Language models assessed for restoring power with incomplete grid data
FLARE: Physics-Grounded Evaluation of Language-Model Agents for Feeder Restoration with Degraded Observations
Machine Learning
Summary
Restoring electricity after power grid faults is tricky because operators must make decisions with incomplete or conflicting information while making sure the electric system stays safe. The authors created FLARE, a testing system that combines real electrical rules with scenarios where data is limited or uncertain. They used FLARE to test several language-based AI models on many restoration tasks and found that while most decisions were safe, there were some important cases where mistakes could cause violations. This work helps safely evaluate AI tools that might support electric grid restoration.
What this means in practice
- •For power grid operators: Evaluate and improve AI assistance for restoring grid service safely after faults with limited measurement data.
- •For energy system software developers: Test AI models under realistic electrical constraints and partial observations to build safer grid restoration tools.
Authors
BHVSP Subrahmanyam, Neeraj Choudhary
Abstract
Restoring service after distribution-grid faults requires decisions from incomplete and conflicting information while respecting electrical constraints. Addressing this need, we introduce FLARE (Feeder-Level Agent Restoration Evaluation), a physics-grounded benchmark for post-isolation feeder restoration. FLARE con- tains 800 AC-feasible cases across five observation settings and allows agents to gather evidence, evaluate switches, and perform restoration actions under fixed operational limits. We evaluate five instruction-tuned models over 78,750 episodes, including topology, distributed-generation, and loading shifts. Mean normalized regret ranges from 0.082 to 0.148. Although every model has zero median regret, P95 regret ranges from 0.814 to 1.000, revealing an important failure tail. Shadow replay shows that 96.4% of 15,706 guard-rejected actions would have caused a physical violation. FLARE enables reproducible evaluation of language-model based restoration support while emphasizing the need for external safety checks.