基于网格聚类的峡谷风电短期功率预测方法  

Short term power prediction method of canyon wind power based on grid clustering

作　　者：唐冬来 周强 宋卫平 何鹏[3] 黄璞 叶鸿飞 TANG Donglai;ZHOU Qiang;SONG Weiping;HE Peng;HUANG Pu;YE Hongfei(Aostar Information Technology Co.,Ltd.,Chendu 610041,China;State Grid Information and Communication Industry Group Co.,Ltd.,Beijing 100000,China;Chongqing University of Posts and Telecommunications School of Communication and Information Engineering,Chongqin 400065,China)

机构地区：[1]四川中电启明星信息技术有限公司,四川成都610041 [2]国网信息通信产业集团有限公司,北京100000 [3]重庆邮电大学通信与信息工程学院,重庆400065

出　　处：《供用电》2023年第1期80-87,共8页Distribution & Utilization

基　　金：四川省科技计划项目(2021GFW0021)。

摘　　要：为解决峡谷风电受峡谷尾流影响而造成的风电短期功率预测准确率低的问题,提出了一种基于网格聚类的峡谷风电短期功率预测方法。对峡谷风电站的历史气象、预测功率、实际功率数据进行预处理,以提升预测模型的训练效果。按照经纬度进行峡谷风电站地域网格聚类,获得峡谷风电网格特征。在此基础上,采用长短期记忆网络对每个网格进行功率预测,并进行网格功率叠加与误差修正,峡谷风电站短期功率预测实际准确率达到88.35%,其运行效果表明:所提出的预测方法能够有效地提高峡谷风电短期功率的预测精度。In order to solve the problem of low accuracy of canyon short-term wind power prediction caused by canyon wake,a shortterm wind power prediction method based on grid clustering is proposed.The historical meteorological,predicted power and actual power data of canyon wind power station are preprocessed to improve the training effect of the prediction model.Then,the regional grid clustering of canyon wind power station is carried out according to the longitude and latitude,and the grid characteristics of canyon wind power are obtained.On this basis,the long-term and short-term memory network is used to predict the power of each grid,and the grid power superposition and error correction are carried out.The actual short-term power prediction accuracy of a canyon wind power station is 88.35%.The operation results show that the prediction method proposed in this paper can effectively improve the short-term power prediction accuracy of Canyon wind power.

关 键 词：网格聚类 峡谷风电 短期功率预测 长短期记忆网络 误差修正 

分 类 号：TM614[电气工程—电力系统及自动化]