Impact of solvent type and injection sequence on Enhanced Cyclic Solvent Process (ECSP) for thin heavy oil reservoirs

Impact of solvent type and injection sequence on Enhanced Cyclic Solvent Process (ECSP) for thin heavy oil reservoirs

Author's Accepted Manuscript Impact of Solvent Type and Injection Sequence on Enhanced Cyclic Solvent Process (ECSP) for Thin Heavy Oil Reservoirs Be...

2MB Sizes 2 Downloads 35 Views

Author's Accepted Manuscript

Impact of Solvent Type and Injection Sequence on Enhanced Cyclic Solvent Process (ECSP) for Thin Heavy Oil Reservoirs Benyamin Yadali Jamaloei, Mingzhe Dong, Nader Mahinpey, Daoyong Yang

www.elsevier.com/locate/petrol

PII: DOI: Reference:

S0920-4105(13)00204-0 http://dx.doi.org/10.1016/j.petrol.2013.08.028 PETROL2476

To appear in:

Journal of Petroleum Science and Engineering

Received date: 12 July 2012 Accepted date: 1 August 2013 Cite this article as: Benyamin Yadali Jamaloei, Mingzhe Dong, Nader Mahinpey, Daoyong Yang, Impact of Solvent Type and Injection Sequence on Enhanced Cyclic Solvent Process (ECSP) for Thin Heavy Oil Reservoirs, Journal of Petroleum Science and Engineering, http://dx.doi.org/10.1016/j.petrol.2013.08.028 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Impact of Solvent Type and Injection Sequence on Enhanced Cyclic Solvent Process (ECSP) for Thin Heavy Oil Reservoirs Benyamin Yadali Jamaloeia, Mingzhe Donga*, Nader Mahinpeya, Daoyong Yangb a

Department of Chemical and Petroleum Engineering, University of Calgary, Calgary, AB, Canada T2N 1N4 b Petroleum Systems Engineering, University of Regina, Regina, SK, Canada S4S 0A2

*

To whom correspondence should be addressed. Telephone: +1 (403) 210-7642. E-mail: [email protected]

Abstract A considerable portion of the western Canada’s heavy oil resides in thin formations. In this situation, thermal methods cannot be used due to heat loss to overburden and underburden. Vapor extraction (VAPEX) fails because of inefficient gravity drainage and low initial production rate. Studies have been done on the cyclic solvent process (CSP) in an attempt to speed up the oil production rate in the solvent injection process. CSP performs poorly because the presence of continuous free methane saturation at the start of production cycles results in high gas mobility, and, consequently, quick methane production, quick pressure depletion, and a significant loss of oil viscosity reduction. As a result, the drive energy becomes depleted by methane production. Also, if low or intermediate initial production pressures are used, the methane solubility in the oil is not high, and the viscosity reduction is not significant. To resolve the above problems of CSP, Yadali Jamaloei et al. (2012) introduced a new process for thin reservoirs – Enhanced Cyclic Solvent Process (ECSP). In ECSP, two types of hydrocarbon solvents are injected separately, in a cyclic manner; one slug is more volatile (methane) and the other is more soluble (propane or ethane) in heavy oil and bitumen. The 1

focus of this study is finding the optimum solvent injection sequence; this will be accomplished through examining the impact of the solvent injection sequence on the performance of ECSP, using different solvent pairs. The experimental results obtained from four series of ECSP tests, each consisting of six cycles, show higher oil recovery and production rate, and lower gas requirement and drawdown when methane is injected before ethane or propane. Wabiskaw formation in the Pelican oilfield in northern Alberta with 17 wells was chosen for performing simulation of ECSP. History matching was conducted for field-scale cumulative oil, gas and water production, and average reservoir pressure. Injection rate and injection time of methane and propane, soaking time and minimum well bottom-hole pressure in the methane-propane ECSP scheme were optimized to predict the field production performance of ECSP. Field-scale simulation revealed that the proposed methane-propane ECSP scheme is a highly effective method for improving heavy oil recovery in thin reservoirs. Keywords: Solvent type; injection order; cyclic hydrocarbon injection; solution gas drive; foamy oil.

1 Introduction Oil reservoirs in Saskatchewan account for almost 62% of Canada’s total heavy oil resource, including 1.7 billion m3 of proven and 3.7 billion m3 of probable reserves. According to Reservoir Annual (Saskatchewan Energy and Mines, 2000), of the province's proven initial heavy oil-in-place, 97% is contained in reservoirs with a less than 10 m pay zone, and 55% in reservoirs less than 5 m thick. Primary and secondary methods, combined, recover, on average, less than 10% of the initialoil-in-place. Hence, the incentive is strong to develop an appropriate oil recovery technique, which will maximize the recovery potential of these thin heavy oil reservoirs (Dong et al., 2006). Heavy oil in thick-pay reservoirs is commonly extracted using thermal recovery methods, such as steam injection and its variants. These methods are generally not suitable for thin heavy oil

2

reservoirs due to heat loss to overburden/underburden, or bottom water zones (Fairfield and White, 1982; Dyer et al., 1994). Hence, large, untapped thin heavy oil resources remaining after recovery by conventional technology have the potential to be tapped into by improved non-thermal recovery techniques. Among the non-thermal methods, vapor extraction (VAPEX) has gained considerable attention over the past two decades (Butler and Mokrys, 1991; Jha et al., 1995). The main shortcomings of traditional VAPEX are its extremely low production rate and unsuitability for thin reservoirs, due to the lack of efficient gravity drainage. Besides VAPEX, cyclic solvent process (CSP) has been studied and tested in laboratories and the field to find a way to speed up the production rate of the solvent injection process (Dong et al., 2006; Ivory et al., 2010; Lim et al., 1995, 1996; Yadali Jamaloei et al., 2012). Different gas injection processes for heavy oil recovery have been examined in the literature. These modes can generally be categorized as slug displacement, water alternating gas (WAG), and CSP (huff-n-puff). Different types of gas can be used for slug displacement, WAG, and CSP, with the most common ones being CO2, flue gas and produced reservoir gas. Previous studies that have tested these gases include those of: Huang et al. (1987), Olenick et al. (1992), Issever et al. (1993), Srivastava et al. (1993), Ma and Youngren (1994), Srivastava et al. (1994), Lim et al. (1995), Lim et al. (1996), Srivastava and Huang (1997), Srivastava et al. (1999), Dong et al. (2006), Ivory et al. (2010), and Yadali Jamaloei et al. (2012). Among the aforementioned works, Olenick et al. (1992) reported the field test of CO2 huff-n-puff with a significant enhancement of heavy oil production. While technically attractive, CO2 huff-n-puff is not considered to be the most economically attractive option for most heavy oil reservoirs, due to the scarcity of natural CO2 sources. Among CO2, flue gas and produced reservoir gas, the produced reservoir gas is the most readily available, cheap, and abundant. Conducting CSP utilizing the produced reservoir gas, therefore, 3

seems to be a more economically viable option for heavy oil recovery. CSP can be implemented in heavy oil reservoirs to recreate the primary-production conditions after termination of either primary depletion or water flood. The idea, therefore, is to make use of viscosity reduction by solvent and restore the solution-gas-drive mechanism through injecting ethane/propane and methane, and then repressurizing the system back into approximate initial reservoir pressure (Dong et al., 2006). Traditional CSP performs poorly because, during the production cycle, oil regains its high viscosity as a significant amount of solvent leaves the oil. Other reasons for the poor performance of the methane CSP are that: (1) if the initial production pressures used are low or intermediate, the methane solubility in the oil is not high and the viscosity reduction is not significant, and (2) the presence of continuous free methane saturation, created by methane injection, at the start of production cycles results in high gas mobility and the pressure can be quickly depleted by methane production. As a result, the drive energy becomes depleted by methane production and very little is left for oil production. To resolve the problems encountered in CSP, Yadali Jamaloei et al. (2012) introduced a new process for in-situ heavy oil and bitumen recovery in thin reservoirs – Enhanced Cyclic Solvent Process (ECSP). ECSP illustrates: (1) How to effectively distribute the viscosityreducing solvent into the oil; and (2) How to keep the oil viscosity low by maintaining some portion of the viscosity-reducing solvent in the oil during the production cycle. In ECSP, two types of hydrocarbon solvents are injected in a cyclic manner, but in two separate slugs; one slug is more volatile (methane) and the other is more soluble (propane or ethane) in heavy oil and bitumen. The volatile solvent in the first slug fingers into the oil to provide paths for the more-soluble solvent slug to mix with oil. For ECSP production cycles, the volatile gas provides a driving force (expansion) by reducing the reservoir pressure, and some portion of the more-soluble solvent stays in the oil to keep the oil viscosity low. 4

The focus of this study is to find the optimum solvent injection sequence. This will be done through examining the impact of the solvent injection sequence on the performance of ECSP using different solvent pairs. Four series of ECSP tests, each consisting of six cycles, were conducted in a visual long sandpack partially filled with methane and a heavy crude oil (with a viscosity of 2246 mPa·s at 22 °C) to mimic the primary depletion phase. After the primary depletion, methane was deployed as the volatile hydrocarbon gas slug, followed or led by propane or ethane as the more soluble solvent slug. Each of the following solvent pairs and injection sequences were used in one series of six ECSP cycles: (1) ethane─methane, (2) methane─ethane, (3) propane─methane, and (4) methane─propane. The experimental results show that the oil recovery and production rate of ECSP increases

in

this

order:

ethane─methane,

methane─ethane,

propane─methane,

and

methane─propane. The results overwhelmingly indicate that the optimum injection sequence in ECSP is to inject the volatile solvent slug before the more-soluble solvent slug, in order to effectively use the viscosity reduction and solvent gas drive mechanisms.

2 Experimental 2.1 Set-up The experimental set-up is comprised of an oil injection system, a gas injection system, a back pressure regulator (BPR) unit for adjusting the production pressure, an oil-gas separator, an effluent gas meter, a visual sandpack, and the flow pressure gauges. Figure 1 shows the schematic and photograph of the experimental set-up and the long visual sandpack. The oil injection system includes a large transfer vessel, which contains the crude oil. A high-precision, low-rate HPLC pump is used to inject the warm crude oil into the sandpack. The gas injection system consists of gas cylinders, two-stage regulators, and a digital programmable mass flow meter and totalizer with display (Omega FMA 4000- Digital Mass Flow Meters). The BPR unit consists of a BPR

5

specifically designed for systems producing heavy oil, a nitrogen cylinder and regulator, sample vessels, and pressure gauges. The oil-gas separator and effluent gas meter are used to measure the amount of oil recovery and produced gas. The visual sandpack, flow pressure gauges and monitoring system form the central unit of the set-up. The visual sandpack is a relatively long, porous medium which simulates a thin formation. Heating belts are connected to the sandpack and flow lines in order to heat up the system while saturating it with the heavy oil. Physical and hydraulic properties of the visual sandpack are given in Table 1. Further details on the experimental set-up can be found in Yadali Jamaloei et al. (2012).

2.2 Materials Physical properties of the heavy crude oil are given in Table 2. Methane, ethane, propane, and nitrogen were supplied by Praxair with purities of 99.99%, 99.0%, 99.99%, and 99.99%, respectively. Crystalline silica (CAS# 14808-60-7) has been used as the sand material to pack the visual porous medium.

2.3 Procedures In order to prepare the sandpack for the primary depletion test and ECSP cycles, the visual sandpack was first packed with sand (crystalline silica). The sandpack was then tested for leaks using methane. After the leakage test, warm crude oil and methane were sequentially injected into the sandpack. At the end of this stage, the warm sandpack was left to cool to the experimental temperature (22 °C). The initial oil and gas saturations in the sandpack, for each run, are given in Table 1. These sets of runs were conducted in dry sandpacks. A separate set of ECSP tests was conducted in the presence of initial oil, gas, and water saturations. The results of ECSP in wet sandpacks have been published by Yadali Jamaloei et al. (2012).

6

Once the sandpack reached the experimental temperature (22 °C), the desired volumes of methane or methane + propane (for primary depletion) or methane/ethane or methane/propane (for ECSP cycles), were injected into the system while all the valves connected to the sandpack were closed, except the injection valve. After termination of gas injection, the injection valve was closed and the sandpack was left to soak. Once the sandpack pressure reached a steady level, production from the sandpack was commenced through the inlet. That is, a single-well ECSP is conducted. The amount of oil and gas production was measured continuously during the production cycle. This procedure was repeated for each ECSP test. As reported in Tables 2 to 5, before conducting ECSP cycles using ethane-methane (Table 2) and methane-ethane (Table 4), the primary depletion was conducted using only methane. As it can be seen in Tables 2 and 4, the two primary depletion tests using methane were not efficient, partly because the sandpack was not initially saturated with live oil. In order to better create a live oil condition, propane was used along with methane in depletion tests before conducting ECSP cycles in propane-methane run (Table 3) and methane-propane run (Table 5). Because the major focus of this study is to compare the results of ECSP cycles using ethane-methane with those using methaneethane, and ECSP cycles using propane-methane with those using methane-propane, applying propane in depletion tests before methane-propane ECSP cycles does not impact the comparison for finding the optimum injection sequence and strategy of the two solvents.

3 Results and Discussion Tables 2 to 5 show summaries of the experimental conditions and results for ECSP, using the solvent

pairs

ethane-methane,

propane-methane,

methane-ethane,

and

methane-propane,

respectively. In order to provide a starting point for suggesting an optimum solvent pair and injection strategy, a PVT study and phase behaviour simulation using CMG-WinProp were

7

conducted for different heavy oil samples and different feeds of solvent mixtures. The detailed results of the PVT study and phase behaviour simulation have been reported in Yadali Jamaloei (2013). For space reason a short summary of PVT study and phase behaviour simulation is given as follows. The experimentally measured saturation pressure, swelling factor and viscosity of different heavy-oil solvent mixtures with respect to the feed mole fraction of methane, ethane, propane and their mixtures were used to tune the Peng-Robinson EOS (PR-EOS) using CMG-WinProp. Then, validated PVT model was used to simulate and predict the behaviour of different heavy oil-solvent systems at various operating conditions. In summary, PVT and phase behaviour simulation for different heavy oil-solvent systems shows that the viscosity of heavy oil samples (with viscosities in the range of 2,000 mPa·s at room temperature) can be reduced to approximately 500 mPa·s by saturating these oil samples with methane or with blends of methane and propane to a pressure of approximately 1,000 psia. However, if only methane is used, oil viscosity increases rapidly with decreasing pressure. On the other hand, if propane is used along with methane in the solvent blend, the increase in oil viscosity with depressurization becomes much slower. Adding propane in the solvent blend can also greatly improve solvent solubility in the oil phase, which in turn maintains a lower oil viscosity during depressurization. The soaking time given in Tables 2 to 5 was selected on the basis of the time needed for reaching equilibrium condition in the sandpack. The recorded data enabled us to determine the recovery factor, recovery rate, solvent requirement, and drawdown for different solvent types and injection sequences. All of this experimental data helps determine the optimum solvent type and injection sequence for ECSP. It is noted that a constant injection pressure for ethane (500 psig), propane (105 psig), and methane (500 psig) has been used, as it is given in Tables 2 to 5. Injection pressure of propane cannot be increased to 500 psig because this is a much greater pressure than 8

propane’s saturation pressure at experimental temperature, meaning that propane will be injected in liquid state. This can adversely affect propane penetration and distribution in the sandpack, and thus, significantly decrease its solubility in oil. Also, the injection pressure of methane, ethane and propane should be as close as possible to and below their corresponding saturation pressure.

3.1 Recovery Factor Figure 2 shows the ultimate oil recovery factor in terms of fraction of original oil in place (OOIP) for different solvent types and injection sequences. The maximum and minimum oil recovery factor was obtained for the solvent pairs methane─propane and ethane-methane. This suggests that the oil recovery factor in ECSP is a strong function of the more soluble solvent type and the injection sequence. The performance of ECSP is significantly enhanced when ethane is replaced with propane. Also, when methane is injected ahead of propane or ethane, they are better distributed in the oil phase. It is therefore essential, when operating an ECSP, that the methane slug is injected first, so that it fingers in the oil to provide paths for the propane or ethane slug to mix with oil. This suggests that the optimum injection sequence in ECSP is to inject methane before propane or ethane, in order to effectively use the viscosity reduction and solvent gas drive mechanisms. Figure 2 also shows the impact of the initial oil saturation of the cycle on the ultimate oil recovery factor for different solvent types and injection sequences. In general, the oil recovery factor of the ECSP tests peaks during the early cycles, and then drops during later cycles. Because all of the ECSP cycles were operated at an equal initial sandpack pressure before soaking (485 psig), the decrease in the oil recovery factor for later cycles is partially attributed to the significant reduction in the initial oil saturation of the cycles, and subsequent decrease in the degree of gas-oil contact, especially around the injection site. Also, the pressure data in Tables 2 to 5 generally

9

indicate that the rate of pressure drop during the soaking time decreases for the later ECSP cycles because the rate of gas dissolution decreases. This lowers the degree of oil swelling and viscosity reduction for the later ECSP cycles. All of these factors contribute to the decrease in the oil recovery factor for later ECSP cycles. 3.2 Recovery Rate As shown in Figure 3, the oil production rate of ECSP increases in this order: ethane─methane, methane─ethane, propane─methane, and methane─propane. This trend confirms that the optimum injection sequence in ECSP is to inject methane before ethane or propane in order to speed up the oil production rate. This injection sequence ensures that the viscosity reduction and solvent gas drive mechanisms are simultaneously utilized, in the most efficient manner. The oil production rate hits a maximum value during the early ECSP cycles and then it decreases for the later cycles. This is caused by the same factors which contribute to the decrease in the oil recovery factor for later ECSP cycles. Another way of speeding up the production rate in ECSP is to replace ethane with propane. In fact, if a sufficient amount of propane is injected, it is very likely to have the foam formation. If the oil creates foam as the gas comes out of the solution, the gas relative permeability is dramatically reduced, leading to a more efficient foamy solution gas drive. Additionally, when ethane is replaced with propane, there will be more efficient gas dissolution, more reduction in the oil viscosity, and a higher degree of oil swelling. All of these help reduce the gas saturation and gas relative permeability during the production cycles of ECSP, which contribute to enhancing the oil production rate.

3.3 Solvent Requirement

10

In general, lower gas requirements are observed in ECSP when a methane slug is injected before ethane or propane. This is implied in the the data in Figures 4 to 6. According to Figure 4, except the first cycle using methane/propane, the amount of injected ethane or propane in ECSP is lower when methane is injected before ethane or propane. Figure 5 also shows that the amount of injected methane decreases when methane is injected first, except for the last ECSP cycles. Combining the amount of gas in the first and second slugs injected, the overall amount of gas required is reduced when a methane slug is injected first. This is because, when a methane slug is injected first, a smaller drawdown occurs, as compared to the case when a methane slug is injected after propane or ethane. When a smaller drawdown occurs, a smaller gas volume is required to repressurize the sandpack back to an equal initial sandpack pressure (485 psig) before the soaking period. In conclusion, lower gas requirements and higher oil recoveries are observed when methane slug is injected first.

3.4 Drawdown, Pressure Profiles, and Cumulative Oil Production Figure 7 shows the drawdown in ECSP cycles, with different solvent types and injection sequences. Drawdown shown on the vertical axis is the difference between the sandpack pressures before soaking and at the end of production divided by the initial sandpack pressure before soaking. Figures 8 and 9 illustrate the pressure profiles during injection, soaking, and production for each ECSP run (including primary depletion and six ECSP cycles) using different solvent pairs and injection sequences. By injecting methane before ethane or propane in ECSP cycles, a relatively smaller drawdown was observed, in comparison to the injection sequence when methane is injected after propane or ethane (Figure 7). Over these smaller drawdown ranges, higher oil recovery factor and production rate were observed (Figure 10).

11

Additionally, the pressure data in Figures 8 and 9 indicate that the rate of pressure drop during the soaking time increases for the ECSP cycles when methane is injected before ethane or propane because of the increase in the rate of gas dissolution. The increase in the rate of gas dissolution occurs because of a better distribution of the ethane or propane in the oil when the methane slug is injected first. For all injection sequences, the rate of pressure drop during the soaking time decreases for the later ECSP cycles, because the rate of gas dissolution decreases due to a decrease in gas-oil contact. This is partially attributed to a significant reduction in the available oil saturation for later ECSP cycles, in particular, around the injection site.

3.5 Oil and Free Gas Flow and Distribution during ECSP Production Cycle The solvent injection sequence has a significant impact on the oil and gas flow during the early stages of the production cycle of ECSP tests. In Figure 11, the dark color represents the developed oil bank and the bright color represents the gas phase released from the oil during the early stages of ECSP production cycle. The results demonstrate that the size of the oil bank developed during the production cycle increases when a methane slug is injected prior to an ethane or propane slug. Furthermore, by injecting a methane slug before ethane or propane, a more uniform distribution of the more-soluble solvent phase is created in the oil. Also, the gas volume that is released from the oil by pressure reduction, during the early production cycle, decreases when a methane slug is injected before ethane or propane. All of these factors contribute to decreasing the free gas phase relative permeability during the production cycle. The fundamental reason behind an inefficient solution gas drive mechanism can be explained in terms of how the released gas saturation from the oil is formed during the early production cycle (Figure 12a), develops and propagates afterwards (Figure 12b), and channels into the producer (Figure 12c) without an effective contribution to the oil production. Figures 12a to 12c depict the

12

dynamics of an inefficient gas drive mechanism, which were observed during the production cycle of ECSP when ethane was injected prior to the methane slug. These dynamics are somewhat improved when ethane is replaced with propane (Figure 12d), partly because of a higher rate of dissolution of propane than that of ethane, and mainly due to a more effective foamy solution gas drive mechanism. In this case, a less continuous, and relatively smaller, gas phase bank forms during the early production cycle (see Figure 12d), which helps decrease the gas phase permeability relative to prior cases depicted in Figures 12a to 12c. The formation of this less continuous, and relatively smaller, gas phase bank (Figure 12d) during the early production cycle does not occur when ethane is replaced with propane and methane is injected prior to propane (Figure 12e). The oil and free gas distribution shown in Figure 12e is what is desired and is ideal for an efficient solution gas drive mechanism in ECSP production cycles.

3.6 Foamy Solution Gas Drive Mechanism Regardless of the solvent injection sequence, the performance of ECSP cycles is significantly enhanced when ethane is replaced by propane. The impressive performance of ECSP cycles using propane is due to a higher degree of the reduction in oil viscosity and a more efficient foam formation, caused by the propane slug (Figures 13c and 13d). Figure 13 depicts samples of the produced foamy oil in ECSP cycles using ethane (images (a) and (b)) and propane (images (c) and (d)). The presence of propane makes the oil foam in a more efficient manner than ethane does during depletion, as seen when Figures 13a and 13b and Figures 13c and 13d are compared. If the oil creates foam as the gas comes out of the solution, the gas relative permeability is dramatically reduced, leading to a more efficient foamy solution gas drive. Additionally, the dissolution of propane not only reduces the oil viscosity more significantly than ethane does, it also causes a higher degree of oil swelling as compared to the dissolution of an ethane slug. All of these factors

13

help reduce the gas saturation and gas relative permeability more efficiently when ethane is replaced with propane. It should be noted that in the experiments presented here, a crude oil with a viscosity of 2,246 mPa·s at 22 ºC was used. Although the oil is less viscous than some Canadian heavy oils which have viscosities ranging from 10,000 to 100,000 mPa·s, it is expected that in the presence of low mobile water saturations the gas injection type and sequence used in this study can potentially keep the evolved solution gas bubbles in the state of foamy oil. The reason for this behavior lies in that the solution gas bubbles that evolve can have a lower coalescence rate in the oil phase due to both higher oil viscosity and low mobile water saturation. Dong et al. (2006) have suggested this solution gas drive behavior for methane pressure-cycling process for heavy oils with different viscosities. The foamy oil would be more stable in more viscous oils than in the less viscous oil. Therefore, the ECSP can be effective for more viscous oils if enough volatile and intermediate solvents are dissolved in the oil for both solution gas drive and viscosity reduction.

3.7 Field-scale Reservoir Simulation A field-scale reservoir simulation was conducted using the injection sequence that resulted in the best ECSP experimental performance (i.e. Methane-Propane). Field-scale oil rate, time, oil, gas, and water saturation distributions, and cumulative productions are provided to demonstrate the field-scale performance of ECSP.

3.7.1. PVT Experiments and Fluid Model Validation In order to validate the PVT model for reservoir simulation, experimental PVT data of methane-propane-heavy oil system including saturation pressure, liquid densities and swelling factor at saturation pressures were used to regress the parameters in the Peng-Robinson Equation of State. The regressed parameters included the molecular weight of the last two pseudo-components, 14

binary interaction parameters between methane and each pseudo-component, and the binary interaction parameters between propane and each pseudo-component. Physical properties of the pseudo-components of heavy oil after lumping are given in Table 6. The validated PVT model was then incorporated in the CMG-STARS to perform field-scale history-matching, operational optimization and field-scale predictions.

3.7.2 Simulation Model Dispersion is a combination of convection and molecular diffusion. In most of the published literature, the diffusion coefficient is considered to be a constant. Yang and Gu (2006) showed that the diffusion coefficients of propane in Lloydminister heavy oil are in the range of 0.9-6.8×10-10 m2/s in the pressure range of 58-130.5 psia (0.4-0.9 MPa) and at temperature of 23.9°C. Dispersion coefficient also depends on concentration, though it is often regarded as a constant. Nghiem et al. (2001) stated that the average value of the convective dispersion coefficients is around 1.4×10-7 m2/s. In this simulation study, an oil-phase dispersion coefficient of 5.0×10-8 m2/s was used in the simulation model. To include the foamy solution gas drive in the simulation model, the non-equilibrium gas dissolution of methane and propane in heavy oil is represented by the following equations (Ivory, et al. 2010), respectively:

(CH 4 ) G + (CH 4 ) L → 2(CH 4 ) L

(1)

(C3 H 8 ) G + (C3 H 8 ) L → 2(C3 H 8 ) L

(2)

where (CH4)L is dissolved methane in oil phase, (C3H8)L is dissolved propane in oil phase, (CH4)G is methane in gaseous phase, and (C3H8)G is propane in gaseous phase. Moreover, the non-equilibrium gas exsolution with the formation of foamy oil in oil is represented by:

15

(CH 4 ) L → BubCH4 → (CH 4 ) G

(3)

(C3 H 8 ) L → BubC3 H 8 → (C3 H 8 ) G

(4)

where BubCH4 and BubC3H8 are methane and propane bubbles in the oil phase, respectively. The following values were used in the numerical simulations: the reaction frequency factor of 0.2 day-1 kPa-1 for gas dissolution, 2 day-1 for gas exsolution from oil phase to bubble, and 2×10-3 day-1 for gas exsolution from bubble to gaseous phase.

3.7.3 Description of Reservoir and Simulation Target Area Pelican oilfield is located 250 km north of Edmonton, Alberta, covering an area of 530 km2, which lies in the middle of the Wabiskaw oil sands deposit (Fontaine et al., 1993; Fossey et al., 1997). The reservoir depth is about 400 m, with in-situ fluid temperature of 17.9-19.9 ºC (291-293 K) and oil gravity of 9°API to 15°API. The Wabiskaw "A" sand is the primary heavy-oil bearing formation with an average net pay of 5.0 m (Fontaine et al., 1993). The initial reservoir pressure, permeability, average porosity and initial average oil saturation of the Wabiskaw "A" formation are 261-348 psia (1800-2400 kPa), 300-3000 mD, 30%, and 70%, respectively. The viscosity of the dead oil produced from the Wabiskaw "A" formation is in the range of 1000-25000 mPa·s (West, 2012). An area of 6.4 km2 enclosing 17 wells was chosen as the target area (Figure 14). The geological model was constructed based on petrophysical properties of the Wabiskaw formation. The pay zone depth and thickness in the target area were extracted from AccuMap, which were interpolated to construct the geological model of the entire formation with a grid block of 32×20×10 (see Figure 15). The reservoir temperature and initial pressure were set at 19.9 ºC (293 K) and 348 psia (2400 kPa), respectively. The oil properties used are the same as those reported in Section 7.3.1. The 16

porosity, permeability, and oil saturation were obtained using the core analysis data for 5 wells from AccuMap, including W-00/02-10, W-00/04-10, W-00/10-10, W-00/11-10, and W-00/16-11. The formation properties of other locations in the reservoir geological model were interpolated based on the data from these 5 wells by applying the Gaussian geostatistical simulation method.

3.7.4 Field-scale History-Match The target area considered in the study includes 17 wells, among which wells W-00/15-11, W04/16-11, W-03/09-11, and W-00/10-11 commenced producing after September 31, 2011. Based on the production data for these 4 wells, it was found that gas injection is initiated to enhance heavy oil recovery from October 1, 2011. The monthly production data of the other 13 wells, including cumulative oil, water, and gas productions were obtained from AccuMap. The production data of these 13 wells from December 1, 2010 to September 31, 2011 show that the fluids from these 13 wells were produced via primary recovery. Therefore, in this study, history match was performed only for these 13 wells which experienced natural production via reservoir-pressure depletion. The material balance method was used to determine the average reservoir pressure (Dake, 2010) as these data were not available in AccuMap. The field oil production rates were used as the input constraints for the production wells. The oil-water and liquid-gas relative permeability curves used in the simulations are plotted in Figure 16. Figures 17 and 18 show the history-matching results of the average reservoir pressure, and the cumulative oil, water and gas productions, respectively. There exists a good agreement between the simulated production histories and the observed field data. The upper 5 layers in Wabiskaw "A" formation are the major producing pay zones. Among the 5 layers, the first layer is not perforated in most of the 13 wells. As such, Figure 19 shows the oil saturation distribution in the second upper layer before and after primary-depletion production,

17

respectively. Because of the low recovery factor (about 5%) achieved in the primary recovery stage, the oil saturation distribution in the second layer remains almost unchanged. Figure 20 depicts the pressure distribution in the second upper layer before and after depletion production, respectively. Although the primary oil recovery is only about 5% of OOIP, the reservoir pressure has experienced a significant decline from 348 psia (2400 kPa) to the bubble point pressure (about 203 psia or 1400 kPa), indicating that there is only a small amount of solution gas dissolved in heavy oil which can only provide limited natural reservoir energy.

3.7.5 Optimization of ECSP Operational Parameters by Orthogonal Design The performance of the ECSP was simulated by using the CMG-STARS simulator (Version 2011). Synergetic mechanisms for improving heavy oil recovery were included, i.e., solution-gasdrive caused by methane injection, viscosity reduction due to propane dissolution and the foamy oil behavior due to the released solvent from heavy oil. Based on the laboratory experiments, the methane-propane-based ECSP scheme was chosen as the suitable scheme for the targeted area. The orthogonal design method (Yang et al., 2011) was used to optimize the operational parameters in the methane-propane ECSP scheme. The following parameters were chosen as the tobe-optimized variables: methane and propane injection rate and time, soaking time and minimum well bottom-hole pressure. Figure 21 shows the five-level values for each of these six parameters. The objective function is the cumulative oil production from December 1, 2010 to September 31, 2012. The simulation results based on the orthogonal design are given in Figure 21. The highest recovery efficiency can be achieved when these six parameters are optimized. Variance analysis of these parameters was also considered in the orthogonal design to understand the relative sensitivity of the cumulative oil production to these six parameters (Yang et al., 2011). The variances for the injection rate and injection time of methane, the injection rate and

18

injection time of propane, the soaking time and the minimum well bottom-hole pressure are calculated to be 5.988×105, 48.175×105, 14.045×105, 65.222×105, 78.160×105 and 1953.656×105, respectively. Therefore, the minimum well bottom-hole pressure is found to be the most sensitive parameter. The injection time of methane and propane, and soaking time are also subject to relatively large sensitivities. The injection rates of methane and propane are the less sensitive parameters since they have smaller variances.

3.7.6 Field-scale ECSP Performance Prediction Based on the orthogonal-experiment design (Figure 21), the following optimum operational parameters for the ECSP were selected to be used in the 13 wells: (1) methane injection rate and time are 200 m3/day and 2 days, respectively, (2) propane injection rate and time are 100 m3/day and 3 days, respectively, (3) soaking time is 3 days, and (4) the minimum well bottom-hole pressure is 72.5 psia (500 kPa). These values were incorporated into the history-matched reservoir model to predict the future field production using methane-propane ECSP scheme. Figure 22 presents the prediction results of cumulative oil and water production during the 30 cycles of the ECSP from October 1, 2011 to May 5, 2017. The cumulative oil production increases quickly with time. Afterwards, the increasing rate in cumulative oil production slows slightly after two years. Meanwhile, the cumulative water production increases monotonically with time. Figures 23 to 25 show the oil, water, and gas distribution in the second layer before (i.e., right after primary-depletion production) and after the ECSP, respectively. The oil saturation decreases in the entire second layer even in areas far from the production wells due to the exsolution of solution gas from heavy oil. Moreover, oil saturation decreases significantly near the wellbore because of the dominancy of near wellbore inflow mechanism, which triggers a good oil production in this area (Figure 23). The water saturation in the second layer after ECSP is much lower than that prior to the

19

ECSP initialization (Figure 24) as a large amount of water is produced during the ECSP (Figure 22). Since gas solvents are cyclically injected and produced from the same well, the gas saturation adjacent to the wellbore experiences a substantial increase (Figure 25). To sum up, this field-scale simulation indicates that the proposed ECSP scheme is a highly effective method for improving the heavy oil recovery.

Conclusions • Oil recovery and production rate of ECSP increases in this order: ethane─methane, methane─ethane, propane─methane, and methane─propane. • Higher oil recovery and production rate, as well as lower gas requirement and drawdown, were observed in ECSP when a methane slug is injected before ethane or propane. • The optimum injection sequence in ECSP is to inject the volatile solvent slug before the moresoluble solvent slug, in order to effectively use the viscosity reduction and solvent gas drive mechanisms. • ECSP performance is enhanced significantly when ethane is replaced with propane, partly because of a more efficient foamy solution gas drive and a higher degree of oil swelling. • Wabiskaw formation in the Pelican oilfield with 17 wells was chosen for implementing the ECSP. History matching was conducted for cumulative oil, gas and water production, and average reservoir pressure. The operational parameters including injection rate and injection time of methane and propane, soaking time and minimum well bottom-hole pressure in the methane-propane ECSP scheme were optimized and incorporated into the history-matched reservoir model to predict the field production performance for the enhanced cyclic solvent process.

20

• The highest oil recovery is obtained when: (1) methane injection rate and time are 200-500 m3/day and 2 days, respectively, (2) propane injection rate and time are 100 m3/day and 3 days, respectively, (3) soaking time is 3 days, and (4) minimum well bottom-hole pressure is 72.5 psia (500 kPa). • Field-scale simulation revealed that ECSP triggers a sharp increase in cumulative oil production over two years and a gradual increase afterwards. Also, the minimum bottom-hole pressure was found to be the most sensitive parameter that should be carefully set in each scenario.

Acknowledgements The authors acknowledge the financial support provided by the Petroleum Technology Research Centre (PTRC), Regina, Canada. A portion (which includes Abstract, Introduction, Section 2, and Subsections 3.1 to 3.6) of this article has been reproduced from the first author’s Ph.D. Thesis (Yadali Jamaloei, 2013). The first author has granted a non-exclusive, for the full term of copyright protection, license to Library and Archives Canada: (a) to reproduce, publish, archive, preserve, conserve, communicate to the public by telecommunication or on the Internet, loan, distribute and sell his thesis (the title of which is set forth above) worldwide, for commercial or non-commercial purposes, in microform, paper, electronic and/or any other formats; (b) to authorize, sub-license, sub-contract or procure any of the acts mentioned in paragraph (a).

References Reservoir Annual, Saskatchewan Energy and Mines, Regina. SK, Canada, 2000. Butler, R.M., Mokrys, I.J. 1991. A New Process (VAPEX) for Recovering Heavy Oils using Hot Water and Hydrocarbon Vapour. J. Can. Petrol. Tech. 30(1). Dake, L. P. Fundamentals of Reservoir Engineering. Elsevier Publishing Company: 2010.

21

Dong, M., Huang, S., Hutchence, K. 2006. Methane Pressure-Cycling Process with Horizontal Wells for Thin Heavy-Oil Reservoirs. SPE Reservoir Eval. Eng. 9(2): 154–164. Dyer, S.B., Huang, S.S., Farouq Ali, S.M., and Jha, K.N. 1994. Phase Behaviour and Scaled Model Studies of Prototype Saskatchewan Heavy Oils with Carbon Dioxide. J. Can. Petrol. Tech. 33(8): 42 48. Fairfield, W.H., White, P.D. 1982. Lloydminster Fireflood Performance, Modifications Promise Good Recoveries. Oil & Gas J. 80(6): 101–102. Fontaine, T., Hayes, L., Reese, G., 1993. Development of Pelican Lake Area Using Horizontal Well Technologies. J. Can. Pet. Technol. 32 (9): 44-49. Fossey, J.P., Morgan, R.J., Hayes, L.A., 1997. Development of the Pelican Lake Area: Reservoir Considerations and Horizontal Technologies. J. Can. Pet. Technol. 36(6): 53-56. Huang, S.S., Pappas, E.S., and Jha, K.N. 1987. The Carbon Dioxide Recovery Process and Its Potential for Saskatchewan Reservoirs. Paper No. 1 presented at the Second Petroleum Conference of the South Saskatchewan Section, Petroleum Soc. of CIM. Regina, 6–8 October. Issever, K., Pamir, A.N., and Tirek, A. 1993. Performance of a Heavy-Oil Field under CO2 Injection, Bati Raman, Turkey. SPE Reservoir Eval. Eng. 8(4): 256–260. Ivory, J., Chang, J., Coates, R., Forshner, K. 2010. Investigation of Cyclic Solvent Injection Process for Heavy Oil Recovery. J. Can. Petrol. Tech. 49(9): 22–33. Jha, K.N., Butler, R.M., Lim, G.B., Oballa V. 1995. Vapour Extraction (VAPEX) Process for Recovery of Heavy Oil and Bitumen. 6th UNITAR International Conference, Houston Texas, 12–17 February. Lim, G.B., Kry, R.P., Harker, B.C., Jha, K.N. 1995. Cyclic Stimulation of Cold Lake Oil Sand with Supercritical Ethane. Paper SPE 30298, International Heavy Oil Symposium, Calgary, Alberta, Canada. 22

Lim, G.B., Kry, R.P., Harker, B.C., Jha, K.N. 1996. Three Dimensional Scaled Physical Modeling of Solvent Vapour Extraction of Cold Lake Bitumen. J. Can. Petrol. Tech. 35(4): 32–40. Ma, T.D. and Youngren, G.K. 1994. Performance of Immiscible Water- Alternating-Gas (WAG) Injection at Kuparuk River Unit, North Slope, Alaska. Paper SPE 28602 presented at the SPE Annual Technical Conference and Exhibition, New Orleans, 25–28 September. Nghiem, L.X., Kohse, B.F., Sammon, P.H., 2001. Compositional Simulation of the VAPEX Process. J. Can. Pet. Technol. 40(8): 54-61. Olenick, S., Schroeder. F.A., Haines, H.K., and Monger-McClure, T.G. 1992. Cyclic CO2 Injection for Heavy-Oil Recovery in Halfmoon Field: Laboratory Evaluation and Pilot Performance. Paper SPE 24645 presented at the SPE Annual Technical Conference and Exhibition, Washington, DC, 4–7 October. Padamsey, R. and Railton. J. 1993. CO2 Capture and Use for EOR in Western Canada, 4. Economic Results and Conclusions. Presented at the IEA Carbon Dioxide Disposal Symposium, U. of Oxford, U.K., 29–31 March. Peng, D., Robinson, D.B., 1976. A New Two-Constant Equation of State. Ind. Eng. Chem. Fundam. 15(1): 59-64. Srivastava, R.K. and Huang. S.S., 1997. A Laboratory Evaluation of Suitable Operating Strategies for Enhanced Heavy Oil Recovery by Gas Injection. J. Can. Petrol. Tech. 36(2): 33–41. Srivastava, R.K., Huang, S.S. and Dong, M. 1999. Comparative Effectiveness of CO2, Produced Gas, and Flue Gas for Enhanced Heavy-Oil Recovery. SPE Reservoir Eval. Eng. 2(3): 238–247 Srivastava, R.K., Huang, S.S., Dyer, S.B., and Mourits, F.M. 1993. A Scaled Physical Model for Saskatchewan Heavy Oil Reserves Design, Fabrication and Preliminary CO2 Flood Studies. Paper No. 33 presented at the Fifth Petroleum Conference of the South Saskatchewan Section, Petroleum Soc. of CIM, Regina, 18–20 October. 23

Srivastava, R.K., Huang. S.S., Dyer, S.B., and Mourits, F.M. 1994. Heavy Oil Recovery by Subcritical Carbon Dioxide Flooding. Paper SPE 27058 presented at the SPE Latin American and Caribbean Petroleum Engineering Conference, Buenos Aires, 27–29 April. West, J., 2012. Performance Review of In-Situ Oil Sands Scheme Approval 9404P. Calgary, AB. Yadali Jamaloei, B., 2013. Enhanced Cyclic Solvent Process (ECSP) for Thin Heavy Oil Reservoirs. Ph.D. Thesis, University of Calgary, Calgary, AB. Yadali Jamaloei, B., Dong, M., Mahinpey, N., Maini, B. B., 2012. Enhanced Cyclic Solvent Process (ECSP) for Heavy Oil and Bitumen Recovery in Thin Reservoirs. Energy & Fuels. DOI: 10.1021/ef300152b. Yang, C., Gu, Y., 2006. Diffusion Coefficients and Oil Swelling Factors of Carbon Dioxide, Methane, Ethane, Propane, and Their Mixtures in Heavy Oil. Fluid Phase Equilib. 243(1-2): 64-73. Yang, X.Y., Ma, C., Deng, D.D., 2011. Orthogonal Test of Injection and Production Parameter Optimization of N2 Simulation. Xinjiang Oil & Gas. 7(2): 63-65.

24

Table 1. Summary of materials and physical properties of the sandpack used in ECSP tests Materials Dead oil viscosity at 22 °C, mPa·s Dead oil density at 22 °C, kg/m3 Dead oil molecular weight, g/mol

2246 970.9 389

Physical properties of visual sandpack in each series of ECSP runs

Length, m Width, m Depth, m Porosity Permeability, Darcy Initial oil saturation, % Initial gas aturation, %

ECSP (Ethane + Methane)

ECSP (Propane + Methane)

ECSP (Methane + Ethane)

ECSP (Methane + Propane)

1.013 0.049 0.032 0.40 43.6 80.8 19.2

1.013 0.049 0.032 0.40 43.6 70.0 30.0

1.013 0.049 0.032 0.39 42.9 82.6 17.4

1.013 0.049 0.032 0.39 43.2 73.9 26.1

Table 2. Summary of experimental conditions and results for ECSP using ethane-methane Run Depletion Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Ethane injected, scm3 1573.7 1494.7 1058.1 1180.3 1180.8 Ethane injection pressure, 500 500 500 500 500 psig 4126.9 2126.8 2464.4 2674.7 2980.4 3128.9 Methane injected, scm3 Methane injection 500 500 500 500 500 500 pressure, psig Initial oil, g 498.2 491.0 471.9 460.0 442.7 434.3 Soaking time, hr 72.0 21.0 22.0 22.5 23.0 22.3 Sandpack pressure before 485.0 485.0 485.0 485.0 485.0 485.0 soaking, psig Sandpack pressure after 323.5 386.0 422.5 448.7 451.6 452.7 soaking, psig Final sandpack pressure, 75.3 106.8 156.6 156.5 160.9 192.5 psig Production time, min 15 15 15 15 15 15 Oil recovered, g 7.2 19.1 11.9 17.3 8.4 5.8

25

Cycle 6 787.0 500 3231.4 500 428.5 23.0 485.0 463.9 199.8 15 4.7

Table 3. Summary of experimental conditions and results for ECSP using propane-methane Run Depletion* Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Cycle 6 3 Propane injected, scm 844.4 402.5 588.9 657.9 717.7 744.9 711.8 Propane injection 105 105 105 105 105 105 105 pressure, psig 3 Methane injected, scm 5223.8 5941.8 6819.1 7557.5 8177.1 8715.6 9158.6 Methane injection 500 500 500 500 500 500 500 pressure, psig Initial oil, g 431.0 406.2 375.9 350.4 329.0 310.4 295.1 Soaking time, hr 22.3 22.5 21.5 22.7 22.5 22.5 22.5 Sandpack pressure before 485.0 485.0 485.0 485.0 485.0 485.0 485.0 soaking, psig Sandpack pressure after 462.3 439.4 446.6 448.9 451.2 452.5 455.2 soaking, psig Final sandpack pressure, 60.9 53.1 54.5 55.3 57.3 61.7 64.1 psig Production time, min 15 15 15 15 15 15 15 Oil recovered, g 24.8 30.3 25.5 21.4 18.6 15.3 14.0 *

Prior to 22.3 hours of soaking before depletion, sandpack was left to soak for 72 hours and pressurized by methane to 485 psig.

Table 4. Summary of experimental conditions and results for ECSP using methane-ethane Run Depletion Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Ethane injected, scm3 1481.7 1133.1 828.5 845.0 790.7 Ethane injection pressure, 500 500 500 500 500 psig 3684.9 1933.9 2221.9 2577.2 2904.1 3137.4 Methane injected, scm3 Methane injection 500 500 500 500 500 500 pressure, psig Initial oil, g 496.8 488.4 472.1 452.0 433.5 420.3 Soaking time, hr 72.0 21.5 21.8 22.4 22.5 22.3 Sandpack pressure before 485.0 485.0 485.0 485.0 485.0 485.0 soaking, psig Sandpack pressure after 319.2 377.1 413.4 429.5 436.9 441.4 soaking, psig Final sandpack pressure, 69.1 129.6 174.1 181.3 190.5 209.5 psig Production time, min 15 15 15 15 15 15 Oil recovered, g 8.4 16.3 20.1 18.5 13.2 10.3

Cycle 6 569.4 500 3319.4 500 410.0 22.8 485.0 451.6 228.4 15 6.1

Table 5. Summary of experimental conditions and results for ECSP using methane-propane Run Depletion Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Cycle 6 Propane injected, scm3 1146.5 497.1 505.8 568.3 601.0 623.1 608.5 Propane injection 105 105 105 105 105 105 105 pressure, psig Methane injected, scm3 4448.2 5227.1 6171.0 7181.5 8012.5 8695.8 9257.5 Methane injection 500 500 500 500 500 500 500 pressure, psig Initial oil, g 444.2 417.3 384.7 349.8 321.1 297.5 278.1 Soaking time, hr 72.0 23.5 23.1 22.8 22.7 22.5 22.3 Sandpack pressure before 485.0 485.0 485.0 485.0 485.0 485.0 485.0 soaking, psig Sandpack pressure after 309.8 423.2 431.5 434.5 438.1 440.3 441.9 soaking, psig Final sandpack pressure, 63.1 68.2 69.3 70.9 72.2 74.5 78.5 psig Production time, min 15 15 15 15 15 15 15 Oil recovered, g 26.9 32.6 34.9 28.7 23.6 19.4 16.3

26

Table 6. Physical properties of the pseudo-components of heavy oil after lumping Pseudo-component C8-11 C12-15 C16-21 C22-29 C30-41 Composition, Mole %

15.64

19.48

21.64

15.21

12.20

7.78

C57+ 8.05

Critical temperature, ºC

339

405.8

485.2

562.2

638.6

709

782.7

Critical pressure, psi

305.8

238.5

191.6

154.6

130.4

117.6

114.5

Molecular weight

137.42

193.00

260.36

358.21

495.25

541.92

781.53

27

C42-56

Temperature Controller

Pressure Gauge Mass Flow Meter Gas Detector

Oil Vessel

Sand Pack

BPR N2 Gas

Liquid–Gas Separator C2H6

CH4

Wet Gas Meter

C3H8

Digital Scale Air Bath

(a) Schematic of the experimental set-up

(b) Actual photograph of the experimental set-up including the oil, brine and gas injection systems

(c) Close-up of the long visual sandpack Figure 1. Experimental set-up: (a) Schematic of set-up, (b) Photograph of set-up, (c) Visual sandpack.

28

R ec ov ery  fac tor, %  OOIP

9 8 7 6 5 4 3 2 1 0

Depletion

Cyc le 1

Cyc le 2

C yc le 3

C yc le 4

Cyc le 5

Methane + Propane

6.06

7.34

7.86

6.46

5.31

4.37

Cyc le 6 3.67

Propane+ Methane

5.75

7.03

5.92

4.97

4.32

3.55

3.25

Methane + Ethane

1.69

3.28

4.05

3.72

2.66

2.07

1.23

Ethane + Methane 

1.45

3.83

2.39

3.47

1.69

1.16

0.94

Figure 2. Ultimate oil recovery factor for ECSP using different solvent types and injection sequences. A total of 14.93%, 18.70%, 34.79%, and 41.07% of OOIP were recovered through ethane + methane, methane + ethane, propane + methane, and methane + propane, respectively.

A v erag e rec ov ery  rate, g /m in

2.5 2 1.5 1 0.5 0

Depletion

Cyc le 1

Cyc le 2

C yc le 3

C yc le 4

Cyc le 5

Methane + Propane

1.79

2.17

2.33

1.91

1.57

1.29

Cyc le 6 1.09

Propane+ Methane

1.65

2.02

1.7

1.43

1.24

1.02

0.93

Methane + Ethane

0.56

1.09

1.34

1.23

0.88

0.69

0.41

Ethane + Methane 

0.48

1.27

0.79

1.15

0.56

0.39

0.31

Figure 3. Ultimate oil recovery rate for ECSP using different solvent types and injection sequences

29

Figure 4. Amount of injected ethane or propane in ECSP with different solvent types and injection sequences

Figure 5. Amount of injected methane in ECSP using different solvent types and injection sequences

30

Figure 6. Amount of total injected gas in ECSP with different solvent types and injection sequences

Figure 7. Drawdown in ECSP with different solvent types and injection sequences. Drawdown is the difference between the sandpack pressures before soaking and at the end of production divided by the initial sandpack pressure before soaking.

31

Figure 8. Pressure profiles in depletion test and ECSP cycles with solvent pair methane-propane

Figure 9. Pressure profiles in depletion test and ECSP cycles with solvent pair methane-ethane

Figure 10. Impact of solvent type and injection sequence on the cumulative oil recovery factor in ECSP

32

(a) Injection sequence: (1) Ethane, (2) Methane

(b) Injection sequence: (1) Methane, (2) Ethane

(c) Injection sequence: (1) Propane, (2) Methane

(d) Injection sequence: (1) Methane, (2) Propane Figure 11. Impact of solvent type and injection sequence on the oil and gas distribution during the early time of the production cycle (flow is from right to left). Dark color is the developed oil bank in the vicinity of the injection-production site on the left-hand side and the bright color is the continuous gas phase bank formed during the early time of production cycle. The size of the developed oil bank increases from (a) to (d). The gas volume released from oil and the gas relative permeability decrease from (a) to (d).

33

(a) Injection sequence: (1) Ethane, (2) Methane. Early time of production cycle.

(b) Injection sequence: (1) Ethane, (2) Methane. Intermediate time of production cycle.

(c) Injection sequence: (1) Ethane, (2) Methane. Late time of production cycle.

(d) Injection sequence: (1) Propane, (2) Methane. A less continuous and smaller free gas phase bank.

(e) Injection sequence: (1) Methane, (2) Propane. Ideal oil and free gas distribution for an effective gas drive. Figure 12. Impact of solvent type and injection sequence on the oil and gas distribution in production cycle (flow is from right to left). Dark color is the developed oil bank and the bright color is the released gas phase formed during the early time of production cycle. Images (a) to (c) depict an inefficient gas drive mechanism.

34

(a) Produced foamy oil in ECSP using ethane

(b) Produced foamy oil in ECSP using ethane

(c) Produced foamy oil in ECSP using propane

(d) Foamy oil in separator in ECSP using propane

Figure 13. Samples of the produced foamy oil in ECSP cycles using ethane (a and b) and propane (c and d). Propane makes the oil foam during depletion in a more efficient manner than ethane does (higher degree of foaming is evident in Figures 13c and 13d than in Figures 13a and 13b)

35

Figure 14. Well patterns used in numerical simulation in the Pelican oilfield

Figure 15. 3D reservoir geological model of the targeted area in the Pelican oilfield

36

Figure 16. Relative permeability curves of oil-water and oil-gas systems used in the simulation model

Figure 17. History matching results of the average reservoir pressure in the targeted area of the Pelican oilfield

Figure 18. History matching results of the cumulative oil, water and gas productions in targeted area of Pelican

37

Figure 19. Oil saturation distribution in the second layer before (LHS) and after (RHS) primary depletion

Figure 20. Pressure distribution in the second layer before (LHS) and after (RHS) primary-depletion production

Figure 21. Factor index for the orthogonal design used for optimization of operational parameters in ECSP

38

Figure 22. Predicted cumulative oil and water productions due to ECSP initialization in Pelican oilfield

Figure 23. Oil saturation distribution in the second layer before (LHS) and after (RHS) the ECSP initialization

Figure 24. Water saturation distribution in the second layer before (LHS) and after (RHS) ECSP initialization

Figure 25. Gas saturation distribution in the second layer before (LHS) and after (RHS) the ECSP initialization

39

Research Highlights • Four series of enhanced cyclic solvent process (ECSP) cycles were conducted; • Optimum injection sequence in ECSP is to inject methane before ethane or propane; • Higher recovery, and lower gas requirement and drawdown were observed when methane is injected before ethane or propane; • ECSP performance is enhanced significantly when ethane is replaced with propane; • Replacing ethane with propane enhances foamy solution gas drive and oil swelling.

40