Spontaneous Imbibition with Matched Liquid Viscosities

Abstract

Abstract Laboratory results of oil recovery through spontaneous imbibition are commonly scaled-up to forecast oil recovery from fractured reservoirs. This study addresses oil recovery from cylindrical sandstone cores by spontaneous imbibition at very strongly water-wet conditions for viscosity ratios of unity (a special case of the Mattax and Kyte scaling condition that viscosity ratios be matched). Combinations of mineral oils and aqueous solutions of glycerol were used to obtain matched viscosities ranging from 4 to 172 cp. In all, 25 imbibition data sets are reported for boundary conditions of all core faces open, two ends closed (radial imbibition), and one end open (linear imbibition). The data sets for individual boundary conditions were satisfactorily correlated by the Mattax and Kyte scaling group. Overall correlation of results was obtained after compensating for different boundary conditions by means of a characteristic length. An obvious limitation of correlations based on linear scaling of time is that differences in the shapes of imbibition curves cannot be eliminated. The correlated data provides clear distinction between the overall shape of recovery curves for linear versus radial imbibition. Final oil recoveries for radially dominated imbibition were independent of viscosity whereas recoveries for linear imbibition were consistently lower and decreased by up to 2.5 % PV with increase in viscosity. Shortly after the onset of imbibition, oil recovery for linear flow is generally close to linearly proportional to the square root of time until the imbibition front reaches the end of the core. Introduction In naturally fractured reservoirs, spontaneous imbibition of brine can be the dominant mechanism of oil production from the rock matrix. Rock properties, fracture geometry, fluid properties, and rock fluid interactions govern the rate and extent of recovery.[1–2] Oil recovery from laboratory measured spontaneous imbibition is commonly scaled to obtain mass transfer functions that are used to predict oil production from fractured reservoirs. Many questions remain unanswered in regard to the mechanism of spontaneous imbibition and valid scaling of laboratory imbibition data. Mattax and Kyte[3] introduced a scaling group based on the theoretical analysis of Rapoport and Leas.[4–5] (1) where µw is the aqueous phase viscosity and µo is the oil phase viscosity. Identical viscosity ratio was one of six listed conditions for valid scaling. The complexities of spontaneous imbibition in porous media are such that perfect correlation by linear scaling of time should not be expected. This study tests aspects of scaling for matched oil and aqueous phase viscosities. Results are presented for cores with all faces open, one end open (linear imbibition) and two ends closed (radial imbibition) for nearly two orders of magnitude variation in viscous forces but only small change in capillary forces (caused by small differences in interfacial tension).