Formation of Power Consumption Forecasting Models for the Day-ahead Electricity Market Purposes

This article examines the dependence of electricity consumption on a set of climate factors in the power energy system of the Irkutsk region. To determine the forecast value, a linear regression model is used, which is evaluated by the Fisher F -criteria, the Student t-criteria, and the autocorrelation of the model rests was evaluated by the Darbin-Watson criteria. To obtain a meaningful linear regression for electricity market participants (energy retail organizations), sets of independent variables were determined. To eliminate the positive autocorrelation of the rests, the day of the week and the electricity consumption for the previous hour were taken into account - additional factors that are not related to climate factors. The obtained models for forecasting the volume of electricity consumption were used to determine the amount of demand for electricity, for which auctions were conducted on the day-ahead market in the software “DAM Simulator” by SKM PRO. The results of trading on the day-ahead market showed that it is important for the Irkutsk Region's power system to optimize the regime on a daily basis due to the enormous potential of hydroelectric power plants.