A parametric study of reservoir cooling for enhanced recovery by CO₂ flooding

Slim-tube experiments and analytical calculations show that minimum miscible pressure (MMP) can significantly decrease with a relatively modest reduction in temperature. Compositional simulation, however, is often made under isothermal conditions even though a prior waterflood may have reduced reservoir temperature in the swept zones of the reservoir. This paper examines how cooling by a prior waterflood can impact recovery during a CO₂ flood by lowering the MMP in the swept zones. The results show that for the cases considered injection of cooler water can increase incremental oil recovery significantly owing to MMP reduction in the zones swept by the solvent. A parametric simulation study demonstrates how injection temperature, initial reservoir pressure, formation heterogeaneity, formation thickness, heat transfer with the over and under burden formations, and WAG ratio affect the incremental oil recovery. The simulations are conducted by performing a long waterflood of up to 2.0 PVI prior to CO₂ injection. The water during the secondary recovery is injected at two temperatures for the 1-D and 2-D flow models. CO₂ solvent is then injected continuously or in WAG at the same waterflood injection temperature. The increase in incremental recoveries (above what would have been obtained by a standard CO₂ flood at original reservoir temperature) varied greatly depending on the flow dimension, initial reservoir pressure, level of heterogeneity, formation thickness, the degree of energy gain from the surroundings, and the injection temperature. Increases in recovery by CO₂ flooding varied from a few %OOIP to over 20% with the highest recoveries occurring in 1-D flow. For the same flow dimension, the largest increase in recoveries is achieved when the MMP is sufficiently reduced by temperature so that an otherwise immiscible or near miscible flood becomes a multicontact-miscible (MCM) flood. The results demonstrate that including temperature variations in the simulations is important for floods that are near miscible as recoveries are most impacted in that region. Further, including temperature variations could be very important to improve the quality of history matches used to understand the reservoir. Lastly, we give a new MMP correlation for CO₂ injection based on the most recent slim-tube data from the literature.