摘要
在光伏发电功率预测中,地表太阳辐射量的准确获取十分重要,卫星遥感能够定量获得区域气象和环境信息,为此提出结合卫星遥感数据的光伏发电功率预测方法。文章基于高时间分辨率卫星数据,采用深度学习方法,获取大范围内的地表太阳辐射数据,进而结合光伏发电功率数据和气象数据,进行光伏发电功率预测。结果表明,采用文章提出的预测方法,光伏发电功率4h预测平均功率精度为R2=0.92,均方根误差ERMSE=0.99MW,24h预测平均功率精度为R2=0.78,ERMSE=0.60MW。文章研究可为大范围光伏功率预测提供参考。
In the photovoltaic power prediction, it is very important to accurately acquire surface solar radiation data. Satellite remote sensing can quantitatively obtain regional meteorological and environmental information. Therefore, a photovoltaic power prediction method combined with satellite remote sensing data is proposed. Based on the satellite data with high temporal resolution, a wide range of surface solar radiation data is obtained by the in-depth learning method, and then combined with the photovoltaic power data and meteorological data to predict the photovoltaic power. The results show that the average power accuracies of the prediction method proposed in this paper are R^(2)=0.92, E_(RMSE)=0.99 MW, and R^(2)=0.78, E_(RMSE)=0.60 mw for 4-hour and 24-hour prediction of photovoltaic power respectively. This study can provide a reference for large-scale photovoltaic power prediction.
作者
刘周斌
徐崇斌
王鑫磊
陈前
左欣
吴俣
徐丹露
LIU Zhoubin;XU Chongbin;WANG Xinlei;CHEN Qian;ZUO Xin;WU Yu;XU Danlu(Center of Mass Entrepreneurship and Innovation State Grid Zhejiang Electric Power Co.,Ltd.,Hangzhou 310051,China;Beijing Institute of Space Mechanics&Electricity,Beijing 100094,China;Beijing Aerospace Innovative Intelligence Science and Technology Co.,Ltd.,Beijing 100076,China;Beijing Engineering Technology Research Center of Aerial Intelligence Remote Sensing Equipments,Beijing 100094,China;Institute of Aerospace Information Innovation,Chinese Academy of Sciences,Beijing 100094,China)
出处
《航天返回与遥感》
CSCD
北大核心
2021年第5期85-95,共11页
Spacecraft Recovery & Remote Sensing
基金
国家电网有限公司总部管理双创孵化培育基金项目(SGZJSC00XMJS2000027)。
关键词
太阳辐射
光伏发电功率
短期预测
深度学习
航天遥感应用
solar radiation
photovoltaic power generation
short-term prediction
deep learning
space remote sensing application
