Combination Forecasting of Power Load Based on Polynomial Trend Extrapolation and ARIMA Model

The power load forecasting is the core component of the early warning system for fuel storage margin in power system and an important guarantee to the early warning function to achieve. In this paper, one province's 2008 load data is chosen to forecast the electricity consumption in 2009. Firstly the two forecasting models of polynomial trend extrapolation and ARIMA are established, and then the combined model of them is used to forecast, that is, the final result is equal to the sum of the trend value by polynomial extrapolation and the non-trend D-value’s forecasting result by ARIMA. The results indicate that the combination forecasting make the forecast accuracy significantly improved and ensure the effective operation of the early warning system.