Modelling and Forecasting of Photovoltaic Generation for Renewable Energy Communities: A Narrative Review
Abstract
This narrative review examines the state of the art in modelling solar energy production in energy communities, with a particular focus on photovoltaic systems. It explores a wide range of approaches, from classical parametric models to intelligent techniques such as machine learning and deep learning. It identifies key methods, their applications and limitations, with an emphasis on the transition from static models linked to physical system parameters to dynamic and data-driven approaches using weather data and historical data inputs. It further identifies a specific gap in the current literature: the predominance of short-term forecasting over real-time estimation and the limited integration of intelligent techniques with dynamic sharing coefficients and peer-to-peer exchange schemes. Synthesising advances in peer-to-peer energy exchange models, distributed generation frameworks, and predictive optimisation systems, this review highlights the integration of intelligent techniques as a promising direction for improving the management of renewable energy communities, since such techniques are data-driven and decoupled from the physical structure of the photovoltaic system.