Effect of Interfacial Tension on Water/Oil Relative Permeability on the Basis of History Matching to Coreflood Data

Abstract

Summary The effect of interfacial tension (IFT) on the displacement of the nonwetting and wetting phases has been investigated by the use of simulations/history matching of flooding experiments. In surfactant flooding, a conventional assumption is to neglect the effect of capillary pressure (Pc) on measured two-phase properties. The methodology applied in this paper allows improved interpretation of experimental results by correcting for the influence of capillary end effects on the measured capillary desaturation curve (CDC) and on the estimated relative permeability (kr). Three fluid systems of different IFTs were prepared by use of a solvent system (CaCl2 brine/iso-octane/isopropanol) rather than a surfactant system with the assumption that both systems have similar flood behavior at reduced IFT. Three coreflood cycles, including multirate oil injection (drainage) followed by multirate water injection (imbibition), were carried out at each IFT in water-wet Berea cores. The kr functions corrected for capillary end effects were derived by numerically history matching the experimental production and pressure-drop (PD) history. A typical CDC is observed for the nonwetting phase oil, with a roughly constant plateau in residual oil saturation (ROS), Sor, below a critical capillary number (Ncc) and a declining slope above Ncc toward zero Sor. No influence of Pc was found for the nonwetting-phase CDC. The results from the displacement of the wetting phase formed an apparent CDC with a declining slope and no Ncc. Analyzing the wetting-phase results, we find that the wetting-phase CDC is not a true CDC. First, it is a plot of the average remaining water saturation (Sw) in the core which, in all the experiments, is higher than residual water saturation, Swr, obtained from Pc measurements. Second, we find that the remaining Sw is only partly a function of Nc. At low Nc, the water production (WP) is limited by capillary end effects. Rate-dependent WP observed with the high-IFT system is fully reproduced in simulations by use of constant kr and Pc. The remaining wetting-phase saturation at a low capillary number (Nc) is a result of the core-scale balance between viscous and capillary forces and would, for example, depend on the core length. At a higher Nc, the WP is found to be limited by the low kr tail, typical for wetting phases. However, we find that the kr functions become rate dependent at a higher Nc, and we assume that this rate dependency can be modeled as a function of Nc. The remaining wetting-phase saturation at a higher Nc would then be a function of Nc and the number of pore volumes (PVs) injected. The observed Nc dependency in the flow functions indicates a potential for the accelerated production of the wetting phase by use of surfactant. Assuming that the results obtained here for the wetting phase also apply to oil in a mixed-wet system, it is strongly recommended to evaluate the effect of both Pc and Ncc when designing a surfactant model for a mixed-wet field.