Diagnosis of Power Device Failures Using Discrete Fourier Transform for DC-AC Flying Capacitor Multilevel Converters

The fault detection of the power switching device enables the multilevel converter to enter the fault-tolerant operation mode in time, so that the converter can still operate in the case of a fault, and the reliability of the converter is improved. According to the characteristics of flying capacitor multilevel converter (FCMC), a fault diagnosis method for power switch based on voltage harmonics is proposed. This method uses Discrete Fourier Transform (DFT) to continuously analyze the switching node voltage harmonics, and uses a look-up table to identify the type and location of the fault according to different fault characteristic values. Use DFT in FPGA controller to analyze harmonics of specific frequency. The method proposed in this paper can accurately and quickly identify the fault type and location, and only requires an additional sensor, which greatly reduces the cost. The proposed method is verified by simulation and experimental results of a SiC-based 3-Level 100kHz DC-AC flying capacitor buck converter (FCBC).