Rolling element bearing fault diagnosis based on multi-scale global fuzzy entropy, multiple class feature selection and support vector machine. (October 2019)

• Record Type: Journal Article
• Authors: Zhu, Keheng; Chen, Liang; Hu, Xiong
• Is Part Of: Transactions of the Institute of Measurement and Control. Volume 41:Number 14(2019)
• Page Start: 4013
• Page End: 4022
• Publication Date: 2019-10
• Subjects: Multi-scale global fuzzy entropy -- multiple class feature selection -- rolling element bearing -- fault diagnosis; Automatic control -- Periodicals; Measuring instruments -- Periodicals; Commande automatique -- Périodiques; Mesure -- Instruments -- Périodiques; 681.2
• DOI: 10.1177/0142331219844555
• Languages: English
• ISSNs: 0142-3312

Abstract

Multi-scale fuzzy entropy (MFE) is a recently developed non-linear dynamic parameter for measuring the complexity of vibration signals of rolling element bearing over different scales. However, the calculation of fuzzy entropy (FuzzyEn) in each scale ignores the sequence's global characteristics while the bearing vibration signals' global fluctuation may vary as the bearing runs under different states. Therefore, in this paper, the multi-scale global fuzzy entropy (MGFE) method is put forward for extracting the fault features from the bearing vibration signals. After the feature extraction, multiple class feature selection (MCFS) method is introduced to select the most informative features from the high-dimensional feature vector. Then, a new rolling element bearing fault diagnosis approach is proposed based on MGFE, MCFS and support vector machine (SVM). The experimental results indicate that the proposed approach can effectively fulfill the fault diagnosis of rolling element bearing and has good classification performance.