Analysis of Heavy-Oil Immiscible CO2 Tertiary Coreflood Data

Abstract

Summary. This paper describes the results of a series of tertiary, immiscible, CO2 corefloods of Wilmington field Pliocene reservoir rock containing heavy oil ( 14 degrees API [ 0.97 g/cm3] and 480 cp [ 480 mP.s]). An initial set of corefloods defined the recovery potential of the CO2 injection, while a series of later tests served potential of the CO2 injection, while a series of later tests served to define the process more accurately as applied in the field. In an attempt to understand the displacement mechanism, simulator matching of one of the later, more refined groups of corefloods was performed. The corefloods and simulator work indicate that the incremental recovery is more than can be accounted for by oil-viscosity reduction and crude-oil swelling. The improved performance is attributed to more favorable displacement characteristics and the presence of a free gas saturation in the cores. Introduction The literature has documented that immiscible CO2 solution can swell heavy crudes and reduce their viscosity. Results of several immiscible heavy-oil CO2 floods have been reported, and additional laboratory test results have indicated favorable CO2 solution behavior. As a result, the availability in the Wilmington area of an impure CO2 refinery effluent and the U.S. DOE's Tertiary Incentive Program encouraged investigation of the application of immiscible CO2 flooding in the more easterly portions of the Wilmington field in 1979–80. The City of Long Beach's Dept. of Oil Properties and its two field contractors, THUMS Long Beach Co. and Long Beach Oil Development Co., conducted a series of six corefloods in preserved Wilmington field, Ranger zone, Fo sand cores at 125 degrees F [52 degrees C] confined at 2,450 psi [16 892 kPa] with a pore pressure of 850 psi [5861 kPa] (1,600- psi [11 032 kPa] net simulated overburden). The initial oil saturating the psi [11 032 kPa] net simulated overburden). The initial oil saturating the cores was gas free. The cores were first waterflooded to a 100:1 WOR, which is beyond the secondary-recovery economic limit. The tertiary-recovery part of the floods followed, starting with injection of pure CO2 to breakthrough. After standing 24 hours with inlet and outlet pressures fixed at 850 psi [5861 kPa], carbonated water was injected until it reached the outlet core face (breakthrough). Each core was shut in again for 24 hours, after which carbonated water was injected until depletion. These corefloods provided the laboratory justification for the field test begun by the City of Long Beach and Long Beach Oil Development Co. in the Tar zone of Fault Block V in the Wilmington field. The second series of five corefloods, which more closely approximated field conditions, was completed in 1985. Tar zone cores from Fault Block V of the Wilmington field were used. Test temperature again was 125 degrees F [52 degrees C], and the simulated hydrostatic reservoir stress was 1,300 psi [8963 kPa] (confining pressure of 2,300 psi [15 858 kPa] and pore pressure of 1,000 psi [6895 kPa]). The initial oil contained 50 scf/STB [9.0 std m3/stock-tank m3] natural gas. The CO2 slug contained 16 mol% N2, which is representative of the field supply composition. The constant pressure (1,000 psi [6895 kPa]) static period after CO2 breakthrough in the second series of floods was 36 hours. In three of the core samples, pure CO2 was used for carbonated water, and in the other plugs, two simultaneous injections of synthetic field brine and a mixture of 84 mol% CO2 and 16 mol% N2 occurred. The results of these tests continued to show the same order of magnitude incremental tertiary recovery increase as the initial set of samples. The presence of the N2 did not have a detrimental effect on the recovery. To define the process further, the results of one of the second series of corefloods were simulator-matched. This work indicates that other mechanisms, in addition to crude swelling and viscosity reduction, are responsible for the improved oil recovery from the immiscible CO2 process. Description of Coreflood Testing Initial (1981) Group. Six core plugs were flooded. These were obtained from preserved plastic-sleeve full cores taken in an inert water- based mud. The plugs averaged 4.8 in. [12.2 cm] long, 3.2 in. [20.6 cm2] in cross section with a PV of 3.3 in. 154 cm3] at simulated overburden. The plugs were cut parallel to the full cores' axes while the cores were kept frozen with liquid N2. The still- frozen plugs were jacketed in plastic tubing and confined at simulated overburden conditions. Each plug was then heated to 125 degrees F [52 degrees C] and saturated with a simulated reservoir brine. It was then driven to the initial saturation conditions by filtered crude. Once all free water had been displaced, the temperature was increased to 180 degrees F [82 degrees C] for 24 hours with oil flow maintained to stabilize the initial saturation. This elevated temperature was to dissolve bitumen precipitated from the crude when the cores were frozen. The temperature was then reduced to that used for the rest of the test, [25 degrees F [52 degrees C], and initial oil permeability was measured. The overburden and internal pore pressures were adjusted to the values given in the Introduction. Each core was flooded with filtered injection water until breakthrough at a rate that gave a scaling factor of 1 to minimize core end effects. At that point, the flow rate was reduced to 1 ft/D [0.3 m/d]. Flooding continued until a WOR of about 100:1 was reached. The core was then saturated with pure CO2 and left pressurized by CO2 at the inlet at 850 psi [5861 kPa] for 24 hours. Carbonated injection water was pushed to the outlet face of the core; the plug was allowed to stand for another 24 hours. It was then driven with the carbonated water at a 1-ft/D [0.3-m/d] rate to about a 100:1 WOR. Five of the six plug tests yielded reliable incremental oil recoveries with increased WOR. The effluent from the sixth sample was so emulsified that only a final incremental oil value was obtained. The flood behavior of each of the five plugs was plotted and then graphically averaged to give the result shown in Fig. 1. After the gas (CO2) injection, the WOR decreased from about 100:1 to an estimated average minimum of 1.7 with an oil recovery of 4.5% PV. From that point, WOR increased to more than 100:1 on a slope PV. From that point, WOR increased to more than 100:1 on a slope parallel with the waterflood curve. Inspection of Fig. 1 indicates parallel with the waterflood curve. Inspection of Fig. 1 indicates that the irreducible oil saturation after CO2/carbonated waterflooding should be substantially less than for waterflooding only. SPERE p. 69