Peak Load Forecasting using Wavelet Neural Network with RPROP

Y. Feng and L. Jin (PRC)


Peak load characteristics, peak load forecasting, wavelet neural network, RPROP algorithm


In this paper, a new peak load forecasting (PLF) method for power system is proposed based on a wavelet neural network (WNN) with the resilient back propagation (RPROP) algorithm. In order to make full use of the information of the past load data, the model selects the load samples under the Euclid scale. The method then carries out peak load forecasting using a wavelet neural network which combines the wavelet theory with the artificial neural network. Here, the RPROP algorithm is used to regulate the parameters of the WNN for accelerating the training process. The proposed algorithm shows more accurate and more quickly converge to the peak load forecasting by actual data of a power grid in China.

Important Links:

Go Back