期刊论文

Prediction Model for Short-term Wind Speed Based on Improved EMD and RBFNN

中文

广东电力

1007-290X

2016

29

4

34-38+44

10.3969/j.issn.1007-290X.2016.04.006

本校为第一机构

为提高短期风速预测精度,提出改进经验模态分解法（empirical mode decomposition,EMD）与径向基函数神经网络（radial basis function neural network,RBFNN）相结合的短期风速预测模型。首先,利用极值点对称延拓法对预处理过的风速序列进行处理,以抑制传统EMD在分解过程中所引起的边缘效应,并引用分段三次埃米特插值法解决传统EMD包络线的过冲或欠冲问题;然后,利用改进EMD将风速序列分解成各本征模态（intrinsic mode function,IMF）分量,再针对各分量分别构建各自的RBFNN模型进行预测;最后,将各分量的预测结果进行重构、叠加,得到最终的原始风速预测值。实验结果表明,改进的EMD-RBFNN预测模型...

In order to improve precision of prediction on short-term wind speed,a prediction model for short-term wind speed combining improved empirical model decomposition（ EMD） and radial basis function neural network（ RBFNN） is proposed. Firstly,extreme point symmetric extension is used for processing on preprocessed wind speed sequences so as to restrain fringe effect in decomposition caused by traditional EMD,and piecewise cubic Hermite interpolation method is used to solve overshoot or undershoot of traditional EMD envelope lines. Then,improved...