基于小波奇异值和支持向量机的高压线路故障诊断

Fault diagnosis of HV transmission lines based on wavelet singular value and support vector machine

提出了基于小波奇异值(WSV)和支持向量机(SVM)的电力系统故障类型识别的新方法。利用WSV来量化故障特征,再与SVM结合进行故障类型识别。对故障线路三相电流信号进行小波包变换分解,获取故障信号的小波细节系数;利用相重构技术将小波细节系数向量形成系数矩阵,并对该矩阵作奇异值分解,获取小波奇异值;将小波奇异值向量输入到SVM分类器进行故障类型识别。仿真表明,对于不同的故障类型,其小波奇异值分布明显不同,而对于同一类型故障,其小波奇异值分布在不同的故障位置、过渡电阻的情况下仍保持很大的相似性。SVM具有训练样本少、训练时间短、识别率高等优点。

Based on the wavelet singular value（WSV） and support vector machines（SVM）,a new fault diagnosis method in HV transmission lines is proposed.The new method uses wavelet singular value to quantify the fault signature and combines it with support vector machines for the fault type identification.First of all,using wavelet to decompose the three-phase fault current and obtain the wavelet detail coefficient of the fault signal.Secondly,according to the phase space reconstruction theory,forming the coefficient matrix with the wavelet detail coefficient and obtaining the wavelet singular value by using singular value decomposition to the coefficient matrix.Thirdly,inputing the wavelet singular value into the SVM classifier and identifying the fault type.The simulation results show that the wavelet singular value distribution is obviously different to different faults,and to the same faults,the wavelet singular value distribution is similar under different fault transition resistance and location.SVM has the advantages of less training samples,short training time and high recognition rate.