A novel method to classify bearing faults by integrating standard deviation to refined composite multi-scale fuzzy entropy

Abstract

A new method is proposed in the present work for identifying fault severity in the ball bearings. Proposed method named as multi-scale refined composite standard deviation fuzzy entropy is based on the improvement in the existing method called refined composite multi-scale fuzzy entropy. The acquired vibration signal is initially decomposed into numerous mode functions by ensemble empirical mode decomposition method. To investigate the performance of new method, methodology is split into two phases - detection and identification. In detection phase, response of a healthy system in comparison to the faulty system under different operating conditions are examined while estimation of fault severity in the inner and outer race of bearing is analyzed in identification phase. Accuracy of classifying fault severity by the proposed method has been verified by three well-established classifiers. Proposed methodology can be reliably used for fault diagnosis because of the remarkable results obtained. © 2019 Elsevier Ltd