Research on Forecasting of Monthly Residential Electricity Consumption Considering the Decomposition of Quarterly Variables and Stochastic Variables

With the acceleration of the power market reform, China has introduced the proxy power purchase policy, and then the users' monthly electricity consumption is needed to be forecasted separately for residential electricity consumption. To solve the problem that time series model is difficult to predict accurately, this paper proposes an improved ARIMA forecasting method for residential electricity consumption considering seasonal variable decomposition, decomposes monthly residential electricity consumption into time trend component, seasonal cycle component and stochastic component through the X12 seasonal adjustment method, and combines with the Eviews software to design and realize the decomposition of monthly residential electricity consumption. Finally, an example analysis is carried out by combining the actual residential electricity consumption data of provincial power grids, and the results show that the prediction error is relatively small, and it can accurately predict the monthly residential electricity consumption, which can have a positive effect on the proxy power purchase business of power grid agents.