Machine learning for performance prediction in smart buildings: Photovoltaic self-consumption and life cycle cost optimization

The application of Photovoltaic (PV) system in buildings is growing rapidly response to the need for clean energy sources and building decarbonization targets. Nonetheless, enhancing PV self-consumption through technical solutions such as Energy Storage Systems (ESS) getting higher importance increase profitability plants, by minimizing building-grid interaction. In this context, analyzing different storage configurations becomes more relevant crucial modeling although it heavily time-consuming complicated, particularly within a multi-objective optimization related ESS design. As solution resolve issue, paper evaluates accuracy, training, prediction speed 24 Machine Learning (ML) models be used surrogate smart buildings. Furthermore, performance short-term Thermal (TES) assessed presented using ML models. results showed Gaussian Process Regression (GPR), Neural Networks (NN) including bilayered trilayered NN models, Support Vector Machines (SVM) fine gaussian cubic SVM Ensembles Trees (EoT) superior also revealed that TES systems can efficiently equipped with electric heat pumps provide heating, cooling, domestic hot water. Moreover, size regarding Life Cycle Cost (LCC) LCC-based optimum yield 7.1% savings 30 years service life. novelties research are first reference select most suitable predicting self-consumption, second implement machine learning thermal enhance buildings, third carry out an on ML-based