基于EMD的冲击信号提取方法及其在设备故障诊断中的应用

EMD based impulse identification and its application in friction fault diagnosis of rotating machine

提出一种基于经验模式分解(Empirical Mode Decom position,EMD)的冲击信号提取方法,利用该方法首先将含有周期性冲击的信号进行EMD分解,在分解后的高频段IMF中,存在着类似冲击响应信号的成分,这些成分是由原始信号中的周期性冲击引起的,通过包络解调方法,可以得到冲击响应信号出现的频率,该频率对应原信号中冲击信号出现的频率。由于碰摩故障发生时,往往伴随着周期性冲击信号的产生,故该方法可以应用于旋转设备碰摩故障诊断中。仿真信号和试验数据的分析结果表明,这种方法正确有效,可以应用于工程实际。

A novel impulse identification method based Empirical Mode Decomposition（ EMD） is proposed. With this method,the signal containing impulses is processed by EMD. According to definition of Intrinsic Mode Function（ IMF）, the impulse signal can not be decomposed as IMF,whereas it can be decomposed as impulse response. Then demodulation technique is engaged to identi- fy impulse period from gained IMF. Simulations and experiments verified that the proposed methods are accurate and efficient and can be applied in engineering practice effectively.