Laura Boca de Giuli, doctoral candidate in Information Technology from DEIB Systems and Control section, has won the "Si può fare di più" national award by Fondazione Cogeme for the best thesis in the energy sector, assigned to her M.Sc. thesis "Physics-based Neural Network modelling, Predictive Control and Lifelong Learning applied to District Heating Systems”.
The thesis, carried out with the supervision of Alessio La Bella and Riccardo Scattolini, proposes innovative machine learning techniques to model and control complex and large-scale energy systems, such as district heating networks. In particular, the thesis proposes a novel modelling methodology that allows integrating physical system knowledge, such as the topology of energy networks, with advanced neural network models, guaranteeing significantly superior performance with respect to existing modelling methods. The developed models are particularly suitable for the design of predictive control systems, capable of optimally managing district heating networks with significant energy savings and reduced computational times. The thesis proposes also an innovative continuous learning algorithm, capable of monitoring the developed models and automatically adapting them in case of changes in the system or in the operating conditions.
The “Si può fare di più” national award by the Cogeme Foundation was given to Laura Boca de Giuli during the award ceremony which was held on December 15, 2023 at the Cogeme headquarter in Rovato (BS).