小波神经网络在电力负荷预测中的应用

Application of Wavelet Neural Network in Power Load Forecasting

针对BP神经网络的问题与不足,提出小波神经网络电力负荷预测模型。仿真实验结果表明,小波神经网络可根据信号的时频特性进行动态调整,有利于减少权值参数,缩短训练时间,在隐节点数目和所需训练样本等方面都表现出很大的优越性,对于电力负荷预测具有较高的精度,是可行的。

Aiming at the problems and shortages of BP neural network, it proposes a neural network model of power load forecasting based on wavelet neural network. Simulation experimental results show that the wavelet neural networks can be dynamically adjusted by the time-frequency characteristic of signal, and is conducive to reducing the weight parameters, shorten the training time, in the number of hidden nodes. Required training samples showing great superiority for electric power load forecasting with high accuracy, is feasible.