Appears in collection : School/Workshop: Energy, mathematics, and theoretical challenges

Due to the Earth's rotation and cloud passage, photovoltaic (PV) power generation is inherently variable and stochastic, limiting its predictability. Grid operators typically curtail PV power when it exceeds grid demand, resulting in energy losses. Maintaining grid stability requires instantaneous balance between electricity consumption and production, posing significant challenges for PV integration. Solar power forecasting plays a crucial role in optimizing PV grid integration by enabling strategies such as energy compensation, storage sizing, demand-side management, and electricity trading. However, for intraday time horizons, accurate solar irradiance forecasting remains a complex task due to the challenges of modeling cloud dynamics and life cycles in traditional meteorological models. In recent years, alternative techniques leveraging geostationary satellite and all-sky camera observations have shown promise in providing more accurate intraday forecast, addressing the needs of emerging grid configurations, including microgrids self consumption optimisation, and aggregator portfolios valorisation for intradays markets.