A Fault Detection and Diagnosis Method Based on the Currents Entropy Indexes for the SRM Drive with a Fault Tolerant Multilevel Converter

Neutral Point Clamped Asymmetric-Half-Bridge (NPC-AHB) has been proposed as one of the power converter topologies for the Switched Reluctance Machine (SRM) drive. This topology is characterized by multilevel operation and the capability to operate in fault tolerant mode, which is most indicated for use in applications that require high reliability. However, one fundamental aspect associated with fault tolerance operation is the necessity to diagnose and identify a fault in the power semiconductors of the converter. Thus, this paper proposes a new approach for the detection and diagnosis of multiple faults in the power semiconductors of this converter. The proposed approach is based on an image analysis, namely through the discrimination of the different eccentricities that will appear. This identification is done through the use of proposed normalized indexes that are developed from the entropy analysis. The proposed approach will allow discriminating multiple power semiconductors in fault, as well as between open and short circuit conditions. The performance and capability of the proposed approach will be tested using a laboratory system.