On the Evaluation of Wind Characteristics in a Low-Wind Location Utilizing Weibull Distribution: Case Study of the Lviv Airport (Lviv, Ukraine)

Abstract

In a most common case, weather data for wind power analysis at a location is analyzed to obtain (1) most probable wind speed (critical for wind turbine selection) and (2) wind power density (critical for wind power yield analysis). 30- minutes 2000-2017 dataset collected at the Lviv airport (Lviv, Ukraine), was analyzed using Weibull probability distribution that is most commonly utilized for the wind speed representation. The dataset is an example of a low-wind location; in particular, the dataset contains a couple of years characterized by high zero-wind-speed probability (up to 0.46 in 2006). Weibull parameters were estimated utilizing a variety of the most commonly used methods, as well as Matlab built-in tools for Weibull and Rayleigh (a partial case of Weibull) distributions. The analysis has shown that the maximum likelihood method and Matlab Weibull estimation provide the worst representation of wind characteristics at the location, while other methods provide similar acceptable results if the possibility of zero winds is low. The energy pattern factor method provides the best wind power density estimation. Most probable wind speed for high-zero-wind-speed-probability years is best estimated by Rayleigh distribution. It has been demonstrated that interannual variation of the most probable wind speed $(1.5-2m/s)$ is insignificant in terms of a wind turbine selection. On the other hand, annual wind power density varies from 8.8 $W/m^{2}$ in 2006 to 15 $W/m^{2}$ in 2002; therefore, one-year or three-year period is insufficient for wind power yield analysis.