An Automatic Diagnosis of Bearing Faults of an Induction Motor Based on FFT-ANN

Abstract

The present paper proposes a diagnosis and monitoring method for detecting and locating bearing faults in an induction motor based on vibration signal processing. The proposed method served to combine Fast Fourier Transform as an advanced signal-processing tool with the Artificial Neural Network (ANN). This study starts in a first stage with the application of the FFT in order to extract frequencies characterizing the fault of the three vibration signals VX, VY and VZ. These frequencies will be used in a second stage as inputs of the proposed ANN to locate the bearing’s undamaged components. The features extracted in this study for training the ANN model are the fr, 2fir, 4fir, 2for and 3for. Therefore, the results generated by ANN indicate a satisfactory outcome with a higher classification rate of 98.93 %. The suggested FFT-ANN method successfully demonstrates its effectiveness, and the acquired results are completely validated by experiments carried out in the CWRU.