【摘要】 在采用分段预测方法的基础上,利用小规模BP(back propagation)神经网络学习时间短和径向基函数(radial basis function,RBF)神经网络自身训练速度快的优点,提出了基于BP和RBF网络的级联神经网络日负荷预测模型,将影响日负荷变化的非负荷因素(气象、日类型等)与历史负荷因素分别加入BP和RBF网络中分开考虑,进一步简化了预测模型。计算实例表明,该模型较一般级联神经网络模型收敛更快速、高效,预测精度有了很大提高。更多还原

【Abstract】 Based on subsection forecasting and utilizing the advantages of back propagation(BP) neural network and radial basis function(RBF) neural network,such as short learning time of small-scale BP network and quick training of RBF network itself,a daily load forecasting method based on cascaded neural networks(CNN) is put forward.In this model non-load factors,i.e.,weather factors,day styles and so on which affect the changes of daily load,and historical load factors are added into BP and RBF neural networks and considered separately,thus the forecasting model is simplified further.The results of calculation example show that the proposed model converges quicker and is more efficient than common CNN,the forecasting accuracy is evidently improved.更多还原