摘要
在神经网络负荷预测实际应用中,突出的问题是训练样本大、训练时间长、收敛速度慢。针对负荷预测样本代表性问题,建立了基于蚁群聚类的Elman神经网络预测模型。对负荷历史数据进行蚁群聚类预处理,将聚类后的数据作为神经网络的训练样本。其目的是使输入样本具有代表性,改善网络训练时间和收敛速度,有效提高预测精度。通过某发电厂负荷数据的验证,该模型的预测结果精度较好。
In application of neural networks based short-term load forecasting model, the main problems are over many training samples, thus resulting long training time and slow convergence speed. For the representativeness of training samples, an ant colony clustering based Elman neural network forecasting model was presented in this paper. First, historical load data were pro-processed by using ant colony clustering method. Then the clustered data were chosen as the training samples for the network. The objects are to make the input samples representative, decrease training time and increase convergence speed, thus improving forecasting accuracy. Based on daily load data in Datang electric power plant, this model can obtained more accurate forecasting results.
出处
《中国电力》
CSCD
北大核心
2006年第7期49-51,共3页
Electric Power
