Research On The Characteristics Of Series Arc Fault In Multi-loads Circuits

In the low voltage power system,when the series arc fault occurs in the multi-loads circuits,the characteristics of the arc are concealed because of the complex circuits.The arc fault circuit interrupter is not easy to detect the series arc fault of the next stage line.In this paper,the series arc fault in the multi-loads circuits of low voltage power system is studied.The characteristics of the series arc fault in each line point are analyzed.The effective arc signal which can reflect the characteristics of the series arc fault current in the multi-loads circuit is extracted and the timefrequency characteristic is studied.This research is of great significance for the safe and reliable operation of low voltage power system.First,the series arc fault experiment of multi-loads circuits is built,and the different fault points are selected.The experiment of the single load circuit,double loads circuits,four loads circuits and six loads circuits is carried out,and the trunk current signal of the arc fault at each point are obtained.The characteristics of the time domain and frequency spectrum of the current are analyzed.Then,the wavelet packet transform is used to the 5 layer decomposition and reconstruction of the current signal of the trunk line,the wavelet packet energy entropy is used to measure the degree of the change of the reconstructed wavelet packet coefficients before and after the arc fault occur,the change rate of the energy entropy of the wavelet packets in different experimental groups is counted,and the effective signal of the series arc fault is extracted.Finally,the variational mode decomposition(VMD)and Winger distribution(WVD)are used to analysis the time-frequency of the extracted effective signals.The characteristic modal components that can characterize the characteristics of the arc fault was found,the arc fault characteristic signal is constructed and the feature extraction is combined with the extreme learning machine(ELM)to identify arc fault.The results show that the extracted characteristic modal components have significant changes after the arc fault occuring.The time domain characteristics and the WVD edge characteristics of the arc fault characteristic signal can describe the series arc fault characteristics of the multi-loads circuit accurately and effectively,which can be used to provide evidence for the diagnosis of multi-loads series arc fault.