Research database

The core idea is to identify a small but statistically significant set of “sentinel” PV plants, strategically distributed across different climatic and morphological zones. These plants will serve as proxies for the entire PV population, enabling the inference of national-scale production trends through limited but high-quality and real-time data. The identification of these sentinel systems will rely on multi-criteria clustering methods supported by geospatial analysis. In addition to geographic location, the selection will consider historical production profiles, meteorological variability, and plant-specific technical characteristics. The forecasting system will integrate physical models (e.g., solar irradiance estimation, panel temperature corrections, geometrical and electrical configuration) with advanced data-driven approaches such as neural networks, ensemble learning, and probabilistic forecasting. Particular attention will be paid to localized meteorological anomalies – such as Saharan dust transport – that may severely affect solar irradiance and are often poorly represented in conventional forecasting models. One of the expected outcomes is the identification of the optimal spatial resolution for clustering, enabling a balance between forecast accuracy and monitoring effort. A web-based demo software platform will be developed to support two levels of access: • Professional interface for system operators and stakeholders requiring accurate quantitative forecasts; • Public-facing GIS map, providing day-ahead PV production forecasts for communities, energy users, and prosumers. The demo will be implemented and validated in a selected pilot region, chosen for its geographical and morphological relevance. The platform will be designed for scalability and replicability in broader national and European contexts. The project is structured in six Work Packages: (1) review of state-of-the-art and critical forecasting challenges; (2) design of the hybrid modelling architecture and selection of sentinel systems; (3) development of the web platform; (4) quality control and dissemination through publications and public events; (5) preparation of a follow-up European collaborative project; and (6) overall management and coordination.