Fault Recovery and Reconfiguration of Distribution Networks Based on Membrane Computing Multi-Objective Optimization Algorithm

To improve post-fault service restoration and network reconfiguration in distribution networks with distributed generation, this paper proposes a fault recovery and reconfiguration method based on a membrane computing multi-objective optimization algorithm. The proposed method formulates the restoration problem as a multi-objective optimization model that simultaneously considers load-restoration maximization, switching-operation minimization, network-loss reduction, and voltage-deviation minimization, while prioritizing the restoration of critical loads. Within the membrane-computing multi-objective optimization algorithm framework, the hierarchical parallel structure and evolutionary mechanisms of membrane systems are employed to enhance global search capability. Non-dominated sorting and crowding-distance calculation are incorporated to generate a well-distributed Pareto solution set, providing decision-makers with multiple candidate schemes for island partitioning and network reconfiguration. A weighted decision-making strategy is then used to select the optimal restoration scheme from the Pareto solution set. The proposed method is validated on the IEEE 33-bus distribution system. Simulation results show that, under distributed generation integration, the proposed method can effectively partition electrical islands, improve critical-load restoration, reduce network losses and voltage deviations, and support efficient post-fault restoration and reconfiguration of distribution networks. Compared with BWO, the best-performing benchmark method, the proposed method increases the total load-restoration rate from 92.1% to 95.3%, representing an improvement of 3.2 percentage points. It also reduces active power losses from 112.6 kW to 89.4 kW, a reduction of 20.6%, and decreases the maximum voltage deviation from 0.036 p.u. to 0.023 p.u., a reduction of 36.1%.