机械加工过程中轴承故障诊断方法研究  

Bearing Fault Diagnosis Method in the Machining Process

作　　者：杨婧婷 YANG Jing-ting(Engineering Training Center,Shenyang Aerospace University.Liaoning Shenyang 110136,China)

机构地区：[1]沈阳航空航天大学工程训练中心,辽宁沈阳110136

出　　处：《机械设计与制造》2021年第2期112-116,共5页Machinery Design & Manufacture

基　　金：国家自然科学基金项目(51675450)。

摘　　要：为了提高机械加工过程中滚动轴承故障诊断准确度,提出了基于新的解析能量算子的轴承故障诊断方法。在分析Teager能量算子缺陷基础上,提出了新的能量算子,命名为解析能量算子;解析能量算子无需满足Teager能量算子的使用条件,且能够更好地跟踪故障信号的冲击瞬态特征;使用EMD算法分解原始振动信号,给出了多指标融合的IMF分量筛选方法;对筛选出的IMF分量,计算其解析能量谱样本熵作为特征参数;在分析二叉树SVM缺点基础上,建立了投票法SVM多分类器;经试验验证,基于Teager能量谱样本熵的故障诊断准确率为78%,而基于解析能量谱样本熵的故障诊断准确率为100%,充分证明了解析能量算子在轴承故障诊断中的有效性。To improve rolling bearing fault diagnosis accuracy in the machining process,a new energy operator named analytic energy operator used to diagnose bearing fault is proposed.By analyzing shortcoming of Teager energy operator,a new energy operator named analytic energy operator is put forward.Analytic energy operator does not need to satisfy using conditions of Teager energy operator,and can track impact transient characteristics of fault signal.EMD algorithm is used to decompose original vibration signal,and IMF component screening method is given by fusing multi-index.Compute analytic energy spectrum sample entropy of IMF component as characteristic parameters.By analyzing defect of binary tree SVM,voting SVM multiple classifiers is built.Clarified by trial,fault diagnosis accuracy based on Teager energy spectrum sample entropy is 78%,and 100%based on analytic energy spectrum sample entropy,which proves validity of analytic energy operator on bearing fault diagnosis.

关 键 词：轴承 故障诊断 解析能量算子 样本熵 投票法支持向量机 

分 类 号：TH16[机械工程—机械制造及自动化] TH133