A baseline prediction method for building demand response based on historical similarity day recognition

The conventional method for calculating the demand response baseline involves selecting the load data from the days preceding the demand response event and computing the average value. However, this approach may lead to significant inaccuracies when the load experiences substantial fluctuations. This paper proposes an improved calculation method for the demand response baseline of temperaturecontrolled loads based on identifying environmental similarity factors. The proposed method enhances baseline prediction accuracy by analyzing current electricity consumption patterns in buildings. The performance of the improved process is evaluated and validated through practical case studies.