Research On Low Voltage Series Fault Arc Diagnosis Technology

Nowadays,electrical fires have become the primary cause of fire accidents,and arc faults are one of the main causes of electrical fires.At present,distribution line fault protection devices cannot provide comprehensive protection for arc faults.Fault arcs have become a loophole in power protection,which has important theoretical significance and engineering value for the study of fault arc detection technology.This article takes low-voltage series fault arc as the research object,and studies its diagnostic technology.The main work is as follows:Firstly,the basic concept of arc,the cause of fault arc and the basic characteristics of series fault arc in this article are introduced.Among them,the basic concept of arc involves the arc generation mechanism,the composition of the arc and its voltage distribution and common arc classification;the main causes of fault arc include insulation carbonization,air ionization caused by the outside and short circuit,which lays a theoretical foundation for the analysis of low-voltage series fault arc.Secondly,the Cassie arc model was used to conduct a low-voltage series fault arc simulation experiment on resistive load,inductive load,single-phase induction motor load and nonlinear load,and the fault current signal and arc voltage signal were analyzed.According to UL1699,a low-voltage series fault arc test platform was built,and a variety of different types of electrical loads were selected as test loads.Eight sets of series fault arc tests under single load operation and four sets of series fault arc tests under parallel operation were designed,and the current signals in series arc fault state and normal working state are collected.Then,the collected test current signal was analyzed in time domain,frequency domain and time-frequency domain.In the time domain,the zero-break time,rate of change,absolute value of the average value and kurtosis coefficient of the test current signal are calculated.Calculate the harmonic factor,total harmonic distortion rate,subband energy ratio and frequency centroid of the test current signal in the frequency domain.In the time-frequency domain,the test current signal is decomposed into five layers to calculate the maximum value of the wavelet transform modulus of each detail layer,the energy of the frequency band of each detail layer and the the wavelet Shannon entropy.Finally,based on the standard PSO,the dynamic parameter adjustment strategy and the particle dispersion mutation strategy are introduced to improve it.The improved PSO is used to optimize the standard BP neural network,and an improved PSO-BP neural network model is established.A low-voltage series fault arc detection algorithm based on the improved PSO-BP neural network is proposed,and the low-voltage series fault arc detection algorithm is imulated and verified.There are 46 figures,9 tables and 82 references in this thesis.