Machine Learning-Based Fault Prediction of Electromechanical System with Current and Vibration Signals

Abstract

This paper describes the development of intelligent fault diagnosis for electromechanical systems (EMS) and examines various combinations of mechanical and electrical faults in such systems. Specifically, the study focuses on a three-phase asynchronous motor (IM) with an outer rotor bearing system. The faults investigated in this study include a healthy system (F1), healthy motor with outer bearing faults (F2), healthy motor with unbalance in outer rotor (F3), bearing fault in the motor with healthy outer rotor (F4), inner motor bearing and outer rotor bearing fault (F5), motor bearing fault with outer unbalanced rotor (F6), motor stator fault with outer healthy rotor (F7), motor stator fault with outer unbalanced rotor (F8), motor stator fault with outer bearing fault (F9), and motor bearing fault with outer bearing fault and unbalanced rotor (F10). In order to detect combined faults in the motor-rotor-bearing assembly, the paper proposes using wavelet characteristics extracted from the current and vibration signals, which are used to develop a support vector machine (SVM)-based defect detection system. Finally, the paper concludes with a discussion of the research results. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.