基于小波分解与支持向量机的风速预测模型  

Wind speed forecasting model based on wavelet decomposition and support vector machine

作　　者：张华[1] 郁永静[1] 冯志军[2] 孙科[1] 

机构地区：[1]华北电力大学可再生能源学院,北京102206 [2]新华水利水电投资公司,北京100053

出　　处：《水力发电学报》2012年第1期208-212,共5页Journal of Hydroelectric Engineering

基　　金：国家自然科学基金项目(50579083)

摘　　要：采用小波分析与支持向量机数据挖掘技术,建立风速预测模型(WD-SVM)。首先对原始风速时间序列进行小波分解,然后对分解后各层数据序列分别用支持向量机进行预测,最后将各层预测值进行小波重构得到预测风速序列。将WD-SVM模型应用于某风电场,进行了风速数据时间间隔为10min的提前4小时,即提前24步滚动预测,将一天分为21个时段进行了预测,预测的均方根误差平均值为11.71%。与仅使用支持向量机进行预测的模型(SVM)进行比较,组合模型预测精度明显高于单一模型。A wind speed prediction model（WD-SVM） that uses data mining techniques of wavelet analysis and support vector machine,was developed.In this model,a given wind speed time series are decomposed by wavelet analysis into various layers that are predicted with support vector machines,and then by reconstructing the predicted values of each layer a prediction of wind speed is obtained.The model was applied to a wind farm for a forecast of its 10-minute average wind speed four hours in advance,i.e.,a forecast 24 steps ahead.By dividing one day into 21 intervals for such a prediction,an average root-mean-square error of 11.71% was resulted.The accuracy of WD-SVM model is much higher than that of support vector machine model（SVM）.

关 键 词：风功率预测 多步预测 支持向量机 小波变换 风电场 

分 类 号：TV11[水利工程—水文学及水资源]