Controlled Comparison of Machine Learning Models for Fault Classification and Localization in Power System Protection
A new paper details a controlled comparison of machine learning models for fault classification and localization in power systems. The study used identical datasets, sensing assumptions, and decision horizons to ensure comparability across models. For fault classification, top models achieved F1 scores above 0.98 even with short 10ms decision windows, indicating early transient data is informative. For fault localization, the best models achieved a stable error of approximately 10% of the normalized line length, with accuracy varying by grid segment rather than solely by temporal context. AI
IMPACT Provides a standardized reference for evaluating machine learning models in power system protection, potentially accelerating adoption and improving grid reliability.