基于CEEMD和FastICA的滚动轴承故障诊断研究

由于滚动轴承振动信号易受噪声干扰的影响、故障特征提取较为困难。为此,提出了一种基于互补集合经验模态分解(CEEMD)和快速独立分量分析(Fast ICA)的轴承故障特征提取方法。该方法首先利用CEEMD算法将原故障振动信号进行分解运算,得到一系列模态分量(IMF);然后依据峭度准则选取一些模态分量来完成观测信号的重构,剩余其他的模态分量完成虚拟噪声通道信号的重构;再利用Fast ICA方法对重构信号进行降噪;引入Teager能量算子(TKEO)对降噪后的信号进行解调处理;最后对解调后的信号进行快速傅里叶变换(FFT)运算,分析变换后信号的频谱特征,提取出原信号的故障特征频率。将该方法应用到滚动轴承故障实际数据中,实验结果表明,该方法可以有效提取出滚动轴承故障的基频和倍频特征信息。

Research on fault diagnosis of rolling bearing based on CEEMD and FastICA

Because the rolling bearing vibration signal is susceptible to noise interference,it is difficult to extract fault features.Therefore,A bearing fault feature extraction method based on CEEMD and Fast ICA is proposed in the study of rolling bearing fault diagnosis method. Firstly,the original fault vibration signal is decomposed by CEEMD algorithm,a series of intrinsic mode function components are obtained. Then select some modal components according to the kurtosis criterion to reconstruct the observed signal and the rest modal components to reconstruct the virtual noise channel signal. Then use Fast ICA algorithm to denoise the reconstructed signal.Then use Teager kaiser energy operator analysis method to demodulate the denoised signal. Finally,the FFT transform operation is performed on the demodulated signal,the spectral characteristics of the transformed signal are analyzed,and the fault characteristic frequency of the original signal is extracted. This method is applied to the actual data of rolling bearing faults,The experimental results show that the proposed method can extract the fundamental frequency and doubling frequency characteristics of rolling bearing faults effectively.