摘要
提出了一种基于邻近电站和综合相似日的BP网络光伏输出功率异常数据修复方法。考虑了地理位置、温度以及1日类型等影响出力的因素,通过皮尔逊积距法选取与待修复电站功率相关度高的邻近电站,综合使用灰色关联度和曲线相似度来分析相似日,找出与待修复日相符的周边电站的综合相似日数据,然后建立BP网络模型,用自适应调节学习速率的方法修复不良数据。对青海地区实际光伏预测功率中的异常数据进行修复的结果表明,该方法有较高的修复精度。
This paper presents a repairing method for abnormal data of BP network PV output power based on adjacent power plants and integrated similar days.Factors such as geographical position,temperature and day type which influence power generation are considered and Pearson product-moment method is used to select the adjacent power plant which has high power correlation with the plant to be repaired,and the combination of Grey Relational Analysis with curve similarity is adopted to select similar days and find out the integrated similar day data of the adjacent power plant that is in conformity with the day to be repaired,and then the corresponding BP neural network model is built,and the diverse learning speed algorithm is employed to repair the abnormal data.The result of abnormal data repairing in the actual PV power prediction in Qinghai indicates that the proposed method has better repairing accuracy.
作者
郭辉
杨国清
姚李孝
张舒捷
GUO Hui;YANG Guoqing;YAO Lixiao;ZHANG Shujie(College of Water Resource and Hydro-Electric Engineering,Xi’an University of Technology,Xi’an 710048,Shaanxi,China;Electric Power Research Institute,State Grid Qinghai Electric Power Company,Xining 810000,Qinghai,China)
出处
《电网与清洁能源》
2018年第9期52-58,共7页
Power System and Clean Energy
基金
国家电网公司科技项目:光伏电站采集数据分析治理及发电计划优化研究(NYB11201704444)~~
关键词
不良数据
邻近电站
综合相似日
功率相关性
abnormal data
adjacent power plant
integrated similar days
power correlations