A normalized Hilbert-Huang transform technique for bearing fault detection

Bearings are commonly used in rotary machinery, whereas up to half of machinery malfunctions could be related to bearing defects. Unfortunately reliable fault detection systems still remain a challenging task, especially when bearing defect-related features are nonstationary. A new normalized Hilbert-Huang transform (NHHT) technique is proposed in this paper for vibration-based bearing fault detection. The NHHT for bearing fault detection takes two processes: firstly the vibration signal is denoised to highlight defect-related impulses; and secondly representative features are extracted for bearing fault detection. Vibration signal denoising is carried out by the use of the maximum kurtosis deconvolution filter to reduce impedance effect of transmission path of the measured vibration signal. A novel strategy based on D’Agostino-Person normality is suggested to enhance the distinctive intrinsic mode functions for representative features extraction and formulation for bearing fault detection. The effectiveness of the proposed NHHT technique is verified by a series of experimental tests corresponding to different bearing health conditions, and its robustness in bearing fault detection is examined by the use of data sets from a different experimental setup.