Methods for energy price prediction in the Smart Grid

This paper compares different strategies for the prediction of energy prices. This feature is very important to support the Energy Management System for the computation of optimal power flows in a smart grid framework, e.g., in a Virtual Power Plant. The paper compares simple strategies like the typical one based on the assumption that the prices of the following day will remain the same of the current day, with more complicated approaches, like the Kalman Filter and empirical strategies that also include the information of the current day of the week. The performance of the different algorithms are thoroughly discussed and compared on real data taken from the Italian energy market.