基于S变换和遗传算法优化SVM的GIS机械故障诊断  

GIS Mechanical Fault Diagnosis Based on S-transform and SVM Optimized by Genetic Algorithm

作　　者：陈明 马宏忠[1] 潘信诚 张利[2] 屈斌[2] CHEN Ming;MA Hongzhong;PAN Xincheng;ZHANG Li;QU Bin(College of Energy and Electrical Engineering,Hohai University,Nanjing 211100,China;Electric Power Research Institute,State Grid Tianjin Electric Power Company,Tianjin 300041,China)

机构地区：[1]河海大学能源与电气学院,江苏南京211100 [2]国网天津市电力公司电力科学研究院,天津300041

出　　处：《电力信息与通信技术》2020年第5期1-6,共6页Electric Power Information and Communication Technology

基　　金：国家自然科学基金项目(51577050);国网天津市电力公司重点科技项目资助(SGTJDK00DYJS1900056)。

摘　　要：为实现对气体绝缘金属开关设备(GIS)机械故障在线监测,文章提出一种基于S变换和遗传算法优化SVM的GIS机械故障诊断方法。首先用振动传感器采集GIS中断路器动作期间的表面振动信号,然后对采集到的振动信号进行S变换得到时频谱矩阵,再利用改进的奇异值分解法对时频谱矩阵进行降维分解得到一系列的奇异值即可形成特征向量,最终利用遗传算法优化的二叉树支持向量机实现不同工况的准确分类。实验结果表明:GIS在不同机械故障状态下运行时振动信号发生不同畸变,文章所提的GIS机械状态诊断方法可准确提取故障特征,能够为GIS机械故障诊断提供必要的参考。In order to realize on-line monitoring of GIS mechanical fault,this paper proposes a method of GIS mechanical fault diagnosis method based on S-transform and SVM optimized by genetic algorithm.Firstly,the vibration sensor is used to collect the surface vibration signal during the operation of the circuit breaker in GIS,then the S-transform of the collected vibration signal is used to obtain the time spectrum matrix,and the improved singular value decomposition method is used to reduce the dimension of the time spectrum matrix to obtain some columns of singular values,which can form the eigenvector.Finally,the support vector machine optimized by genetic algorithm is used to realize the accurate categorization under different conditions.The experimental results show that the vibration signal of GIS has different distortion when it runs under different mechanical fault conditions.The mechanical state diagnosis method proposed in this paper can accurately extract the fault features and provide necessary reference for the mechanical fault diagnosis of GIS.

关 键 词：气体绝缘金属开关设备(GIS) 振动信号 S变换 改进的奇异值 故障诊断 

分 类 号：TM561[电气工程—电器]