Jun Liu, Guojiang Xiong, Xiaofan Fu, Ali Wagdy Mohamed
Weibull distribution is extensively used for wind resource assessment in fitting wind speed frequency distribution because of its reliability and simplicity. Its precision is highly dependent on the parameter estimation methods and different methods may result in varied results. This paper implements ten numerical methods to estimate Weibull's parameters in fitting wind speeds of nine years from January 1, 2015 to December 31, 2022 at nine locations in Guizhou province, China. Firstly, the distribution characteristics including average value, skewness, kurtosis, and direction of wind speeds at these nine locations are analyzed. Secondly, to ascertain the optimal parameters, ten numerical methods are implemented in accordance with the Weibull distribution. Four goodness-of-fit indicators and the Friedman test are used to reveal the difference of their performance. Finally, with the optimal parameters, the wind power density is determined in every location. The findings indicate that the empirical method of Lysen performs the best. The study can provide empirical data analysis and detailed descriptions of wind speed characteristics for local wind energy development and resource assessment. Besides, it can also help clarify the strengths and weaknesses of different methods in practical applications and provide a reliable basis for evaluating the potential of wind energy.