A Predictive Network Model for Three-Phase Flow in Porous Media

Abstract

Abstract We reconstruct three-dimensional (3D) Berea sandstone samples by stochastically modeling the results of the main sandstone forming processes. The morphology of the reconstructed microstructure is compared quantitatively with a microtomography image of an actual sample of Berea sandstone. The comparison shows that the process based reconstruction adequately reproduces important intrinsic properties such as the local porosity distribution and the degree of connectivity. The reconstructed pore space is transformed into a pore network that is used directly as input to a network model. The model simulates two-phase flow and drainage dominated three-phase flow. Computed relative permeabilities for reconstructed Berea sandstone samples are compared with experimental data. The predicted relative permeability curves for both two-phase and three-phase flow compare favorably with the experimental results. Both the experiments and the simulations show that continuous oil layers present in the corners of gas invaded pores may allow low oil saturations to be reached during drainage dominated three-phase flow. However, the residual oil saturation and the oil permeability are strongly dependent on the saturation history.