Fault diagnostic algorithm for three-level neutral point clamped AC motor drives, based on the average current Park's vector

This study presents a new fault diagnostic technique for a semiconductor open-circuit fault in a neutral-point clamped inverter, based on the average current Park's vector approach. This method relies on the measurement of two motor line currents and is able to perform a complete diagnosis in less than half of the inverter output current period. The implementation of this diagnostic technique requires no additional current or voltage sensors besides the ones already used by the control algorithm of the converter. The diagnostic method uses the modulus and angle of the normalised average current Park's vector to detect the fault and identify the faulty insulated gate bipolar transistor (IGBT) pair, respectively. Simultaneously, the normalised average values corresponding to the positive and negative parts of the motor supply currents waveforms are used to identify the inner faulty IGBTs with even and odd numbers, respectively. The diagnostic method is also robust to transient conditions imposed by load and speed variations. The performance and effectiveness of the proposed diagnostic technique are validated both by the simulation and experimental results.