Rotating speed isolation and its application to rolling element bearing fault diagnosis under large speed variation conditions

During the past decades, the conventional envelope analysis has been one of the main approaches in vibration signal processing. However, the envelope analysis is based on stationary assumption, thus it is not applicable to the fault diagnosis of bearings under rotating speed variation conditions. This constraint limits the bearing diagnosis in industrial applications significantly. In order to extend the conventional diagnosis technique to speed variation cases, a rotating speed isolation method is proposed. This method consists of four main steps: (a) a low-pass filter is used to separate the rotating speed components and the resonance frequency band from the original signal; (b) the trend line of instantaneous rotating frequency (IRF) is extracted by ridge detection from the short-time spectrum of the low-pass filtered signal; (c) the envelope signal is obtained by fast kurtogram based resonance demodulation; (d) the trend line of instantaneous fault characteristic frequency (IFCF) is extracted by ridge detection from the short-time spectrum of the envelope signal; (e) the rotating speed is isolated and the instantaneous fault characteristic order (FCO), which is obtained by simply dividing the IFCF by IRF, can be used to identify the fault type. By rotating speed isolation, the bearing faults under speed variation conditions can be detected without additional tachometers. The effectiveness of the proposed method has been validated by both simulated and experimental bearing vibration signals. The results show that the proposed method outperforms the conventional envelope analysis method and is effective in bearing diagnosis under speed variation conditions.