全矢LMD能量熵在齿轮故障特征提取中的应用

齿轮故障信号具有非线性、非平稳特征,齿轮发生故障时,信号的能量结构随之改变,在不同的频带内能量不同。传统方法采用局部均值分解(LMD)提取振动信号的能量熵,将能量熵指标作为故障评判标准进行故障分类,依靠单一传感器信息源进行故障诊断，因而容易造成误诊、漏诊。全矢LMD能量熵法融合了双通道同源信息的回转能量，可降低故障误判率。通过实验模拟齿轮正常、齿根裂纹、断齿、缺齿等4种状态，验证了全矢LMD能量熵作为故障特征能达到很好的故障分类效果。

Full Vector LMD Energy Entropy in Gear Fault Feature Extraction

Gear vibration  signals  in the events  of failure were  often  non-stationary,non-linear.Energy  structure  would change in the fault signals,leading to different energy in different  frequency bands.LMD  was used  to  extract energy entropy of the vibration signals,and energy entropy was used as failure evaluation standards  for fault classification.It is easy to be misdiagnosed with the traditional single channel signal diagnostic method.Full vector LMD energy entropy integrated two-channel homologous  informations  of vibration signals, and reduced the misdiagnosis rate.Through  experiments  the gear  normal state,tooth root crack,broken teeth,missing teeth were simulated,and it is proved that with full vector LMD energy entropy as fault feature can achieve good fault classification results.