Forecast load impact from demand response resources

Abstract

To improve forecasting accuracy for baseline load and load impact from demand response resources, this paper develops three innovative statistical models. These models are regression spline fixed effect model, fixed effect change point model and mixed effect change point model. The models developed are applied to forecast baseline load and load impact from air conditioning cycling demand response program in Southern California. All three forecasting models yield accurate forecasts for baseline load and load impact from demand response events. Noticeable rebound effect from demand response events are observed from the dataset.