Benchmarking Machine Learning Models for Fault Classification and Localization in Power System Protection

2025-10-01

AI Summary

Traditional fixed-threshold protection schemes exhibit insufficient reliability in dynamic short-circuit fault identification and localization under high penetration of distributed energy resources. Method: This study conducts the first systematic, comprehensive benchmarking of classical machine learning models for real-time power grid protection. Using electromagnetic transient simulation data, voltage and current waveform features are extracted via 10–50 ms sliding time windows; model performance is rigorously evaluated across fault classification (F1-score) and fault distance estimation (R²) tasks, with emphasis on accuracy, robustness, and real-time inference latency. Contribution/Results: The best-performing classifier achieves an F1-score of 0.992 ± 0.001; the top regressor attains an R² of 0.806 ± 0.008. Average inference time per sample is only 0.563 ms—well within the millisecond-level response requirement for protective relaying. This work establishes the first unified, reproducible, real-time-constrained benchmark for ML-based protection models, addressing a critical gap in the literature.

Technology Category

• Machine Learning: Hardware-aware ML
• Humans and AI: Human-in-the-loop Machine Learning
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

The increasing integration of distributed energy resources (DERs), particularly renewables, poses significant challenges for power system protection, with fault classification (FC) and fault localization (FL) being among the most critical tasks. Conventional protection schemes, based on fixed thresholds, cannot reliably identify and localize short circuits with the increasing complexity of the grid under dynamic conditions. Machine learning (ML) offers a promising alternative; however, systematic benchmarks across models and settings remain limited. This work presents, for the first time, a comparative benchmarking study of classical ML models for FC and FL in power system protection based on EMT data. Using voltage and current waveforms segmented into sliding windows of 10 ms to 50 ms, we evaluate models under realistic real-time constraints. Performance is assessed in terms of accuracy, robustness to window size, and runtime efficiency. The best-performing FC model achieved an F1 score of 0.992$pm$0.001, while the top FL model reached an R2 of 0.806$pm$0.008 with a mean processing time of 0.563 ms.

Problem

• Benchmarking ML models for fault classification in power systems
• Evaluating machine learning for fault localization under dynamic conditions
• Comparing classical ML performance on real-time protection challenges

Innovation

• Benchmarking machine learning models for power protection
• Using sliding window segmentation of voltage waveforms
• Evaluating models under real-time performance constraints