Numerical investigations on the effect of initial state CO2 topology on capillary trapping efficiency

Abstract

This study numerically investigates the effects of initial CO2 saturation distribution on residual saturation by a digital rock approach. The pore geometry in Berea sandstone is extracted by high-resolution microcomputed tomography (CT) scanned images. Pore scale displacement simulations are performed using a novel lattice Boltzmann numerical simulation method. To deeply understand the effect of initial fluid topology on capillary trapping efficiency we investigate the relationships between the trapping curves and various characteristics of initial CO2 clusters (e.g. cluster number N, volume V, sphericity index S, connectivity C, wall contacting area ratio A, and length L). In this study, to generate fluid distribution with different topological characteristics, we considered four initial saturation distribution schemes: (1) random distribution, (2) drainage distribution, (3) boundary coating with wetting fluid, and (4) patchy saturation. Imbibition simulations were performed on each of these four initial distribution models. We found that the capillary trapping curve strongly depends on initial CO2 distribution. Results for random and boundary coating patterns show nearly linear relationships between initial and residual saturation, and those for the other two initial schemes show strong nonlinear relationships. Correlation analyses demonstrate that the connectivity C and sphericity index S of the initial CO2 clusters significantly impact such nonlinear behavior. It is found that higher initial state CO2 connectivity C and sphericity index S result in decreased capillary trapping owing to the increasing pressure instability and inertial effects during ganglion disconnection. The CO2 injection patterns can be manipulated to achieve low C and S for maximizing the capillary trapping efficiency.