Configuration optimization of an off-grid multi-energy microgrid based on modified NSGA-II and order relation-TODIM considering uncertainties of renewable energy and load

This study develops a two-stage hybrid decision framework to configure an off-grid multi-energy microgrid (MEMG) while considering uncertainties in renewable energy resources and load demand. In the first stage, a non-dominated sorting genetic algorithm (NSGA)-II modified by reinforcement learning is employed to obtain Pareto solutions. Its three objectives are minimum total annual cost (TAC), loss of energy supply possibility (LESP), and energy abandonment rate (EAR). In the second stage, the weights for TAC, LESP, and EAR are obtained using the order relation method, and the best solution is chosen from Pareto solutions using a ranking method. Then, a case study is carried out to validate the rationality and feasibility of the proposed framework. The results show that when uncertainties and electricity storage are simultaneously considered, the respective values of TAC, LESP, and EAR are 81.148 million yen , 16.146%, and 1.486%, respectively. A comparative analysis illustrates how the proposed method is better than NSGA-II, because the values of spacing metric and diversity are 0.0004 and 2.3546 obtained with the proposed method, and the former is smaller than 0.0013 and the latter is larger than 2.2713 obtained by employing NSGA-II. Therefore, the proposed framework is helpful to construct an off-grid renewable MEMG.