SONDER - Service Optimization of Novel Distributed Energy Regions

The project aims at effectively improving the accuracy of the grid energy balance prediction for subsequent optimization in several ways: 1) by validating sensors data (improved input quality); 2) as the grid is a networked system at different abstraction levels, novel advanced graph-based machine learning techniques will be used to design predictors (graph representations suitably models grid sensor datastreams and their functional interdependences); 3) Increase grid flexibility potential by advanced battery control algorithms; 4) lifelong online adaptation for algorithms he to track the grid evolution over time. Improved prediction helps better matching between forecasted and real energy balance at the local level enabling better utilization of flexible loads as well as more efficient energy trading at regional level and creation of efficient predictive maintenance of the grid.