唐雅洁,龚迪阳,倪筹帷,王波,张雪松,朱耿
TANG Yajie,GONG Diyang,NI Chouwei,WANG Bo,ZHANG Xuesong,ZHU Geng

• 1:国网浙江省电力有限公司电力科学研究院
• 2:国网浙江省电力有限公司宁波供电公司

摘要(Abstract)：

在全球碳中和目标下,光伏等清洁零碳新能源成为碳减排的关键,分布式光伏也成为该行业发展的重要方向。由于分布式光伏出力具有较强的随机性和波动性,精准功率预测是其参与电网调度运行的基础。在当前分布式光伏所获取信息有限的情况下,提出了一种基于邻域前向时序最优组合的分布式光伏超短期功率预测模型,考虑邻近区域分布式光伏站点间实时出力关联性,挖掘生成区域性多气象状态时空耦合预测场景。模型扩展有效信息维度,在不依赖于地表气象观测装置的情况下进行高效训练建模,具有经济、轻量与易部署的特点,提升了有限信息下的分布式光伏超短期功率预测精度。采用某区域分布式光伏系统实采数据进行算例分析,分别与基于功率时序推移、基于多站点相似的经典分布式光伏超短期功率预测模型相比较,结果表明所提模型具有更高的预测精度。
In the context of global carbon neutrality, redoubling the development of PV and other clean zero carbon new energies is the master key to carbon emission reduction, and distributed PV now has become the course of PV industry. Distributed PV features output power randomness and fluctuation, precise power forecasting technology is the base of its participation in power grid dispatching and operation. In the context of distributed PV generations with limited acquired information, this paper proposes an ultra-short-term PV power forecasting model based on neighboring optimal forward sequential combination considering the relevance of real-time power outputs in distributed PV stations in neighboring areas and mining and generating regional multi-meteorological and spatio-temporal coupling forecast scene. The model expands the effective dimension of acquired information, provides an efficient training method without the dependency of land meteorological observative service. Meanwhile, it is also economical and light and can be easily deployed, thus increasing the accuracy of ultra short-time distributed PV power forecasting with limited information. Through experimental verification of the measured data of regional distributed PV generations in a district, Zhejiang Province,the proposed model, compared to typical models based on times series and similar multi-stations forecasting,is more accurate in ultra-short-term distributed PV power forecasting.

关键词(KeyWords)： 超短期功率预测;分布式光伏;邻域;时间序列;前向选择;组合模型
ultra short-term power forecasting;distributed PV generation;neighboring area;time series;forward selection;combined model

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金资助项目（51807026）

作者(Author): 唐雅洁,龚迪阳,倪筹帷,王波,张雪松,朱耿
TANG Yajie,GONG Diyang,NI Chouwei,WANG Bo,ZHANG Xuesong,ZHU Geng

DOI: 10.19585/j.zjdl.202110012

参考文献(References)：

• [1]卓振宇，张宁，谢小荣，等.高比例可再生能源电力系统关键技术及发展挑战[J].电力系统自动化，2021,45(9):171-191.
• [2]赵伟鹏.分布式光伏发电与区域电网间的协调发展[J].电气时代，2021(4):20-21.
• [3]国家能源局.光伏发电功率预测系统功能规范：NB/T32031—2016[S].北京：新华出版社，2016.
• [4]车泉辉，吴耀武，祝志刚，等.基于碳交易的含大规模光伏发电系统复合储能优化调度[J].电力系统自动化，2019,43(3):76-82.
• [5]张江林，高红均，王家怡，等.计及预测精度及拓扑结构的光伏电站储能经济配置[J].电力自动化设备，2019,39(6):115-121.
• [6]王晶，黄越辉，李驰，等.考虑空间相关性和天气类型划分的多光伏电站时间序列建模方法[J].电网技术，2020,44(4):1376-1384.
• [7]陈志宝，丁杰，周海，等.地基云图结合径向基函数人工神经网络的光伏功率超短期预测模型[J].中国电机工程学报，2015,35(3):561-567.
• [8]刘卫亮，刘长良，林永君，等.计及雾霾影响因素的光伏发电超短期功率预测[J].中国电机工程学报，2018,38(14):4086-4095.
• [9]胥芳，童建军，蔡世波，等.面向分布式光伏超短期功率预测的云团特征建模[J].太阳能学报，2016,37(7):1748-1755.
• [10]王粟，江鑫，曾亮，等.基于VMD-DESN-MSGP模型的超短期光伏功率预测[J].电网技术，2020,44(3):917-926.
• [11]王昕，黄柯，郑益慧，等.基于PNN/PCA/SS-SVR的光伏发电功率短期预测方法[J].电力系统自动化，2016,40(17):156-162.
• [12]唐雅洁，林达，倪筹帷，等.基于XGBoost的双层协同实时校正超短期光伏预测[J].电力系统自动化，2021,45(7):18-27.
• [13]国家市场监督管理总局，国家标准化管理委员会.分布式光伏发电系统集中运维技术规范：GB/T 38946—2020[S].北京：中国标准出版社，2020.
• [14]YAO S Y,PAN L,YU Z B,et al.Hierarchically non-continuous regression prediction for short-term photovoltaic power output[C]//2019 IEEE 16th International Conference on Networking,Sensing and Control(ICNSC).Banff,AB,Canada:IEEE,2019:379-384.
• [15]彭曙蓉，郑国栋，黄士峻，等.基于XGBoost算法融合多特征短期光伏发电量预测[J].电测与仪表，2020,57(24):76-83.
• [16]甄皓.有限信息下基于深度学习模型的小型分布式光伏电站功率预测[J].上海节能，2020(4):302-308.
• [17]OKADA S,OHZEKI M,TAGUCHI S.Efficient partition of integer optimization problems with one-hot encoding[J].Scientific Reports,2019,9(1):13036.