基于改进EMD与滑动峰态算法的滚动轴承故障特征提取

Fault feature extraction of rolling element bearing based on improved EMD and sliding kurtosis algorithm

针对滚动轴承故障特征往往被强背景噪声淹没的特点,提出基于改进经验模态分解(Empirical ModeDecomposition,EMD)与滑动峰态算法的滚动轴承故障特征提取方法。利用EMD方法分解原故障信号得到一组平稳固有模态分量(Intrinsic Mode Function,IMF)后采用互信息和广义相关系数筛选法消除传统EMD分解结果中虚假分量,运用滑动峰态算法对真实IMF分量处理得到滑动峰态时间序列。计算滑动峰态序列频谱提取故障特征频率。实例研究结果表明:该方法能有效提取滚动轴承故障特征,可取得较直接滑动峰态算法及传统包络解调分析更好的效果。

Considering the feature of rolling element bearing fault signal with strong noise, a rolling element bearing fault feature extraction method was proposed based on improved EMD and sliding kurtosis algorithm. Original fault signal was decomposed with EMD to get a finite number of stationary intrinsic mode functions（IMFs）. Then, mutual information and general correlation coefficient were together used to get rid of pseudo-components in the traditional EMD results, and real IMF components were processed with sliding kurtosis algorithm to obtain a sliding kurtosis time series. Finally, the frequency spectrum of the time series was calculated with Fourier transformation to extract the fault feature frequency. Example results showed that the proposed method can effectively extract the fault feature of rolling element bearing, and is more effective than the direct sliding kurtosis algorithm and the traditional envelope demodulation in fault feature extraction.