CO2 WAG Performance Sensitivity to Relative Permeability and Wettability Assumptions in Compositional Simulation

Abstract

Water alternating gas (WAG) injection is widely applied in CO2 EOR to improve sweep efficiency and mobility control, yet forecasted WAG performance is often dominated by uncertainty in relative permeability, three-phase behavior, and cycle hysteresis. This study isolates the impact of relative permeability and wettability assumptions on predicted WAG outcomes using a controlled compositional simulation workflow where the reservoir, fluid system, well pattern, and WAG schedule are held constant across cases. A CMG GEM model was constructed from a standard WAG template using a simplified 3D stratified reservoir with uniform porosity and layered permeability to represent vertical heterogeneity. The fluid description is a Peng Robinson EOS compositional model with 11 components (CH4 to C7+, CO2, N2) based on recombined PVT from the Maui-1 system, with CO2 as the injected gas. A single corner injector and an offset producer were operated under an identical WAG design across all cases, allowing direct attribution of performance differences to rock fluid flow functions. Four relative permeability scenarios were designed to represent base, more water wet, less water wet, and oil wet behavior by systematically varying endpoint saturations and curve shapes consistent with published WAG and wettability literature. Key performance metrics included oil recovery factor, cumulative oil production, water cut and water breakthrough timing, cumulative water production, and gas recycling behavior. Results show that plausible wettability driven changes in relative permeability can generate large spreads in WAG forecasts, including an approximately 20 percentage point range in ultimate oil recovery between the base and strongly oil wet descriptions. Water wet cases exhibited earlier water breakthrough and higher water recycling, while oil wet cases delayed water production and delivered higher cumulative oil. Gas recycling trends also shifted with wettability, with higher produced gas fractions in the base and oil wet scenarios and lower gas recovery in more water wet cases, consistent with increased gas trapping. These findings emphasize that reliable CO2 WAG design and evaluation require careful relative permeability selection, explicit uncertainty bounds, and fit-for-purpose hysteresis treatment when forecasting incremental recovery and operational risks.