The base of power system operating and programming, namely load forecasting, will enable us, by accurate forecasting, to enhance the systematic operating reliability and economize the whole system. Thus, short-term load forecasting ranks as one of the most important means to guarantee that the power system run under safety and at a low cost.This paper detect and revise the deficient and abnormal data that affect load forecasting precision through the Vertical contrast method and the Grade method. Based on this theory, analysis of optimal window-width to all kinds of forecasting models is given and iterative method is applied to get the optimal window-width in different forecasting models at different time. Combined forecasting model advantageously embraces the merits of many existing models and could get more accurate forecasting results. This paper emphasizes on the implementation and application of this combined method. Moreover, algorithms is applied to the power choice of the combined forecasting model. The analysis of the results of the calculation examples may be available for short-term load forecasting.
|