Wind regimes representation modeling by using multivariable distributions

Abstract

The electricity produced from one wind turbine is almost entirely dependent on the intensity of the upcoming wind. On the other hand, while making analysis of electricity production of a single wind field, equal attention should be paid to wind direction, since it defines the overall wind flow in the wind field. Accordingly, when making estimations of the expected annual production of energy from one wind field, the correlations of wind speed and direction should be appropriately taken into account. The usual practice of using Weibull distributions by a number of sectors is not the most suitable for optimization processes for defining the layout of one wind filed. For this purpose, the paper proposes a method that is simple and easy to implement, which is based on a multivariable distribution function of the parameters of the wind. For the analysis of the proposed model the measured wind data used in this paper are from one existing wind field. In addition, three stage functional distributions are obtained, which satisfactorily match the measurement data. In order to determine the matching degree of the adapted multivariable distributions with the measurement data, the quantitative parameter for error estimation - R2 was used. Various sizes of data sampling intervals are also examined in order to reduce the generation time of multiple distributions, which will not affect the quality of matching between generated distributions and measured data.