灰色Elman神经网络的电网中长期负荷预测  

Prediction of Mid-long Term Load Based on Gray Elman Neural Networks

作　　者：张健美[1] 周步祥[1] 林楠[2] 张勤[1] 陈杰[1] 

机构地区：[1]四川大学电气信息学院,成都610065 [2]四川电力职业技术学院,成都610071

出　　处：《电力系统及其自动化学报》2013年第4期145-149,共5页Proceedings of the CSU-EPSA

摘　　要：为了降低原始负荷数据突变对Elman神经网络预测精度的影响,考虑电网负荷预测样本时变性强、不确定因素影响多的特点,利用Elman神经网络计算和适应时变特性的能力强、误差可控以及灰色理论所需计算数据少、计算量小,在样本较少的情况下也能达到较高预测精度的优点,建立灰色Elman神经网络的负荷预测模型,首次将灰色Elman神经网络模型在中长期负荷预测中应用。实例结果表明,该预测方法提高了预测精度、取得了较快的收敛速度,说明该模型是可行而有效的。In order to reduce the prediction errors of Elman neural network caused by the sudden load change, a pre- diction model of load was established by combining the gray model and the Elman neural network prediction model, considering the characteristics of load prediction sample, which is a periodical manner with many uncertain factors. The Elman neural network has the advantages of high capability to adapt the time-varying characteristic, be error-con- trollable, less data requirement, and easy application of the grey theory. A Gray-Elman network model for prediction of mid-long term load based on the gray theory and neural network is put forward firstly in this paper. This approach was used to test some historical data from a power network. The results are quite satisfying and the model is efficient and accurate.

关 键 词：ELMAN神经网络 灰色理论 中长期负荷 负荷预测 

分 类 号：TM744[电气工程—电力系统及自动化]