An improved time-frequency signal decomposition method and its application in bearing incipient fault diagnosis

Aiming at the problem that incipient fault signals of rolling bearings are often drowned by strong noise, a rolling bearing incipient fault diagnosis method is proposed by combing an ARIMA based empirical mode decomposition algorithm and an improved variational mode decomposition algorithm (AEMD-IVMD). Firstly, the so-called AEMD algorithm was introduced to deal with the end effect problem of traditional EMD by extending the length at both ends of the decomposed IMF components, and a denoising vibration signal could be obtained by selecting and restructuring these selected sensitive IMF components. Secondly, an improved particle swarm optimization (IPSO) algorithm was proposed to solve the problem of difficult parameter selection of VMD algorithm, and which was the above IVMD algorithm. Then, the IVMD algorithm was used to decompose these de-noising vibration signal by AEMD and several new IMF components could be obtained to get secondary denoising vibration signals by restructuring these selected new sensitive IMF components once again. Finally, the experimental verification of the proposed method was carried out by using the collected vibration signals, and the Hilbert envelope analysis was compared with the traditional EMD and EEMD. The experimental results showed that the proposed method is more suitable for incipient fault diagnosis of rolling bearings than the traditional EMD and EEMD.