摘要
提出了一种基于多分辨率分析下的短期风速预测方法。利用小波分解将原始风速序列分解成低频信号分量和高频信号分量,将低频信号分量作为时间序列模型的输入,将高频信号分量作为最小二乘支持向量机的输入输出未来时间段的各分量预测值。最后将各分量的预测值重构为风速序列的预测值。以内蒙古风电场为例进行仿真,结果表明文中方法显著提高了超前风速预测的精度。
A short-term wind speed forecasting method based on multi-frequency analysis is proposed.The original wind speed data is decomposed into low frequency signal component and high frequency signal components by wavelet decomposition.Taking the lowfrequency components as time series model input and high-frequency components as LS-SVM model input,then the components of future wind speed data is output.Finally the prediction of wind speed by re-constructing the components is acquired.Serving wind farms in Mongolia region as an example,the simulation is carried out.The results show that the proposed method can improve the accuracy of one-step ahead wind speed forecasting.
出处
《陕西电力》
2014年第1期27-31,36,共6页
Shanxi Electric Power
基金
国家自然科学基金资助项目(50973002)
关键词
短期风速预测
多分辨率分析
时间序列
小波分解
最小二乘支持向量机
short-term wind speed forecasting
multi-resolution analysis
time series method
wavelet decomposition
least square support vector machine
