The integration of solar energy into the built environment is undergoing a major transition. Building-Integrated Photovoltaics (BIPV) and smart microgrids require advanced operational strategies to handle variable weather conditions, complex urban shading, and dynamic loads. While Artificial Intelligence (AI) and data-driven methods offer promising tools for PV forecasting and system control, current challenges lie in model robustness, physical consistency, and computational efficiency.

This Special Session focuses on advanced AI and data-driven solutions designed to optimize the lifecycle of building energy systems, from initial potential assessment to real-time grid integration. A key emphasis is placed on bridging the gap between the physical constraints of PV devices and modern machine learning, ensuring that AI models remain physically consistent, computationally sustainable (frugal AI), and highly integrable with smart microgrids and Electric Vehicle (EV) infrastructures.

The enhancement of electric mobility solutions as a major transition being deployed on public roads is reshaping the whole transportation sector. This revolution is helpful not only in reducing the carbon footprint of the mobility sector but also induces some changes of habits and needs to be addressed deeply considering scenarios, vehicles and human factors and interactions. Model-based (MB) and Data-driven (DD) approaches can represent valuable tools to address the topic under different perspectives and based on a diversity of boundary conditions. Digital Twins (DTs) can integrate advantages of both approaches to consider interactions in a Human-in-the-Loop architecture (HITL), without losing meaningfulness of human interaction.

This Special Session focuses on how DTs can reshape the analysis of mobility systems, ranging from Heavy-duty Electric Vehicles (HDEVs) to soft-micromobility solutions (e-scooters, e-bikes) in urban and extra-urban contexts. DT can merge MB and DD approaches to increase the level of problem formulation and analysis without losing the interaction between the human being and the surrounding environment, usually neglected by classical simulation environments. This last aspect represents the key point of the implementation of DT, bridging the gap between the physical constraints of real-world scenarios and simulation environments.