An Experimental Investigation of Viscous Oil Recovery Efficiency as a Function of Voidage Replacement Ratio

Abstract

Recovery processes with a voidage replacement ratio (VRRϭinjected volume/produced volume) of 1 rely solely upon viscous forces to displace oil whereas a VRR of 0 relies upon solution-gas drive. Activating a solution gas drive mechanism in combination with waterflooding using periods of VRR less than 1 (VRR Ͻ1) may be optimal for recovery. Laboratory evidence suggests that recovery for VRR Ͻ1 is enhanced by emulsion flow and foamy oil at pressures under bubble point for some crude oils. This paper investigates the effect of VRR for two crude oils referred to as A1 (88 cP and 6.2 wt% asphaltene) and A2 (600 cP and 2.5 wt% asphaltene) in a sandpack system (18 Љ long and 2 Љ dia.). The crude oils are characterized using viscosity, asphaltene fraction, and acid/base numbers. A high-pressure experimental sandpack system (1 Darcy and S wi ϭ0) was used to conduct experiments with VRR's of 1.0, 0.7, and 0 for both oils. During waterflood experiments, we controlled and monitored the rate of fluid injection and production to obtain well-characterized VRR. Based on the production ratio of fluids, the gas-oil and-water relative permeabilities were estimated under two-phase flow conditions. For a VRR of 0, the gas relative permeability of both oils exhibited extremely low values (10 -6~1 0 -4 ) due to internal gas drive. Water floods with VRR Ͻ 1 displayed encouraging recovery results. In particular, the final oil recovery with VRRϭ0.7 (66.2 % OOIP) is more than 15% greater than that with VRRϭ1 (55.6 % OOIP) using A1 crude oil. Results for A2 with VRRϭ0.7 (60.5 % OOIP) were identical to the sum of oil recovery for solution gas drive (19.1 % OOIP) plus waterflooding (40.1 % OOIP). An in-line viewing cell permitted inspection of produced fluid morphology. For A1 and VRR ϭ 0.7, produced oil was emulsified and dispersed as gas bubbles, as expected for a foamy oil. For A2 and VRR Ͻ1, foamy oil was not clearly observed in the viewing cell. In all cases, the water cut of VRRϭ1 is clearly greater than that of VRRϭ0.7. Finally, three-phase relative permeability was explored based upon the experimentally determined two-phase oil-water and liquid-gas relative permeability curves. Using well known algorithms for three-phase relative permeability, however, did not result in good history matches to the experimental data. Numerical simulations matched the experimental recovery versus production time acceptably after modification of the measured k rg and k row . The degree of agreement with experimental data is sensitive to the details of gas (gas-oil system) and oil (oil-water system) mobility.