Development and application of a machine learning based multi-objective optimization workflow for CO2-EOR projects

Carbon dioxide-Enhanced Oil Recovery (CO2-EOR) is known as one of techniques for hydrocarbon production improvement as wells as an important candidate to reduce greenhouse gas emissions. Thus, an ideal development strategy for a CO2-EOR project would consider multiple objectives including to maximize oil recovery, CO2 storage volume and project economic outcomes. This work proposes a robust computational framework that couples artificial neural network (ANN) and multi-objective optimizers to optimize the aforementioned objectives in CO2-EOR processes simultaneously. Expert ANN systems are trained and employed as surrogate models of the high-fidelity compositional simulator in the optimization workflow. The robustness of the development optimization protocol is confirmed via a synthetic injection-pattern-base case study. Afterward a field implementation to Morrow-B formation to optimize the tertiary recovery stage of the field development is discussed. This work compares the optimum solution found using an aggregate objective function and the solution repository (Pareto front) generated by the multi-objective optimization process. The comparison indicates the existence of potential multi-solutions satisfying certain criteria in a CO2-EOR project designing, which cannot be found using traditional weighted sum method. The optimization results provide significant insight into the decision-making process of CO2-EOR project when multiple objective functions are considered.