Simulating wind characteristics through direct optimization procedures: illustration with three Russian sites

Abstract

Wind energy assessment of a territory where a wind park is planned to be built is important. This can be performed through an appropriate evaluation of the wind characteristics in this territory. To simulate the wind speeds, a Weibull function is recommended whose parameters are classically determined either applying logarithms or using one of the formulas proposed in the literature. In the present study, direct optimization procedures are applied, which consist to minimize the squared difference between the experimental and simulated densities or probabilities. These procedures are applied on the wind characteristics collected from the ERA5 website during 41 years at three Russian sites close to Arkhangelsk. These direct optimization procedures are proved to give lower errors than the classical one or the formulas of the literature. They also lead to lower values of the estimated Annual Energy Production for a Vestas V90-2.0 wind turbine. Direct optimization procedures are also applied to determine the optimal parameters associated with a unique or a superposition of two von Mises distribution functions to simulate the wind directions in these three Russian sites.