Broken rotor bar fault detection using Hilbert transform and neural networks applied to direct torque control of induction motor drive

• Published in: IET power electronics 2020-11, Vol.13 (15), p.3328-3338
• Main Authors: Ramu, Senthil Kumar, Irudayaraj, Gerald Christopher Raj, Subramani, Saravanan, Subramaniam, Umashankar
• Format: Article
• Language: English
• Subjects: ANN; artificial neural networks; broken rotor bar fault detection; direct torque control; DTC; fast Fourier transforms; fault diagnosis; frequency components; fundamental frequency; Hilbert transform; Hilbert transforms; induction motor drive; induction motor drives; induction motors; industrial drives system; least mean squares methods; machine control; neural nets; Research Article; rotors; sideband frequency; stator current envelope; stators; suitable control technique; torque control
• ISSN: 1755-4535; 1755-4543
• DOI: 10.1049/iet-pel.2019.1543

This study proposes a new approach for the detection of broken rotor bar (BRB) fault in three phase induction motor drive using Hilbert transform (HT) and artificial neural networks (ANNs), where the machine is controlled by direct torque control (DTC). HT is preferred to develop the stator current envelope. The sideband frequency and its amplitude of the samples are the input for the ANN. By using fast Fourier transform, the amplitude and frequency components are extracted and the severity of fault is determined by comparing the magnitude of an average of sideband frequency with the fundamental frequency. High accuracy identification of fault is found by ANN, where the results are trained and tested to a minimum mean square error that will detect the number of BRB in the induction motor. DTC is adopted for a suitable control technique in the industrial drives system to maintain good performance in torque control. The performance of the proposed method is verified by using MATLAB/SIMULINK and experimental tests.