Predavateljica: Aleksandra Rashkovska Koceva (IJS)

Povzetek:
Many critical operations in the energy sector — from determining the transmission capacity of power lines to managing the thermal loading of transformer stations — are directly dependent on weather conditions. Accurate, localized weather forecasts are therefore essential for safe and efficient grid management, particularly for applications such as Dynamic Line Rating (DLR) of power lines and Dynamic Thermal Rating (DTR) of transformer stations, where real-time environmental data directly influences operational decisions.

In this contribution, we present our work on AI-enhanced weather forecasting tailored for weather-dependent energy applications, carried out across two complementary projects and the SLAIF initiative. Within the Horizon Europe project HEDGE-IoT, we develop and deploy edge-based DTR/DLR systems integrated into substations across the Slovenian demo site. AI-powered tools process real-time environmental sensor data to support predictive grid planning and operational automation. Through the SLAIF initiative (Slovenian AI Factory), we are building AI-based weather forecasting services as part of the Green Transition vertical, providing industry-ready models to support DTR and broader transmission network management. The newly funded ARIS project AIDTR (L2-70125) complements these efforts with fundamental research on predictive models for DTR, addressing challenges such as spatial and temporal modelling, concept drift, probabilistic forecasting, and multimodal learning with LiDAR data. Together, these projects and initiatives form a coherent pipeline from research to operational deployment, demonstrating how AI-driven weather forecasting can help grid operators unlock hidden capacity in existing infrastructure and improve system resilience.

Keywords: dynamic thermal rating, weather forecasting, AI, power grid, transmission lines, transformer stations