Impact of wind power production in a European Optimal Power Flow

Wind power production is very important and, due to its variability, it seems clear that a single wind production scenario is not sufficient to characterize correctly the net demand. For this reason, we develop in this paper a technique that, starting from historical time series aggregated at the country level, leads to the selection of a shortlist of wind production scenarios statistically relevant. We show how to use Principal Component Analysis and clustering techniques to obtain the shortlist of scenarios to be simulated. Different techniques are compared, analysing the statistical impact on relevant results (transits, cost). The results follow a discrete distribution dependent on the wind scenarios selected, and thus allow to capture the correlations with wind production in different countries.