变电站高电压试验设备故障诊断方法研究  

Research on Fault Diagnosis Method of High Voltage Test Equipment in Substation

作　　者：孙立群 张灿 SUN Liqun;ZHANG Can(Nanjing Power Supply Branch of State Grid Jiangsu Electric Power Co.,Ltd.,Nanjing 210000,Jiangsu,China)

机构地区：[1]国网江苏省电力有限公司南京供电分公司,江苏南京210000

出　　处：《电气传动自动化》2022年第6期30-32,29,共4页Electric Drive Automation

摘　　要：由于高电压试验设备的数据具有线性不可分割的属性特征,导致对其故障诊断的难度较大,为此,本文提出一种变电站高电压试验设备故障诊断新方法。该方法将高电压试验设备运行信号包络谱作为特征提取基准数据,将频率信息作为特征参量,通过Hilberx变换将数据映射到高维空间中,在超平面实现高电压试验设备数据特征提取,以提取的设备运行数据特征参量为基础建立关联关系,结合决策树理论划分当前数据集中的最优特征,并根据各个特征节点与其的距离,从而实现对其故障的诊断。实验测试结果表明,该新方法诊断故障的精确率始终稳定在94.00%以上,召回率和F值也稳定在90.00%左右。Since the data of the high-voltage test equipment has linear indivisible characteristics,it is difficult to diagnose its faults. Therefore,this paper proposes a new method for fault diagnosis of high-voltage test equipment in substations. In this method,the envelope spectrum of the operating signal of the high-voltage test equipment is used as the benchmark data for feature extraction,and the frequency information is used as the feature parameter.The data is mapped into the high-dimensional space through Hilberx transformation,and the feature extraction of the high-voltage test equipment data is realized in the hyperplane. The association relationship is established based on the extracted operating data feature parameters of the equipment,the optimal features in the current data set are divided by combining the decision tree theory,and according to the distance between characteristic nodes,thus the fault diagnosis is realized. The experimental results show that the accuracy of the new fault diagnosis method is always above 94.00%,and the recall rate and F value are also stable at about 90.00%.

关 键 词：变电站 高电压试验设备 故障诊断 惩罚因子 决策树理论 最优特征参量 

分 类 号：TM63[电气工程—电力系统及自动化]