Mechanical Fault Diagnosis Based on Local Mean Decomposition Method

Abstract

With the effect of periodical impulses, most machinery fault vibration signals are multicomponent modulation signals. The effective decomposition approach is the key to demodulate signals and extract fault characteristics. Local mean decomposition (LMD) is a new kind of time-frequency analysis approach, which can decompose the signals adaptively into a set of product function (PF) components. The envelope of PF is the instantaneous amplitude and instantaneous frequency can be calculated by the derivative of the unwrapped phase of modulation signal with uniform amplitude. The method bypasses the Hilbert transform totally. Therefore, it involves no question about negative frequency of no physical meaning. Through the successful mechanical fault diagnosis examples, it shows that LMD has high theoretical value and engineering practicability.