A Fault Diagnosis Scheme Based on the Normalized Indexes of the Images eccentricity for a Multilevel Converter of a Switched Reluctance Motor Drive

This paper addresses the fault detection and diagnosis of a fault in the switches of the Switched Reluctance Machine (SRM) power electronic converter. Due to the advantages of using multilevel converters with these machines, a fault detection and diagnosis algorithm is proposed for this converter. The topology under consideration is the asymmetric Neutral Point Clamped (ANPC), and the algorithm was developed to detect open and short circuit faults. The proposed algorithm is based on an approach that discriminates eccentricity of the images formed by the converter voltages. This discrimination is realized through the development of normalized indexes based on the entropy theory. Besides the different fault type the algorithm is also able to detect the transistor under fault. The possibility to implement the proposed approach will be verified through simulation tests.