基于盒维数序列变异系数的轴承外圈失效特征  

Bearing Outer Race Failure Feature Based on C.V of Box Dimension Series

作　　者：王华中[1] 魏晖[2] 高志刚[1] WANG Hua-zhong;WEI Hui;GAO Zhi-gang(School of Automotive Engineering,Jiangxi University of Technology,Jiangxi Nanchang 330098,China;Collaborative Innovation Center,Jiangxi University of Technology,Jiangxi Nanchang 330098,China)

机构地区：[1]江西科技学院汽车工程学院,江西南昌330098 [2]江西科技学院协同创新中心,江西南昌330098

出　　处：《机械设计与制造》2018年第10期149-152,共4页Machinery Design & Manufacture

基　　金：江西省重点研发计划(工业领域)(20161BBE50079;20142BBE50061);江西省教育厅科学技术研究项目(GJJ151157)

摘　　要：特征向量选择是轴承失效检测与诊断的关键。作为解释复杂系统的动力学行为机制的重要工具,分形理论及分形维数常被用来描述轴承的故障特征,然而某状态下的分形维数难以刻画轴承失效的过程。针对轴承振动的非线性,首先介绍了分形理论,然后尝试选择盒维数及其变异系数作为轴承故障特征,并以IMS的轴承全寿命失效实验数据为样本,对盒维数的轴承外圈失效诊断进行了验证。认为:外圈正常工况与失效工况的盒维数均值差异不大,但失效工况的盒维数变异系数有着显著的增加。通过盒维数的变异性或可为滚动轴承的外圈失效诊断提供依据。Selection of feature vector is the critical issue of bearing failure detection and diagnosis.As an important mathematical tool for explaining the dynamic behavior of complex system,fractal theory and fractal dimensions are usually used to describe the fault feature of bearing.But fractal dimension of certain status is unfit for charactering the failure process.At first,it introduced fractal theory in consideration of the nonlinearity of bearing vibration,then selected box dimension as feature vector for bearing fault feature.Finally,taken the bearing entire life cycle data generated by IMS as test samples,simulation validation was processed.Results suggested that:regardless of whether the outer race is in normal condition or failure condition,mean values of box dimensions are almost equal;but in failure condition,the C.V of box dimensions are significantly increased,maybe the variability could provide evidence for outer race failure.

关 键 词：轴承 分形 盒维数 失效 变异系数 

分 类 号：TH16[机械工程—机械制造及自动化] TH165.3