The Harmogram: A periodic impulses detection method and its application in bearing fault diagnosis. (15th February 2022)

• Record Type: Journal Article
• Authors: Zhang, Kun; Chen, Peng; Yang, Miaorui; Song, Liuyang; Xu, Yonggang
• Is Part Of: Mechanical systems and signal processing. Volume 165(2022)
• Publication Date: 2022-02-15
• Subjects: Harmonic spectral kurtosis -- Periodic pulses -- Harmogram -- Fault diagnosis -- Signal processing; Structural dynamics -- Periodicals; Vibration -- Periodicals; Constructions -- Dynamique -- Périodiques; Vibration -- Périodiques; Structural dynamics; Vibration; Periodicals; 621
• DOI: 10.1016/j.ymssp.2021.108374
• Languages: English
• ISSNs: 0888-3270

Abstract

Highlights: Harmonic spectral kurtosis is proposed to quantify the periodic pulses. Harmogram is proposed to diagnose bearing faults. The trend component obtained can realize the spectral division. The variable trend component interception can realize multi-level decomposition. Abstract: Bearings are important and easily damaged parts in equipment. Collecting and analyzing the vibration signals of damaged bearings is a common fault diagnosis method. The vibration signal of the damaged bearing under constant speed conditions contains periodic pulses, and may also contain equipment operating vibration information, environmental noise or occasional pulses. In order to extract the periodic pulse information in the signal and weaken the influence of the interference signal, this paper proposes a periodic pulse detection indicator and calls it Harmonic spectral kurtosis (HSK). HSK can extract the harmonic information in the envelope spectrum, quantify the periodic pulses in the signal, and suppress the influence of interference such as random pulse. In order to expand the applications of this indicator, a tower-shaped boundary distribution diagram called Harmogram similar to Fast Kurtogram was established. Harmogram with HSK not only optimizes the spectral segmentation method, but also obtains a better center frequency and bandwidth. The filtered frequency bands can contain more periodic pulse information, and the proposed method has greater advantages in decomposing noisy signals. The simulation signal shows that the proposed method is accurate and effective. The data of bearing inner ring, outer ring and compound faults prove that the method can be applied to bearing fault diagnosis.