计及风速相关性的风电功率预测方法研究  

Research on Wind Power Prediction Method by Considering Wind Speed Correlation

作　　者：刘显茁 邓韦斯 谢恩彦 LIU Xianzhuo;DENG Weisi;XIE Enyan(Power Dispatching Control Center,China Southern Power Grid Co.,Ltd.,Guangzhou 510623,China;NARI-TECH Nanjing Control Systems Co.,Ltd.,Nanjing 215200,China)

机构地区：[1]中国南方电网有限责任公司,电力调度控制中心,广东广州510623 [2]国电南瑞南京控制系统有限公司,江苏南京215200

出　　处：《微型电脑应用》2024年第5期81-84,共4页Microcomputer Applications

基　　金：中国南方电网有限责任公司科技项目(ZDKJXM20210051)。

摘　　要：风电功率的准确预测对于风电场的运行控制及电力系统的安全稳定具有重要的意义,但风电功率具有典型的随机波动特性,风电功率预测存在较大的难度。针对风电功率预测准确率不理想的现状,提出一种计及风速相关性的风电功率预测模型,利用风速时空相关特性建立风电场间的功率相关性模型,由集合经验模态法和游程法求得不同变化趋势的模态分量,利用自回归条件异方差模型获得预测功率。通过风电功率预测实例的对比分析,表明考虑风速相关性可使预测误差降低35.67%,而自回归条件异方差模型可使预测误差进一步降低5.26个百分点,因此所提出的风电功率预测模型具有很好的的优越性。The accurate prediction of wind power is of great significance to the operation control of wind farm and the safety and stability of power system,but the wind power has typical random fluctuation characteristics,and it is difficult to forecast wind power.In view of the unsatisfactory prediction accuracy of wind power,a wind power prediction model that takes into account the correlation of wind speed is proposed,the power correlation model between wind farms is established based on the temporal and spatial correlation characteristics of wind speed,the modal components of different variation trends are obtained by empirical mode method and run-length method,and the predictive power is obtained by autoregressive conditional heteroscedasticity model.Through comparative analysis of wind power prediction examples,the results show that considering the wind speed correlation can reduce the prediction error by 35.67%,the autoregressive conditional heteroscedastic model can further reduce the prediction error by 5.26%,the wind power prediction model has good advantages.

关 键 词：风力发电 功率预测 风速相关性 集合经验模态分解 自回归条件异方差模型 

分 类 号：TM721[电气工程—电力系统及自动化]