Evaluating Fault Detection Techniques in Real Electrical Transformers: A Comparative Case of Study

Transformers are integral to the reliability of electrical networks, necessitating robust diagnostic methods for fault detection. This study conducts a comparative evaluation of three fault detection techniques using real-world data from distribution transformers. The methods analyzed include differential current analysis, correlation-based techniques, and flux linkage increments. Results demonstrate that differential current analysis exhibits the highest sensitivity (93.33\%), detecting faults at 4.41\% of the short-circuit current. Correlation-based methods follow with 86.67\% sensitivity, while flux linkage increments offer lower sensitivity but robust performance at high current levels. This comparative analysis provides actionable insights for enhancing transformer reliability through effective monitoring strategies.