基于OOMP与A-T谱的滚动轴承故障特征提取  

Fault feature extraction for rolling bearings based on OOMP and A-T spectrum

作　　者：夏均忠 郑建波 白云川 吕麒鹏 杨刚刚 XIA Junzhong;ZHENG Jianbo;BAI Yunchuan;L Qipeng;YANG Ganggang(Department of Military Vehicle Engineering,Academy of Military Transportation,Tianjin 300161,China)

机构地区：[1]陆军军事交通学院军用车辆工程系

出　　处：《振动与冲击》2019年第21期86-90,96,共6页Journal of Vibration and Shock

摘　　要：为解决变转速下正交匹配追踪(OMP)过度匹配和非正交投影的问题,提出优化正交匹配追踪(OOMP)。根据轴承故障振动信号的特性,构建组合时频原子字典与OMP匹配;将鲸鱼优化算法引入到OMP中选择与残差信号匹配的最优原子,实现信号重构和故障特征增强。为避免阶次追踪的缺陷,引入角度-时间(A-T)谱提取故障特征。试验验证,OOMP可有效增强轴承故障特征,A-T谱用于变转速下轴承故障特征提取效果良好。In order to solve problems of orthogonal matching pursuit(OMP)’s over-matching and non-orthogonal projection under variable rotating speed, an optimized orthogonal matching pursuit(OOMP) was proposed. According to characteristics of bearing fault vibration signals, the combined time-frequency atom dictionary was constructed to match OMP. The whale optimization algorithm was introduced into OMP to choose the optimal atom matching residual signals, and realize signal reconstruction and fault feature enhancement. To avoid defects of the order tracking, an angle-time(A-T) spectrum was introduced to extract fault features. The test results showed that the OOMP can effectively enhance bearing fault characteristics;using the A-T spectrum has a good effect on bearing fault feature extraction under variable rotating speed.

关 键 词：滚动轴承 故障特征增强 特征提取 优化正交匹配追踪 角度-时间谱 

分 类 号：TN911.23[电子电信—通信与信息系统] TB53[电子电信—信息与通信工程]