AI-ASSISTED CONDITION MONITORING OF POWER ELECTRONIC CONVERTERS

Abstract

Power electronic converters are key enabling components in renewable energy systems, electric vehicles, and industrial automation. Continuous operation under high electrical and thermal stress accelerates aging of semiconductor devices, capacitors, and magnetic components, leading to reduced reliability and unexpected failures. Conventional condition monitoring approaches rely on thresholdbased measurements and periodic inspections, which are often insufficient for early fault prediction. This paper presents an AI-assisted condition monitoring framework for power electronic converters. The proposed system integrates intelligent sensing, signal preprocessing, and machine learning–based diagnostics to identify anomalies and predict component degradation. MATLAB/Simulink-based evaluation demonstrates improved fault detection accuracy, early warning capability, and enhanced system reliability compared to traditional monitoring techniques. The results confirm that artificial intelligence provides an effective solution for predictive maintenance of modern power electronic converters.