Evaluation of Robust Estimation Methods in Estimating Weibull Parameters for Wind Energy Application

Abstract

Two-parameter Weibull distribution has been widely-used reference distribution in wind energy studies and thus its parameter estimation methods have been well-studied in the literature. However, the literature have generally focused on non-robust methods which produce unreliable results in the cases of wind speed data with outliers. In this study, we deal with robust estimation methods of the Weibull distribution for wind energy applications. The considered robust methods are evaluated for both clear and contaminated real wind data cases. It was found that the considered robust methods provides reliable results when it is taken into account in the case of real wind speed data cases. Also, the certain robust methods for the Weibull distribution yield less mean power density error than classical methods in the case of wind speed data with outliers. As a result, it is deduced from analysis that robust methods can be simultaneously used with efficient estimators to check the estimated reliability of the results on wind power.