Optimal Control Based on Deep Learning Techniques for a Hybrid Solar-Biomass System for Residential Buildings

Thumbnail Image
Zsembinszki, GabrielZsembinszki, Gabriel - ORCID ID
Fernàndez Camon, CésarFernàndez Camon, César - ORCID ID
Borri, EmilianoBorri, Emiliano - ORCID ID
Cabeza, Luisa F.Cabeza, Luisa F. - ORCID ID
Other authors
Journal Title
Journal ISSN
Volume Title
The EU-funded H2020 project SolBio-Rev aims to develop an energy system based on the use of solar energy and biomass with the objective to increase the share of renewable energy needed to meet heating, cooling, domestic hot water, and electricity demand in buildings. In this study, deep reinforcement learning techniques were applied for a SolBio-Rev system designed for a residential building (multi-family house) located in a Mediterranean climate in order to define an optimal control policy able to minimize the operating cost during summer and winter. Results showed that the smart control is able to reduce the cost of operation during summer compared to a standard rule-based strategy. Nevertheless, in winter the operating cost results slightly higher suggesting a further optimization for future studies.
Journal or Serie
EuroSun 2022. ISES and IEA SHC International Conference on Solar Energy for Buildings and Industry. 25 - 29 September 2022, Kassel, Germany