Daily load curve forecasting using factor analysis and RBF neural network based on load segmentation

The short-term daily load forecasting plays an important role in active and reactive power scheduling for power system. Aiming at the shortcomings that minute load fluctuation is not considered, a new daily load curve forecasting using factor analysis and RBF neural network based on load segmentation is proposed. Based on the precise prediction of 288-point daily load curve, normalized load curve segmentation method is used to segment the original load curve which reduces the number of segments and the demand of samples. To reduce the dimension, factor analysis is utilized to develop a few and independent common factors which characterize common features of daily load curve. Considering the load correlation, the RBF neural network model is established for each common factor, and then the minute daily load curve can be restored from the forecasted common factors. Case studies based on the historical load data of the actual substation in Shantou are provided to illustrate the effectiveness of the proposed method.