A Fuzzy Logic Approach to Evaluating Mechanical and Insulation Faults in Transformers for Electric Arc Furnaces

Electric arc furnaces (EAFs) subject transformers to substantial thermal and mechanical loads, often resulting in accelerated insulation wear and mechanical defects. If not diagnosed and mitigated in time, these degradations can significantly reduce the service life of transformers and cause operational disruptions in industrial environments. This work introduces a fuzzy logic-based diagnostic framework aimed at evaluating the mechanical and insulation-related health of EAF transformers. Assessment of condition of transformers requires intricate approaches beyond simple fixedthreshold methods as transformers operate under a wide range of operational conditions and generalizing operation of transformers isn't recommendable. Therefore, a fuzzy logic-based technique is used which skillfully incorporates expert knowledge and excels at handling ambiguous or incomplete data. The system employs a system of if-then rules and defines membership functions for examining crucial indicators such as heat stress, partial discharge signals, vibrations, and dielectric performance to rank different components on the basis of their extent of degradation. This ranking allows maintenance teams to prioritize their actions and devise a cost-performance optimized maintenance plan.