Analysis of steam–solvent–bitumen phase behavior and solvent mass transfer for improving the performance of the ES-SAGD process

Journal of Petroleum Science and Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Contents lists available at ScienceDirect

Journal of Petroleum Science and Engineering journal homepage: www.elsevier.com/locate/petrol

Analysis of steam–solvent–bitumen phase behavior and solvent mass transfer for improving the performance of the ES-SAGD process Dongqi Ji a, Mingzhe Dong a,b,n, Zhangxin Chen a a b

Department of Chemical & Petroleum Engineering, University of Calgary, Calgary, AB, Canada T2N1N4 College of Petroleum Engineering, China University of Petroleum (East), Qingdao, China 266555

art ic l e i nf o

a b s t r a c t

Article history: Received 5 August 2014 Accepted 2 April 2015

Steam-Assisted Gravity Drainage (SAGD) has been the preferred thermal method for bitumen recovery from reservoirs in western Canada. To save energy and to be more environmentally friendly, ExpandingSolvent SAGD (ES-SAGD) is herein proposed by adding solvent into the injection vapor. The addition of solvent gives rise to different phase behavior (solvent–steam–bitumen) characteristics than that of the steam-only injection (steam–bitumen) process. Early steam condensation, solvent accumulation in the vapor phase, and convective oil flow near the steam boundary are critical mechanisms of the ES-SAGD process. In this paper, the phase behavior of the steam–solvent–bitumen system and solvent mass transfer in oil are studied through a numerical simulation method. Results show that the dominant mechanism of solvent dissolution in oil is by gas–oil equilibrium, rather than condensate mixing. The dissolved solvent is further convectively delivered into the mobile oil zone by gravity drainage. Under high solvent injection concentration, oil production rate is improved by the significant amount of solvent dissolution in oil. The high injection pressure enhances oil production rate through enlarging the mobile oil zone. In addition, this study proposes a steam–solvent injection strategy to improve the ES-SAGD process. & 2015 Elsevier B.V. All rights reserved.

Keywords: oil sands thermal-solvent recovery phase behavior mass transfer operation strategy

1. Introduction The main property of bitumen is its extremely high viscosity under reservoir conditions. To lower bitumen's viscosity sufficiently so that it becomes mobile, two methods are applicable: increasing the temperature of the bitumen by steam injection, or diluting the bitumen by light hydrocarbon component (solvent) dissolution (Gates and Chakrabarty, 2008). To take advantage of temperature and gravity drainage, SAGD was proposed as an insitu bitumen recovery method by Butler and his colleagues in the 1970s (Butler et al., 1981; Butler and Stephens, 1981). With recent advances, SAGD has become commercially viable and has been an extensively deployed method as a bitumen recovery operation in Alberta, Canada. Typically, SAGD consists of a pair of horizontal wells drilled into a bituminous formation. The spacing between a well-pair is primarily governed by the reservoir conditions of net pay thickness, bitumen viscosity, permeability, and heterogeneity (Al-Bahlani and Babadagli, 2009). In a SAGD operation, the production well is usually located about 2 m above the base of the reservoir, with

n

Corresponding author. Tel.: þ 1 403 210 7642; fax: þ 1 403 284 4852. E-mail address: [email protected] (M. Dong).

the injection well drilled parallel to and about 5 m above the producer (Butler et al., 1981; Butler, 1994; Edmunds, 1999; Komery et al., 1999). Steam is introduced into the reservoir through the injection well and a steam chamber is formed within reservoir at the saturated steam temperature. Steam flows and condenses when it comes in contact with cold oil sands at the boundary of the steam chamber. The latent heat of the steam transfers to the surrounding formations and thereby warms up the bitumen. Under the action of gravity, the heated oil and condensate flow to the production well (Butler, 1987; Gates and Leskiw, 2010). As the oil is produced, the evacuated pore space is occupied by injected steam, which results in steam chamber growth (Butler, 1997). The main cost of the SAGD process is the extensive amount of steam consumption as indicated by the Cumulative Steam–Oil Ratio (cSOR), where the amount of steam is expressed in Cold Water Equivalents (CWE). The cSOR indicates the amount of consumed steam per unit of produced bitumen. It has been demonstrated that heavy oil or bitumen can be produced at rates from 100 to 400 m3/day, with the cSOR varying from 2 to 10 m3/m3 (Butler, 1998). Costs are more adverse for a highly fractured reservoir with a typically high cSOR and low oil recovery factor (Zendehboudi et al., 2014). Vapor Extraction (VAPEX), which utilizes a similar well configuration as SAGD, is a process for heavy oil recovery with injection of a mixture of hot water and a low boiling point vaporized solvent

http://dx.doi.org/10.1016/j.petrol.2015.04.005 0920-4105/& 2015 Elsevier B.V. All rights reserved.

Please cite this article as: Ji, D., et al., Analysis of steam–solvent–bitumen phase behavior and solvent mass transfer for improving the performance of the ES-SAGD process. J. Petrol. Sci. Eng. (2015), http://dx.doi.org/10.1016/j.petrol.2015.04.005i

2

D. Ji et al. / Journal of Petroleum Science and Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

into the reservoir. In VAPEX, a solvent vapor chamber is formed and the diluted bitumen flows towards the lower producer by gravity drainage along the boundary of the chamber. The dissolved solvent acts as a carrier of heat when it is boiled off back to the chamber as the drained oil interacts with hot water at the bottom of the reservoir (Butler and Mokrys, 1991, 1993, 1989). Although this process exhibits high energy efficiency, the major drawback of VAPEX is the typically low production rate (Deng et al., 2010). ES-SAGD was proposed by Nasr et al. (2003) to combine the advantages from both SAGD and VAPEX. In the ES-SAGD process, the injection into the reservoir of a hydrocarbon additive, at low concentration in stream, has the potential benefits of both heat introduction and solvent dissolution for reducing the in-situ bitumen viscosity. Solvent is vaporized under the condition of saturated steam and is delivered into the reservoir through the injection well. Near the boundary of the steam chamber, solvent dissolves into oil to enhance oil mobility. The effect of solvent dilution on oil viscosity is apparent in Shu's correlation (Shu, 1984). For example, at 100 1C, the viscosity of bitumen is around 200 cp. When the C6 mole fraction is higher than 0.48, the viscosity of the solvent–bitumen mixture can drop to as low as 10 cp, at the same temperature (Shu, 1984). As a result, oil recovery is effectively improved, as indicated by experimental evidence based on an Athabasca oil sands reservoir, which showed a 20% increase in oil production rate when compared to the SAGD process (Mohammadzadeh et al., 2012). The addition of solvent into the vapor turns the ES-SAGD operation into a more complex process. Previous studies have reported little on the gas–oil equilibrium in the steam–solvent– bitumen system. In particular, there is little information available regarding the phenomenon of solvent dissolution into the bitumen under the action of the equilibrium state in the thermal-solvent gravity drainage process. Previous studies show that hydrocarbon additives with similar saturation properties as steam should be chosen. At the boundary of the steam chamber, solvent condenses with steam simultaneously (Li and Mamora, 2011; Jha et al., 2013). However, Dong investigated the phase behavior of the steam– solvent system and proposed an algorithm to estimate the equilibrium temperature and solvent fraction in the vapor phase at the boundary of the steam chamber (Dong, 2012). It was found that, in a large range of solvent fractions in vapor, steam condenses first from the vapor, and that the condensation of solvent can occur only when the solvent concentration in the vapor phase is extremely high. As further described by Dong (2012), at pressures lower than 2000 kPa, the first condensation from the mixture, which occurs at 210.9 1C, is steam, in a mixture which contains a 0.03 mol fraction of C6 and a 0.97 mol fraction of steam. It was also found that the solvent fraction in the vapor phase increases as steam condenses by means of heat transfer to the cold oil sands. The lowest temperature of vapor is reached at the boundary of the steam chamber. For heavier solvents, at the steam chamber boundary, there would be a lower solvent concentration and a higher temperature, compared to lighter solvents (Dong, 2012). Thus, further study is needed to achieve an improved understanding of steam–solvent–bitumen phase behavior near the boundary of the steam chamber. Furthermore, detailed studies of the mechanism of solvent distribution in the mobile oil zone is scarce. Previous studies established that the diffusion of solvent into bitumen is the mechanism for solvent–bitumen mixing beyond the steam chamber (Gates, 2007; Leyva-Gomez and Babadagli, 2013). However, solvent diffusion into bitumen, which is a very slow process, cannot deliver much solvent into the oil sands under the fast growing steam chamber (Mohammadzadeh et al., 2010). Pore-level studies of the mass transfer mechanism reveal that the mobile oil region thickness of ES-SAGD is greater than that of VAPEX, in which

molecular diffusion is the dominant solvent mass transfer mechanism. In ES-SAGD, the solvent–bitumen mixing is supported by direct gravity drainage (Mohammadzadeh et al., 2010). Simulation studies also show that there is little effect on oil production by varying the solvent diffusion coefficient (Ivory et al., 2008). However, the understanding of how the solvent is distributed in the mobile oil zone needs further study. Examples from industrial operations show the varying success of ES-SAGD. Nexen tested ES-SAGD on Pair 3 of the Long Lake pilot in 2006 (Orr, 2009). From the results, for the co-injection of Jet B (C7 to C12), Keyera condensate (C5 and C6), and C6, each solvent can improve oil production rate to a similar degree: from 17% to 24%. However, the co-injection of butane was found to reduce bitumen production (Orr, 2009). A project managed by Suncor with naphtha co-injection in SAGD in the Firebag area found that naphtha addition has no effect on bitumen production (Nasr and Ayodele, 2006).To achieve a better understanding of ES-SAGD, reservoir simulation was employed to examine the profiles of different reservoir properties, including temperature, gas phase composition, oil phase composition, oil and water saturations, oil viscosity, and oil flow rate in regions near the boundary of the steam chamber. Figures are presented to show important phenomena occurring in those regions and their consequences on oil flow. The solvent distribution and solvent mass transfer in the mobile oil zone were also investigated through reservoir simulation. Furthermore, the effects of operating parameters such as solvent injection concentration and steam injection pressure are studied to investigate improvements in the ES-SAGD process.

2. Reservoir model In this study, a right half 2-D regular Cartesian grid simulation model was generated to investigate the ES-SAGD process using CMG STARS (2011). 2.1. Oil sands model The key reservoir properties are summarized in Table 1. This model is of a single well pair and is homogeneous with respect to permeability, porosity, and initial oil saturation. The modeled reservoir has a depth of 300 m, a thickness of 20 m, and a half width of 30 m. The initial water saturation and oil saturation are 0.2 and 0.8, respectively, and the model has a porosity of 35%, a horizontal permeability of 4 darcies, and a vertical permeability of 2.4 darcies. In the thermal-solvent recovery process, the wettability alteration of a heavy oil reservoir is important and needs to be included (Rao, 1999). Experimental results show that reservoir rock becomes more water-wet with temperature increase from 100 to 500 1F (Sola et al., 2007). Owing to the observation that water-wetness is preferred for high oil recovery (Hascakir et al., 2009), the water-wetness change was examined by (1) alternating end-points of water, oil, and steam saturations, and (2) using various temperature-dependent relative permeability curves in the reservoir simulation (Hascakir and Kovscek, 2010). In addition, experiments demonstrate that wettability can also be changed from oil-wet to preferred water-wet by solvent dissolution in bitumen (Mohammed and Babadagli, 2014). Due to the mostly water-wet reservoir conditions induced by steam and solvent injection, a water-wet rock–fluid system was used in the simulation (Zhao et al., 2013). The initial reservoir temperature is 10 1C and pressure is determined by a hydrostatic method, with a reference pressure of 1210 kPa specified at the top of the reservoir. The bitumen properties are modeled by tuning the Peng–Robinson equation of state (1978) through CMG Winprop (2011), based on the experimental results from Khan et al. (1984), using a similar

Please cite this article as: Ji, D., et al., Analysis of steam–solvent–bitumen phase behavior and solvent mass transfer for improving the performance of the ES-SAGD process. J. Petrol. Sci. Eng. (2015), http://dx.doi.org/10.1016/j.petrol.2015.04.005i

D. Ji et al. / Journal of Petroleum Science and Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

3

Table 1 Key simulation parameters used in the model. Items

Values

Depth to reservoir top (reference depth), m Net pay, m Reference pressure, kPa Initial water saturation, % Initial oil saturation, % Initial reservoir temperature (reference temperature), 1C Horizontal absolute permeability, darcy Vertical absolute permeability, darcy Formation compressibility, 1/kPa Formation heat capacity, J/(m3 1C) Rock conductivity, J/(m day 1C) Water conductivity,J/(m day 1C) Oil conductivity, J/(m day 1C) Gas conductivity, J/(m day 1C) Overburden/underburden volumetric heat capacity, J/(m3 1C) Overburden/underburden thermal conductivity, J/(m day 1C) C6 K-value ¼ (Kv1/P)exp(Kv4/(Tþ Kv5)) Kv1, kPa Kv4,1C Kv5, 1C Effective molecular diffusion coefficient of C6, m2/day Effective dispersivity of C6, m

300 20 1210 20 80 10 4 2.4 7.0E–06 2.0E þ 06 6.6E þ 05 5.4E þ 04 1.2E þ 04 2880 2.4E þ 06 1.7E þ 05 1.0062E þ 6  2697.55  224.37 4.32E  5 0.0002

Bitumen and C6 viscosity versus temperature, cp mC6

mbitumen

T (1C) 10 50 90 130 170 210 250

1,021,354 3657.3 212.6 50.5 13.4 4.9 2.4

Oil–water relative permeability

0.360 0.237 0.178 0.141 0.112 0.100 0.087 Gas–liquid relative permeability

Sw

krw

kro

Sl

krg

krog

0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00

0.0000 0.0002 0.0016 0.0055 0.0130 0.0254 0.0440 0.0698 0.1040 0.1480 0.2040 0.2710 0.3520 0.4470 0.5590 0.6870 0.8340 1.0000

0.9920 0.9790 0.9500 0.7200 0.6000 0.4700 0.3500 0.2400 0.1650 0.1100 0.0700 0.0400 0.0150 0.0000 0.0000 0.0000 0.0000 0.0000

0.15 0.20 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00

1.0000 0.9500 0.8400 0.7200 0.6000 0.4700 0.3500 0.2400 0.1650 0.0930 0.0750 0.0450 0.0270 0.0200 0.0100 0.0050 0.0000 0.0000

0.0000 0.0002 0.0016 0.0055 0.0130 0.0254 0.0440 0.0698 0.1040 0.1480 0.2040 0.2710 0.3520 0.4470 0.0559 0.6870 0.8340 0.9920

method as the one used by Yazdani and Maini (2010). The thermal properties and heat loss parameters of rock and fluids are the same as the data published by Butler (1997). The production well is located 2 m above the bottom of the formation and the injection well is located parallel to and 5 m above the producer. The analysis of grid sensitivity was evaluated and results are shown by the oil production rate versus time plot in Fig. 1. It is seen that the 0.2 m grid system has similar oil rate characteristics as the 0.1 m system, but that the larger grid dimensions give rise to severe deviations. Thus, in both the lateral and vertical directions, grid block dimensions were set to 0.2 m, which appears to be sufficiently accurate for modeling the phase behavior of the solvent–steam–oil system and mass transfer in the region near the boundary of the steam chamber. The model contains one 500 m long grid along the horizontal well.

Fig. 1. Grid sensitivity analysis by the comparison of oil production rates.

Please cite this article as: Ji, D., et al., Analysis of steam–solvent–bitumen phase behavior and solvent mass transfer for improving the performance of the ES-SAGD process. J. Petrol. Sci. Eng. (2015), http://dx.doi.org/10.1016/j.petrol.2015.04.005i

D. Ji et al. / Journal of Petroleum Science and Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

4

Simplifications include that there are no geo-mechanical effects and that both gas cap and bottom water zones are neglected in this model. 2.2. Cases For the SAGD case, steam is injected at 212 1C, with a quality of 90%. The injection well is constrained to a maximum bottom-hole pressure of 2000 kPa and the maximum available surface water rate of 500 m3/day in CWE. The production well is operated at a maximum steam rate of 5 m3/day in order to prevent extensive steam loss. To initialize SAGD, a start-up procedure is simulated for heating up both the injection and production wellbores to ensure a high steam quality at the sand surface (Nasr et al., 2000).For the ES-SAGD case, C6 is injected into the reservoir as a surrogate of solvent with steam (Li and Mamora, 2011; Mohebati et al., 2010), with other parameters the same as the SAGD case. The k-values of C6 at different pressures and temperatures, provided by STARS User's Manual (2011), are used in the simulation. The diffusion coefficient of C6 in oil is approximated based on the work of Butler and Mokrys (1989). The dispersivity is approximated according to the review done by Perkins and Johnston (1963). The cases including steam–solvent–bitumen analysis and subsequent sensitivity studies are summarized in Table 2. 2.3. Phase behavior and oil flow studies in SAGD The mechanisms of SAGD have been extensively studied and thoroughly reported in the literature (e.g., Edmunds and Gittins, 1993; Gates and Leskiw, 2010). A description of the SAGD case in this study establishes a reference for comparison with ES-SAGD, which helps to illustrate the effect of solvent on the SAGD process.

To better understand the SAGD process, the profiles of temperature, saturations, oil viscosity, and oil flow rate versus distance in a lateral section of a SAGD operation are analyzed in detail. Fig. 2a shows the oil saturation profile through a cross section of SAGD. Fig. 2b shows the schematic of four zones (A, B, C, and D) along the study location from the left end to its right end (from inner region of the steam chamber to the original cold oil sand), which are summarized in Table 2 (Mohammadzadeh et al., 2010; Jha et al., 2013). To be more specific, the middle location of the cross section depicted in Fig. 2a is selected as the representative region through which to study SAGD and ES-SAGD. A. Non-condensation zone: this zone has constant temperature and pressure (Fig. 2c); as steam flows, the pores of rock are filled with steam, water, and residual oil. There is no steam condensation. B. Steam condensation zone: as vapor moves toward cold oil sands, steam condenses as the hot vapor comes into contact with the cold oil sands. As shown in Fig. 2c, temperature remains at a constant value up to the chamber boundary. Water saturation increases due to the accumulation of condensate. Gas saturation becomes zero at the boundary. C. Mobile oil zone: with heat transferred to the cold oil sands ahead of the boundary of the steam chamber, oil viscosity is effectively reduced and oil is drained downwards by gravity. As the oil viscosity increases with distance into the mobile zone, oil flow rate decreases (Fig. 2d). There is no gas phase in this zone. The thickness of the mobile oil zone (C) is important because most of the oil drainage occurs in this zone, as shown by the oil flow rate curve in Fig. 2d. For a given thickness, oil flow rate is mainly determined by the amount of heat transferred to the cold oil sands. With a higher temperature at the boundary, heat can be delivered further into the cold oil sands,

Table 2 Cases summary. Operation parameters

SAGD ES-SAGD Solvent diffusion analysis

Solvent dispersion analysis

Solvent injection concentration analysis

Injection pressure analysis

Injection strategy comparison

Case 1 Case 2 Case 3 Case 2 Case 4 Case 5 Case 6 Case 7 Case 2 Case 8 Case 9 Case 10 Case 11 Case 2 Case 12 Case 13 Case 14 Case 15 Case 16 Case 17 Case 2 Case 18 Case 19 Case 20 Case 2 Case 20 Case 13 Case 21 Case 22 (various pressure and solvent concentration)

Injection pressure (kPa)

Solvent injection concentration (mole fraction)

Solvent dispersivity (m)

Solvent diffusion coefficient (m2/day)

2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 2000 1600 2000 2400 2800 3200 2000 3200 2000 3000 3000, 2000, 1400

0.000 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.005 0.010 0.020 0.030 0.040 0.050 0.060 0.010 0.010 0.010 0.010 0.010 0.010 0.010 0.030 0.030 0.000, 0.030, 0.000

/ 0.00020 0.00020 0.00020 0.00020 0.00020 0.00020 0.00002 0.00020 0.00200 0.02000 0.20000 0.00020 0.00020 0.00020 0.00020 0.00020 0.00020 0.00020 0.00020 0.00020 0.00020 0.00020 0.00020 0.00020 0.00020 0.00020 0.00020 0.00020

/ 4.32E  5 4.32E  6 4.32E  5 4.32E  4 4.32E  3 4.32E  2 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5 4.32E  5

Please cite this article as: Ji, D., et al., Analysis of steam–solvent–bitumen phase behavior and solvent mass transfer for improving the performance of the ES-SAGD process. J. Petrol. Sci. Eng. (2015), http://dx.doi.org/10.1016/j.petrol.2015.04.005i

D. Ji et al. / Journal of Petroleum Science and Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

5

Fig. 2. Properties along the middle location of SAGD (Case1) and ES-SAGD (Case 2) at 300 days based on simulation models. (a) Oil saturation distribution in ES-SAGD. (b) Schematic of four zones (A, B, C, and D) along the middle location in ES-SAGD. The blue droplets represent steam condensate and the black droplets represent mobile oil. (c) Gas saturation, oil saturation, water saturation, and temperature along the middle location (5–11 m) in SAGD. (d) Oil viscosity and oil flow rate along the middle location (5–11 m) in SAGD. (e) Gas saturation, oil saturation, water saturation, and temperature along the middle location (7–13 m) in ES-SAGD. (f) Oil viscosity and oil flow rate along the middle location (7–13 m) in ES-SAGD. (g) C6 mole fractions in gas and oil along the middle location (7–13 m) in ES-SAGD. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

D. Ji et al. / Journal of Petroleum Science and Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

6

which enlarges the mobile oil zone (Birrell, 2003; Sharma and Gates, 2011; Irani and Ghannadi, 2013; Irani and Gates, 2013). D. Immobile oil zone: this zone is far from the steam chamber and the temperature is too low to mobilize oil. The slightly higher oil saturation in this zone is due to the oil expansion caused by the somewhat elevated temperature (Fig. 2c). 3. Results and discussion In this section, the process of ES-SAGD is discussed in comparison with the SAGD process. Later, the effects of solvent concentration and steam injection pressure are investigated and a steam– solvent injection strategy to improve ES-SAGD performance is proposed.

3.1. Analysis of the ES-SAGD process In the ES-SAGD case, the C6 concentration in the injected stream is 0.01 mol fraction (on a CWE basis). To better understand the roles of steam and solvent in this process, an analysis similar to the one performed in the SAGD case is done for ES-SAGD as depicted in Figs. 2e–g. Fig. 2e shows the oil, water, and gas saturations, along with temperature at 300 days versus distance along the middle location. The oil viscosity and oil flow rate at the same location are shown in Fig. 2f. The C6 mole fractions in the gas and oil phases are shown in Fig. 2g. In the non-condensation zone (A in Figs. 2e–g, spanning 0–7.6 m), the vapor mixture of steam and C6 flows upwards and laterally at a constant temperature and pressure. In Fig. 2e, water and oil are kept at a residual level. As illustrated in Fig. 2g, most of the C6 remains in the vapor phase; a very minor amount of the C6 is dissolved in the residual oil, under the action of gas–oil equilibrium. The oil is immobile at residual level, as shown by the oil flow rate curve in Fig. 2f. It is obvious that the thickness of the steam condensation zone (B in Fig. 2e–g, spanning 7.6–9.2 m) is greatly enlarged by the coinjection of C6 as compared to the SAGD case (Fig. 2c and d, 6.2– 6.6 m), and is due to the phase behavior of the steam–solvent– bitumen system (Table 3). This is because of the different saturation pressure–temperature relationships between the steam–solvent–bitumen and steam–bitumen systems. In the steam–solvent– bitumen system, partial pressure and solvent solubility in oil play important roles in establishing an equilibrium state. Firstly, steam condensation induces a lowered steam mole fraction and an increased solvent mole fraction in the vapor phase. Consequently, the steam condensation temperature is decreased because of its lowered partial pressure, and there is significant solvent dissolution into the oil because of its solubility. In the steam condensation zone, steam condenses over a certain range of temperature and a large range of solvent concentrations (Dong, 2012). As the latent

heat of vapor is released, water continues condensing out in the direction from the inner steam chamber to the chamber boundary. The temperature of the steam–C6 vapor mixture starts to decrease until the equilibrium state at the boundary is reached, as shown in Fig. 2e. In this zone, steam condensation is revealed by the gradually decreasing gas saturation curve. The difference between the ES-SAGD and the SAGD processes also includes gradually increased oil saturation through C6 dissolution in oil towards the boundary in ES-SAGD. This condensation process results in a gradual increase of the C6 mole fraction in the vapor phase, as illustrated in Fig. 2g. At the boundary of steam chamber, the C6 mole fraction in the vapor phase reaches the maximum value of 0.38. Correspondingly, the C6 mole fraction in oil, which is determined by the gas–oil equilibrium, increases gradually and reaches the highest level of 0.51. Compared to the SAGD case, the oil viscosity is further reduced by solvent dissolution in ES-SAGD. Moreover, the total oil flow rate is increased with the addition of C6 in the condensation zone. This is evidenced by the area beneath the oil flow rate curve in Fig. 2f, which is larger than the area in the SAGD case (Fig. 2d). In the mobile oil zone (C in Figs. 2e–g, spanning 9.2–11.4 m), the condensate and mobilized oil containing dissolved C6 flow downwards. C6 is distributed far from the boundary of the chamber into the mobile oil zone and helps to reduce oil viscosity. The benefits of the solvent in this zone is explained schematically in Fig. 3a and b using the similar drawing methods of Butler (1997) and Irani and Gates (2013), illustrating solvent transfer from the vapor phase further into the mobile oil zone. Firstly, since the steam–solvent mixture is in direct contact with bitumen at the boundary, solvent is dissolved into the oil under the action of gas– oil equilibrium (Fig. 3a). The solubility of solvent in oil is determined by the temperature and the partial pressure of the solvent in the vapor phase near the boundary. Secondly, as oil flows downwards within the sloping mobile oil zone, C6 is convectively transferred further into the oil sands, establishing a solvent rich zone. As the force analysis Fig. 3b shows, gravity (G) on the oil containing dissolved C6 can be represented in terms of two components, F1 and F2. F1 is parallel to the boundary of the steam chamber and F2 is perpendicular to the boundary. As a result, on the one hand, the oil has the potential to flow along the boundary (υ1 ). On the other hand, the oil along with the dissolved C6 flow into the mobile oil zone (υ2 ) perpendicular to the moving boundary of the steam chamber. This flow process is significantly different from the traditional laminar flow between two parallel plates, in which the flow streamline is strictly parallel to the boundary plates. Following the confirmation of mobile water within cold oil sands by field observations and laboratory experiments (Aherne and Maini, 2008; Chan et al., 2012), the flow of water in the direction perpendicular to the boundary of the steam chamber to cold oil sands is indicated in SAGD (Irani and Ghannadi 2013; Irani and Gates 2013). With the multi-phase flow of oil and water, there is the potential for oil flow perpendicular to the

Table 3 Temperatures, thicknesses of the steam condensation and mobile oil zones, C6 mole fractions in gas and oil, and total oil flow rate at 300 days at the top, middle, and bottom locations of the SAGD and ES-SAGD cases. Study location

Temperature at chamber boundary (1C) Steam condensation zone thickness (m) Mobile oil zone thickness (m) C6 mole fraction in vapor at chamber boundary C6 mole fraction in oil at chamber boundary Total oil flow rate (m3/day)

SAGD (Case 1)

ES-SAGD (Case 2)

Top

Middle

Bottom

Top

Middle

Bottom

212 0.4 2.8 / / 59

212 0.4 2.8 / / 72

212 0.4 2.8 / / 104

183 2.6 2.2 0.46 0.70 267

189 1.6 2.2 0.38 0.50 213

200 0.6 2.4 0.20 0.28 187

Please cite this article as: Ji, D., et al., Analysis of steam–solvent–bitumen phase behavior and solvent mass transfer for improving the performance of the ES-SAGD process. J. Petrol. Sci. Eng. (2015), http://dx.doi.org/10.1016/j.petrol.2015.04.005i

D. Ji et al. / Journal of Petroleum Science and Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

7

Fig. 3. Illustration of solvent dissolution and transfer in oil: (a) is the process of solvent vapor contacting oil at a sloping interface; (b) is the movement of dissolved C6 under the analysis of force and velocity and (c) is the process of solvent vapor contacting oil at a vertical interface.

Fig. 4. Cross sectional oil flow streamlines in mobile oil zones based on data from CMG STARS: (a) is oil flow in the sloping mobile oil zone; and (b) is oil flow in the vertical mobile oil zone.

boundary of the steam chamber through the sloping mobile oil zone under the action of gravity. Fig. 4 depicts oil streamlines in the mobile oil zone based on data from CMG STARS. In Fig. 4a, as the oil flows in the sloping mobile oil zone, there is a small angle (θ) between the oil streamline and the chamber boundary. Under the action of F2, the dissolved solvent can be moved deeper into the mobile oil zone. However, in Fig. 4b, the oil streamline is parallel to the boundary when oil flows in the vertical mobile oil zone. One concludes that not much dissolved solvent can be delivered to the mobile oil zone without the component of gravity perpendicular to the chamber boundary. In ES-SAGD, the lateral steam chamber growth rate is accelerated mostly near the top of the reservoir (Deng et al., 2010;

Mohammadzadeh et al., 2010). To understand this process, parameters of the steam chamber, including the C6 mole fractions in gas and oil at the chamber boundary, the thicknesses of the steam condensation and mobile oil zones, and the oil flow rate along the top, middle, and bottom locations of the reservoir at 300 days (Fig. 2a) are summarized in Table 3. At the top, as the steam–C6 mixture flows into the steam condensation zone, there is much steam condensation. This is a result of the fact that, since there is much heat lost to the overburden and in transfer to the cold oil sands, much steam condenses with a high C6 fraction in the vapor (0.46) and also, at the equilibrium state of a high C6 mole fraction in the vapor, a high C6 mole fraction in the oil (0.70) at the chamber boundary. At the

Please cite this article as: Ji, D., et al., Analysis of steam–solvent–bitumen phase behavior and solvent mass transfer for improving the performance of the ES-SAGD process. J. Petrol. Sci. Eng. (2015), http://dx.doi.org/10.1016/j.petrol.2015.04.005i

8

D. Ji et al. / Journal of Petroleum Science and Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

bottom, the C6 mole fraction in the vapor is low (0.20) and accordingly the C6 mole fraction in the oil is also low (0.28) at the boundary, due to the small amount of heat release and steam condensation. As a result, there is more C6 at the top than in the bottom of the mobile oil zone. From top to bottom, even though the thickness of the mobile oil zone slightly increases from 2.2 m to 2.4 m with the increase of boundary temperature from 183 to 200 1C, the steam chamber grows faster at the top (higher solvent dissolution) than at the bottom (lower solvent dissolution) because of the more significant effect of C6 dissolution on oil mobility. Near the boundary of the steam chamber, C6 returns to the vapor phase from the oil phase as the oil flows downwards in the mobile oil zone. This is due to the phase behavior change in character from the top to the bottom of the reservoir. As a result, some of the mobile oil becomes immobile as it drains into the lower part of the reservoir. At 300 days, the total oil rate decreases from 267 to 187 m3/day because of oil flowing downwards from the top to the bottom (Table 3). In conclusion, in addition to heat transfer, which is one mechanism for oil viscosity reduction in ES-SAGD, the oil production rate is significantly improved by solvent dissolution. Since there is essentially no solvent condensation (Dong, 2012), the amount of dissolved solvent is determined by the gas–oil equilibrium state near the boundary of the steam chamber. Deep in the oil sands, where the gas saturation is zero, the solvent is convectively delivered through the oil drainage process in the mobile oil zone. From the top to the bottom of the reservoir, the amount of dissolved solvent in oil is gradually reduced, controlled by gas–oil equilibrium. 3.2. Sensitivity analysis to improve ES-SAGD To achieve a better understanding of the mechanisms involved in ES-SAGD, the effect of solvent diffusion and dispersion in oil are examined by a simulation method. The operating parameters which have a close relationship with and influence steam–solvent–bitumen phase behavior, including solvent concentration and injection pressure, are investigated in this section. In addition, a steam–solvent injection strategy to improve ES-SAGD performance is proposed using an alternating solvent injection period, solvent concentration, and pressure. 3.2.1. Effect of diffusion and dispersion on ES-SAGD An analysis was performed to understand the effect of solvent diffusion on the ES-SAGD process. The typical diffusion coefficient of solvent in bitumen is estimated to be in the range of 8.64  10  6 –4.32  10  5 m2/day, according to the test of propane and butane diffusion in Peace River bitumen (Das and Butler, 1996). An estimation of the solvent diffusion distance is proposed as follows (Butler and Mokrys, 1989): Z 1 C s;max Ds ξ¼ dC s U C s;min C s where ξ is the distance of solvent diffused, U is the advance velocity of the boundary of the steam chamber, C s is the solvent concentration, and Ds is the diffusion coefficient of solvent in the oil phase. Assume that the steam chamber is moving at 1.7 cm/day (U) (Irani and Gates, 2013; Irani and Ghannadi, 2013), C s is 0.51 at the vapor–liquid interface (C s;max ) and 0.01 at end of diffusion (C s;min ), as shown in Fig. 2g, and that the C6 diffusion coefficient is 4.32  10  5 m2/day (Ds ). As a result, the diffusion distance is estimated to be only 0.01 m, which is two orders of magnitude less than the thickness of the mobile oil zone observed in Fig. 2g (2.2 m). In this study, a simulation is done to examine whether there is a significant effect on oil recovery by either the diffusion or

the dispersion process. Fig. 5a shows the simulated oil recovery factor at 600 days (end of production) with different C6 diffusion coefficients. It is noted that the recovery factor is increased by only 2.1% (i.e., from 69.4% to 71.5%) when the C6 diffusion coefficient is increased by a factor of 10,000 (i.e., from 4.32  10  6 to 4.32  10  2 m2/day). Fig. 5b shows the oil recovery factor at 600 days with a different dispersivity. It is seen that the oil recovery factor changes little when the dispersivity is increased from 0.00002 to 0.2 m. Thus, diffusion and dispersion appear to have negligible effects in the ES-SAGD process. 3.2.2. The impact of solvent concentration on production performance Solvent concentration is an important parameter, which has a significant effect on oil production rate (Li and Mamora, 2011). In this study, C6 is injected at 0.005, 0.010, 0.020, 0.030, 0.040, 0.050, and 0.060 mol fractions in the stream (Table 2), with the other parameters remaining the same as in the ES-SAGD base case. The oil recovery factor at 600 days (end of production) is greatly improved with an increase in C6 concentration, as shown in Fig. 5c. Table 4 summarizes the temperature at the boundary of the steam chamber, thicknesses of the steam condensation and mobile oil zones, and C6 mole fractions in vapor and oil. The thickness of the steam condensation zone is enlarged with an increase in C6 concentration, since more steam condenses under the lowered saturation temperature of the vapor mixture and greater distance to reach equilibrium at the boundary. The thickness increases from 0.8 to 2.4 m with the change of C6 injection mole fraction from 0.005 to 0.040. Furthermore, the C6 mole fraction in the vapor at the boundary increases from 0.23 to 0.46 and the C6 mole fraction in oil also rises from 0.26 to 0.80. The thickness of the mobile oil zone decreases slightly from 2.4 to 2.2 m as a result of the temperature at the boundary decreasing from 199 to 183 1C. When the C6 injection mole fraction varies from 0.005 to 0.040, the increased solvent dissolution in the oil more than compensates for the reduced thickness of the mobile oil zone, resulting in the total oil flow rate being increased. As illustrated by Fig. 5c, when the C6 mole fraction is higher than 0.030, the oil recovery factor improvement is small because the oil viscosity cannot be further reduced to any significant degree because of the prevailing very high C6 concentration in the oil. Fig. 5e shows the bitumen-C6 viscosity as a function of C6 mole fraction based on Shu's correlation (1984). At any given temperature, this correlation, which is particularly applicable to heavy oil–solvent systems without excessive asphaltene precipitation, is dependent only on the densities and viscosities of crude oil and solvent (Shu, 1984). Although this method correlates mixture viscosity of bitumen and solvent based on a limited temperature range of 25–82.2 1C (Shu, 1984; Barrufet and Setiadarma, 2003), it has been found that this correlation can be reasonably used for bitumen–solvent viscosity estimation at a wide range of temperatures based on experimental evidence of insensitivity of this correlation to temperature (Shu, 1984; Gates, 2007). From Fig. 5e, oil viscosity changes very little when the C6 concentration increases from 0.4 to 0.7 mol fraction at a relatively high temperature (180 1C), which is the temperature near the chamber boundary. However, since some of the injected steam is replaced by solvent, during a high solvent concentration injection regime, the significant decrease of injected heat may lead to a decrease in the oil flow rate. 3.2.3. The impact of pressure on production performance Pressure is another important operating parameter as it affects steam chamber growth and oil production rate (Law et al., 2003). Five scenarios (Table 2), including 1600, 2000, 2400, 2800, and

Please cite this article as: Ji, D., et al., Analysis of steam–solvent–bitumen phase behavior and solvent mass transfer for improving the performance of the ES-SAGD process. J. Petrol. Sci. Eng. (2015), http://dx.doi.org/10.1016/j.petrol.2015.04.005i

D. Ji et al. / Journal of Petroleum Science and Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

9

Fig. 5. Sensitivity analysis results of ES-SAGD at 600 days: (a) is oil recovery factor versus C6 diffusion coefficient in ES-SAGD with 2000 kPa injection pressure; (b) is oil recovery factor versus dispersivity in ES-SAGD with 2000 kPa injection pressure; and (c) is oil recovery factor versus C6 injection concentration in ES-SAGD with 2000 kPa injection pressure; and (d) is oil recovery factor versus injection pressure in ES-SAGD with 0.01 mol fraction C6 co-injection; and (e) is oil viscosity versus C6 mole fraction in oil at 140, 160, 180, and 200 1C in ES-SAGD with 2000 kPa injection pressure (Shu’s correlation, 1984).

3200 kPa, are considered, with a 0.01 mol fraction of C6 injected. As shown in Fig. 5d, the oil recovery factor at 600 days (end of production) increases from 65.5% to 75.0% when the injection pressure is changed from 1600 to 3200 kPa. To explain how the oil recovery factor is improved by a higher injection pressure, the temperature, thicknesses of zones, and C6 mole fractions in vapor and oil along the middle location (Fig. 2b) are summarized in Table 4. Since C6 is injected at the same concentration, the thickness of the steam condensation zone is similar among the various injection pressure cases due to the similar phase behavior in this zone. Furthermore, C6 concentration in oil at the chamber boundary is also similar. With the injection pressure increasing from 1600 to 3200 kPa, temperature at the boundary increases from 181 to 219 1C because of an increase in the steam partial pressure. As a result, the thickness of the mobile oil zone is enlarged from 2.0 to 2.6 m by virtue of more heat being transferred to the oil sands. Therefore, it can be concluded that the

oil flow rate increase under high injection pressure is mainly the result of an enlarged mobile oil zone and lowered oil viscosity. From the sensitivity analysis of solvent injection concentration and pressure, it is found that solvent concentration has the greatest effect on the phase behavior of the steam–solvent–bitumen system, and that pressure has the greatest effect on heat distribution ahead of the boundary of the steam chamber. Due to the change of solvent fraction in the vapor, the variations in partial pressures of steam and solvent, which are related to condensation temperatures, result in significant differences in the amount of steam condensation and solvent dissolution in oil. For example, as solvent concentration changes from 0.005 to 0.06, the thickness of the condensation zone increases from 0.8 to 2.6 m (greater steam condensation) and solvent fraction in oil increases from 0.26 to 0.80 at the boundary (Table 4). The injection pressure mostly affects the thickness of the mobile oil zone owing to the relationship of saturated steam temperature and pressure at the boundary. Under high pressure (without a change in

Please cite this article as: Ji, D., et al., Analysis of steam–solvent–bitumen phase behavior and solvent mass transfer for improving the performance of the ES-SAGD process. J. Petrol. Sci. Eng. (2015), http://dx.doi.org/10.1016/j.petrol.2015.04.005i

D. Ji et al. / Journal of Petroleum Science and Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

10

Table 4 Temperature, thickness of the steam condensation and mobile oil zones, C6 mole fractions in gas and oil at the boundary of steam chamber at 300 days at the middle location in ES-SAGD with varies solvent concentrations and injection pressures. Properties at the chamber boundary

Solvent injection concentration analysis

Injection pressure analysis

Case Case Case Case Case Case Case Case Case Case Case Case

11 2 12 13 14 15 16 17 2 18 19 20

Temperature C6 mole fraction in vapor (1C)

C6 mole fraction in oil

199 189 188 184 183 180 179 172 184 192 201 214

0.26 0.50 0.64 0.65 0.70 0.76 0.80 0.67 0.65 0.68 0.68 0.64

0.23 0.38 0.40 0.44 0.46 0.49 0.51 0.47 0.44 0.45 0.43 0.35

Steam condensation zone thickness

Mobile oil zone thickness

(m)

(m)

0.8 1.6 2.0 2.2 2.4 2.6 2.6 2.2 2.2 2.0 1.8 1.8

2.4 2.2 2.2 2.2 2.2 2.0 2.0 2.0 2.2 2.4 2.6 2.6

Fig. 6. Saturation and oil flow properties along the middle location of ES-SAGD (140 days). (a) Oil saturation distribution in cross section of ES-SAGD with 2000 kPa injection pressure and 0.01 mol fraction C6 co-injection. The dashed line is at the middle of the reservoir. (b) Gas saturation, C6 mole fractions in vapor and oil, and oil flow rate at 140 days along the middle location (0.5–4.5 m) in ES-SAGD with 2000 kPa injection pressure and 0.01 mol fraction C6 co-injection.

the solvent mole fraction in the vapor), the high temperature at the boundary of the steam chamber induces more heat to be transferred to the oil sands and thereby results in a larger mobile oil zone. Small variations are found in solvent dissolution in the oil. Thus, for the ESSAGD improvement strategy, varying the solvent injection concentration is a means of improving solvent efficiency and varying the injection pressure is a means of improving the heat efficiency.

3.2.4. Proposed steam–solvent injection strategy to improve ES-SAGD 1. To better utilize steam and solvent so as to improve ES-SAGD production performance, an extensive sensitivity study was

carried out to investigate the effects of solvent concentration, pressure, and solvent injection start time on the process. As a result, the following injection strategy steps are proposed: Vertical growth of steam chamber: introduce steam only into the reservoir under high pressure (such as 3000 kPa), and continue until the steam chamber reaches the overburden of the reservoir. High pressure increases vertical steam chamber growth and oil rate, and the “steam only” condition acknowledges that solvent addition provides little benefit to oil recovery improvement in this stage. Fig. 6a shows oil saturation at 140 days during the vertical steam chamber growth period (ES-SAGD case). As before, the study location is selected at the middle of the reservoir. Fig. 6b shows gas saturation, C6 mole fractions in

Please cite this article as: Ji, D., et al., Analysis of steam–solvent–bitumen phase behavior and solvent mass transfer for improving the performance of the ES-SAGD process. J. Petrol. Sci. Eng. (2015), http://dx.doi.org/10.1016/j.petrol.2015.04.005i

D. Ji et al. / Journal of Petroleum Science and Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

vapor and oil, and oil flow rate along the middle location. In the mobile oil zone (C, 2.0–2.6 m in Fig. 6b), the transfer distance of the dissolved C6 through the vertical chamber boundary is much shorter (0.4 m, from 2.0 to 2.4 m as shown by the curve of C6 mole fraction) than that from the sloping vapor–liquid interface (2.2 m in Fig. 2g). This is because, through the vertical chamber boundary, C6 cannot be transferred deeply by diffusion without gravitationally convective flow. 2. Lateral growth of steam chamber: A co-injectant should then be injected with the steam when the chamber starts to grow laterally. The recommended co-injectant is C6, which has been proven as the best choice for a steam–solvent process by Li and Mamora (2011), and it should be injected at a relatively high concentration (such as 0.03 mol fraction) with steam, under a moderate pressure (such as 2000 kPa). With such a high C6 injection concentration, the oil production rate can be significantly enhanced by much solvent dissolution in the sloping mobile oil zone. A moderate injection pressure, which by equilibrium considerations implies a moderate temperature, can prevent excessive heat loss to the overburden. 3. Continuous production: As the slope of the chamber boundary becomes gradually flat, only steam is injected under a relatively low pressure (such as 1400 kPa) to continue the production. In this process, most of the solvent can be recovered at the surface by dissolved solvent vaporization (lowered partial pressure of solvent).

To demonstrate the advantages of the proposed injection strategy, several cases are compared as summarized in Table 2. In Case 22, the proposed injection strategy is used. Fig. 7 compares oil recovery factors, cSOR, lost C6 per unit of produced bitumen at 600 days (end of production), and net profits calculated based on the data in Table 5 of the five cases. Case 22 has the second highest oil recovery factor, the second lowest cSOR, extremely low solvent loss, and the highest economic efficiency. Due to the key challenge of low heat efficiency in existing SAGD operations, typified by a heterogeneous reservoir, thin reservoir thickness, and thief zones, such as top and bottom water layers,

11

the proposed strategy, based on the new findings of this study, would provide a direction for improving heat and economic efficiency by reducing steam consumption, decreasing steam chamber temperature, regulating steam chamber pressure, and minimizing solvent loss within the reservoir. Due to the limitations of numerical simulation, more extensive studies need to be conducted, such as including oil sands wettability alterations under drastic temperature variations during steam injection, and compositional changes brought on by solvent injection (Sola et al., 2007; Hascakir and Kovscek, 2010; Mohammed and Babadagli, 2014). Further, since the formation of emulsions result from steam condensate contacting oil near the boundary of the steam chamber (Mohammadzadeh et al., 2012; Zendehboudi et al., 2014; Kar et al., 2014), the role of emulsions on oil flow and solvent mass transfer also needs more study.

4. Conclusions 1. A new insight and greater understanding are herein provided for steam–solvent–bitumen system phase behavior near the boundary of the steam chamber in ES-SAGD. Solvent dissolves into the mobile oil by gas–oil equilibrium, rather than by condensate mixing. 2. An innovative theory is presented which indicates that the dissolved solvent propagates further into the mobile oil zone mainly through convective oil flow under the influence of gravity. 3. Solvent can only be significantly transported to the mobile oil zone through a sloping steam chamber boundary; the mass transfer of solvent into the oil through a vertical boundary is negligible due to the lack of fluid convection. 4. A proposed ES-SAGD stepwise strategy, namely high pressure steam injection, solvent–steam co-injection, and low pressure steam injection, can effectively improve SAGD operating performance to achieve a high economic efficiency.

Acknowledgments The financial support from the Natural Sciences and Engineering Research Council (NSERC) of Canada and the Doctoral Program of Higher Education of China (20110133110007) are gratefully acknowledged. An acknowledgment and appreciation are expressed to Dr. Zhaowen Li for his assistance in this study. References

Fig. 7. Oil recovery factor, cSOR, C6 loss per unit of produced bitumen, and net profitability of cases with various injection strategies.

Table 5 Details of typical costs (Deng, 2005). Operation costs Steam generator operating Fuel for steam generator Treatment of water production C6 price Heavy oil price

$2.83/m3 (CWE) $10.1/m3 (CWE) $1.96/m3 $1100/m3 $703/m3

Aherne, A.L., Maini, B.B., 2008. Fluid movement in the SAGD process: a review of the Dover Project. J. Can. Pet. Technol. 47 (1), 31–37. Al-Bahlani, A.M., Babadagli, T., 2009. SAGD laboratory experimental and numerical simulation studies: a review of current status and future issues. J. Pet. Sci. Eng. 68 (3–4), 135–150. Barrufet, M.A., Setiadarma, A., 2003. Reliable heavy oil–solvent viscosity mixing rules for viscosities up to 450 K, oil–solvent viscosity ratios up to 4  105, and any solvent proportion. Fluid Phase Equilib. 213 (1–2), 65–79. Birrell, G., 2003. Heat transfer ahead of a SAGD steam chamber: a study of thermocouple data from phase b of the Underground Test Facility (Dover Project). J. Can. Pet. Technol. 42 (3), 40–47. Butler, R.M., McNab, G.S., Lo, H.Y., 1981. Theoretical studies on the gravity drainage of heavy oil during in-situ steam heating. Can. J. Chem. Eng. 59 (4), 455–460. Butler, R.M., Stephens, D.J., 1981. The gravity drainage of steam-heated heavy oil to parallel horizontal wells. J. Can. Pet. Technol. 20 (2), 90–96. Butler, R.M., 1987. Rise of interfering steam chambers. J Can. Pet. Technol. 26 (3), 70–75. Butler, R.M., Mokrys, I.J., 1991. A new process (VAPEX) for recovering heavy oils using hot water and hydrocarbon vapour. J. Can. Pet. Technol. 30 (1), 97–106. Butler, R.M., Mokrys, I.J., 1993. Recovery of heavy oils using vapourized hydrocarbon solvents: further development of the VAPEX process. J Can Petrol Technol 32 (6), 56–62.

Please cite this article as: Ji, D., et al., Analysis of steam–solvent–bitumen phase behavior and solvent mass transfer for improving the performance of the ES-SAGD process. J. Petrol. Sci. Eng. (2015), http://dx.doi.org/10.1016/j.petrol.2015.04.005i

12

D. Ji et al. / Journal of Petroleum Science and Engineering ∎ (∎∎∎∎) ∎∎∎–∎∎∎

Butler, R.M., 1994. Steam-assisted gravity drainage: concept, development, performance and future. J. Can. Pet. Technol. 33 (2), 44–50. Butler, R.M., 1997. Thermal Recovery of Oil and Bitumen, third edition GravDrain Inc., Canada. Butler, R.M., 1998. SAGD comes of age!. J. Can. Pet. Technol. 37 (7), 9–12. Butler, R.M., Mokrys, I.J., 1989. Solvent analog model of steam-assisted gravity drainage. AOSTRA J. Res. 5 (1), 17–32. Chan, S., Chen, Z., Dong, M., 2012. Experimental and numerical study of initial water mobility in bitumen reservoirs and its effect on SAGD. J. Pet. Sci. Eng. 92–93, 30–39. CMG STARS Users’ Manual, version 2011.10. Computer Modeling Group Ltd., Calgary, Alberta, Canada. CMG Winprop Users’ Manual, version 2011.10. Computer Modeling Group Ltd., Calgary, Alberta, Canada. Das, S.K, Butler, R.M., 1996. Diffusion coefficients of propane and butane in Peace River bitumen. Can. J. Chem. Eng. 74 (6), 985–992. Deng, X., 2005. Recovery performance and economics of steam/propane hybrid process. In: Paper SPE 97760 presented at SPE/PS-CIM/CHOA International Thermal Operations and Heavy Oil Symposium, Calgary, Alberta, Canada, November 1–3. Deng, X., Huang, H., Zhao, L., Law, D.H.S., Nasr, T.N., 2010. Simulating the ES-SAGD process with solvent mixture in Athabasca reservoirs. J. Can. Pet. Technol. 49 (1), 38–46. Dong, L., 2012. Effect of vapour–liquid phase behaviour of steam-light hydrocarbon systems on steam assisted gravity drainage process for bitumen recovery. Fuel 95, 159–168. Edmunds, N.R., 1999. On the difficult birth of SAGD. J. Can. Pet. Technol. 38 (1), 14–17. Edmunds, N.R., Gittins, S.D., 1993. Effective application of steam assisted gravity drainage of bitumen to long horizontal wells. J. Can. Pet. Technol. 32 (6), 49–55. Gates, I.D., 2007. Oil phase viscosity behaviour in expanding-solvent steam-assisted gravity drainage. J. Pet. Sci. Eng. 59 (1–2), 123–134. Gates, I.D., 2008. Chakrabarty, N. Design of steam and solvent injection strategy in expanding-solvent steam assisted gravity drainage. J. Can. Pet. Technol. 47 (9), 12–20. Gates, I.D., Leskiw, C., 2010. Impact of steam trap control on performance of steamassisted gravity drainage. J. Pet. Sci. Eng. 75 (1–2), 215–222. Hascakir, B., Acar, C., Akin, S., 2009. Microwave-assisted heavy oil production: an experimental approach. Energy Fuels 23, 6033–6039. Hascakir, B., Kovscek, A.R., 2010. Reservoir simulation of cyclic steam injection including the effects of temperature induced wettability alteration. In: Paper SPE 132608 presented at SPE Western Regional Meeting, Anaheim, California, USA, May 27–29. Irani, M., Gates, I.D., 2013. Understanding the convection heat-transfer mechanism in steam-assisted-gravity-drainage process. SPE J. 18 (6), 1202–1215. Irani, M., Ghannadi, S., 2013. Understanding the heat-transfer mechanism in the steam-assisted gravity-drainage (SAGD) process and comparing the conduction and convection flux in bitumen reservoirs. SPE J. 18 (1), 134–145. Ivory, J., Zheng, R., Nasr, T., Deng, X., Beaulieu, G., Heck, G., 2008. Investigation of low pressure ES-SAGD. In: Paper SPE 117759 presented at SPE/PS/CHOA International Thermal Operations and Heavy Oil Symposium, Calgary, Alberta, Canada, October 20–23. Jha, R.K., Kumar, M., Benson, I., Hanzlik, E., 2013. New insights into steam/solventcoinjection-process mechanism. SPE J. 18 (5), 867–877. Kar, T., Williamson, M., Hascakir, B., 2014. The role of asphaltenes in emulsion formation for steam assisted gravity drainage (SAGD) and expanding solvent – SAGD (ES-SAGD). In: Paper SPE 171073 presented at SPE Heavy and Extra Heavy Oil Conference, Latin America, September 24–26.

Khan, M.A.B., Mehrotra, A.K., Svrcek, W.Y., 1984. Viscosity models for gas-free Athabasca bitumen. J. Can. Pet. Technol. 23 (3), 47–53. Komery, D.P., Luhning, R.W., O’Rourke, J.C., 1999. Towards commercialization of the UTF project using surface drilled horizontal SAGD wells. J. Can. Pet. Technol. 38 (9), 36–43. Law, D.-H.S., Nasr, T.N., Good, W.K., 2003. Field-scale numerical simulation of SAGD process with top-water thief zone. J. Can. Pet. Technol. 42 (8), 32–38. Leyva-Gomez, H., Babadagli, T., 2013. Numerical simulation of heavy-oil/bitumen recovery by solvent injection at elevated temperatures. J. Pet. Sci. Eng. 110, 199–209. Li, W., Mamora, D.D., 2011. Solvent-type and ratio impacts on solvent-aided SAGD process. SPE Reserv. Eval. Eng. 14 (3), 320–331. Mohammadzadeh, O., Rezaei, N., Chatzis, I., 2010. Pore-level investigation of heavy oil and bitumen recovery using solvent-aided steam assisted gravity drainage (SA-SAGD) process. Energy Fuels 24 (12), 6327–6345. Mohammadzadeh, O., Rezaei, N., Chatzis, I., 2012. Production characteristics of the steam-assisted gravity drainage (SAGD) and solvent-aided SAGD (SA-SAGD) processes using a 2-D macroscalephysical model. Energy Fuels 26 (7), 4346–4365. Mohammed, M., Babadagli, T., 2014. Alteration of matrix wettability during alternate injection of hot-water/solvent into heavy-oil containing fractured reservoirs. In: Paper SPE 170034 presented at SPE Heavy Oil ConferenceCanada, Alberta, Canada, June 10–12. Mohebati, M.H., Maini, B.B., Harding, T.G., 2010. Numerical evaluation of hydrocarbon additives to steam in the SAGD process. J. Can. Pet. Technol. 49 (9), 42–55. Nasr, T.N, Law, D.H.S., Golbec, H., Korpany, G., 2000. Counter-current aspect of the SAGD process. J. Can. Pet. Technol. 39 (1), 41–47. Nasr, T.N., Ayodele, O.R., 2006. New hybrid steam–solvent processes for the recovery of heavy oil and bitumen. In: Paper 101717 presented at Abu Dhabi International Petroleum Exhibition and Conference, Abu Dhabi, UAE, November 5–8. Nasr, T.N., Beaulieu, G., Golbeck, H., Heck, G., 2003. Novel expanding solvent-SAGD process ‘ES-SAGD. J. Can. Pet. Technol. 42 (1), 13–16. Orr, B., 2009. ES-SAGD: Past, present and future. In: Paper SPE 129518 presented at SPE Annual Technical Conference and Exhibition held in New Orleans, Louisiana, USA, October 4–7. Perkins, T.K., Johnston, O.C., 1963. A review of diffusion and dispersion in porous media. SPE J. 3 (1), 70–84. Rao, D.N., 1999. Wettability effects in thermal recovery operations. SPE Reserv. Eval. Eng. 2 (5), 420–430. Sola, B.S., Rashidi, F., Babadagli, T., 2007. Temperature effects on the heavy oil/water relative permeabilities of carbonate rocks. J. Pet. Sci. Eng. 59, 27–42. Sharma, J., Gates, I.D., 2011. Convection at the edge of a steam-assisted-gravitydrainage steam chamber. SPE J. 16 (3), 503–512. Shu, W.R., 1984. A viscosity correlation for mixtures of heavy oil, bitumen and petroleum fractions. SPE J. 24 (3), 277–282. Yazdani, A., Maini, B.B., 2010. Measurements and modelling of phase behaviour and viscosity of a heavy oil/butane system. J. Can. Pet. Technol. 49 (2), 9–14. Zendehboudi, S., Rajabzadeh, A.R., Bahadori, A., Chatzis, I., Dusseault, M.B., Elkamel, A., Lohi, A., Fowler, M., 2014. Connectionist model to estimate performance of steam-assisted gravity drainage in fractured and unfractured petroleum reservoirs: enhanced oil recovery implications. Ind. Eng. Chem. Res. 53 (4), 1645–1662. Zhao, D.W., Wang, J., Gates, I.D., 2013. Optimized solvent-aided steam-flooding strategy for recovery of thin heavy oil reservoirs. Fuel 112, 50–59.

Please cite this article as: Ji, D., et al., Analysis of steam–solvent–bitumen phase behavior and solvent mass transfer for improving the performance of the ES-SAGD process. J. Petrol. Sci. Eng. (2015), http://dx.doi.org/10.1016/j.petrol.2015.04.005i