Short Term Pv Forecasting Using Satellite Data For Austria

For the grid integration as well as for the integration into smart grids of renewable energies accurate forecasting methods for renewal energies are important. In this paper we consider intraday forecasting of photovoltaic (PV) power production and focus on methods which can be used with satellite data. After a literature review we tested the methods cross correlation (CC), the Bayes method (BM), Variational optical flow (VOF) and persistence (PER) in more detail. With these methods we predict the PV-output for approximately 1500 PV systems in Austria. Finally we present some ideas how the existing methods can be improved and provide first results.