换流站阀水冷系统主泵轴承故障诊断方法研究  

Methods for Bearing Fault Diagnosis of Converter Station's Valve Cooling System Main Circulating Pump

作　　者：黄山[1] 方宇[1] 傅坚[2] 周孝法[2] 胡定玉[1] 

机构地区：[1]上海工程技术大学城市轨道交通学院,上海201620 [2]国网上海市电力公司检修公司,上海200063

出　　处：《测控技术》2016年第5期37-40,共4页Measurement & Control Technology

基　　金：国网上海市电力公司科技项目(5209501350S9);上海工程技术大学研究生科研创新项目(EI-0903-14-01026)

摘　　要：对换流站阀水冷系统主泵轴承的故障诊断方法进行了比较研究。采用支持向量机作为分类工具,分别利用时域特征值和小波包分解取样本熵作为样本训练,比较分类准确率,并择优用于换流站阀水冷系统主泵的轴承故障诊断。首先,利用轴承故障试验台的数据,对采用时域特征值和小波包分解取样本熵作为样本训练的分类准确率进行了比较,结果表明小波包分解取样本熵值比时域特征参数更适合用于特征故障分类。然后将小波包分解取样本熵值用于换流站阀水冷系统主泵的轴承故障诊断,结果显示分类准确率达98%,完全满足工程运用需求。Method for the fault diagnosis of the bearing of converter station＇ s valve cooling system main circulating pump is studied.Using SVM as the classification tool,the time domain eigenvalue and the sample entropy of wavelet packet decomposition are respectively utilized for training sample to compare the classification accuracies.First,a validation experiment is conducted to implement the comparison,and the results show that sample entropy of wavelet packet decomposition is more suitable for classification than time domain eigenvalue.Then the wavelet packet decomposition is used for the fault diagnosis of the bearing of converter station＇ s valve cooling system main circulating pump,and an accuracy of 98%for the classification is achieved,which means that the method completely meets the engineering requirements.

关 键 词：小波包分解 样本熵 支持向量机 轴承故障诊断 

分 类 号：TH165.3[机械工程—机械制造及自动化]