Abstract: | This paper presents a short_term load forecasting(STLF)system based on artificial neural network(ANN). Load data of Guangzhou electric power network are used in numerical simulations to give both the one_hour ahead forecast and the following 24_hours forecast. The results show the promising foreground of the applications of ANN in STLF. In addition, the construction of the every_hour model, the improvement on the selection method of training cases and the adoption of the highly effective Levenberg_Marquardt(LM) training algorithm greatly accelerate the training of the ANN model, making possible real_time implementation. |