滚动轴承外圈失效的盒维数与信息维数特征  

Failure Feature Extraction Based on Box-Counting Dimension and Information Dimension of Rolling Bearing Outer Race

作　　者：张美芸[1] 魏晖[2,3] 樊富起[3] ZHANG Mei-yun;WEI Hui;FAN Fu-qi(School of Mechanic Engineering,Jiangxi University of Technology,Jiangxi Nanchang330098,China;Collaborative Innovation Center,Jiangxi University of Technology,Jiangxi Nanchang330098,China;School of Automotive Engineering,Jiangxi University of Technology,Jiangxi Nanchang330098,China)

机构地区：[1]江西科技学院机械工程学院,江西南昌330098 [2]江西科技学院协同创新中心,江西南昌330098 [3]江西科技学院汽车工程学院,江西南昌330098

出　　处：《机械设计与制造》2019年第9期227-230,共4页Machinery Design & Manufacture

基　　金：江西省重点研发计划(工业领域)(20161BBE50079);江西省教育厅科学技术研究项目(GJJ151157);南昌市指导性科技计划项目(2017-KJZCCXY-014);江西省汽车服务工程及产业升级协同创新中心开放基金(18 XTKFYB01)

摘　　要：旋转机械故障诊断的关键在于故障特征的提取,由于轴承振动信号的非线性与非平稳性,传统平稳信号处理方法在故障特征提取中存在不足,而多重分形等非平稳方法存在计算冗余。认为故障状态下轴承振动信号的空间的占布与概率密度均发生变化,因此可采用盒维数与信息维数描述故障特征。在此基础上,借助美国辛辛那提大学IMS中心公开的全寿命轴承实验数据,分别分析了滚动轴承外圈在正常和故障情况的信息维数和盒维数,认为:在故障状态下,轴承振动信号的信息维数与盒维数均有降低,这一特点有利于提高故障模式的可分性。The key to fault diagnosis of rotating machinery is the extraction of fault features.Because of the non-linearity and non-stationaritycharacters of bearing vibration signals,traditional stationary signal processing methods have some shortcomings in fault feature extraction,and non-stationary methods such as multi-fractal method have computational redundancy.It suggested that the distribution and probability density of bearing vibration signals in fault state are changed,so the box dimension and information dimension can be used to describe the fault characteristics.On this basis,the information dimension and box dimension of rolling bearing outer rings under normal and fault conditions are analyzed respectively with the help of the life-cycle bearing experimental data published by IMS Center of University of Cincinnati in the United States.It is concluded that the information dimension and box dimension of bearing vibration signals are reduced under fault conditions,which is beneficial to the improvement of bearing vibration signals.

关 键 词：滚动轴承 信息维数 盒维数 故障诊断 

分 类 号：TH16[机械工程—机械制造及自动化] TH212