Common Mode Voltage-based Open-switch Fault Diagnosis for Inverters with Model Predictive Control in IM Drive System

This paper proposes a fast fault diagnosis strategy for the induction motor (IM) drive system using model predictive control (MPC). This method cleverly utilizes the common mode voltages (CMVs) and the characteristic of MPC that it has a fixed switching state in each switching cycle. The CMVs are used as fault feature quantity in this diagnosis strategy and can be obtained according to the switching state and sampling current information in each switching cycle. The CMVs will show different characteristics under normal conditions and fault conditions, and these characteristics are utilized to detect the fault switch of the inverter. This proposed diagnosis method does not need to add extra sensors into the origin system, avoiding the increase of system cost. Meanwhile, this method can diagnose and position the fault switch within only one switching cycle. Simulation results are presented to verify the effectiveness of the proposed fault diagnosis strategy.