Short-Circuit Fault Diagnosis Based on Rough Sets Theory for a Single-Phase Inverter

Abstract

The short-circuit (SC) fault diagnosis in inverters is an important procedure for the continuity of the performance and the extension of its useful life. The methods of diagnosis of SC failures produce good results, however, they present unfavorable aspects: they detect only one of the faults of SC, that is to say, the hard switch fault (HSF) or the fault under load (FUL); depend on switch parameters; and use artificial intelligence (AI) techniques in their algorithms, which are executed simultaneously with the inverter operation. This article presents a method of diagnosing SC faults performed with a digital circuit. The proposed method identifies short-circuit faults: HSF and FUL; can be used with any switch, regardless of its parameters; and does not use AI algorithms and techniques concurrently with inverter operation. The digital diagnostic circuit is obtained with the use of rough sets theory (RST), which optimizes and defines a minimum set of variables necessary to diagnose faults. Applying RST to the variables obtains a set of diagnostic rules. These rules are performed with basic logic functions and, for this reason, a digital diagnostic circuit is obtained. The diagnostic variables are the command signals and the voltage source inverter switches currents. The simulation and experimental results validate the shown diagnostic method.