Weather forecast based scheduling for demand response optimization in smart grids

Demand side management is one of the most promising techniques to reduce energy losses in future electricity distribution networks. Although this topic has been addressed in literature already, the existing approaches remain on the level of mathematical and analytical models rather than deployable solutions that would actually utilize controllable demand. In this paper, we present an algorithm for scheduling water heaters' activity intervals with the purpose of balancing fluctuations of total power demand and uncontrollable photovoltaic production, and designed for real deployment. To do so, our method uses historical metering data and weather forecasts. We specify the environment the algorithm was designed for, describe predictions we did and present the steps of choosing optimal switching times.