On-line Monitoring of Stator Inter-Turn Failures in DTC driven Asynchronous Motors using Mathematical Morphological Gradient

This study presents an in-service non-intrusive diagnosis methodology for the detection of incipient faults in the induction motor's (IM) stator windings. The failure identification (FI) algorithm is based on the application of mathematical morphological gradient (MMG) technique on the stator's current in the time domain. The IM under study is driven by the direct torque control (DTC). Comprehensive investigation of harmonics, inter-harmonics and sub-harmonics content is conducted to compare between the FI based on the classical motor current signature analysis (MCSA), and the proposed technique. The proposed MMG based FI routine is fast, online, and easy to implement on digital controllers. The FI technique was validated through the co-simulation between the physics based finite element (FE) model of IM, and the voltage source inverter (VSI), DTC controller, and FI logic. The co-simulation have the merit of simulating the IM with all the harmonic content to improve the reliability of the FI. Moreover, it represents a trustworthy alternative to the experimental platform, which accelerate the ageing rate of the IM. Simulation reinforced by experimental results depicted the robustness of the FI method.