Deep Convolutional Neural Network Based Fault Detection and Diagnosis Method for Three-Phase T-Type Converter

Power electronics industry has relied heavily on recent developments in the study of power converters. In recent years, multilevel converter has gained a lot of attention due to its simple structure and control for high voltage and high power applications. The reliability of the converter is a vital sign of performance that needs to be taken into consideration. A 3L-T-type converter is vulnerable to a variety of failures due to a large number of components (switches and their associated gate driver circuits). The fault-tolerant ability of the topologies is very crucial for the reliable operation of the overall system. Identification of the faults is the first step toward the reliable operation of the overall system. This paper provides an analytical study of the fault-tolerant ability of the 3L-T-type converter with an open-circuit fault (OCF). The effect of OCF on the power loss distribution has been carried out using PLECS software. The analysis has been used for the detection and identification of components using less computational complex deep convolutional neural network. The proposed identification technique has been trained and tested for various OCFs that occurred due to gate driver failure and the results have been discussed.