MTGreen

The increased use of photovoltaic (PV) energy production plays a crucial role in Europe's strategy to decarbonize its energy sector. PV power production is, however, highly dependent on atmospheric conditions. To maintain an efficient and stable electric grid, energy providers and grid operators therefore need reliable weather and PV production forecasts, as well as information on forecast uncertainties. These uncertainties are derived from forecast ensembles and allow planning mitigation measures in case of a strong increase or decrease in power production. MTGreen will build upon the following four current developments to improve radiation forecasts and PV forecasts over Austria:

1. The launch of the third generation of Meteosat weather satellites (MTG) with increased resolution and accuracy.
2. The upgrade of the GeoSphere Austria ensemble forecast system (C-LAEF) from 2.5 to 1.0 km resolution.
3. Advancements in data assimilation allow for the incorporation of cloud-affected satellite observations to improve initial conditions and subsequent forecasts of clouds and radiation.
4. The rapid progress of machine learning (ML) tools enables more efficient satellite observation simulation and more accurate site-specific PV power predictions.

MTGreen aims to develop a fully probabilistic forecast chain of PV power production by assimilating the next-generation MTG observations into the upcoming 1-km resolution ensemble forecast system, resulting in more accurate radiation forecasts. These improved forecasts will feed into an ML tool to forecast site-specific probabilistic PV power production.