Renewable Energy Communities (RECs) are characterized by the presence of multiple actors operating autonomously, making it difficult to share data necessary for improving energy forecasting systems. This thesis project proposes the use of federated learning techniques, which allow the training of predictive models without centralizing data, safeguarding the privacy of individual participants and leveraging the unique characteristics of each REC in the process of forecasting energy consumption and production.
The student will benefit from synergies with LEAP – Laboratorio Energia e Ambiente Piacenza, already active on several projects on the topic.
Area di competenza
Smart Energy Systems
Responsabili
Filippo Bovera (filippo.bovera@polimi.it) – Politecnico di Milano
Matteo Zatti (matteo.zatti@polimi.it) – LEAP
Gaia Martoriello (gaia.martoriello@polimi.it) – LEAP
Marco Gabba (marco.gabba@polimi.it) – LEAP
Competenze richieste
Energy communities, machine learning, renewable energy sources
Background in Energy/Electrical Engineering
Data di inizio presunta
As soon as possible, deadline for applications by Nov 30, 2024
Durata
9-12 months. Full thesis only.
Maggiori informazioni
Weekly meeting online or in presence.
