Evaluation of simultaneous water and gas injection using CO2

Miscible CO2 injection is the second largest contributor to global enhanced oil recovery, as it has successfully undergone extensive laboratory tests and field applications for recovering residual oil left behind after waterflooding. Prolific incremental recoveries have been obtained for some fields. Although miscible CO2 injections generally have excellent microscopic displacement efficiency they often result in poor sweep efficiency. In order to address sweep problems and maximize recoveries, other schemes of gas injection have been developed. Two such processes are water-alternating-gas (WAG) and simultaneous water-and-gas (SWAG) injection. WAG and SWAG have been successfully used to minimize poor sweep. Improved gas utilization and oil recovery have been reported for SWAG injection at Joffre Viking, Kapurak River, and Rangley Weber fields. There are very little published data evaluating the performance of simultaneous water and gas injection under miscible conditions and very little published data exists that compares enhanced recovery processes conducted under consistent experimental conditions. This is especially true when the gas is CO2. In this work a sequence of experiments were conducted to evaluate core flood behavior of Continuous Gas Injection (CGI), 1:1 Water Alternating Gas (WAG) with a slug size of 0.25 pore volumes, and Simultaneous Water-and-Gas (SWAG) injection at four fg values. The experiments were conducted at rock wettability, flow rates and pressures that were as consistent as possible in order to make meaningful comparisons. After 2 PV of CO2 injection the SWAG flood with fg = 0.4 recovered about 0.9692 of waterflood residual oil. CGI had the second best recovery of about 0.8998 followed by WAG with 0.8602. The SWAG flood with fg = 0.6 recovered about 0.8300 of waterflood residual oil and SWAG with fg = 0.8 and fg = 0.2 recovered about 0.7507 and 0.7253 respectively. The gas utilization was the least for SWAG with fg = 0.4 at 15.54 Mscf/bbl followed by CGI with 16.13 Mscf/bbl. The remaining experiments utilized over 17.20 Mscf/bbl