Expert models for electric load forecasting of power system

The paper addresses the problem of expert models application for electric load forecasting of the typical days a week. The methodologies under consideration found their practical application in day-ahead electric energy market for X-1-day electric load forecasting given the loading curve for a typical X day, excluding holidays, the days before and after the holidays as well as any other non-typical days. The expert models approach includes the formulation of fictitious perspective loading curves, based on which the forecasting procedure is carried out. The developed algorithm was verified using real data of Ural Power System loading conditions. The calculations of the total forecasting error showed that the proposed methodology has high-quality accuracy characteristics, which makes its application possible for real power system loading forecasting.