A Two-terminal Fault Location Method Based on Particle Swarm Optimization BP Neural Network

This paper proposes a two-terminal fault location method for distribution network based on particle swarm optimization BP neural network (PSO-BP). First, the traveling waves (TWs) of the aerial mode voltage detected are aligned according to the modulus maximum of the initial TWs wavefront, which is defined as the first coincident wavefront. Then the other coincident wavefronts and the corresponding coincidence time will be obtained. The analysis shows that the fault distance can be accurately reflected by the time difference between the first coincident wavefront and the second coincident wavefront at the sampling frequency of 10MHz. In order to reduce the sampling frequency to 1MHz, the paper combines that time difference with the modulus maxima ratio of the first and second coincident wavefronts at both ends respectively as the input of the PSO-BP neural network to construct the model of fault location. Simulation results show that the proposed method has high accuracy and is not affected by fault location and fault resistance. Keywords—distribution network, fault location; wavelet transform; particle swarm optimization algorithm; two-ended fault location; modulus maximum ratio, sample frequency