摘要
高压断路器机械故障特征都极为类似,缺少必要分级分类过程会导致识别精度大幅降低。本文提出基于KPCA-SVM的高压断路器机械故障诊断技术。采集高压断路器机械故障样本数据,使用核主元分析方法提取样本中故障的特征向量,将其输入到支持向量机内,完成故障多级分类,通过3个支持向量机训练与分类设备正常状态以及拐臂润滑不足、分闸弹簧脱落两个典型机械故障和其它故障,实现高压断路器机械故障准确、高效诊断。实验结果表明:该技术将正则化参数和核函数参数分别设置为30,15,能获得更优异的诊断性能;诊断各类型缺陷的准确度高达91%,且诊断用时均低于40 s,效率较高。
The mechanical fault characteristics of high voltage circuit breakers are very similar.The lack of the necessary classification process will lead to a significant reduction in identification accuracy.The mechanical fault diagnosis technology of high voltage circuit breaker based on KPCA-SVM is proposed.Collect the mechanical fault sample data of high-voltage circuit breaker,extract the fault feature vector in the sample by using the kernel principal component analysis method,input it into the support vector machine,and complete the multi-level fault classification.Through the training and classification of three support vector machines,the normal state of the equipment,two typical mechanical faults:insufficient lubrication of the crank arm,falling off of the opening spring,and other faults.Realize accurate and efficient diagnosis of mechanical fault of high voltage circuit breaker.The experimental results show that this technique can obtain better diagnostic performance by setting the regularization parameters and kernel function parameters to 30 and 15,respectively.The accuracy of diagnosing various types of defects is as high as 91%,and the diagnosis time is less than 40s,with high efficiency.
作者
张迅
黄军凯
赵超
许逵
吴建蓉
陈沛龙
ZHANG Xun;HUANG Junkai;ZHAO Chao;XU Kui;WU Jianrong;CHEN Peilong(Electric Power Research Institute of Guizhou Power Grid Co.,Ltd,Guiyang 550002,China)
出处
《测试技术学报》
2023年第2期158-164,共7页
Journal of Test and Measurement Technology
基金
贵州电网科技资助项目(GZKJXM20200528)。
关键词
KPCA-SVM
高压断路器
机械故障诊断
核主元分析
支持向量机
KPCA-SVM
high voltage circuit breaker
mechanical failure diagnosis
kernel principal component analysis
support vector machine