Summary
This paper presents an AI-driven approach to forecasting electricity demand and variable renewable generation in the Faroe Islands’ isolated power system, supporting the transition to 100% renewable electricity by 2030. The study employs automated machine learning to leverage SCADA (Supervisory Control and Data Acquisition) and weather data to generate accurate short-term forecasts, enabling efficient dispatch of generation and storage resources and facilitating dynamic pricing strategies that encourage electricity use when renewable output is high. The methodology addresses the technical challenges of balancing supply and demand in a high-renewables scenario, demonstrating that automated machine learning can simplify model development and improve forecast accuracy compared with classical statistical methods.
Read more Read lessResults show that ensemble machine learning models generally outperform classical statistical approaches, providing more reliable predictions essential for safe and sustainable grid operation. The paper highlights the limitations of the European Centre for Medium-Range
Weather Forecasts (ECMWF) fifth-generation reanalysis (ERA5) due to its relatively coarse spatial resolution. It outlines plans to integrate high-resolution local data from the Faroese
Weather Office in future work. Overall, the findings support the broader adoption of artificial intelligence-based forecasting tools in isolated power systems, contributing to global efforts to build decarbonized, resilient power systems.
Additional informations
| Publication type | Session Materials |
|---|---|
| Reference | D2_11774_2026 |
| Publication year | |
| Publisher | CIGRE |
| Country | Denmark |
| Study committees | |
| File size | 1 MB |
| Price for non member | 30 € |
| Price for member | 30 € |
Authors
TRÓNDHEIM Helma Maria - Electrical Power Company SEV; BAK Claus Leth - Aalborg University; GISLASON Hannes - University of the Faroe Islands; KAMBAN Høgni C. - University of the Faroe Islands
Keywords
Artificial Intelligence, Generation forecasts, Load forecasts, Isolated Power Systems
