Abstract: Vibration signal which is of multiple components and nonstationary, is susceptible to noise interference in bearing fault diagnosis. In order to find a better method for bearing fault diagnosis, adaptive chirp mode decomposition (ACMD) and its application are studied.Firstly, the fault signal of bearings was decomposed adaptively into several components by ACMD. Secondly, the envelope spectrum of the component with biggest kurtosis was obtained by Hilbert transform and calculated. Finally, the fault was diagnosed by figuring out the characteristic frequency in the envelope spectrum. The method can extract the characteristic frequency of fault signals effectively and be of a good application prospect.

Key words: rolling bearing, fault diagnosis, adaptive chirp mode decomposition, kurtosis, envelope spectrum