Recursive analysis and forecast of superimposed generation and load time series

摘要

The scope of this survey is the forecast of mixed power time series characterized by the superimposition of various generation and load components. This can support planning tasks for operational management in grid systems with high proportion of renewable energies. For this study real measured data of several time series as well as measured and forecasted values of weather and PV feed-in data are used. Within the forecast process the number of signal components is stepwise reduced by repetitive analysis and model construction. By that a component-wise modelling is pursued. Different model variations are compared and first results are shown.