On-Line Detection of Stator Faults in DTC-Driven IM Using SC Impedance Matrix Off-Diagonal Term

Abstract

Stator's turn fault is one of the major causes of induction motor's (IM) failures. IMs are typically driven by direct torque control (DTC) on industry applications and more recently, on electric vehicles. In this article, an in-service nonintrusive stator turn fault detection (FD) technique for a DTC inverter fed IM is proposed. The presented FD strategy is based on the on-line estimation of the off-diagonal term in the IM sequence components (SCs) impedance matrix. The FD does not require prior knowledge of IM's parameters, and it is robust and immune against sensor's inherent errors. Furthermore, the proposed FD routine is able to detect incipient failure of approximately 1% fault severity factor. Two practical methodologies are applied for the computation of the SC: Symmetrical components and Savitzky-Golay filter. Simulation reinforced by experimental results demonstrated the success of the FD in a closed-loop controlled inverter driven IM.