Monthly Load Forecasting Model and Seasonal Characteristic Effect Analysis under the Background of Energy Internet

Abstract

A monthly load forecasting method based on load trend is proposed for monthly load data, which has dual characteristics of long-term trend and periodic fluctuation. Taking the monthly power generation from August 2012 to July 2017 as the research object, the monthly load data are decomposed into long-term trend and cyclic variation sequence, seasonal factor sequence and error sequence by seasonal decomposition. This paper focuses on the monthly cycle component characteristics of the four high energy-consuming industries, and deep analyses the characteristics of the monthly cycle component of the sub-industries electricity consumption and its impact on the electricity consumption of the industry. The monthly power generation from August 2017 to July 2018 is predicted by ARIMA model. The results show that the seasonal fluctuation law of monthly power generation is significant, and the relative errors of forecasting results are less than 3%, which verifies the validity and applicability of this method.