A Virtual Power Plant (VPP) is a cloud-based cluster of intermittent energy generators such as wind turbines, solar panels, power-to-x facilities, vehicle-to-grid facilities, etc. that can be controlled from a central point. IANOS will demonstrate a VPP that uses Artificial Intelligence (AI) to optimise the generation of energy and balance demand and supply of energy on the islands. The AI is based on meta-learning predictive methods, fog computing, and a smart grid/advanced metering infrastructure integration to maintain stable system operation by taking into account several uncertainties and technical constraints such as ramp-up and down times, and compensating imbalances within various timeframes.

The core functionalities of the VPP include:

• efficient forecasting of renewable energy sources uncertainties and demand profiling through novel meta-learning, AI-based predictors and incorporating physical and statistical models based on the specified response-to-accuracy ratio;
• local-to-global energy assets’ management, through the use of edge-to-cloud infrastructure to deliver intelligence at low voltage prosumers and empower the local energy communities towards self-consumption and tackling local-to-regional area supply and demand balancing;
• incorporation of transactive mechanisms based on DLTs and smart contracts to support energy credits’ exchange between prosumers, aiming at enhancing transparency, immutability and localised energy balance incentivisation (also enabling prosumers’ active participation in flexibility services.