Smart Gic - Resilient Power Grids Through Geophysical Sensing and Graph-Based Modeling

Geomagnetically induced currents (GICs), which are ground-level quasi-direct currents induced by space-weather phenomena, pose a serious threat to high-voltage power transmission grids, leading to transformer saturation, reactive power loss, and cascading outages during geomagnetic storms. This paper presents a smart, graph-theory-based framework that bridges real-time geophysical sensing with topology-informed grid modeling to protect power networks proactively. We represent the power grid as a graph data structure: substations and transformers are nodes, and transmission lines are edges; we then embed real-time geoelectric field measurements and space-weather forecasts. This feeds into a graph-based GIC algorithm motivated by sparsity principles of sparse-matrix computations and graph neural networks (GNNs), and trained against comparison grids, such as IEEE test systems and China's 500 kV Shandong network. Validation compared performance and accuracy using real storm instances in the UK and Poland. Our method, enhanced with GNN, detects 20% more high-risk grid sections than traditional approaches and enables optimized GIC mitigation hardware placement (e.g., series capacitors, DC-blockers), maintaining computation speed of approximately 18× owing to graph-parallelization. The models also estimate economic trade-offs, enabling cost-efficient investment in grid hardening and operating decisions in real-time during geomagnetic disturbances. This geophysics, electrical engineering, graph theory, and data science-backed cross-disciplinary approach provides an integrated, scalable solution for GIC threat prediction, mitigation, and management. Our findings verify its promise to enhance the resilience of the grid and ensure a secure electric supply in the age of rising space-weather uncertainty.