摘要
短期风速预测对并网风力发电系统的运行有重要意义。该文简述了短期风速预测的价值和方法,分析了小时风速的日变化特点。在此基础上,提出将单变量小时风速时间序列向量化,以消除日周期非平稳,进而建立了向量自回归(vector autoregression,VAR)模型,并用于小时风速预测。算例表明,正常天气条件下,该模型可以预测提前72 h的短期风速。该文提出的方法和模型具有一定的普适性,可用于其它领域的时间序列建模与预测。
Short-term wind speed forecasting is important to the operation of grid-connected wind power generation systems. This paper outlined the value and methods of short-term wind speed forecasting. The characteristic of diurnal hourly wind speed variation has been thoroughly analyzed. A method named as vectorization of univariate hourly wind speed time series has been presented for eliminating diurnal nonstationary, and vectorized hourly wind speed was expressed as a vector autoregression (VAR) model. Finally, the VAR model was applied to forecast hourly wind speed. The results showed that the presented VAR model can yield satisfactory hourly wind speed forecast as long as 72 h ahead under normal weather conditions. The proposed method is suitable for eliminating the periodic nonstationary, and can be used in other time series modeling and forecasting.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2008年第14期112-117,共6页
Proceedings of the CSEE
基金
国家自然科学基金项目(60775047)
国家863高技术基金项目(2007AA04Z244)
高校科技创新工程重大项目(706043)~~
关键词
风力发电
小时风速
向量自回归
预测
wind power generation
hourly wind speed
vector autoregression
forecast
作者简介
孙春顺(1965-),男,湖南东安人,博士研究生,主要研究方向为电力系统运行与控制、风力发电,suncs65@csust.edu.cn;
王耀南(1957-),男,云南昆明人,教授,博士生导师,主要研究方向为人工智能、智能控制、图像处理和模式识别等;
李欣然(1957—),男,湖南娄底人,教授,博士生导师,主要研究方向为电力系统分析与控制、电力系统辨识与建模等。