The purpose of AI-4ENERGY is to optimize the trading, management and sharing of distributed generated (DG) energy from renewable energy (RE) resources in future digitized energy markets by applying AI and ML algorithms.
Project objectives
- Design ML algorithms able to identify and predict the uncertainty of production and consumption of individual prosumers having RE resources and furthermore, the correlation of multiple DG from RE resources in a local microgrid, where similar demand and supply patterns may emerge. The algorithms will be tested and the prediction quality from hours to day ahead will be evaluated using operational measurement data available through partner Glava (from PV system in operation since 2009 and from solar radiation since 2013).
- Develop advanced energy trading and energy management schemes based on AI for the new smart grid architecture, including individual prosumers for RES. We will consider the supply–demand balance to guarantee the safe operation of the smart grid and leverage the predictions of energy demands and supply and associated uncertainties. This will result in optimization problems that we aim to solve using fast algorithms based on ML techniques. We expect that our algorithms will lead to cost savings of prosumers compared to fixed pricing schemes by at least 20% and prosumer willingness to engage in flexible DR schemes will improve leading to significant CO2 reduction increasing with the penetration of DG and DS.
- Integrate ML algorithms into OpenSource Cloud and Edge platforms. We plan to integrate the algorithms for prediction of individual supply and demand into OpenSource edge platforms, that run locally at the facilities of the small-scale RE prosumers. In addition, we integrate the AI based energy management schemes into multi-tenant Cloud platforms and implement demonstrator applications that showcase project results.
Workpackages
WP1: Machine Learning for Robust Supply/Demand Forecasting
WP2: Machine Learning for Robust Energy Management
WP3: Cloud/Edge/AI Integration
WP4: Performance Assessment