集中式新能源功率预测系统设计与实现  

DESIGN AND APPLICATION OF CENTRALIZED NEW ENERGY POWER PREDICTION SYSTEM

作　　者：袁兴德 曾垂宽 彭喆 杨东升 梁卉林 Yuan Xingde;Zeng Chuikuan;Peng Zhe;Yang Dongsheng;Liang Huilin(New Energy Technology Department,China Resources Power Technology Research Institute Co.,Ltd,Shenzhen 518000,China)

机构地区：[1]华润电力技术研究院有限公司新能源技术部,深圳518000

出　　处：《太阳能》2025年第1期60-69,共10页Solar Energy

基　　金：华润电力科技研发项目(CRP-R&D-LX-2023-003)。

摘　　要：新能源电站功率预测的稳定性、准确性直接影响电网的安全调度、场站的生产运营等,为解决行业痛点,设计了针对风电和光伏发电的集中式新能源功率预测系统。首先构建系统架构,其具有云端、集控端、场站端应用层级,3个安全区的网络环境且采用双链路传输机制的网络架构,集中式应用架构,以及微服务技术架构;然后开展了软件开发技术选型,设计了高可用方案和埋点监控方案;最后提出了5种核心算法。应用效果表明:该功率预测系统运行稳定,具有良好的经济价值。某省6个风电场实际应用该预测系统后,短期、超短期功率预测准确率分别提高了6.2%、2.2%,且数据上报率达100%。The stability and accuracy of power forecasting in new energy power stations directly impact the safe scheduling of the power grid and the production and operations of the stations.To address industry challenges,a centralized new energy power prediction system targeting wind power and PV power generation has been designed in this paper.Initially,the system architecture is constructed,which includes application levels with cloud end,centralized control end,and station end,a network environment with three security zones,and using duallink transmission mechanism of network architecture,centralized application architecture,and microservice technology architecture.Subsequently,software development technical selection is carried out,studying highly available solutions and buried point monitoring solutions is designed.Finally,five core algorithms are proposed.The application effect shows that the power prediction system operates stably and has good economic value.After the practical application of the prediction system in six wind farms in a certain province,the short-term and ultrashort-term power prediction improved by 6.2%and 2.2%respectively,and the data reporting rate reached 100%.

关 键 词：集中式 新能源功率预测 风电 光伏发电 双链路 高可用性 埋点监控 预测算法 

分 类 号：TM614[电气工程—电力系统及自动化] TM615