Physically Scaled Model Studies Simulating the Displacement of Residual Oil by Miscible Fluids

Abstract

Abstract The displacement of waterflood residual oil by a solvent drive has been studied with scaled physical models of a line drive. This work was a first step in an overall study of the displacement of residual oil by CO2. Some results of the subsequent study are presented for comparison. The usefulness and limitations of scaled physical models for studying the behavior of solvent displacement is analyzed, and it is concluded that such models are very useful for elaborating the mechanism of such a recovery process. Also, because the results obtained with such models correspond with field results, we believe that certain factors, which of necessityhave been neglected in the scaling, play only a minor role in the performance of the process. When the displacement of the mobile water by the solvent is very stable, the residual oil is produced after a large fraction of the moveable water has been first produced. There is very little evidence of an oil bank being developed, with virtually all the residual oil being produced in solution with the displacing fluid. When the displacement of the mobile water by the solvent is inefficient, because of gravity and/or viscous instability, both the residual oil and solvent begin to be produced much earlier. The recovery of the residual oil parallels the recovery of the mobile water just as in the case of the stable displacement. The efficiency of residual-oil recovery (volumes of oil produced per volume of injected solvent) is much higher at low recoveries when there is marked instability, but with increasing recovery this advantage is lost. This study concludes that the efficiency of CO2 to recover residual crude oil is influenced primarily by gravity and viscous instabilities resulting from CO2's low density and viscosity in comparison to water. It does not appear that CO2 is any more efficient, in terms of reservoir volumes, than a hydrocarbon-miscible solvent.