Robustness Evaluation of Machine Learning Models for Fault Classification and Localization In Power System Protection

2025-12-17

AI Summary

To address the degradation of sensor data—such as channel failures, reduced sampling rates, and communication interruptions—in emerging power systems, which undermines the robustness of protective relaying, this paper establishes the first multi-dimensional robustness quantification framework tailored specifically for relay protection tasks. Leveraging high-fidelity electromagnetic transient (EMT) simulations, we generate realistic degraded sensor data and integrate performance degradation analysis with observability impact quantification to systematically evaluate the sensitivity of fault classification and fault location models. Results reveal substantial robustness heterogeneity: fault classification remains relatively robust (e.g., only a 13% accuracy drop under single-phase current loss), whereas fault location exhibits extreme sensitivity to voltage measurement loss (yielding over 150% increase in localization error). This work is the first to explicitly characterize such task-level robustness divergence, identifies critical measurement channels, and provides both theoretical foundations and benchmark support for designing robust protection algorithms and optimizing sensor placement.

Technology Category

• Intelligent Robots: State Estimation
• Machine Learning: Adversarial Learning & Robustness
• Natural Language Processing: Safety and Robustness

Application Category

• Security and Privacy: Large-scale security measurements
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

The growing penetration of renewable and distributed generation is transforming power systems and challenging conventional protection schemes that rely on fixed settings and local measurements. Machine learning (ML) offers a data-driven alternative for centralized fault classification (FC) and fault localization (FL), enabling faster and more adaptive decision-making. However, practical deployment critically depends on robustness. Protection algorithms must remain reliable even when confronted with missing, noisy, or degraded sensor data. This work introduces a unified framework for systematically evaluating the robustness of ML models in power system protection. High-fidelity EMT simulations are used to model realistic degradation scenarios, including sensor outages, reduced sampling rates, and transient communication losses. The framework provides a consistent methodology for benchmarking models, quantifying the impact of limited observability, and identifying critical measurement channels required for resilient operation. Results show that FC remains highly stable under most degradation types but drops by about 13% under single-phase loss, while FL is more sensitive overall, with voltage loss increasing localization error by over 150%. These findings offer actionable guidance for robustness-aware design of future ML-assisted protection systems.

Problem

• Evaluates robustness of ML models for fault classification and localization in power systems
• Addresses reliability under missing, noisy, or degraded sensor data conditions
• Provides framework for benchmarking models and identifying critical measurement channels

Innovation

• Unified framework for evaluating ML model robustness
• High-fidelity EMT simulations model sensor degradation scenarios
• Benchmarking methodology quantifies impact of limited observability