基于ARMA模型和BP神经网络组合优化算法的风电预测模型  

Wind Power Prediction Model Based on the Combined Optimization Algorithm of ARMA Model and BP Neural Networks

作　　者：曾鸣[1] 李树雷[1] 王良[1] 薛松[1] 王睿淳[1] 

机构地区：[1]华北电力大学能源与电力经济研究咨询中心,北京102206

出　　处：《华东电力》2013年第2期347-352,共6页East China Electric Power

基　　金：国家自然科学基金(71271082);国家软科学研究计划(2012GXS4B064);美国能源基金会支持项目(G-1006-12630)~~

摘　　要：从风电场营销可持续发展的角度出发,提出了风电场发电量组合预测的研究角度;选取了两种比较具有代表性的预测模型自回归移动平均和神经网络,以发电场实测数据为模型输入,分别得出各自的预测结果;最后,将两种单一预测模型与组合预测模型结果进行比较,验证了组合预测应用于风电场发电量预测领域的可行性。研究结果表明,该组合预测方法相对于单一预测模型能够提高预测精度。In light of the sustainable development of the wind farm marketing, the combined forecast of wind farm power generation is proposed. Two representative prediction models, based on ARMA and on neural networks respec- tively, were selected; the farm measured data were input to the models and their own prediction results were obtained respectively. Finally, the prediction results of the combined model were compared with those of the two individual models, verifying the feasibility of the combined model to wind farm generation capacity forecast. It is concluded that the combined forecast model can predict more accurately than the individual forecast model.

关 键 词：组合预测 自回归移动平均 神经网络 风电场 发电量预测 

分 类 号：TM743[电气工程—电力系统及自动化] TM614