Evaluation of CO2 Flood Performance: North Coles Levee CO2 Pilot, Kern County, California

Abstract

Summary CO2 injection began on June 1, 1981, in the Main Western (MW) III interval of the North Coles Levee field Kern County CA, to test the tertiary potential of the reservoir to CO2. The pilot test was designed to inject continuously a 40 %-HCPV slug of CO2 into two inverted 5-acre [2.02-ha] 5-spot patterns and one 10-acre [4.05-ha] linedrive pattern. The injection was switched to water near the end of CO2 injection because of the CO2 source loss. Overall pilot test performance varied; oil production increased above secondary decline in some wells, but not in others. A simulation study was initiated to determine the dominant mechanisms controlling pilot performance. The simulation study investigated the influence of reservoir description, gravity override, communication with other producing horizons. and production strategy. Simulation results indicate that a random permeability description with Zone MWIII in communication with upper zones is most appropriate. Gravity override of CO2 into the upper zones dominated displacement efficiency. These results indicate that dual injection of CO2 low in MWIII and of water high in MWIII would greatly enhance sweep efficiency. Further analyses with oil and CO2 production data from individual wells and a post-CO2 tracer test showed that the test was a technical success in wells that responded to CO2 and that grouping and analyzing all pilot wells together masks the true incremental oil response to CO2. This is true mainly because some wells that never responded to CO2 experienced well problems that severely curtailed their production and that offset increased production from wells that did respond to CO2. The pilot test patterns are badly out of balance, possibly because of operational patterns are badly out of balance, possibly because of operational imbalance, directional permeability or a combination of the two. Introduction EOR by CO2 injection in watered-out reservoirs has been proven in several reservoirs throughout the world. This paper documents an analysis of the first application of CO2 injection in the Kern River Valley, Kern County. CA. Although California is well known for EOR by application of thermal methods to viscous oil reservoirs, it also has many deep, light-oil reservoirs discovered in the 1930's and 1940's that are approaching their economic limits from waterflooding. The North Coles Levee field is typical. A pilot study was initiated to test the potential of CO2 in three patterns on field spacing. Many problems have plagued the project, especially mechanical problems in wells that did not respond to CO2. The primary purpose of this study is to determine the technical success of the pilot test and to identify the dominant mechanisms controlling the process. The pilot test was analyzed by attempting to identify mechanisms that could account for the response to CO2 injection by use of reservoir simulation and a detailed well-by-well analysis of actual field production, injection, and tracer data. The lack of data-especially fluid saturations, pressure distribution, and a pilot-wide reservoir description-hinder a detailed simulation, but controlling processes can be qualitatively derived. Field and Pilot History The North Coles Levee reservoir is a large oil pool located in the Stevens sand interval about 10 miles [16.1 km] west of Bakersfield, CA, in the Kern River Valley (Fig. 1). Table 1 summarizes reservoir properties. The reservoir is mature and approaching its economic properties. The reservoir is mature and approaching its economic limit from waterflooding. The field was discovered in Nov. 1938 and produced by gas-cap expansion until Nov. 1942, when fluid withdrawal produced by gas-cap expansion until Nov. 1942, when fluid withdrawal was balanced by gas injection to preserve reservoir pressure. This included injection of make-up gas and gas cycling. A controlled blowdown of gas was started in 1965 because the economics favored gas sales over continued gas injection, allowing reservoir pressure to drop below 1,000 psi [6895 kPa]. A pilot alternating propane with water and dry chase gas was tried in 1964, but was unsuccessful because of early breakthrough and a diminishing supply of propane. A waterflood pilot was started in 1967. A fieldwide waterflood was started in 1970 pilot was started in 1967. A fieldwide waterflood was started in 1970 and continues to the present. Water preferentially enters the gas-swept region, while fluid withdrawals come mainly from the lower regions of the reservoir. This has led to a higher pressure, greater than 5,000 psi [34 475 kPa], in Zone MWI compared with 4,500 psi [31 027 kPa] in Zones MWII and -III. Because of underbalanced withdrawals in Zone MWIII, the waterflood was infill-drilled at 20- and 10-acre [8.09- and 4.05-ha] spacing (Fig. 2). Some patterns were converted to linedrive or irregular configurations later in the waterflood. Rod pumps were used for artificial lift on most producing wells. Gas lift was used on the remaining wells.