Feasibility of Cyclic CO2 Injection for Light-Oil Recovery

Abstract

Summary Field performance trends are presented for cyclic CO2 injection light-oilsandstone reservoirs. Results indicate that the mass of CO2 injected is thebest predictor of stimulated oil prod uction rate and ultimate incrementalrecovery. The correlations prod uction rate and ultimate incremental recovery. The correlations enhance understanding of the single-well process, improvereservoirs election criteria, and suggest operational guidelines. Introduction Since 1984, the applicability Of cyclic CO2 injection (CO2 huff ‘n’ puff) tothe enhanced recovery of light oil has been examined with puff) to the enhancedrecovery of light oil has been examined with encouraging results. Reportedcoreflood data demonstrate that multiple cycles of CO2 injection recoverwaterflood residual oil and that operating at miscible displacement conditionsis disadvantageous. Case histories document low-cost field procedures anddepict production responses with quick payout and attractive CO2 utilizations. History matches of field performance using numerical simulation suggest thatthe principal oil-recovery mechanisms are oil swelling, off-viscosityreduction, and gas relative-permeability hysteresis. This paper is based on original field studies and further analysis ofexisting studies. The purpose was to extend our basic understanding of thecyclic CO2 process, to define reservoir selection criteria, and to formulateoperational guidelines. Three previous reports evaluated light-oil/cyclic-CO2 field data bases toidentify performance trends. An analysis of 28 east and south Texas projectsled Haskin and Alston to conclude that larger slugs recover more incrementaloil, the optimal soak period is 2 to 3 weeks, CO2 injection typically reduceswater production and no relation exists between oil recovery and pretest oilcut. A simple yet fairly accurate predictive method was developed for the Texasreservoirs that assumed oil recovery by oil swelling and oil-viscosityreduction. Monger and Coma I examined 14 south Louisiana tests and found thatresponse improved with larger slogs, thicker pay, and lower pretest water cut. CO2 injection typically reduced water cut, and neither reservoir pressure norsoak duration significantly influenced field performance. Results from 66 east Kentucky tests were used by Monger et al. to highlight favorable processperformance in a pressure-depleted field. Response was found to improve withthicker pay, and CO2 injection typically reduced water cut. Although developed to address cyclic CO2, for heavy-oil recovery, thenumerical simulation work of Patton el al. identified per-formance trends thatmay be relevant to light-oil applications. Unde per-formance trends that may berelevant to light-oil applications. Unde ideal conditions, predicted CO2 utilization was 1 Mscf/STB [180 std m /m], and efficacy decreased with boththe number of cycles and the volume of CO2 injected. The process did notrequire a high initial oil saturation and was thus well-suited to high-watercutreservoirs. Description of the Data Base Projects were implemented by independent and major oil Projects wereimplemented by independent and major oil companies in 14 fields located in Louisiana and Kentucky. The data base contains results for 106 single-well CO2 huff 'n'puff field tests. Of these, 14 were included in a previous evaluationof south Louisiana cyclic CO2 field data. Early response for 66 of the testperformed in Field G were also presented previously. Tests from performed in Field G were also presented previously. Tests from these previous evaluationswere updated, and 26 additional wells were included for the current study. Adetailed description of the data base and the methods used to assess field-testresponse is provided elsewhere. provided elsewhere. Ninety-seven tests, orapproximately 90% of all tests in the data base, showed incremental oil. Eightof the failures were attributed to mechanical difficulties during CO2 injection, and one to apparent CO2 migration along a fault. Each test wasperformed in a light-oil (23 to 38 deg. API [0.92 to 0.83 g/cm]) -bearingsandstone (consolidated, unconsolidated, or dolomitic) reservoir. Otherreservoir conditions were diverse. Table 1 summarizes select reservoirproperties for fields in which CO2 injection recovered incremental oil. Test Conditions. Table 2 lists conditions for the 97 tests that producedincremental oil. Test procedures varied, even for the more produced incrementaloil. Test procedures varied, even for the more successful projects. CO2 wasusually transported in tanker trucks as a 300-psig [2068-kPa], 0 deg. F [255-K]liquid. CO2 was sometimes heated before injection down the annulus, tubing, orworkstring. Injection rates were as fast as pumps could provide withoutdamaging well completions. Injected CO2 typically was displaced from thewellbore with corrosion inhibitor/lease crude. Tubing and casing pressures forthe test well and offsets were monitored during the pressures for the test welland offsets were monitored during the soak period. In some cases, offset wellswere shut in to confine CO2 . The procedure preferred for reopening wellsincluded use of a small choke to increase backpressure on the well to minimize CO2 breakout and isolation of production so that the detrimental effects Of CO2 would not affect surface operations. Soaks were sometimes extended when only CO2 was produced initially. Data Distribution. Ideally, the data base would have included a largernumber of more diverse projects. Although the available field results seemedsufficient for trend identification, it should be noted that data distributionwas nonuniform. For example, in comparison with other fields, test activity in Field G was higher and test conditions were characterized by lower pressure, smaller slugs, and shorter soaks. Evaluation Parameters. Four parameters were used to evaluate field response:incremental oil, CO2 utilization, CO2 reservoir utilization, and stimulationratio. Incremental oil was calculated as the increase in recovery over the baselineproduction forecast. The baseline was established by regression analysis ofproduction decline before CO2 injection. If the forecast reached the welleconomic limit, then baseline production was held constant with no furtherdecline. A best fit was production was held constant with no further decline. Abest fit was also made to post-CO2 production. Project life was terminated when Post-CO2 production neared the extrapolated baseline or if the well was workedover. Incremental oil was calculated from the area between the curves, asillustrated in Figs. 1 through 3. Use of decline curve analysis is an accepted method for estimatingproduction response to a workover. As shown by the examples in Figs. 1 through3, the effectiveness of this approach varied with the quality of the productiondata. Decline-curve analysis was used because oil production data were oftenmore accurate than water production data. In Haskin and Alston's evaluation offield data, production data. In Haskin and Alston's evaluation of field data, the calculation of incremental oil was based on oil cut. Differences between anaverage pretest oil cut and oil cuts measured after CO2 injection weremultiplied by the daily total production rate. Adoption of this approach to thepresent data base would alter some of the values obtained for incremental oil. Because relying on production cuts also has limitations, however, it was notfelt that the production cuts also has limitations, however, it was not feltthat the evaluation would improve.

EOS Simulation for CO₂ Huff-n-Puff Process

“…In practice, after the gas injection, the well is shut-in for a period of time to allow for intense gas-oil interaction. It was reported that extending the soak period up to 4 weeks significantly improved the production response [21]. However, this period cannot be excessively long because it could cause the pressure dispersion in the reservoir that leads to re-release of dissolved gas, and because the CO 2 would dissipate due to compositional gradient leading to reduced CO 2 concentration in the oil.…”

Reservoir Characterization, Production Characteristics, and Research Needs for Fluvial/Alluvial Reservoirs in the United States

“…Thomas and Monger-McClure (1991) evaluated performance trends for light-oil cyclic C02 stimulation treatments and suggested the following well selection guidelines q Desirable reservoir characteristics include shallower reservoirs, operating below the minimym miscibility pressure, a thick pay interval, a mobile-water or free-gas saturation, and an adequate drive mechanism.…”

Optimization of a Mature CO2 Flood — From Continuous Injection to WAG

“…Almost all the CO2 EOR projects in Permian Basin have been designed as miscible or near miscible processes. CO2 injection schemes have evolved from huff n' puff (or cyclic) injection [3] to Tapered Water Alternating Gas (TWAG) injection [2] . Table 1 summaries the major injection schemes developed in the last 30+ years [4,5,6,7,8,9] .…”