Краткосрочный прогноз часового потребления электроэнергии с учетом погоды для субъектов РФ

Известия РАН. Энергетика. 2017. Т. 55. № 5. С. 47–56.

Gordin V. A., Bykov P. L.

The power consumption is determined by various factors. It is affected by time of day, day of week, and of season. The price of electricity is determined by the controller, and can affect the schedule of consumption of various consumers. And this formulae depends essentially on the concrete region. Forecast of consumption may influence both the tactics of consumers and producers of electricity and the strategy of the regulator. The impact will be stronger, when more reliable the forecast. One of the non-recurrent factors is the weather. We developed operational computer technology short-term forecasting of hourly electricity consumption for 63 subjects of the Russian Federation, using the short-term forecast of the air temperature. We evaluate the parameters of our algorithm according to information from archives, which describes electricity consumption and air temperature. The MAPE error obtained forecast the average for the subjects is 3.2% while the forecast for 1 day and 3.7% in the second. The impact of the weather forecast into the error decreasing corresponds to the reduction of the forecast lead-time of 1 day.

Keywords: метод главных компонент, correlation, корреляция, потребление электроэнергии, прогноз погоды, weather forecast, power load, principal component method

Publication based on the results of: Дифференциальные и интегро-дифференциальные уравнения. Включение больших массивов статистической информации. Модели. Аналитические и вычислительные подходы (2016)