The liberalization of the Italian electricity market has led to a situation where the logic of maximizing operators’ profits prevails. In this context, it is essential for electricity producers to be able to formulate supply strategies. Therefore, plants that can buy or store the energy resources they use to produce electricity are advantaged, because they can choose when to produce with more flexibility. This feature is often absent in environmentally sustainable electrical plants, although hydropower storage plants offer flexibility to the electricity system and allow their operators to formulate bidding strategies. This paper provides a method to identify the best predictive variables and the appropriate predictive indexes for an aggregate hydropower storage forecasting model. To this end, we use an entropy-based approach. The optimal subset can be considered complete, meaning it incorporates both market and physical variables, including those influencing the timing of snowpack melting. In particular, we show that the variables that incorporate the expectations of the operators always have a more dominant role than the realized variables in understanding the current hydropower generation. We find evidence of a cyclicality, over the year, of relevant drivers in the representative subset; in particular, we show the mechanism by which, during this cycle, some variables become more or less influential on others.
Condemi, C., Mastroeni, L.C.L., Vellucci, P. (2021). The selection of predictive variables in aggregate hydroelectric generation models. THE JOURNAL OF ENERGY MARKETS, 14(1), 27-60 [10.21314/JEM.2020.215].