Multi-objective optimization workflow for CO2 water-alternating-gas injection assisted by single-objective pre-search
Renfeng Yang, Wei Zhang, Shuaichen Liu, Bin Yuan, Wendong Wang
Abstract
CO 2 Water-Alternating-Gas (CO 2 -WAG) injection is not only a method to enhance oil recovery but also a feasible way to achieve CO 2 sequestration. However, inappropriate injection strategies would prevent the attainment of maximum oil recovery and cumulative CO 2 storage. Furthermore, the optimization of CO 2 -WAG is computationally expensive as it needs to frequently call the compositional simulation model that involves various CO 2 storage mechanisms. Therefore, the surrogate-assisted evolutionary optimization is necessary, which replaces the compositional simulator with surrogate models . In this paper, a surrogate-based multi-objective optimization algorithm assisted by the single-objective pre-search method is proposed. The results of single-objective optimization will be used to initialize the solutions of multi-objective optimization, which accelerates the exploration of the entire Pareto front . In addition, a convergence criterion is also proposed for the single-objective optimization during pre-search, and the gradient of surrogate models is adopted as the convergence criterion. Finally, the method proposed in this work is applied to two benchmark reservoir models to prove its efficiency and correctness. The results show that the proposed algorithm achieves a better performance than the conventional ones for the multi-objective optimization of CO 2 -WAG.