The Enfigram: A robust method for extracting repetitive transients in rolling bearing fault diagnosis

• Published in: Mechanical Systems and Signal Processing Vol. 158; p. 107779
• Main Authors: Xu, Yonggang, Deng, Yunjie, Ma, Chaoyong, Zhang, Kun
• Format: Journal Article
• Language: English; Japanese
• Published: Berlin Elsevier Ltd 01.09.2021; Elsevier BV
• Subjects: Accidental pulse; Enfigram; Fault diagnosis; Fourier transforms; Frequencies; Interference; Low pass filters; Optimal frequency-band extraction; Robust; Robustness; Roller bearings; Segmentation; Signal processing; Vibration; Fault diagnosis; Optimal frequency-band extraction; Enfigram; Robust; Accidental pulse
• ISSN: 0888-3270; 1096-1216
• DOI: 10.1016/j.ymssp.2021.107779

•Enfigram is insensitive to accidental shocks or the rotation frequency interference.•The local minimum is used to divide the boundaries in spectrum trend.•The kurtosis of the spectrum of the component envelope spectrum is calculated to evaluate fault information. In the actual industrial production, the collected vibration signal often contains accidental impact or harmonic interference due to the influence of mechanical vibration or human interference. However, the current popular Fast Kurtogram (FK) method cannot accurately extract the frequency band with more fault information when processing the signal with such two-type interference. Therefore, this paper proposes a robust method called Enfigram to improve the drawback. In Enfigram method, a new statistical index is designed to select the optimal modulation frequency-band. Envelope Fourier index (Enfi) uses two Fourier transforms to transfer the interference of accidental pulse in the signal envelope to the origin, so that accidental pulse can be eliminated by low-pass filtering. Moreover, the frequency-band segmentation method based on spectrum trend can adaptively divide the fault information into a frequency band in Enfigram, which ensures the integrity of fault information. Compared with FK, Infogram and other methods, the result shows that Enfigram has better performance.