基于改进小波包和EEMD的轴承故障诊断

Fault Diagnosis for Bearings Based on Improved Wavelet Packet and EEMD

针对轴承振动信号非平稳性及工作情况下难以获得故障频率,提出一种基于改进小波包和总体经验模态分解(EEMD)的轴承故障诊断方法。首先运用改进小波包对振动信号进行分解,得到按顺序排列的子带频带。然后提取故障频率范围的子带信号并进行EEMD,以互相关系数和峭度准则提取故障分量,避免了固有模态函数(IMF)分量选择的盲目性。仿真和试验分析结果表明,该方法能有效且准确地检测出轴承故障。

In view of the non - stationary of vibration signals of bearings and the difficulty to obtain fault frequencies in practice, a fault diagnosis method for bearings is put forward based on improved wavelet packet and EEMD. Firstly, the vibration signals are decomposed by improved wavelet packet, a number of sub - band frequency bands in order are ob- tained. Then the sub - band signals of fault frequency range are extracted and the EEMD is carried out. The fault component is extracted by cross correlation coefficient and kurtosis criteria, which can avoid the blindness of the IMF component selection. The simulation and test analysis results show that the method can effectively and accurately detect fault of bearings.