Transformer Internal and Inrush Current Fault Detection Using Machine Learning

Preventive maintenance in the transformer is performed through a differential relay protection system, and it protects from internal external faults. However, Current Transformer (CT) system mal-operates during inrush currents. CT saturates due to magnetizing currents causes false tripping of relays. Moreover, identification either current or faults needs be diagnosed. For above problem, continuous monitoring breather terminals with thermal camera helps detect fault. The transformer’s fault leads high breathing process breather, never for During currents, temperature increased. Continuous imaging radiometric pixels saturation differentiates maloperation. Hybrid wavelet threshold image analytics (HWT-IA) based analysis after de-noising provides an accurate result about 95% transformer.