A Residual Current Location Method Based on a Back Propagation Neural Network Optimized by a Genetic Algorithm

With the diversified development of source and load in low-voltage distribution system, the time-frequency characteristics of residual current under multi-work condition interference have strong time-variability, and the leakage fire caused by this is difficult to predict. It is also a hidden danger of electric shock. In this paper, a method of locating the residual current of low voltage distribution system based on a back propagation neural network optimized by genetic algorithm is proposed. The residual current and bus current are determined as key input signals based on the theoretical derivation of the impedance method. Through the simulation model of the direct current system, this method has been validated to accurately predict the fault distance under different single-phase grounding resistances.