High-accuracy fault feature extraction for rolling bearings under time-varying speed conditions using an iterative envelope-tracking filter

Rolling bearings under time-varying speed conditions will induce non-stationary vibration signals. Fault diagnosis of varying-speed bearings is a challenging task. Towards this goal, the most widely used method is to extract the fault characteristic frequency (FCF) of the bearing by detecting ridge curves from a time-frequency representation (TFR) of the vibration signal. However, the resolution of the TFR is restricted by the Heisenberg uncertainly principal, resulting in the limited accuracy of the FCF estimation. In this paper, a high-accuracy FCF extraction method for varying-speed bearing fault diagnosis is proposed based on an iterative envelope-tracking filter (IETF). Besides estimating the signal envelope, the IETF incorporates an iterative frequency-refinement procedure to accurately estimate the FCF of the faulty bearing. In addition, a high-resolution TFR based on the estimated signal envelopes and frequencies is constructed to clearly reveal bearing fault features. Both our simulation and experimental studies have shown that the IETF can generate much more accurate estimation results than using the conventional TFR ridge detection method alone. (C) 2019 Elsevier Ltd. All rights reserved.