Performance Comparison of Ensemble Classifiers Algorithms Used in Transformer Fault Detection

• DOI: 10.52846/aucee.2024.04
• Publication Date: 2025-02-12T09:43:09Z
• Authors: Nițu Maria-Cristina; Aciu Ancuța-Mihaela; ; Nicola Marcel; ; Nicola Claudiu-Ionel
• Keywords: Ensemble Learning

Power transformers are essential elements in the production and distribution of electricity, keeping them optimum operating condition is a constant concern for specialists field. The power mainly determined by mixed insulation system, i.e. solid cellulose paper liquid insulating oil insulation. identification method, described this order to determine fault based on fact that assessment their namely made oil. This why Three Ratio Technique (TRT) used with good results early detection transformer faults. method considered as simple, but at same time efficient interpreting dissolved gas analysis. It uses three new ratios differentiate between thermal electrical In paper, defined TRT train machine learning classifier Ensemble Classifiers using Bagged Trees (random forest), Boosted Trees, RUSBoosted algorithms. validation software application proposed carried out experimental section.