A Seasonal–Trend-Decomposition-Based Voltage-Source-Inverter Open-Circuit Fault Diagnosis Method

Abstract

This article considers power transistor open-circuit fault in a two-level voltage-source inverter and proposes a highly accurate, robust, and simple diagnosis method based on seasonal–trend decomposition. First, the three-phase current signal is decomposed into the trend, seasonal, and remainder components through the simplified seasonal–trend decomposition algorithm. Then, two characteristics “seasonal index” and “trend index” are defined. The theoretical values of the two characteristic quantities under different conditions are derived through simplified current waveform analysis and approximate integral calculation. Finally, the thresholds of characteristic quantities are set, and the diagnosis algorithm is designed. Experiments verify the effectiveness and the merits of the proposed method.