AbstractTo improve the reliability of motor system, this paper investigates the sensor fault diagnosis methods for T-type inverter-fed dual three-phase permanent magnet synchronous motor (PMSM) drives. Generally, a T-type three-level inverter-fed dual three-phase motor drive utilizes four phase-current sensors, two direct current (DC)-link voltage sensors and one speed sensor. A series of diagnostic methods have been comprehensively proposed for the three types of sensor faults. Both the sudden error change and gradual error change of sensor faults are considered. Firstly, the diagnosis of speed sensor fault was achieved by monitoring the error between the rotating speed of stator flux and the value from speed sensor. Secondly, the large high-frequency voltage ripple of voltage difference between the estimated voltage and the reference voltage was used to identify the voltage sensor faults, and the faulty voltage sensor was determined according to the deviation of voltage difference. Thirdly, the abnormal current amplitude on harmonic subspace was adopted to identify the current sensor faults, and the faulty current sensor was located by distinguishing the current trajectory on harmonic subspace. The experiments have been taken on a laboratory prototype to verify the effectiveness of the proposed fault diagnosis schemes.
Artificial Neural Network Application for Current Sensors Fault Detection in the Vector Controlled Induction Motor Drive
