随机共振消噪与ELMD相结合的轴承故障诊断

The Fault Diagnosis Approach for Bearing of Combining Stochastic Resonance De-noising With ELMD

针对实际工况下轴承故障信号容易淹没在强烈噪声中难以提取,局部均值分解（local mean decomposition,LMD）在信号处理方面不可避免出现模态混叠现象的问题,提出1种基于随机共振（Stochastic Resonance,SR）与总体局部均值分解（ensemble local mean decomposition,ELMD）相结合的轴承故障诊断方法。首先将采集到的实际轴承振动信号进行随机共振消燥,以白噪声驱动与双稳系统形成随机共振现象,将噪声的部分能量转移到低频区域集中,提高信号的信噪比;然后将经过随机共振后的信号进行ELMD分解,得到1组乘积函数（product function,PF）,以及1个残余分量,对得到的PF分量进行时频分析,最终发现轴承故障特征频率。实验结果表明,该方法可以提高信号信噪比,以达到微弱信号检测的目的,因此可以很好的应用在轴承故障诊断中。

For the phenomenon that the effective signal is submerged with strong noise at the background during the practical fault diagnosis for bearing,moreover,due to there is presented a mode confusion phenomenon of Local Mean Decomposition（ LMD） during the decomposition process,a fault diagnosis approach for bearing which based on Stochastic Resonance（ SR） and Ensemble Local Mean Decomposition（ ELMD） is proposed. Firstly,the bearing vibration signals are taken stochastic resonance operation,there will be a stochastic resonance phenomenon with the drive of white noise and bi-stable system,it can improve the signal-noise ratio of the efficient signal when transfer part of noise energy to low-frequency area at together; Then,there will be some PFs and a Residual component after ELMD for the signal which after stochastic resonance. The characteristic frequency of bearing fault can be obtained by doing time-frequency analysis of the obtained PF component. This experimental result shows that this approach can improve the signal-noise ratio of the signal and achieve the goal of weak signal detection,so it can be used well in the fault diagnosis for bearing.