The optimal management of both transmission and distribution electricity networks requires precise knowledge of the current and future state of the flows passing through them. Accurate and reliable indications of both production from renewable sources and demand are key elements to reduce the demand for reserves, which are expensive and not always available, as well as to limit network congestion (e.g., via dynamic airline management). The uncertainty and variability due to renewable sources is linked to the meteorology and climatology of their location, as is the demand for electricity for domestic use. It is thus necessary to make forecasts on different spatial and temporal horizons of photovoltaic and wind production and energy needs. The activities carried out were aimed at defining innovative and high-performance predictive methodologies for both generation and demand, building hybrid multi-model meteorological and post-processing systems based on statistical and machine learning methods. The time horizons under examination ranged from the short term (3 days, with hourly step) to the very short term (3 hours, with quarter-hourly step), whereas the tested techniques ranged from intelligent persistence to statistical methods (Analog Ensemble, functional Principal Component Analysis), machine learning (Support Vector Regression, Random Forest, Quantile Random Forest, Neural Network with 1-dim and 2-dim series) and regression models (ARIMAX with exogenous forcing provided by short-term and satellite forecasts). Furthermore, the increasingly widespread presence of innovative photovoltaic systems, such as concentrated ones, azimuthal tracking planes, bifacial and tandem cells, require information at the spectral level of the direct and diffuse component. The ARTEMIS (Automatic Radiative TransfEr Model for Irradiation with Satellite data) methodology, which is currently being developed, is a hybrid method which, using information provided by the geostationary meteorological satellite Meteosat Second Generation (MSG) and some numerical meteorological models, reconstructs high spatial and temporal resolution and, for the entire Italian territory, the radiation spectra under clear sky conditions. The system was validated using data from some RSE solar spectrophotometers installed in Milan, Piacenza and Cagliari (at Sardegna Ricerche).