基于小波变换和奇异值分解的串联电弧故障检测方法

根据线路中电流信号的变化来检测电弧故障,小波变换是一种常用的检测方法,但是单纯利用小波变换对于正常情况和电弧故障的区分并不明显,而且其结果存在很大的冗余。针对这一问题,提出了采用一种基于小波变换和奇异值分解的串联电弧故障检测的方法。利用电弧模拟发生装置产生串联故障电弧,采集在多种负载下线路正常工作和发生串联电弧故障时的电流。首先对采集的电流信号进行离散小波变换,得到离散小波系数序列,构造特征矩阵;然后对特征矩阵进行奇异值分解,并定义电流信号的特征参数,利用特征参数作为串联电弧故障检测的依据。试验结果表明:正常情况和电弧故障下的特征参数区分明显且没有交叉,易于确定阈值,利用该方法进行串联电弧故障检测的准确率较高,且大大压缩了小波变换结果的冗余性。

Detection Method of Series Arcing Fault Based on Wavelet Transform and Singular Value Decomposition

Wavelet transform was a commonly used method to detect the arcing fault according to the change of the current signal in the circuit.However,it was not easy to distinguish the normal condition from arcing fault when simply using wavelet transform,and there was a lot of redundancy in the results.In order to solve this problem,a new detection method of series arcing fault which based on wavelet transform and singular value decomposition is proposed.An arc generator is used to generate series arcing fault,currents in normal condition and arcing fault are collected under multiple loads.Discrete wavelet transform is firstly used in the collected current signal,and the discrete wavelet coefficient sequence is obtained.Then,based on singular value decomposition of characteristic matrix,the characteristic parameters of current signal are defined and used as the basis of the series arcing fault detection.The experimental results show that it is easy to distinguish the characteristic parameters and there is no cross under normal condition and series arcing fault,thus it is easy to determine the threshold value.The accuracy of the series arcing fault detection is high,and the redundancy of the wavelet transform is greatly compressed.