Dynamic evolution of displacement patterns during thickened supercritical carbon dioxide flooding in heavy-oil reservoirs

During heavy-oil displacement by supercritical carbon dioxide (sc-CO2), the large viscosity contrast between sc-CO2 and heavy oil induces fingering instability and gas channeling. Thickened sc-CO2 improves mobility control and partly suppresses fingering; however, CO2 dissolution and diffusion still alter the oil-phase viscosity near the displacement front. Thus, the classical Lenormand phase diagram cannot fully describe displacement-pattern evolution in this variable-viscosity system. To address this gap, the present study incorporates the oil-viscosity-reduction effect induced by the dissolution of thickened sc-CO2 into a coupled flow–mass transfer–diffusion mathematical model. The model is solved using a phase-field method to investigate the evolution of displacement patterns in thickened sc-CO2–heavy-oil systems within heterogeneous porous media, with particular emphasis on the combined effects of injection velocity and CO2 thickening ratio. Relative to the classical Lenormand phase diagram, the boundary between the viscous fingering (VF) region and the transition region shifts toward higher log Ca values, whereas the boundary between the transition region and the capillary fingering (CF) region shifts toward lower log M values. Consequently, the VF and transition regions contract, whereas the CF region expands. Comparison with the constant-oil-viscosity reference model indicates that the phase-boundary redistribution is not controlled by dissolution alone, but by coupled mobility control, CO2 diffusion and dissolution, local oil-viscosity variation, and pore-scale heterogeneity. The displacement-pattern evolution is further characterized using the degree of viscosity improvement, fractal dimension, finger number, and frontal gas saturation. These analyses show that injection velocity and CO2 thickening ratio jointly regulate front stability, mass-transfer distribution, and displacement performance.