Bearing fault diagnosis based on improved LMD

A new time-frequency analysis method named Local mean decomposition (LMD) was studied, which can adaptively decompose complicated component signal into a set of product functions (PF). The PF whose instantaneous frequency has physical significance was the product of an envelope signal and a frequency modulated signal. In order to solve the shortage moving average algorithm in LMD method, the author puts forward to using hermite interpolation instead of moving average algorithm. Through the analysis on simulation signals, we found that the improved LMD method is better than the LMD method both in analysis accuracy and computation time. According to the modulating characteristics of the roller bearing fault vibration signals, a fault diagnosis method for bearings based on improved LMD was proposed. The analysis results show the method can be used in the fault diagnosis.