Review of AI in Energy Management System in Isolated Hybrid Power Systems

Authors: Terji Nielsen, Helma Maria Tróndheim, Claus Leth Bak, Hannes Gislason, Høgni C. Kamban

This review paper summarises the status of artificial intelligence (AI) in energy management in isolated hybrid power systems, like the one in the Faroe Islands, dominated by variable renewable energy sources like wind, hydro, solar, etc. AI techniques, particularly machine learning (ML), with its distinctive advantage of solving complex problems in real-time, have gained traction as data availability has grown extensively due to the increasing complexity of modern power systems, with various installed components and sensors. AI techniques are applied in forecasting the energy demand, the potential generation from the different renewable sources, prone to spatiotemporal variations, optimisation and dispatch of generation, storage assets, and ancillary services, to ensure the striking balance between supply and demand in a stable, secure, and efficient islanded power grid. A critical aspect of integrating variable renewable sources in the power system is their intermittent nature, compromising grid stability, frequency control, and voltage regulation. AI has emerged as a transformative tool in the operation and optimisation of power systems. ML algorithms and predictive models enable accurate load and demand forecasting and real-time decision-making, enhancing grid efficiency and reliability.