Power systems have various structures, and the occurrence of faults is inevitable. In this project, the recovery and reconstruction problem of new energy distribution network faults is studied, the mathematical model of distribution network reconstruction is constructed, the particle swarm algorithm is chosen as the intelligent optimization algorithm for solving the recovery and reconstruction problem of distribution network, and it is optimized by combining the parameter improvement and the genetic algorithm. The improved hybrid particle swarm algorithm in this paper has better optimization searching effect, and the number of evolutionary generations to reach the best fitness value is much smaller than that of the traditional particle swarm algorithm. Taking the IEEE33 node system as an example for case analysis, it is found that this paper’s method has good universality, and when accessing the DG and reconfiguring, the node voltages of the whole system are significantly improved (6.09% and 4.37%) to ensure that all node voltage distributions satisfy the confidence constraints, and at the same time the system’s active network losses are also significantly reduced (57.35% and 40.59%), which reflects the superior fault self-healing performance, and has a very good performance. The performance of fault self-healing has good practical value.
