基于多元信息融合的风电功率预测模型研究  

Research on Wind Power Prediction Model Based on Multivariate Information Fusion

作　　者：董鹏辉 崔世文 苏正雄 张齐 郭彦飞[2] DONG Penghui;CUI Shiwen;SU Zhengxiong;ZHANG Qi;GUO Yanfei(Qinghai Branch of China Three Gorges Renewables(Group)Co.,Ltd.,Xining 810000,China;Shenzhen Branch of NARI Technology Co.,Ltd.,Shenzhen 518000,China)

机构地区：[1]中国三峡新能源(集团)股份有限公司青海分公司,青海,西宁810000 [2]国电南瑞科技股份有限公司深圳分公司,广东,深圳518000

出　　处：《微型电脑应用》2025年第2期153-156,共4页Microcomputer Applications

摘　　要：随着风力发电的并网规模越来越大,风电功率的准确预测对风电的优化控制和电力系统的安全经济运行至关重要,但风电功率值为典型的随机变量,它具有很强的间歇性和波动性。针对风电功率难于准确预测的问题,提出基于多元信息融合的风电功率预测模型,利用滑动平均法对原始风电场和相邻风电场的功率时间序列进行分解后,结合物理数值天气预报(NWP)预测的数据构成多元信息,并运用快速相关过滤(FCBF)算法对输入特征量进行筛选,再由自适应模糊神经网络实现特征量与风电功率的非线性映射。通过风电功率预测算例的比较分析,结果验证了所提风电功率预测模型的有效性和优越性。As the increasing scale of wind power connected to the grid,the accurate prediction of wind power becomes very important for optimal control of wind power and safe economic operation of power system,but wind power is a typical random variable,and it is highly intermittent and volatile.In view of the problem that wind power is difficult to predict accurately,this paper proposes a wind power prediction model based on multivariate information fusion.After decomposing the power time series of the original wind farm and adjacent wind farms by the moving average method,combining with data predicted physical numerical weather prediction to constitute multivariate information,and using fast correlation-based filter(FCBF)algorithm to filter the input feature quantity.The nonlinear mapping between feature quantity and wind power is realized by adaptive fuzzy neural network.Through comparison and analysis of wind power prediction examples,the results verify the effectiveness and superiority of the proposed wind power prediction model.

关 键 词：风电功率预测 多元信息融合 快速相关过滤法 滑动平均法 自适应模糊神经网络 

分 类 号：TM721[电气工程—电力系统及自动化]