Graph Neural Networks Show Promise for Fault Location in Power Grids

A new research paper explores the effectiveness of Spatio-Temporal Graph Neural Networks (STGNNs) for fault location in power distribution networks with increasing distributed energy resource (DER) penetration. The study benchmarks STGATv2 against temporal and spatial models, finding that STGATv2 performs best, achieving 92-94% accuracy. Notably, the model trained at higher DER penetration levels generalizes better to lower levels than vice-versa, and it maintains robust performance even under realistic measurement noise. AI

IMPACT This research could improve the reliability of power grids by enabling more accurate fault detection in the face of complex, modern energy systems.

RANK_REASON Academic paper detailing a new methodology and benchmark results. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Distributed Energy Resources Takes Center Stage: A Renewed Spotlight on the Distribution Planning Process
• gated recurrent unit
• GATv2
• graph neural networks
• IEEE 123-bus feeder
• Spato-Temporal Graph Neural Networks
• STGATv2

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →