Short-Circuit Fault Diagnosis for Three-Phase Inverters Based on Voltage-Space Patterns

Abstract

This paper introduces a fault detection and isolation (FDI) method for faulty metal-oxide-semiconductor field-effect transistors in a three-phase pulsewidth-modulated (PWM) voltage source inverter. Short-circuit switch faults are the leading cause of failure in power converters. It is extremely vital to detect them in the early stages to prevent unwanted shutdown and catastrophic failures in motor drives and power generation systems. Against the common FDI methods for power electronic inverters that use phase currents and PWM gate control signals, the proposed method only uses the inverter output voltages. This method analyzes the PWM switching signals in a time-free domain that is called the voltage space. For a healthy inverter, the projection of the state transitions in the voltage space results in a cubic pattern. Each short-circuit switch fault uniquely changes the voltage-space pattern that allows isolating the faulty switch. The fault detection time is only within one PWM carrier period, which is significantly faster than current-based conventional methods. The FDI result does not depend on the load, the PWM switching frequency, and the feedback loop. This method can address the reliability problem of multilevel inverters in renewable electrical generation systems and can dramatically reduce the number of required sensors.