In this article, a robust machine-learning based computational framework that couples multi-layer neural network (MLNN) proxies and multi-objective particle swarm optimizer (MOPSO) to design wateralternative-CO2 injection (CO2-WAG) projects is presented. The proposed optimization protocol considers various objective functions including oil recovery and CO2 storage volume. Expert MLNN systems are trained and employed as surrogate models of the high-fidelity compositional simulator in the optimization workflow. A large volume of blind testing applications is employed to confirm the validities of the proxies. When multiple objective functions are considered, two approaches are employed to treat the objectives: the weighted sum method and Pareto-front-based scheme. A field scale implementation to Morrow-B formation at Farnsworth Unit (FWU) to optimize the tertiary recovery strategy is discussed. In this work, investigations will focus on comparing the optimum solution found by the aggregative objective function and the solution repository covered by the Pareto front, which considers the physical and operational constraints and reduces uncertainties involved by the multi-objective optimization process. Necessary trade-offs need to be decided using the solution repository to balance the project economics and CO2 storage amount.
Keywords optimization, carbon dioxide, sequestration, proxy models, multi-objective