基于NARX神经网络的智能电网短期负荷预测  

Short-term load forecasting of smart grid based on NARX neural network

作　　者：王雪菲 李勇[2] 丁柏宏 程晨[1] 杨辉军[1] WANG Xue-fei;LI Yong;DING Bai-hong;CHENG Chen;YANG Hui-jun(Anhui Institute of International Business,Anhui Hefei 231131,China;Institute of Statistics and Applied Mathematics,Anhui University of Finance&Economics,Anhui Bengbu,233030,China;School of Electronics and Information Engineering,Anhui University,Anhui Hefei 230601,China)

机构地区：[1]安徽国际商务职业学院,合肥231131 [2]安徽财经大学统计与应用数学学院,安徽蚌埠233030 [3]安徽大学电子信息工程学院,合肥230601

出　　处：《齐齐哈尔大学学报（自然科学版）》2020年第6期26-33,共8页Journal of Qiqihar University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61672032);安徽省科技重大专项(18030901018)。

摘　　要：为了进一步完善预测模型,将实时电价、天气和历史负荷数据作为新的参考因子加入到对负荷波动规律特性的分析中,从而提出了基于NARX神经网络的智能电网短期负荷预测模型。该模型首先对负荷波动规律特性进行分析,并采用两种相关性分析方法综合考虑实时电价、天气等因素对于负荷波动的影响,确定了预测模型的输入特征。其次,由于前馈型神经网络在处理负荷序列间关联信息方面存在缺陷,本文建立的预测模型以NARX神经网络为基础进行训练。在完成训练后,输入两组真实数据集,进行最终数据分析比对,结果表明,新建立的模型在实时电价下的短期负荷预测方面具有明显优势,取得了良好的预测精度。In view of such situation,a short-term load forecasting model of smart grid based on NARX neural network is proposed,which considers the effects of weather,historical load and real-time electricity price on the basis of analyzing the characteristics of load fluctuation.Firstly,the model analyses the characteristics of load fluctuation and uses two correlation analysis methods to comprehensively consider the influence of real-time electricity price,weather and other factors on load fluctuation,and determine the input features of the prediction model.Secondly,the feedforward neural network can not effectively deal with the correlation information between load series,In view of this problem,NARX neural network is used to establish the prediction model.Finally,two sets of real data sets are used for example analysis,which proves that the model proposed in this paper has higher accuracy and can effectively solve the short-term load forecasting problem under real-time electricity price.

关 键 词：智能电网 实时电价 负荷预测 NARX神经网络 输入特征 

分 类 号：TM715[电气工程—电力系统及自动化] TP183[自动化与计算机技术—控制理论与控制工程]