Forecasting Electricity Consumption Using Neural Networks

Abstract

The problem of early and accurate forecasting of electricity consumption is acute for the unified energy system of Ukraine. With successful forecasting of consumption, which is based on many aspects, it is possible to buy electricity/losses in different market segments much more profitably, saving large amounts of money, which can then be directed to the development and modernization of electricity networks. This has always been an urgent issue, but today, when a large part of Ukraine's energy equipment has been destroyed by Russian missiles, it has become even more painful. The use of the method of artificial neural networks (ANN) for short-term forecasting of electricity consumption is considered. It was established that ANN can be used to make a forecast of electricity consumption a day ahead with an error of 4.86% compared to the actual amount of electricity consumption. Performing a comparison of forecast values with actual values allows us to talk about the adequacy of the selected forecasting model and its application in practice for the successful operation of energy supply companies in the electricity market.