Demodulated Multisynchrosqueezing S Transform for Fault Diagnosis of Rotating Machinery

Abstract

Time–frequency analysis (TFA) plays an important role in fault diagnosis of rotating machinery. However, a time–frequency representation (TFR) with lower resolution is not conducive to extracting fault features of the equipment. In this article, a novel method, termed demodulated multisynchrosqueezing S transform (DMSSST), is proposed to achieve a highly energy-concentrated TFR, which combines demodulation and multireassignment (MT) techniques for S transform (ST). The proposed method integrates the adaptability of demodulated ST to the frequency with the iterative procedure of MT to optimize the instantaneous frequency (IF) estimation. Compared with the traditional approaches, the presented method has better robustness with respect to noise and can be well applied to the discrimination of several faults. Furthermore, the simulated signal shows that this method can effectively improve the resolution of a TFR and has excellent anti-noise ability. The real examples, including bearing fault and rotor rub-impact, further validate the effectiveness of the DMSSST method in fault diagnosis of rotating machinery.