Rolling bearing fault diagnosis based on multipoint optimal minimum entropy deconvolution adjusted technique and direct spectral analysis

Abstract

This paper introduces a new idea of bearing fault diagnosis, which is to use the Multipoint Optimal Minimum Entropy Deconvolution Adjusted (MOMEDA) Technique to filter the vibration signal, getting the output signal and directly to do spectral analysis, then to observe the fault characteristics for fault diagnosis. MOMEDA technique is an improved MED algorithm proposed by McDonald in 2016, which can enhance the repetition of pulses in the signal. In this paper, MOMEDA is applied to the fault signal processing of rolling bearing. For bearing fault signal with constant speed, we firstly use MOMEDA technology to filter the signal, then do the frequency spectrum analysis. In addition, for variable speed bearing fault signal, this paper uses MOMEDA technology combined with resampling order analysis technology to analyze variable speed signal. The simulation signals and the bearing experimental data of Case Western Reserve University were processed, from the results, the direct spectral analysis after MOMEDA processing can obtain the fault characteristic without traditional envelope analysis, for fault diagnosis of bearing is simple and effective.