Statistical Analysis of Wind Speed Data with Weibull, Lognormal and Gamma Distributions

Abstract

In this study, average wind speed data in Bitlis for the years between 2012 and 2016 is analyzed. Average wind speed estimations for these years are obtained with the Weibull, Gamma and Lognormal distributions. Maximum Likelihood method is used in parameter estimation of these distributions. It is aimed that the most fit distribution is determined with Kolmogorov-Smirnov Goodness of Fit test, coefficient of determination and root mean square error criteria. As a result of evaluating the wind speed data with the program written in MATLAB R2009a, it was determined that average wind speed estimations are similar for each distribution, but Gamma distribution has the lowest standard deviation with the average wind speed value in August (0.15 m/s). In modelling of the average wind speed data between 2012 and 2016, it was seen that Gamma distribution had higher coefficient of determination compared to the other distributions. Similarly, the lowest Kolmogorov-Smirnov Goodness of Fit test statistic and root mean square error value are obtained for Gamma distribution. As a result, it is recommended that Gamma distribution is used in modelling the wind speed data of Bitlis between 2012 and 2016.