The variations of the measured heavy oil production rate, instantaneous GOR, pressure drop, and production pressure with the CSI production time of cycle 1 in test 3 (CSI + WF) are shown in Figure 4a. In general, the CSI production period in each cycle can be roughly divided into three stages. First, the pressures at the two ends of the physical model were declined at the almost same rate of 12.5 kPa/min so that the pressure drop was small. At this stage, the heavy oil production rate was minimal, and the instantaneous GOR reached its peak value because the gas mobility is much higher than the heavy oil mobility. In the second stage, the pressure drop started to increase gradually because the dispersed gas bubbles in the viscous foamy heavy oil grew quickly enough to maintain the reservoir pressure. The heavy oil production rate was increased quickly, and the instantaneous GOR was reduced dramatically due to the increased pressure drop and the controlled gas mobility during the foamy oil flow. In the final stage, the pressure drop continued to increase. More dispersed gas bubbles were coalesced successively to become the so-called free gas, which had a detrimental effect on the foamy oil flow. Accordingly, the heavy oil production rate was decreased, and the gas production rate was increased at the end. It should be noted that the average reservoir pressure was much higher than the production pressure at the outlet even when the latter reached 0.2 MPa. Therefore, a considerable amount of the dissolved and/or dispersed gas still remained in the heavy oil at the end of the CSI production. Figure 4b shows the measured heavy oil production rate, instantaneous WOR, pressure drop, and production pressure versus WF production time of cycle 1 in test 3 (CSI + WF). The measured pressure drops fluctuated considerably so that the smoothed pressure drops were plotted in Figure 4b. At the beginning, the pressure drop reached a maximum, and a small amount of oil was produced with no water production. Then, the WOR rose rapidly, and the pressure drop fell sharply, which indicates that water can break through relatively early in the post-CHOPS reservoirs. It is speculated that the injected reservoir brine first flowed quickly through the free gassaturated and high-permeability pathways. Both the heavy oil production rate and the instantaneous WOR fluctuated because the foamy oil flow and free-gas flow dominated the fluid production alternately. It should be noted that the new foamy oil formed during the WF and/or the residual foamy oil left at the end of the previous CSI production can be stable for several days in the porous medium.20 At a later time, the heavy oil continued to be produced with a high but decreased instantaneous WOR. The growth of the dispersed gas bubbles and brine imbibition pushed the remaining heavy oil into the channels previously occupied by the brine.21 Finally, it is worthwhile to point out that the WF in a post CHOPS reservoir differs from the WF in a light oil reservoir, especially in WF cycle 1 of CSI + WF. For the WF in a heavy oil reservoir, a relatively large amount of heavy oil can still be produced after the water breakthrough (BT), whereas only a small amount of light oil is recovered after the water BT in a light oil reservoir. The measured average heavy oil production rates in CSI, cumulative GORs in CSI, and the cumulative WORs in WF of test 1 (CSI), test 3 (CSI + WF), test 7 (simultaneous CSI + WF), and test 8 (WF + CSI) are plotted and compared in Figures 5a−c. It is found from these three figures that the average heavy oil production rates are decreased and the cumulative GORs and WORs are increased in the late cycles of the combined CSI and WF processes. This is partially

Figure 5. Measured (a) average heavy oil production rates in CSI, (b) cumulative GORs in CSI, and (c) cumulative WORs in WF of different cycles in test 1 (CSI), test 3 (CSI + WF), test 7 (simultaneous CSI + WF), and test 8 (WF + CSI).

attributed to the reduced initial heavy oil saturation in each cycle of CSI or WF. Obviously, the oil relative permeability is reduced when the heavy oil saturation is reduced. At a high water saturation, the heavy oil becomes discontinuous ganglia trapped by the capillary force. In the late WF cycles, the gas H

DOI: 10.1021/acs.energyfuels.6b02596 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels channels were formed because of the free-gas flow in the previous CSI production period. The injected reservoir brine moved preferentially through the high-permeability gas channels, which leads to an earlier water BT and a higher WOR. Apparently, test 7 (simultaneous CSI + WF) has the highest average heavy oil production rate and lowest cumulative GOR and WOR. In this test, more solvent remained in the heavy oil due to the extended reservoir pressure maintenance, and the solvent-diluted heavy oil viscosity was low, though the total oil recovery factor was low. 3.3. Different Pressure Drawdown Rates. Three CSI + WF tests were carried out to investigate how a different pressure drawdown rate affects the production performance of CSI + WF. Tests 2−4 had three different pressure drawdown rates of 25.0, 12.5, and 6.8 kPa/min in the CSI production periods, respectively. The measured average heavy oil production rates and cumulative GORs and WORs of CSI + WF in different cycles of tests 2−4 are compared and plotted in Figures 6a−c, and the measured heavy oil RFs in different cycles of tests 2−4 are listed in Table 5. The heavy oil production rate is higher and the cumulative GOR is lower with the increased pressure drawdown rate, as shown in Figures 6a and b. A higher pressure drawdown rate induces faster bubble nucleation and more dispersed gas bubbles, which results in a stronger foamy oil flow and a lower gas mobility.22,23 Although a higher pressure drawdown rate causes a stronger foamy oil flow, the CSI production period is much shortened. For example, the CSI production period in test 2 was only half of that in test 3 if the pressure drawdown rate was decreased from 25.0 to 12.5 kPa/min. Hence, test 3 recovered the most heavy oil during the CSI production period with a moderate pressure drawdown rate of 12.5 kPa/min among the three tests. Both Table 5 and Figure 6c indicate that the heavy oil RF of WF is increased and that the cumulative WOR in the WF is decreased when the pressure drawdown rate in the CSI production period is increased. In this case, more foamy oil is generated and fewer free gas-saturated channels exist at a higher pressure drawdown rate in the CSI production period. Then, the remaining foamy oil is effectively displaced in the subsequent WF. As given in Table 5, the pressure drawdown rate of 12.5 kPa/min in test 3 leads to the highest joint heavy oil RF of 30.1% in CSI + WF due to a proper balance between the efficient foamy oil formation and extended foamy oil production.24,25 3.4. Two Different Solvents (CO2 and C3H8). In this study, four CSI + WF tests with two different solvents (CO2 for tests 3 and 4 and C3H8 for tests 5 and 6) were performed and compared to differentiate CO2-CSI + WF and C3H8-CSI + WF. Tests 3 and 5 had the respective CSI pressure drawdown rates of 12.5 and 3.0 kPa/min but the same CSI production periods of 200 min. Tests 4 and 6 had the respective CSI pressure drawdown rates of 6.8 and 1.5 kPa/min but the same CSI production periods of 366 min. The measured heavy oil RFs in tests 3−6 are summarized in Table 5, and the measured average heavy oil production rates and cumulative GORs and WORs of CSI + WF in tests 3−6 are plotted in Figures 7a−c. As expected, C3H8 can displace more heavy oil because of its larger heavy oil viscosity and density reductions, higher solubility, and larger oil-swelling effect, as shown in Figures 3a−d. Specifically, CSI is repeated for at least 4 cycles with more than 1.0% heavy oil recovered during each CSI cycle of test 5 or 6 (C3H8-CSI + WF). In contrast, the CSI heavy oil RF of cycle 3 in test 3 or 4 (CO2-CSI + WF) is already lower than 1.0%. Obviously, C3H8diluted heavy oil has a much higher mobility due to its lower

Figure 6. Measured (a) average heavy oil production rates in CSI, (b) cumulative GORs in CSI, and (c) cumulative WORs in WF of different cycles in tests 2−4 (CSI + WF) with the pressure drawdown rates of 25.0, 12.5, and 6.8 kPa/min, respectively.

viscosity, which leads to a higher heavy oil production rate and lower cumulative GOR and WOR, as shown in Figures 7a−c. Moreover, Table 5 also indicates that the last 3 WF cycles of test 6 (C3H8-CSI + WF) produce a total of only 0.7% heavy oil. I

DOI: 10.1021/acs.energyfuels.6b02596 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

3.5. Residual Heavy Oil Saturations. In this study, the residual heavy oil saturations at 12−14 locations of the sandpacked physical model were measured at the ends of test 1 (CO2-CSI), test 3 (CO2-CSI + WF), and test 5 (C3H8-CSI + WF), which had the same CSI production period of 200 min. Figures 8a−c show the digital images of the top surfaces of the

Figure 8. Measured residual heavy oil saturations at different locations of the sandpacked physical model at the ends of (a) test 1 (CO2-CSI), (b) test 3 (CO2-CSI + WF), and (c) test 5 (C3H8-CSI + WF).

sandpacked physical model in these three tests. The sampling locations in each test are marked by circles in the figures. It is found from Figure 8a that the measured residual heavy oil saturations range from 68.6 to 83.6% in test 1 (CO2-CSI). The residual heavy oil saturations at the two ends of the sandpacked physical model are slightly higher than those in the center. Some foamy oil formed in the primary pressure-depletion production and the CSI production does not reach the CSI producer. The foamy oil formed far from the CSI producer is difficult to be mobilized and produced because of the lower pressure gradient.25 It can be seen from Figure 8b that in test 3 (CO2-CSI + WF), the residual heavy oil saturations at the two ends of the sandpacked physical model become lower than those in the middle of the sandpacked physical model, which are in contrast to the residual heavy oil saturations in test 1 (CO2-CSI). This indicates that the WF significantly improves the heavy oil recovery in the post-CHOPS reservoirs. Figure 8c shows that more heavy oil is produced in test 5 (C3H8-CSI + WF) than in test 3 (CO2-CSI + WF) due to a stronger oilswelling effect and a larger viscosity reduction of C3H8-diluted heavy oil. Also, there is a sharp displacement front between the WF injector and the CSI producer. By means of the two-well configuration, the residual heavy oil near the WF injector after the CSI production period was effectively displaced by the subsequently injected reservoir brine.

Figure 7. Measured (a) average heavy oil production rates in CSI, (b) cumulative GORs in CSI, and (c) cumulative WORs in WF of different cycles in test 3 (CO2-CSI + WF and tpro = 200 min), test 4 (CO2-CSI + WF and tpro = 366 min), test 5 (C3H8-CSI + WF and tpro = 200 min), and test 6 (C3H8-CSI + WF and tpro = 366 min).

4. CONCLUSIONS In this paper, a total of eight sandpacked laboratory tests were undertaken to explore the technical synergy of combining CSI and WF. Three different production processes were performed

The WF is not effective in the late cycles due to severe water channeling. J

DOI: 10.1021/acs.energyfuels.6b02596 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

(9) Lin, L.; Ma, H.; Zeng, F.; Gu, Y. A Critical Review of the SolventBased Heavy Oil Recovery Methods. Presented at the SPE Heavy Oil Conference Canada, paper SPE 170098; Calgary, Alberta, June 10−12, 2014. (10) Upreti, S. R.; Lohi, A.; Kapadia, R. A.; El-Haj, R. Vapor Extraction of Heavy Oil and Bitumen: A Review. Energy Fuels 2007, 21 (3), 1562−1574. (11) Bardon, C. P.; Karaoguz, D.; Tholance, M. Well Stimulation by CO2 in the Heavy Oil Field of Camurlu in Turkey. Presented at the SPE/DOE Enhanced Oil Recovery Symposium, Paper SPE 14943; Tulsa, OK, April 20−23, 1986. (12) Olenick, S.; Schroeder, F. A.; Haines, H. K.; Monger, T. G. Cyclic CO2 Injection for Heavy-Oil Recovery in Halfmoon Field: Laboratory Evaluation and Pilot Performance. Presented at the Annual Technical Conference and Exhibition of the SPE, paper SPE 24645; Washington, DC, October 4−7, 1992. (13) Yadali Jamaloei, B.; Dong, M.; Mahinpey, N.; Maini, B. B. Enhanced Cyclic Solvent Process (ECSP) for Heavy Oil and Bitumen Recovery in Thin Reservoirs. Energy Fuels 2012, 26 (5), 2865−2874. (14) Jia, X.; Zeng, F.; Gu, Y. Gas Flooding-Assisted Cyclic Solvent Injection (GA-CSI) for Enhancing Heavy Oil Recovery. Fuel 2015, 140 (1), 344−353. (15) Sun, X.; Dong, M.; Zhang, Y.; Maini, B. B. Enhanced Heavy Oil Recovery in Thin Reservoirs Using Foamy Oil-Assisted Methane HuffN-Puff Method. Fuel 2015, 159 (11), 962−973. (16) Dusseault, M. B. Comparing Venezuelan and Canadian Heavy Oil and Tar Sands. Presented at the Petroleum Society’s Canadian International Petroleum Conference 2001, paper 2001-061; Calgary, Alberta, June 12−14, 2001. (17) Chung, F. T.; Jones, R. A.; Nguyen, H. T. Measurements and Correlations of the Physical Properties of CO2/Heavy-Crude-Oil Mixtures. SPE Reservoir Eng. 1988, 3 (3), 822−828. (18) Luo, P.; Gu, Y. Characterization of a Heavy Oil−Propane System in the Presence or Absence of Asphaltene Precipitation. Fluid Phase Equilib. 2009, 277 (1), 1−8. (19) Lu, T.; Li, Z.; Li, S.; Wang, P.; Wang, Z.; Liu, S. Enhanced Heavy Oil Recovery after Solution Gas Drive by Water Flooding. J. Pet. Sci. Eng. 2016, 137 (1), 113−124. (20) Sheng, J. J.; Maini, B. B.; Hayes, R. E.; Tortike, W. S. Experimental Study of Foamy Oil Stability. J. Can. Pet. Technol. 1997, 36 (4), 31−37. (21) Li, S.; Li, Z.; Lu, T.; Li, B. Experimental Study on Foamy Oil Flow in Porous Media with Orinoco Belt Heavy Oil. Energy Fuels 2012, 26 (10), 6332−6342. (22) Sheng, J. J.; Maini, B. B.; Hayes, R. E.; Tortike, W. S. Critical Review of Foamy Oil Flow. Transp. Porous Media 1999, 35 (2), 157− 187. (23) Bera, A.; Babadagli, T. Relative Permeability of Foamy Oil for Different Types of Dissolved Gas. SPE Res. Eval. Eng. 2016, 19 (4), 1− 16. (24) Alshmakhy, A.; Maini, B. B. Effects of Gravity, Foaminess, and Pressure Drawdown on Primary-Depletion Recovery Factor in HeavyOil Systems. J. Can. Pet. Technol. 2012, 51 (6), 449−456. (25) Zhou, X.; Zeng, F.; Zhang, L.; Wang, H. Foamy Oil Flow in Heavy Oil−Solvent Systems Tested by Pressure Depletion in a Sandpack. Fuel 2016, 171 (5), 210−223.

to study the joint EOR and IOR mechanisms in CSI + WF, simultaneous CSI + WF, and WF + CSI. It is found that CSI + WF has the highest heavy oil RF among the three hybrid production processes and CSI alone. In this case, the remaining foamy oil at the end of the CSI production period is effectively displaced and produced in the subsequent WF. It is also found that a higher pressure drawdown rate in CSI helps to increase the heavy oil production rate and reduce the cumulative GOR in CSI and reduce the cumulative WOR in the WF. The highest heavy oil RF is achieved at a moderate pressure drawdown rate with an intermediate production period. Furthermore, C3H8CSI + WF has a higher heavy oil RF and a high heavy oil production rate with better-controlled gas production and lower water consumption compared to those of CO2-CSI + WF. This is because C3H8-saturated heavy oil has the largest viscosity and density reductions and oil-swelling factor in comparison with CH4/CO2-saturated heavy oil. Finally, the measured residual oil saturations at the end of CO2/C3H8-CSI + WF are much lower than those at the end of CSI alone because of the extended solution-gas drive and foamy oil flow in the WF.



AUTHOR INFORMATION

Corresponding Author

*Tel.: 1-306-585-4630; Fax: 1-306-585-4855; E-mail: Peter. [email protected]. ORCID

Yongan Gu: 0000-0002-2619-4566 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors wish to acknowledge the innovation fund from the Petroleum Technology Research Centre (PTRC) and the discovery grant from the Natural Sciences and Engineering Research Council (NSERC) of Canada to Y.G. The technical assistance of the research group members in carrying out the nonstop CSI + WF tests is highly appreciated.



REFERENCES

(1) Alshmakhy, A.; Maini, B. B. A Follow-Up Recovery Method after Cold Heavy Oil Production Cyclic CO2 Injection. Presented at the SPE Heavy Oil Conference Canada, paper SPE 157823; Calgary, Alberta, June 12−14, 2012. (2) Han, G.; Bruno, M.; Dusseault, M. B. How Much Oil You Can Get from CHOPS. J. Can. Pet. Technol. 2007, 46 (4), 24−32. (3) Ostos, A. N.; Maini, B. B. An Integrated Experimental Study of Foamy Oil Flow during Solution Gas Drive. J. Can. Pet. Technol. 2005, 44 (4), 43−50. (4) Chang, J.; Ivory, J. Field-Scale Simulation of Cyclic Solvent Injection (CSI). J. Can. Pet. Technol. 2013, 52 (4), 251−265. (5) Lu, T.; Li, Z.; Li, S.; Li, B.; Liu, S. Performances of Different Recovery Methods for Orinoco Belt Heavy Oil after Solution Gas Drive. Energy Fuels 2013, 27 (6), 3499−3507. (6) Miller, K. A. Improving the State of the Art of Western Canadian Heavy Oil Waterflood Technology. J. Can. Pet. Technol. 2006, 45 (4), 7−11. (7) Kumar, M.; Hoang, V.; Satik, C.; Rojas, D. High-Mobility-RatioWaterflood Performance Prediction: Challenges and New Insights. SPE Res. Eval. Eng. 2008, 11 (1), 186−196. (8) Dong, M.; Huang, S.; Hutchence, K. Methane Pressure-Cycling Process with Horizontal Wells for Thin Heavy-Oil Reservoirs. SPE Res. Eval. Eng. 2006, 9 (2), 154−164. K

DOI: 10.1021/acs.energyfuels.6b02596 Energy Fuels XXXX, XXX, XXX−XXX