Based on genetic algorithm to optimize learn vector quantization neural network is a new algorithm of distribution network fault line selection

For the problem of low accuracy of using single fault signal feature line selection in distribution network, distribution network grounding fault comprehensive line selection method based on genetic algorithm (GA) to optimize learning vector quantization (LVQ) neural network is proposed. Problem of fault line selection as special classification problem, by fast Fourier transform (FFT) and wavelet analysis WPT algorithm, the corresponding fault characteristics for the zero sequence current signal from the different fault types are extracted as neural network input. LVQ neural network trained by GAs to get the optimal initial weights, the trained LVQ neural network is used to fault line selection. Compared with line selection results of the traditional LVQ neural network, the network performance is improved effectively, and the risk of misjudge is reduced. The simulation results show that the genetic algorithm to optimize LVQ neural network in the application of comprehensive ground fault line selection for distribution network is feasible.