基于RVM与ARMA误差校正的短期风速预测  

Short Term Wind Speed Forecasting Based on RVM and ARMA Error Correcting

作　　者：孙国强[1] 卫志农[1] 翟玮星[1] 

机构地区：[1]河海大学能源与电气学院,南京210098

出　　处：《电工技术学报》2012年第8期187-193,共7页Transactions of China Electrotechnical Society

基　　金：国家自然科学基金资助项目(51107032;61104045)

摘　　要：为提高风速预测的准确性,提出了基于相关向量机(RVM)与自回归滑动平均(ARMA)误差校正的风电场短期风速预测算法。该算法首先在RVM的基础上,建立了影响因素与未来24小时风速的非线性模型,并采用遗传算法(GA)进行优化,从而保证了模型参数最优。然后,针对已建立的RVM预测模型的误差序列,采用ARMA模型对其进行拟合,最后用ARMA模型的误差预测值校正已有的风速预测值。本文对江苏某风电场的风速进行预测,算例结果表明该方法是合理有效的。To improve the accuracy of wind speed forecasting,a method based on relevant vector machine（RVM） and auto-regressive moving average（ARMA） error correcting is proposed.Firstly,the nonlinear model of influencing factors and wind speed of the next 24 hours are built based on RVM,and genetic algorithm（GA） is used to ensure the optimization of model parameters.Secondly,the error series of wind speed forecasted by RVM model are adjusted by ARMA model.Finally,the forecasted wind speed is amended by the forecasting errors,which are generated by ARMA model.The wind speed forecasting results of any day in the future for a wind farm in the Jiangsu province demonstrate that the proposed method is reasonable and effective.

关 键 词：风速预测 误差校正 RVM模型 AMRA模型 

分 类 号：TM614[电气工程—电力系统及自动化]