Research On Arc Detection Method For Low Voltage Line Fault Based On BP Neural Network

In recent years,electrical fire accidents occur frequently,especially in low-voltage lines,because of the lack of effective detection methods for fault arc,it is easier to induce electrical fire and cause significant losses.In view of the above problems,this paper proposes a low voltage line fault arc detection method based on BP neural network.The main work of this paper includes the following points:1.In view of the frequent social phenomenon of electrical fire,the causes of fire are analyzed,and it is pointed out that fault arc detection technology is the focus of electric fire prevention and control.2.A fault arc simulation system based on Cassie arc model is built.Through simulation experiments,a large number of current waveform data of different types and power loads are obtained under normal and arc conditions.3.Using Fourier transform to process and analyze the waveform data,the spectrum of fault arc current is obtained,and the harmonic characteristic of fault arc is extracted.4.Build BP neural network,and get the optimal selection of network parameters through experimental comparative analysis.The extracted harmonic characteristics of fault arc are taken as samples to analyze BP network.After training,additional data samples are used to test the network.The test results show that the detection success rate of fault arc reaches 94%.The simulation results show that the fault arc detection method proposed in this paper is feasible and can provide some reference for the future research of low voltage line fault arc detection technology.