Identification of Bearing Faults Through Vibrational Signal Analysis Using Automated Relative Energy based Empirical Mode Decomposition

Motor bearing failure can cause an undesirable outcome and eventually imbalance the system. To solve this issue, vibrational signal testing (VST) is used in the article for motor fault identification. In this paper, two online datasets (CWRU and Ottawa) have been used for the experimentation. This paper presents the experimentation outcome by fusing the focused dataset to perform the classification of bearing faults after using a pre-processing method referred to as Automated Relative Energy based Empirical Mode Decomposition (AREEMD). In this paper fusion of Cepstral and time features were used. The highest accuracy obtained from Linear Discriminant Classifier fused features is 97.9%.