基于时间序列—状态转移的短期风速预测  

Short Term Wind Speed Predication Based on Time Series-State Transition

作　　者：申永涛[1] 刘宗歧[1] 杨铎[1] 

机构地区：[1]华北电力大学电气与电子工程学院,北京102206

出　　处：《陕西电力》2013年第11期14-16,20,共4页Shanxi Electric Power

基　　金：国家高技术研究发展计划(863项目)(2012AA050201)

摘　　要：风速预测对风电场和含风电的电力系统的运行都具有重要意义。准确的预测风速可以有效地减轻或避免风电场对电力系统的不利影响。目前已经发展了多种风速预测方法,但由于作为预测对象风速受到多种因素影响而具有较大的波动性,导致这些算法预测精度不高。为有效提高预测精度,在进行时间序列预测的基础上,提出一种基于时间序列-状态转移模型的预测方法,即在时间序列预测风速的基础上,利用状态转移模型对传统时间序列预测结果进行修正。利用此预测方法对一个已有的风速预测结果进行了修正,通过误差对比,证明了此方法的应用价值。Wind speed forecasting is of great significance on improving the running stability of wind farm and grid interconnected with wind farm and reducing the adverse effect of the wind farm on the power system.There are a variety of methods for wind speed prediction currently,but because of wind＇ s big fluctuating amplitude,the prediction is not of high accuracy yet.Basing on the idea of improving the accuracy,this paper puts forward a time series-state transition model that makes use of state transition model to correct the results of time series forecasting,and corrects the existing result of wind speed forecasting by using the model.The prediction error comparison shows that the method mentioned above is of application value.

关 键 词：风速预测 误差 时间序列 时间序列-状态转移 

分 类 号：TM614[电气工程—电力系统及自动化]