Proxy Model Development for the Optimization of Water Alternating CO2 Gas for Enhanced Oil Recovery

Optimization studies are an important task in reservoir engineering practices such as production optimization and EOR (Enhanced Oil Recovery) assessments. However, they extensive with many simulations that require huge computational effort resources. In terms of EOR, CO2 injection is one the most common methods employed due to a high recovery potential environmental benefits. To assess feasibility CO2-EOR projects, design study must be conducted before performed. Some have demonstrated advantages employing proxy models perform this saving amounts computer memory space time. study, were developed solve multi-objective problem using NSGA-II (Non-dominated Sorting Genetic Algorithm II) two selected models. The was performed for CO2-WAG (Water Alternating Gas) application, where gas water rates half-cycle lengths assessed maximize oil stored reservoir. One model represents simple geological (the Egg Model), while other complex Gullfaks Model). good performance generated accurate results could improved by increasing amount sampling segmenting behavior (depending on complexity model). proxies average error less than 2% (compared simulation results) concluded robust based blind test results. It has also been found reach maximum CO2-WAG, rate minimum required. configuration may result reduction total