A Fault Diagnosis Method for Silicon Carbide (SiC) Converters Based on ResNet

One of the impediments to the widespread application of Silicon Carbide (SiC) Metal-Oxide-Semiconductor Field-Effect Transistors (MOSFETs) is the insufficient research on their safety and reliability. This paper proposes a ResNet-based fault diagnosis method for SiC converters to address the inadequacies in the research on fault diagnosis of SiC MOSFET converters. Firstly, driven by data, electrical quantities of SiC MOSFET converter are extracted as inputs for the diagnostic model. Subsequently, a one-dimensional ResNet network is constructed to uncover hidden information in the fault data, ultimately yielding diagnostic accuracy for various open-circuit fault modes. The effectiveness of this approach is validated on a simulation platform for SiC MOSFET converters.