Open-Circuit Fault Diagnosis for Three-Level ANPC Inverter Based on Predictive Current Vector Residual

Abstract

This article proposes a novel open-circuit (OC) fault diagnosis method for three-level active neutral-point-clamped (3L-ANPC) inverter. This fault diagnosis method is integrated in the receding optimization of the model predictive control (MPC), where the predictive current model is built to estimate the current vector under the normal and switch OC fault-assumed conditions. The predictive current vector residual between the measured current and predictive current is utilized to distinguish the faulty switch module. Owning to the iterative optimization of the MPC, the predictive current vector residual can be calculated within each control interval, thus achieving fast online diagnosis. Besides, a count function is defined for preventing the misdiagnosis, enhancing the algorithm robustness. Simulations are conducted under different OC faults to evaluate the effectiveness of the defined diagnosis thresholds. Finally, the feasibility of the proposed method is validated experimentally on a 3L-ANPC inverter fed permanent magnet synchronous motor (PMSM) drives.