Wind Park Power Forecasting Models and Comparison

Author

Zhang, Qian ; Lai, Kin Keung ; Niu, Dongxiao ; Wang, Qiang

Author_Institution

Sch. of Econ. & Manage., North China Electr. Power Univ. Baoding, Baoding, China

fYear

2012

fDate

23-26 June 2012

Firstpage

27

Lastpage

31

Abstract

In this paper, four new forecasting models - univariate LS-SVM model and three hybrid models of ARIMA and LS-SVM models are introduced for wind power output forecasting. Historical data of 78 wind farms are used to compare and evaluate the performance of the best models. Empirical analysis indicates that the proposed univariate LS-SVM model and hybrid models can not significantly outperform linear models but they are not inferior to linear models.

Keywords

forecasting theory; least squares approximations; power engineering computing; support vector machines; wind power plants; ARIMA; LS-SVM models; hybrid models; linear models; univariate LS-SVM model; wind farms; wind park power forecasting models; wind power output forecasting; Accuracy; Analytical models; Biological system modeling; Forecasting; Predictive models; Wind forecasting; Wind power generation; ARIMA; LS-SVM; hybrid models; wind power forecasting;

fLanguage

English

Publisher

ieee

Conference_Titel

Computational Sciences and Optimization (CSO), 2012 Fifth International Joint Conference on

Conference_Location

Harbin

Print_ISBN

978-1-4673-1365-0

Type

conf

DOI

10.1109/CSO.2012.15

Filename

6274671