利用倒阶次谱和经验模态分解的轴承故障诊断

Fault Diagnosis of Rolling Bearing Based on Order Cepstrum Analysis and Empirical Mode Decomposition

针对齿轮箱升降速过程中振动信号非平稳的特点,将阶次跟踪分析与希尔波特-黄变换技术相结合,提出了基于倒阶次谱和经验模态分解的滚动轴承故障诊断方法。首先,对齿轮箱加速时测得的瞬态信号进行时域采样,对时域信号进行等角度重采样,转化为角域伪平稳信号,然后对角域信号进行经验模态分解。最后,对包含轴承故障信息的高频固有模态函数进行倒阶次谱分析,就可以提取轴承的故障特征。通过对轴承内圈和外圈故障信号的分析表明,该方法能准确识别轴承的故障类型和部位。

In order to process the non-stationary vibration signals during speedup and speed-down of a gearbox,the order cepstrum technique and the empirical mode decomposition(EMD) methods were presented to diagnose the rolling bearing.This method combined the order tracking technique with the Hilbert-Huang transform.First,the vibration signal was sampled at a constant time increment and resampled at a constant angle increment.Therefore,the transient time domain signal was changed into pseudo-stationary angle domain one.Second,the signal in angle-domain was analyzed with EMD.At last,the high frequency intrinsic mode function(IMF) component,which contained the fault information of the bearing,was studied with the order cepstrum analysis and the fault character of the bearing was extracted.The experimental results show that combining the order cepstrum analysis with the EMD technique can effectively determine the positions and types of bearing faults.