基于自回归滑动平均最小熵反褶积的滚动轴承故障诊断  

Fault Diagnosis for Rolling Bearing Based on Autoregressive Moving Average Minimum Entropy Deconvolution

作　　者：任学平[1] 张玉皓[1] 袁国静 王朝阁[1] 

机构地区：[1]内蒙古科技大学机械工程学院,内蒙古包头014010 [2]河北钢铁集团承德钢铁公司,河北承德067000

出　　处：《机械设计与制造》2017年第2期53-57,共5页Machinery Design & Manufacture

基　　金：内蒙古自治区自然科学基金项目(2012MS0717)

摘　　要：滚动轴承的时域故障信号含有工作部件或轴承元件间微弱碰撞产生非周期性冲击以及工况噪声成分,造成信号中表征故障信息的周期冲击成分难以提取,无法准确有效的对滚动轴承进行故障诊断。针对这一问题,提出自回归滑动平均最小熵反褶积方法。通过自回归滑动平均模型和最小熵反褶积计算得出正逆两组滤波器系数,其中自回归滑动平均模型计算出的滤波器系数用于分离故障信号中的非周期冲击成分,最小熵反褶积计算出的逆滤波器系数用于恢复故障冲击成分。通过仿真和实验的处理结果证明了方法的有效性。Time domain fault signal of rolling bearings contains non periodic impact of weak impact between working parts or bearing components and working condition noise. Cause the periodic impulse component of the fault information is difficult to be extracted and unable to accurately and effectively for rolling bearings for fault diagnosis. To solve this problem, put forward the autoregressive moving average minimum entropy deconvolution method. Two sets of filter coefficients are obtained by calculating the autoregressive moving average model and minimum entropy deconvolution, autoregressive moving average model to calculate the coefftcient of filter is used to separate the aperiodic impact component of fault signal, minimum entropy deconvolution to calculate the inverse filter coefficient is used to restore the fault impact. The effectiveness of the proposed method is demonstrated by simulation and experimental results.

关 键 词：滚动轴承 故障诊断 自回归滑动平均 最小熵反褶积 

分 类 号：TH16[机械工程—机械制造及自动化]