LMD和改进小波阈值去噪的轴承声发射信号故障诊断

针对机械故障声发射信号特征提取的问题,提出了一种局部均值分解(local mean decomposition,LMD)和改进的小波阈值去噪相结合的方法;并应用于滚动轴承的故障诊断。首先,把改进小波阈值与三种小波阈值去噪方法进行比较分析。通过仿真信号表明,改进小波阈值方法能更为有效地去除噪声。其次,采用LMD方法将原始轴承故障的声发射信号分解,分解为若干个乘积函数(production function,PF)的线性组合,通过相关系数原则选取能够反映故障特征的PF分量,利用改进小波阈值去噪法对选出的PF分量进行进一步去噪。最后,对去噪后的声发射信号进行包络谱分析,诊断轴承故障的位置。通过滚动轴承单一故障和耦合故障的声发射实验验证了该方法的有效性。

Fault diagnosis of bearing acoustic emission signals based on improved wavelet threshold denoising and LMD

A method of LMD and improved wavelet threshold denoising was proposed, it is for the acoustic emission signals of mechanical fault feature extraction. Firstly, by comparing the improved wavelet threshold denoising method and three kinds of wavelet threshold. The simulation signals show that the proposed improved wavelet threshold denoising method can remove noise more efficiently. Secondly, the original signal is decomposed into several PF components by LMD, the PF components are selected through correlation coefficient which reflect the fault characteristics. Then the chosen PF components are denoised by the improved wavelet threshold denoising method. Lastly, the fault frequency is gotten from envelope spectrum analysis. Diagnose the fault location of bearing accurately. The experiments of bearing single fault and coupled fault acoustic emission signals show that the method this article proposed is effective.