Fuzzy logic-based diagnosis of inverter switch faults in VSI-Fed induction motors

This paper presents a fuzzy logic-based diagnostic method for detecting and identifying faults in power switches of Voltage Source Inverters (VSIs) feeding induction motors. Problem. Conventional diagnostic techniques often require additional sensors or sophisticated hardware, which increases system cost and complexity. Goal. The objective is to design a reliable, low-cost, and efficient fault detection strategy that ensures high diagnostic accuracy without requiring extra hardware. Methodology. The proposed approach employs only three-phase stator current measurements. It integrates signal normalisation, feature extraction using Park’s vector modulus, and a Sugeno-type fuzzy inference system to classify open-circuit and short-circuit faults affecting IGBT switches. Results. The system was implemented and tested in MATLAB/Simulink under both no-load and load conditions. Simulation outcomes confirm its ability to detect and localise all fault types across the inverter legs, with detection times ranging from 10 ms to 40 ms. The results also demonstrate that the proposed method maintains reliable performance under load variations. Practical value. The proposed solution offers fast response, low computational cost, and high diagnostic accuracy, making it an effective tool to enhance the reliability of inverter-fed motor drives.