A Load Forecasting Algorithm for the Demand Side Resources

Demand side resources (DSRs) are regarded as a perfect choice to provide reserve capacity for the balance service in the power system. Forecasting of DSRs has been well-discussed in recent years. Plenty of researches are conducted on the forecasting algorithms in DSRs. However, the traditional methods take the whole load into consideration, without classifying the controllable and uncontrollable parts among DSRs. The forecasting results may be imprecise for the traditional methods as the controllable and uncontrollable parts have different characteristics. Therefore, this paper divides the DSRs into three parts, the original tendency, the upper dispatchable potential and the lower dispatchable potential. Different methods are separately applied on these three parts. Different time-series load forecasting methods are applied on the prediction for the original tendency of DSRs. The Support vector regression is applied as an intelligent forecasting method on the upper and lower dispatchable potential of DSRs. Various cases are studied to demonstrate the performance of the proposed method.