基于支持向量机与粗糙集的隔爆电动机故障诊断  

Fault diagnosis of explosion proof motor based on SVM and RS

作　　者：马宪民[1] 张兴[1] 张永强[1,2] 

机构地区：[1]西安科技大学电气与控制工程学院,陕西西安710054 [2]神华宁煤集团矿山机械制造维修分公司,宁夏银川750001

出　　处：《工矿自动化》2017年第2期35-40,共6页Journal Of Mine Automation

基　　金：国家自然科学基金项目(51277149);陕西省教育厅科研计划项目(14JK1472)

摘　　要：针对煤矿井下隔爆电动机故障数据获取难且故障数据杂乱、非线性等问题,提出了一种基于支持向量机与粗糙集的隔爆电动机故障诊断方法。该方法采用小波包对隔爆电动机定子瞬时功率进行频谱分析,并提取故障特征量;利用粗糙集的约简特性消除故障特征量冗余数据,将约简后的故障特征量作为支持向量机的输入样本,实现隔爆电动机转子故障诊断和分类。仿真结果表明,该方法故障诊断结果准确率达到92.857 1%。In view of problems that fault data acquisition of explosion proof motor in underground coal mine was difficult and fault data was clutter and nonlinear, a fault diagnosis method of explosion proof motor based on SVM and RS was proposed. Wavelet packet is used to analyze instantaneous power of stator of explosion proof motor, and extract fault feature. Redundant data of fault feature is eliminated using reduction feature of rough set, and the reduced feature is used as input sample of support vector machine, to realize rotor fault diagnosis and classification of explosion proof motor. The simulation results show that fault diagnosis accuracy of the proposed method reaches 92. 8571%.

关 键 词：隔爆电动机 故障诊断 支持向量机 粗糙集 瞬时功率 小波包 

分 类 号：TD614[矿业工程—矿山机电]