摘要
准确预测光伏发电功率对电网调度具有十分重要的意义。该文提出一种基于灰色关联分析和GeoMAN模型的光伏发电功率短期预测方法。首先,利用灰色关联分析对某地区多光伏电站进行空间相关性分析,选取与待预测光伏电站高度相关的周边电站;然后,基于GeoMAN模型动态提取待预测光伏电站的时空特征和外部气象因素,GeoMAN模型采用编解码结构,利用编码器动态提取待预测光伏电站的站内特征和与周边相关电站的站间空间特征,利用解码器提取输入变量的时间特性,并融合晴空指数和数值天气预报动态输出光伏发电预测功率;最后,采用实际光伏电站进行案例分析,结果表明该文所提出的预测方法与传统LSTM模型相比,实现了更高精度的光伏发电功率短期预测。
Accurate forecast of photovoltaic(PV)power is important for power system dispatch.A new short-term PV power forecasting method is proposed based on grey relational analysis and GeoMAN model.Firstly,grey relational analysis is utilized to analyze spatial correlation among multiple PV stations.And several surrounding PV stations highly related with the target PV station are selected.Then,GeoMAN model is established for dynamically extracting the spatiotemporal feature and external meteorological factors.GeoMAN model adopts encoder and decoder structure.The encoder is utilized to dynamically extract the intra-station feature of the target station and the inter-station spatial feature with the related stations.The decoder is utilized to extract the time feature of input variables.Clearness index and numerical weather prediction(NWP)are finally integrated for short-term PV power forecast.A case study is conducted using data collected from practical PV stations.Study results indicate that the proposed method can achieve higher accuracy compared with long short-term memory(LSTM)model.
作者
时珉
许可
王珏
尹瑞
张沛
Shi Min;Xu Ke;Wang Jue;Yin Rui;Zhang Pei(State Grid Hebei Electric Power Co.Ltd,Shijiazhuang 050021 China;Computer Network Information Center Chinese Academy of Sciences,Beijing 100190 China;University of Chinese Academy of Sciences,Beijing 100040 China;Tianjin Hongyuan Smart Energy Co.Ltd,Tianjin 300000 China)
出处
《电工技术学报》
EI
CSCD
北大核心
2021年第11期2298-2305,共8页
Transactions of China Electrotechnical Society
基金
国家电网河北电力有限公司科技项目资助(kj2019-077)。
