Open-Circuit Fault Diagnosis for Three-Level ANPC Inverter Fed PMSM Drives with Model Predictive Control

This paper proposes an open-circuit (OC) fault diagnosis method for three-level active neutral-point-clamped (3L-ANPC) inverter based on model predictive control (MPC). In this method, the predictive current model is used to predict and estimate phase current under normal and faulty conditions. The OC fault can be detected by comparing the residual between estimated phase current and measured phase current with the threshold. The detected fault is further localized through a fault localization matrix, every element of the matrix corresponds to a switch in 3L-ANPC inverter. The matrix is updated by the compared results between actual measured phase current and the estimated phase current considering the possible switch OC faults. The utilized phase current signals are also necessary in the MPC algorithm, so the proposed fault diagnosis method does not require any additional sensors. Moreover, the fast dynamic characteristic is derived from the MPC. Simulation results are obtained from a 3L-ANPC inverter fed permanent magnet synchronous motor (PMSM) drives, verifying the effectiveness of the proposed diagnosis method.