PETROLEUM SOCIETY

PAPER 2008-064

New Investigations into Carbon Dioxide Flooding by Focusing on Viscosity and Swelling Factor Changes
M. ENAYATI
Iranian Offshore Oil Company

E. HEIDARYAN
Islamic Azad University-Masjidsolayman Branch B. Mokhtari Iranian Offshore Oil Company
This paper is accepted for the Proceedings of the Canadian International Petroleum Conference/SPE Gas Technology Symposium 2008 Joint Conference (the Petroleum Societys 59th Annual Technical Meeting), Calgary, Alberta, Canada, 17-19 June 2008. This paper will be considered for publication in Petroleum Society journals. Publication rights are reserved. This is a pre-print and subject to correction.

Abstract
Carbon dioxide ( CO2 ) flooding is an efficacious method of enhanced oil recovery (EOR) that has nowadays become one of the most important EOR processes. It is a very complicated process, involving phase behavior that could increase oil recovery by means of swelling, evaporating and lowering oil viscosity. The present investigation reports the results of extensive experimental and theoretical work (with the aim of computer software, ECLIPSETM) to determine the viscosity and swelling factor changes of the live oil in the Cheshmeh Khoshk reservoir at southern of Iran (Ilam District) and also minimum miscible pressure. In this study we setup a series of slim tube experiments. In order to get representative fluid samples of a reservoir, it was necessary that the right operation of mixing the separator oil and gas samples to match the bubble point pressure be carried out. And, the potential application of the study is that we could have a good estimate of the recovery improvement under CO2 gas injection, which will be the basic input parameters for the economic feasibility study and also a

decision can then be made whether to implement or abandon the prospective project.

Introduction
Carbon dioxide gas as an injection fluid into oil reservoirs has been a recognized well and tested as Enhanced Oil Recovery (EOR) method, because CO2 dissolves easily into oil, it reduces oil viscosity, and it can extract the light components in oil at sufficiently high pressure, and it can become miscible with oil at very low pressure [1]. The injection of carbon dioxide for secondary and tertiary oil recovery has received considerable attention in the industry because of its high displacement efficiency and relatively low cost [2]. It appeared in 1930s and had a great development in 1970s. Over 30 years production practice, CO2 flooding has become the leading enhanced oil recovery technique for light and medium oil. It can prolong the production lives of light or medium oil fields nearing depletion under waterflooding by 15

to 20 years, and may recover 15% to 25% of the original oil in place [3]. Experience gained from CO2 flooding worldwide indicates that enhanced oil recovery by using CO2 as injection gas may result in additional oil ranging from 7 to 15 % of the oil initially in place [4]. CO2 flooding process involves very complicated phase behavior, which depends on the temperature, pressure and fluids properties of a certain reservoir. Many factors have been found contributing to the oil recovery in CO2 flooding. These mainly include: Low interface tensions, Viscosity reduction, Oil swelling, Formation permeability improvement, Solution gas flooding, and Density change of oil and water [3]. In the case of viscosity, there is a reduction for carbon dioxide as long as temperature increases. Appreciable solubility of carbon dioxide in the crude oil reduces its viscosity. At low temperatures, viscosity reduction for light oils is more than heavy oils [5]. Screening criteria have been proposed elsewhere for selecting reservoirs where CO2 may sustain or increase the production of oil. They estimate that upwards of 80% of oil reservoirs worldwide might be suitable for CO2 injection based upon oil-recovery criteria alone. Moreover, the process is widely applicable in both sandstone and carbonate formations with a variety of permeabilities and thickness of hydrocarbon bearing zones. The major factors limiting CO2 injection as an oil recovery process have been availability of CO2 and the cost to build pipelines to carry CO2 into oil producing regions [6]. Over 8,000 Alberta pools were first screened for CO2 -flood suitability, and pertinent reservoir properties were used for the remaining 4,729 pools to calculate oil recovery. The predicted recoveries for all pools ranged from 1.2-13.9%, 6.3-18.7% and 11.8-27.1% at breakthrough and 0.25 and 0.5 Hydrocarbon Pore Volume (HCPV) injection respectively. These values compared well to an average of 13% incremental oil recovery from the field experience of CO2 floods [7]. However, before a further decision be made on whether to apply it in field or not, generally some laboratory experiments need to be done, one of them being a series of slim tube experiments. The experiments on slim tube will indicate the microscopic efficiency of the injection process. They need to be combined with the macroscopic sweep efficiency and invasion efficiency obtained from the reservoir characterization, to have an overall injection efficiency of the process. Since experiments on slim tube at high pressures are costly, time consuming and prone to experimental failures, it is of great interest to simulate those experiments with a numerical simulator. With a numerical simulator, besides the economic and time benefits, we could also calculate results on certain conditions, which otherwise would have never been possible with experiments due to technical constraints [1]. We took a laboratory study of CO2 miscible flooding on one of the Iranian southern reservoir oil. From the slim tube generated and also data calculated finally by ECLIPSETM software it was shown that viscosity reduction and oil swelling by CO2 contributed to oil recovery. The viscosity showed an almost linear decrease with CO2 concentration. The same study taken by R.K Srivastava et al. [8] has shown also a linear decrease of viscosity with CO2 concentration.

Results and Discussion The Preparation of Fluid Sample

Reconstitution of reservoir fluid is a crucial procedure. In fact, it is very difficult to get the representative fluid sample of a reservoir [3]. In this experiment, the separator oil and gas samples were combined to reconstitute the reservoir fluid. For the separator oil and gas samples, there are two important steps which may introduce errors: the sampling operation; and the reconstituting Gas-Oil Ratio (GOR). For the separator oil and gas samples, the most important parameter to reconstitute the reservoir fluid is the Gas-Oil Ratio [3]. Our recombination method was based on gas oil ratio. The reported field GOR has to be matched with the recombined sample. In this work the reservoir fluid has GOR of 105 cc/cc and separator pressure is 320 psia and separator temperature is about 80oC. The recombination method involves Measurement of solution gas existed in separator oil, Measurement of oil formation volume factor of separator oil at specific working pressure (Bo), and Measurement of gas formation volume of separator gas. First of all, we adjust the pressure of both oil and gas cells up to the predetermined working pressure. In this case the separator oil is taken to the PVT cell and increase the pressure up to 1000 psia and room temperature was about 32oC. Then 50 cc of this oil is transferred to another cell and release the pressure to the standard condition and the following result was calculated, GORsep = 24.91 cc/cc, Bo at 1000 psia = 1.128 cc/cc The gas formation volume factor of gas cell was measured by same method, a specific volume of gas was transferred to a cell which has a volume of 368 cc and its pressure was about 230 psia and the result was, Bg at 230 psia and 32 oC = 0.05929 cc/cc Therefore for 176cc oil at 1000 psia, 16346.713cc gas at standard condition is needed. This amount of gas is equivalent to 969.19cc at 230 psia. This procedure associated with some errors and deficiency. Both oil and gas cell was stationary and the segregation of components would cause error. Second error was the temperature variation in location. The variation of temperature would make some condensate from separator gas which could be accumulated at bottom of the cell. We have error of existing air in lines, since we didnt have enough connections to make the lines vacuum.

Validity of Recombined Sample

The bubble point pressure of oil is an important property of oil that is sensitive to the GOR. So the better method to get eligible reservoir fluid is mixing the separator oil and gas samples to match the bubble point pressure. We calculated the bubble point pressure in room temperature and reservoir temperature. The room temperature was tested to see whether the prepared sample is single phase or not. And bubble point at reservoir temperature would confirm the prepared sample for validity with true reservoir oil. Also bubble point in some temperatures between room and reservoir temperature was measured to see its variation with respect to temperature increase. Table 1 shows the composition of oil and gas separator and also reservoir oil. Figure 1 and Figure 2 show the bubble point calculation base on PVT data shown in Table 2 and Table 3. The corresponding measured bubble point pressure at reservoir
2

temperature is about 2542 psia, which is very close to 2650 psia given by the field. Figures 3 in accordance with Table 4 shows that the bubble point will increase due to increasing of temperature and it is nearly linear in our study range (90 F217 F).

show the output result of the software. In a report published by Reeves S.R. [12] the optimized differential swelling factor in CO2 flooding determined is 1.65 which confirms our estimation (S.F.ave = 1.70 ) . The Multiple Contact Miscibility Pressure (MCMMP or just MMP) (Mathiassen O.M. 2003) was also calculated by ECLIPSETM software to compare with the experimental value of MMP. It is 3375.7 psia which is in a good agreement with the slim tube result (3432 psia).

Slim Tube Tests and Its Results

Since the establishment of Multi-Contact Miscibility (MCM) is a dynamic process, a dynamic experiment is required to demonstrate MCM behavior [9]. The slim-tube apparatus has historically been used as the dynamic means of miscibility determination or in the other word to determine the MMP. The MMP is the pressure at which the reservoir fluid is expected to develop multi-contact miscibility with CO2 . It is the most important factor in flooding process and determines the reservoir operating pressure [3]. Figure 4 shows the high pressure slim-tube apparatus which is used in the present investigation. For beginning we washed the slim tube and its lines by first vacuuming the lines and then injecting gasoline into lines. After that the gasoline was displaced with the oil sample at the pressure of 5000 psia and the slim tube became completely filled with the oil sample. Then CO2 gas was charged to the drive gas cell and allowed to equilibrate at the temperature and pressure of the run. The drive gas was then displaced into the saturated glass-bead-packed coil using a mercury injection pump. The fluids flowed through the coil and the sight glass and then broke out at the ambient pressure and temperature. The produced gas and liquid were measured as a function of time using a gasometer and a calibrated collection vessel. The back pressure cell was also charged with N2 and water in the working pressure of the test (5000 psia). The inlet pressure, outlet pressure, differential pressure, oil recovery and amount of gas released from oil were recorded. For each test final recovery was calculated and corresponding figures are shown through Figures 5 to 10. Table 5 represents obtained data from final recovery. The recovery of the 4900 psi displacement was 50.9 cc but as a matter of fact this value is lower than the 4500 psi recovery, and this would be in contradict with the theoretical concept that more pressure so more recovery. Consequently this recovery is less than what would be expected and so correction is necessary. In order to get better and smoother curve, this value was modified to the 51.9 cc with so called awareness. In fact, this correction introduces 50 psi that is not considerable respect to other errors in the work. Figure 11 illustrates the main graph for MMP evaluation. From this chart we can easily specify by quick-look that the MMP is more than 3400 psi and less 3500 psi. But the exact value is 3432 psi. The same study taken by Javadpour et al. [9] for eight systems of Iranian oil reservoir has shown an average MMP of 3887 psia.

Conclusions
1. The better method to get eligible reservoir fluid is recombination of the separator oil and gas samples base on gas/oil ratio of reservoir fluid to match the bubble point pressure. The oil swelling and the reduction in viscosity are two major factors in enhanced oil recovery by CO2 flooding. Knowing the slim tube data, the swelling factor and viscosity in each step of pore volume of CO2 injection can easily calculated by ECLIPSETM software. Figure 5 shows a nearly linear decrease of viscosity during CO2 flooding as the concentration of CO2 increases. EOR with CO2 injection is mainly attributed to multicontact miscibility due to its low MMP in our case of study. (MMP reported by the instructor for separator gas recycled injection of this reservoir is about 5000 psia). A comparison of MMP estimated by slim tube with MMP calculated by ECLIPSETM software shows low error (1.63%). So in the cases that we have a short time only a test can be run by slim tube instead of several runs and the data obtained give to the computer simulator in order to calculate MMP. The results of our study with the results of other gas injection projects (such as N2, methane, and etc.) can be used as a basic input parameter for the economic feasibility study and also a decision can then be made whether to implement or abandon the prospective project or which type of injection lead to a better performance.

2. 3.

4.

5.

6.

7.

Acknowledgements
The authors thank the Pars Special Economy Energy ZoneResearch program for their financial and technical support. Discussions with Dr. Seyed Abdoljalil Razavi, concerning the economic analysis inputs, were much appreciated.

REFERENCES
Abdassah, D., Kristanto, D., The Potential of Carbon Dioxide Gas Injection Application in Improving Oil Recovery, Paper SPE Presented at the SPE Seventh International Oil and Gas Conference and Exhibition held in Beijing, China, November (2000) 64730. [2] Stakup, F.I., Carbon Dioxide Miscible Flooding Miscible Flooding: Past, Present and Outlook for the Future, J. Pet. Tech. August (1978) 1102-1112. [3] Yongmao, H., Zenggui, W., Yueming, J.B.Ch., Xiangjie, L., Laboratory Investigation of CO2 Flooding, Paper SPE Presented at the 28th Annual SPE International Technical
3

[1]

Simulator Results
CO2 injection reduces the viscosity of oil and trend of viscosity reduction can be calculated with Darcy equation also [10,11]. Using PVT apparatus we can find the viscosity reduction of oil and changing in Swelling Factor (S.F) in each step of pore volume injected. But because of the large number of calculations and also time consuming work we used ECLIPSETM software for calculation of viscosity and swelling factor in different amount of gas added. Figure 12 and figure 13

[4]

[5]

[6]

[7]

[8]

Conference and Exhibition in Abuja, Nigeria, August (2004) 88883. Mathiassen, O.M., CO2 as Injection Gas for Enhanced Oil Recovery and Estimation of the Potential on the Norwegian Continental Shelf, MSc. Thesis, Trondheim, May (2003) 2. Paitakhti-Oskouie, Tabatabaei-Nezhad, Mechanisms of Oil Recovery by Non-hydrocarbon Gas Injection, 13th Oil, Gas & Petrochemical Congress with particular emphasize on Improved Oil Recovery, Tehran , Iran, January (2005) 2. Jessen, K., Sam-Olibale, L.C., Kovscek, A.R., Orr, F.M., Increasing CO2 Storage in Oil Recovery, First National Conference on Carbon Sequestration, sponsored by National Energy Technology Laboratory, Washington DC, May (2001) 2. Shaw, J.C., Bachu, S., CO2 Flooding Performance Prediction for Alberta Oil Pools, Paper Presented at the Petroleum Societys Canadian International Petroleum Conference (2002) 2002-026. Srivastava, R.K., Huang, S.S., Dong, M., Laboratory Study of Weyburn CO2 Miscibile Flooding, JCPT, February (2000) 41.

[9] Javadpour, F.G., Jamialahmadi, M., Shadizadeh, S.R., Investigation of Hydrocarbon Miscible Gas Injection by Experimental and Modeling Approaches for Iranian Oil Reservoirs, Paper SPE Presented at the 1998 SPE India Oil and Gas Conference and Exhibition held in New Delhi, India, February (1998) 39552. [10] Dinarvand. N., Mokhtari, B., Enayaty, M. Mahzari, P., Laboratory Investigation of Carbon Dioxide Flooding by Focusing on the Viscosity and the Swelling Factor Changes for one of the Iranian Southwestern Oil Reservoirs, 2nd National Petroleum Congress, Petroleum University of Technology, Ahvaz, Iran, February (2008). [11] Mokhtari, B., Pourabdollah, K., Heydaryan, E., Enayati, M., Laboratory Investigation of CO2 Flooding by Focusing on the Viscosity and Swelling Factor Changes, 2nd Professional Conference of Environment Engineering, The University of Tehran, Tehran, Iran, Accepted for 2122 May (2008). [12] Reeves, S.R., Reservoir Simulation Modeling of the Yubari CO2 -ECBM/ Sequestration Pilot, Ishikari Basin, Japan, Prepared for Japan Coal Energy Center (JCOAL) CO2ECBM Symposium, Tokyo, Japan, February (2007).

Table 1: Separator Oil, Separator Gas and Reservoir Oil Properties

Molecular Weight CO2 44.10 C1 16.04 C2 30.07 C3 44.10 I-C4 58.12 N-C4 58.12 I-C5 72.15 N-C5 72.20 C6 86.17 C7+ 256 Total 736.98 SG. C7+= 0.9015 MW C7+= 256 MW Reservoir Oil= 108.05 (Calculated) Component

1st Stage Separator Gas 0.90 76.47 14.11 5.55 0.65 1.33 0.32 0.29 0.26 0.12 100.00

1st Stage Separator Oil 0.17 6.37 5.88 6.53 1.44 4.93 2.67 2.64 5.33 64.04 100.00

Reservoir Oil 0.51 38.61 9.67 6.08 1.08 3.27 1.59 1.56 3.00 34.63 100.00

Table 2: Bubble Point Data at Room Temperature

Pump Reading (cc) 149.44 149.08 148.67 148.27 147.83 147.40 146.91 146.75 146.61

Pressure Reading (psia) 5960 5520 5010 4550 4040 3550 3040 2830 2710

Pump Reading (cc) 146.42 146.24 146.07 145.86 145.46 145.06 144.26

Pressure Reading (psia) 2520 2320 2130 2050 1990 1930 1820

Table 3: Bubble Point Data at Reservoir Temperature

Pump Reading (cc) 126.1 125.62 125.16 124.7 124.23 123.99 123.74

Pressure Reading (psia) 4700 4250 3840 3435 3030 2810 2615

Pump Reading (cc) 123.49 122.54 121.49 120.49 119.49

Pressure Reading (psia) 2530 2405 2270 2160 2060

Table 4: Bubble Point at Different Temperature

Temperature (F) 90 130 150 180 217

Bubble Point (psia) 2076.6 2200 2287.1 2421.7 2541.5

Table 5: Recovery Data

Injection Pressure (psia) 3000 3300 4000 4200 4500 4900

Recovery (cc) 44.9 47.5 50 50.5 51.3 51.9

Figure 1: Bubble Point Pressure at Room Temperature (90 F)

Figure 2: Bubble Point Pressure at Reservoir Temperature (217F)

2600

Bubble Point (psia)

2400 P = 3.7663T + 1727.6 R2 = 0.9947 2200

2000 80 120 160 Temperature (F) 200 240

Figure 3: Variation of Bubble Point Pressure with Temperature

Figure 4: Schematic Diagram of High Pressure Slim Tube Apparatus [9]

50 40 Produced Oil (cc) 30 20 10 0 0 0.2 0.4 0.6 PV Injected 0.8 1 1.2

Figure 5: Production Profile at 3000 psia

50

40 Produced Oil (cc)

30

20

10

0 0 0.5 PV Injected 1 1.5

Figure 6: Production Profile at 3300 psia

60 50 Produced Oil (cc) 40 30 20 10 0 0 0.2 0.4 0.6 PV Injected 0.8 1 1.2

Figure 7: Production Profile at 4000 psia

60 50 Produced Oil (cc) 40 30 20 10 0 0 0.2 0.4 0.6 PV Injected 0.8 1 1.2

Figure 8: Production Profile at 4200 psia

60 50 Produced Oil (cc) 40 30 20 10 0 0 0.5 PV Injected 1 1.5

Figure 9: Production Profile at 4500 psia

60 50 Produced Oil (cc) 40 30 20 10 0 0 0.2 0.4 0.6 PV Injected 0.8 1 1.2

Figure 10: Production Profile at 4900 psia

Figure 11: Minimum Miscible Pressure Calculation

10