Comparison of Seven Numerical Methods for Estimating Weibull Parameters for Wind Energy Applications

Weibull distribution is widely used for the purpose of wind energy estimation. Their parameters should be precisely estimated since the wind energy is affected by it. The aim of this paper is to compare seven numerical methods to find out which is the most efficient for determining the parameters of Weibull distribution based on wind speed data, collected in Zuwara, Libya during 2007 at three hub heights of 10 m, 30 m, and 50 m above ground, by recording the wind speed every 10 minutes. The selected methods used in this study included graphical method, standard deviation method, empirical method of Justus, empirical method of Lysen, energy pattern factor method, maximum likelihood method, and modified maximum likelihood method. The performance of the seven numerical methods is evaluated by using different statistical criteria including mean absolute percentage error, mean absolute bias error, root mean square error, and correlation coefficient. The presented results indicated if Weibull distribution matches well with observed wind speed data, the empirical methods of Justus and Lysen present favorable efficiency; but if not, maximum likelihood gives the best performance followed by empirical methods of Justus and Lysen. The graphical method shows weak performance.