基于矢功率谱和D-S证据理论分层融合的旋转机械故障诊断方法  

A rotating machinery fault diagnosis method based on fusing vector power spectrum and D-S evidence theory

作　　者：杨春燕[1] 吴超[1] 李宏伟[1] 

机构地区：[1]郑州轻工业学院机电工程学院,河南郑州450002

出　　处：《郑州轻工业学院学报（自然科学版）》2012年第4期30-32,共3页Journal of Zhengzhou University of Light Industry:Natural Science

基　　金：河南省科技攻关计划项目(122102210122)

摘　　要：提出基于矢功率谱和D-S证据理论分层融合的旋转机械故障诊断方法,该方法把转子的2个截面信息分别以矢功率谱进行数据层融合,提取矢功率谱的特征输入到径向基概率神经网络分类器进行故障识别,最后把两截面诊断结果输入D-S证据理论融合中心进行决策层融合.实验结果表明,该方法可降低故障诊断的不确定性,并提高故障决策准确率.A rotating machinery fault diagnosis method based on fusing vector power spectrum and D-S evi- dence theory was presented. The method was that two-section information was fused respectively in data lay- er by vector power spectrum, and then the characteristics which were extracted from the vector power spec- trum were input to the basis probabilistic neural network classifier for fault identification, and finally, the two-section diagnosis results were entered into D-S evidence theory for decision level fusion. The experi- ment results showed that the method reduces the diagnostic uncertainty and had high correct recognition rate.

关 键 词：旋转机械故障诊断 矢功率谱 D—S证据理论 数据融合 

分 类 号：TH165.3[机械工程—机械制造及自动化]