Baseline Load Forecasting for Demand Response Taking Meteorological Factors into Account

Abstract

To conduct baseline load forecasting for commercial building users participating in demand response, this paper proposes a demand response baseline load forecasting method that takes meteorological factors into account. Using the principal component analysis, the meteorological indicators contributing more than 80% to the baseline load are selected as the reference data for forecasting neural network baseline load. The isolated forest algorithm and KNN algorithm are used to identify the abnormal data and to fill in the missing data, and the data with different attributes are partially “decoupled.” The two parts of the decoupled data are separately used for neural network load forecasting, and the total baseline load of users is obtained after merging. From these results, it is obvious that the proposed method can effectively improve the forecasting accuracy of the baseline load of the algorithm.