Fault severity detection of ball bearings and efficiency of one-period analysis in early fault diagnosis of rotating machinery

Abstract

This paper investigates several number of methods: Wavelet Packet Energy (WPE), Time-domain features and Multipoint Optimal Minimum Entropy Deconvolution Adjusted (MOMEDA) which are efficient of extracting features in fault diagnosis of rotating machinery. The database, which is attained via Bearing Data Center of Case Western Reserve University (CWRU), includes signal samples related to the different faulty cases and severity levels of bearing type 6205-2RS JEM. Throughout the research, combination of different faulty sample signals which are segmented into different number of periods, one of which is so called one-period analysis, of rotation of the motor are used in order to classify early faults of bearings and five class severity levels of ball bearings. Upon using proposed approaches, an outstanding classification performance of 100% and 99,7% are observed in specificity of early faults by the use of one-period analysis and five severity level classification of ball faults, respectively.