魏亚军,李开灿,董振

• 1:国网山东省电力公司济宁供电公司

摘要(Abstract)：

针对现有的配网电缆因局部放电随机且复杂导致故障诊断尚存不足的问题，本文提出了一种融合局部放电谱图的矩特征、Tamura纹理特征和方向梯度直方图特征的特征提取方法。该方法从形状特征、全局和局部方向纹理特征3方面更加全面地表达局部放电谱图特征，对局部放电谱图的矩特征、Tamura和方向梯度直方图纹理特征进行提取与融合，结合自适应提升算法、后馈神经网络和支持向量机算法均可以实现识别准确率较高的故障诊断效果。然后，将本文方法与后馈神经网络、支持向量机及卷积神经网络、栈式自编码器深度学习算法进行识别对比，结果表明，自适应提升算法的识别准确率最高、耗时最短，具有良好的通用性和稳定性，为配网电缆的故障诊断提供了新的方法。

关键词(KeyWords)： 配网电缆;矩特征;Tamura-方向梯度直方图纹理特征;自适应提升算法;故障诊断

Abstract:

Keywords:

基金项目(Foundation): 国网山东省电力公司科技项目（5206061800AT）

作者(Author): 魏亚军,李开灿,董振

DOI: 10.19635/j.cnki.csu-epsa.000957

参考文献(References)：

• [1]杨帆，杨旗，程鹏，等(Yang Fan,Yang Qi,Cheng Peng,et al).电缆接头内部气隙放电缺陷下的绝缘劣化程度表征方法(Study of cracking extent for gap discharge in insulating material of power cable joint)[J].电工技术学报(Transactions of China Electrotechnical Society),2017,32(2):24-32.
• [2]Morshuis P H F,Smit J J.Partial discharges at DC voltage:their mechanism,detection and analysis[J].IEEE Trans on Dielectrics and Electrical Insulation,2005,12(2):328-340.
• [3]Seo I J,Khan U A,Hwang J S,et al.Identification of insulation defects based on chaotic analysis of partial discharge in HVDC superconducting cable[J].IEEE Trans on Applied Superconductivity,2015,25(3):1-5.
• [4]Jeroense M J P.Charges and Discharges in HVDC Cablesin Particular in Mass-impregnated HVDC Cables[M].TUDelft,Holland:Delft University of Technology,1997.
• [5]黎珞，曹伟东，余洋(Li Luo,Cao Weidong,Yu Yang).基于图像特征挖掘和深度森林的配网电缆终端故障诊断方法(Fault diagnosis method of distribution network cable terminal based on image feature mining and deep forest)[J].电工技术(Electric Engineering),2021(9):155-157,162.
• [6]苗旺，陈平，任志刚，等(Miao Wang,Chen Ping,Ren Zhigang,et al).电流谐波法在电力电缆状态检测诊断的应用进展(Application progress of current harmonic method in power cable state detection and diagnosis)[J].电力系统及其自动化学报(Proceedings of the CSU-EPSA),2021,33(3):73-80.
• [7]郭蕾，曹伟东，张靖康，等(Guo Lei,Cao Weidong,Zhang Jingkang,et al).基于多尺度纹理特征的EPR电缆终端故障诊断方法(EPR cable terminal fault diagnosis method based on multi-scale texture features)[J].电力自动化设备(Electric Power Automation Equipment),2020,40(11):188-198.
• [8]Ma Hui,Chan J,Saha T K,et al.Pattern recognition techniques and their applications for automatic classification of artificial partial discharge sources[J].IEEE Trans on Dielectrics and Electrical Insulation,2013,20(2):468-478.
• [9]Li Jian,Jiang Tianyan,Harrison R F,et al.Recognition of ultra high frequency partial discharge signals using multiscale features[J].IEEE Trans on Dielectrics and Electrical Insulation,2012,19(4):1412-1420.
• [10]Chang Hong-Chan,Kuo Ying-Piao,Lin Han-Wei.Partial discharge recognition system for current transformer using neural network and 2D wavelet transform[J].IEEJ Transactions on Electrical and Electronic Engineering,2012,7(2):144-151.
• [11]陈帝(Chen Di).基于时频分析的设备故障特征提取方法研究(Research on equipment fault feature extraction method based on time-frequency analysis)[J].中国设备工程(China Equipment Engineering),2021(19):11-13.
• [12]张翼，朱永利(Zhang Yi,Zhu Yongli).图信号与图卷积网络相结合的局部放电模式识别方法(Partial discharge pattern recognition method combining graph signal and graph convolution network)[J].中国电机工程学报(Proceedings of the CSEE),2021,41(18):6472-6481.
• [13]李玉杰，田阳普，赵科，等(Li Yujie,Tian Yangpu,Zhao Ke,et al).基于自编码网络的局部放电信号特征提取与识别(Partial discharge signal feature extraction and recognition based on self-encoding network)[J].电力工程技术(Electric Power Engineering Technology),2021,40(3):148-152.
• [14]高盎然，朱永利，张翼，等(Gao Angran,Zhu Yongli,Zhang Yi,et al).基于边际谱图像和深度残差网络的变压器局部放电模式识别(Partial discharge pattern recognition of transformers based on marginal spectrum image and deep residual network)[J].电网技术(Power System Technology),2021,45(6):2433-2442.
• [15]魏振，齐波，左健，等(Wei Zhen,Qi Bo,Zuo Jian,et al).基于局部放电图像特征的换流变压器油纸绝缘缺陷诊断方法(Diagnosis method for oil-paper insulation defects of converter transformer based on partial discharge image characteristics)[J].电网技术(Power System Technology),2015,39(4):1160-1166.
• [16]秦雪，钱勇，许永鹏，等(Qin Xue,Qian Yong,Xu Yongpeng,et al).基于2D-LPEWT的特征提取方法在电缆局部放电分析中的应用(Application of feature extraction method based on 2D-LPEWT in cable partial discharge analysis)[J].电工技术学报(Transactions of China Electrotechnical Society),2019,34(1):170-178.