A Fault Diagnosis Method Based on Adaptive Threshold Judgment in Zero-Crossing Interval for Aviation Three-Phase Inverter

Most existing inverter fault diagnosis methods simplify actual fault conditions and operating scenarios for research expediency. However, their diagnostic reliability is compromised when confronted with such as nonsimultaneous secondary switch faults or multiple complex operating conditions. To address these issues, this article proposes a novel open-circuit (OC) fault diagnosis method. First, the three-phase current is measured, and a mean filtering is applied to enhance the waveform smoothness. Next, a zero-crossing-driven interval determination method is proposed for the current positive-half and negative-half. An adaptive threshold is then calculated in each interval, and the comparison between the current and this threshold effectively captures the loss of the waveform. Finally, a mapping from the fault signature vector to the fault mode is established to precisely localize the fault switch. Compared to existing advanced methods, the proposed approach can effectively handle the feature aberrance caused by system feedback. It also accurately identifies nonsimultaneous secondary switch fault and demonstrates diagnostic consistency in various load scenarios. Experimental results, comparisons, and analyses confirm the superiority of the proposed method.