Designing aggregation criteria for end-users integration in energy communities: Energy and economic optimisation based on hybrid neural networks models

This study presents innovative methodologies addressing critical challenges in energy community real-cases implementation. The investigation is conducted by exploring and enhancing the concept of Peer-to-Peer energy community, where prosumers interact with consumers by sharing surplus energy to meet their electricity demands. The end-users' connections are optimised by maximizing their energy interactions and the proposed pricing strategies are based on balancing the supply and demand curves for tailored unit costs. This study aims to optimise the design of energy communities' configurations and the applicability in real case scenarios by suggesting novel aggregation criteria. These criteria are focused to select the type and the number of users in an energy community to maximize the economic benefit and the amount of self-consumed renewable energy. A bi-level optimisation approach is at the basis of these aggregation criteria. The first level maximizes self-sufficiency and economic benefits by aggregating prosumers and consumers into subgroups. The second level determines the optimal community configuration, prioritizing total energy self-consumption. This methodology requires the knowledge of users' electrical needs, therefore, to address the problem related to insufficient data, a Hybrid Neural Network model is proposed to simulate building electrical demands. A case study in a residential area of Caserta, Italy, demonstrates the methodology's effectiveness. By identifying user types, predicting demands, and employing optimisation techniques, the study estimates economic benefits for consumers (1.6% to 19.5% savings) and prosumers (return on investment