基于经验模态分解和支持向量机的短期风电功率组合预测模型  

Combined Model Based on EMD-SVM for Short-term Wind Power Prediction

作　　者：叶林[1] 刘鹏[1] 

机构地区：[1]中国农业大学信息与电气工程学院,北京市海淀区100083

出　　处：《中国电机工程学报》2011年第31期102-108,共7页Proceedings of the CSEE

基　　金：国家自然科学基金项目(51077126;51174290);教育部科学技术重点研究项目(109017);教育部新世纪优秀人才支持计划(NCET-08-0543);北京市自然科学基金项目(3113029)~~

摘　　要：针对风速序列随时间、空间呈现非平稳性变化的特征,提出一种基于经验模态分解（empirical mode decomposition,EMD）和支持向量机（support vector machine,SVM）的EMD-SVM短期风电功率组合预测方法。该方法首先利用EMD将风速序列分解为一系列相对平稳的分量,以减少不同特征信息间的相互影响;然后利用SVM法对各分量建立预测模型,针对各序列自身特点选择不同的核函数和相关参数来处理各组不同数据,以提高单个模型预测精度。最后将风速预测结果叠加并输入功率转化曲线以得到风电功率预测结果。研究结果表明,EMD-SVM组合预测模型能更好地跟踪风电功率的变化,其预测误差比单一统计模型降低了5%~10%,有效地提高了短期风电功率预测的精度。A wind power prediction method based on empirical mode decomposition（EMD） and support vector machine（SVM） is proposed to treat with the nonlinearity and nonstationarity of wind speed data.Firstly,the wind speed data is decomposed into a series of components with stationarity by using EMD to reduce the influence between different feature information.Then,different models were built and different kernel functions and parameters were chosen to deal with each group of data by using SVM in order to improve the forecasting accuracy.Finally,short term wind power forecasting was made based on wind speed data through a practical wind power curve.Case study was carried out to investigate the validity of the novel algorithm and the results illustrated that the forecasting error of EMD-SVM combined model decreased by 5%~10% compared to single statistics model.The proposed combined model can improve the short term forecasting accuracy of wind power effectively by tracking the change of wind power.

关 键 词：经验模态分解 支持向量机 风速 短期风电功率预测 组合预测模型 

分 类 号：TM71[电气工程—电力系统及自动化]