Modern control techniques for power distribution systems and microgrids rely on predictive frameworks exploiting the forecast of stochastic resources at time scales ranging from several minutes to seconds. This research subject focuses on the development of machine-learning-based techniques for the ultra short term forecasting of both renewable energy resources as well as heterogeneous loads with small level of aggregation.
Validation of an advanced control algorithms based on short-term forecasting of PV generation (RE Demo)
Operation of the battery storage systems for grid control, feeder dispatching (RE Demo)
Tools for multi time horizon PV point forecasting and prediction intervals
Study of targeted feeders’ (Onnens/Rolle) operational limits (REeL Demonstrator)