Non-linear regression model for wind turbine power curve. (December 2017)

• Record Type: Journal Article
• Authors: Marčiukaitis, Mantas; Žutautaitė, Inga; Martišauskas, Linas; Jokšas, Benas; Gecevičius, Giedrius; Sfetsos, Athanasios
• Is Part Of: Renewable energy. Volume 113(2017)
• Page Start: 732
• Page End: 741
• Publication Date: 2017-12
• Subjects: Wind energy -- Wind power curve -- Non-linear regression -- Cross-validation; Renewable energy sources -- Periodicals; Power resources -- Periodicals; Énergies renouvelables -- Périodiques; Ressources énergétiques -- Périodiques; 333.794
• DOI: 10.1016/j.renene.2017.06.039
• Languages: English
• ISSNs: 0960-1481

Abstract

Abstract: In this article, a study of wind turbine power curve modelling is presented with application to a particular wind turbine of Seirijai wind farm in Lithuania. A non-linear regression model for wind turbine power curve approximation was proposed, which stands out with several advantages, such as fitting physical properties of wind turbine (i.e., power curve does not exceed the highest value of generated power as it is maximum physically possible), lower number of parameters to be estimated, dependency on only one factor. MAPE was used as a measure of approximation method accuracy. Mode approach was introduced as an alternative to typical techniques for modelling power curves of wind turbines with the aim to avoid elimination of the outliers from initial data and the impact of varying concentration of observations in the full range of wind speed. Performed cross-validation analysis demonstrated that the developed power curve model is appropriate for the prediction of wind power and is not directly dependent on the initial data set. Highlights: A new non-linear regression model for WTPC is proposed. Mode approach is introduced as an alternative to typical techniques for WTPC. Cross-validation analysis demonstrates model appropriateness for any data set. Study results contribute to elaboration of more precise wind power prediction methods.