Understanding the Surface Discharge Activity With Nano Oil-Pressboard Insulation Under AC and Lightning Impulse Voltages

The present study investigates the performance of nanofluid-impregnated pressboard material for surface discharges. Surface discharge is simulated as per IEC 60112 standard with nano TiO2 dispersed mineral oil-impregnated pressboard. During surface discharge, multiple activities, such as localized surface discharge inception followed by bright light occurrences, bubble formation, and bridging of electrodes, occur over an extended period of time. The surface discharge activity under ac and transient voltage profiles was monitored by adopting ultrahigh-frequency (UHF) and optical detection sensors. The pressboard material impregnated with nanoparticle dispersed fluid shows that 9.7% improvement is surface discharge inception voltage (SDIV) compared to pressboard impregnated with pure transformer oil. Furthermore, three machine learning (ML) algorithms, support vector machine (SVM), random forest (RF), and k-nearest neighbor (kNN) algorithms were used to evaluate the best-fit model for classifying the various stages of the surface discharge process. Classification accuracy is achieved to be 97.3% with the optimized parameters. Laser-induced breakdown spectroscopy (LIBS) identifies that the elemental composition at the discharge sites on the pressboard and the plasma temperature of carbon traces were evaluated. Plasma temperature for lightning impulse treated is lower due to the instantaneous breakdown occurring in the surface of the pressboard in comparison to the ac voltage profile.