Wind turbine power curve modeling using maximum likelihood estimation method. (June 2019)

• Record Type: Journal Article
• Authors: Seo, Seokho; Oh, Si-Doek; Kwak, Ho-Young
• Is Part Of: Renewable energy. Volume 136(2019)
• Page Start: 1164
• Page End: 1169
• Publication Date: 2019-06
• Subjects: Logistic function -- Wind turbine power curve -- Weibull distribution -- Maximum likelihood estimation method; Renewable energy sources -- Periodicals; Power resources -- Periodicals; Énergies renouvelables -- Périodiques; Ressources énergétiques -- Périodiques; 333.794
• DOI: 10.1016/j.renene.2018.09.087
• Languages: English
• ISSNs: 0960-1481

Abstract

Abstract: Modeling of wind turbine power curve which shows the relationship between wind speed and its power output can be used as an important tool in monitoring and forecasting wind energy. A data-driven approach to find most probable probability distribution function (PDF) for wind speed and turbine power is presented in this study. Equations for the scale and shape parameters in the Weibull wind speed distribution and equations for the four parameters in the logistic function were obtained explicitly by maximum likelihood estimation (MLE) method. With help of a selected data set from the wind speed and the corresponding power output data which was collected over a period of a year, the values of the parameters were obtained by solving the equations by iteration procedures. The predicted powers by the obtained logistic function closely follow the measured turbine powers averaged at 5-min or 10-min. Monitoring turbine power output by the logistic function was also tested for the measured powers in other time duration. Highlights: Modeling of power curve for wind turbine was constructed by data-driven approach. Equations for the four parameters of a logistic function were obtained by maximum likelihood estimation (MLE) method. Most probable probability distribution function (PDF) for wind turbine power was obtained by solving the equations. The predicted power by logistic function closely follows the measured power averaged at 5-min or 10-min.