Evaluation of Parametric Statistical Models for Wind Speed Probability Density Estimation

Abstract

An accurate statistical estimation of wind speed probability density at a given site is crucial when making power network planning decisions involving wind generation resources. The use of parametric probability density functions, such as the Rayleigh, Weibull and Gaussian distributions, can be problematic as it can lead to model mis-specification at a given site. In this paper, the use of the Gaussian Mixture Model (GMM) to estimate wind speed variability is investigated and compared with the above three popular parametric models using wind speed data for six sites in northwest Europe. Results show that the GMM produces the lowest error values with the highest percentage improvements, and is the only model that consistently fails to reject the null hypothesis when conducting the K-S goodness-of-fit test.