A motor fault diagnosis using hybrid binary differential evolution algorithm and whale optimization algorithm with storage space

The most common cause of mechanical failure is bearing failure, and the characteristics of each failure correspond to a certain degree of severity. This paper proposes a fault diagnosis model for detecting motor bearings. The model uses three steps: feature extraction, feature selection, and classification. In feature extraction, empirical mode decomposition, fast Fourier transform, and envelope analysis extract important features from the signals measuring the motor. In feature selection, a binary differential evolution and binary whale algorithm are developed and the storage space is increased to eliminate irrelevant features again. Finally, KNN and SVM are used to determine the stability of the bearing fault diagnosis model.