Bearing fault diagnosis under time-varying rotational speed via the fault characteristic order (FCO) index based demodulation and the stepwise resampling in the fault phase angle (FPA) domain

Time-frequency representation (TFR) based order tracking algorithm can help to realize rolling bearing fault diagnosis under time-varying rotational speed without using a tachometer. Among this TFR based algorithm, however, two important steps, determining the optimal resonance band and realizing the TFR based resampling with high accuracy, are not well addressed. As such, a new criterion called fault characteristic order index (FCOindex) is designed to select the optimal filter band to enhance the fault-related components under time-varying rotational speed, and a new stepwise resampling algorithm is proposed based on the fine instantaneous fault characteristic frequency (IFCF) trend extracted from the SIFT and CWT based envelope TFRs. The proposed algorithm contains five main steps: (a) Band-pass filtering the raw signal with several specially designed filter bands, (b) Constructing the FCOindex as the criterion and selecting the filter band with the highest FCOindex as the optimal filter band, (c) extracting the fine IFCF trend from the CWT based envelope TFR of the filter signal and the STFT based rough IFCF trend, (d) resampling the filter signal with the fine IFCF trend using a new stepwise resampling algorithm, and (e) realizing the fault detection using the FCO spectrum. The effectiveness of the proposed method has been validated using the simulated and experimental signals. (C) 2019 ISA. Published by Elsevier Ltd. All rights reserved.