Análise comparativa da eficiência de sete métodos numéricos para determinação dos parâmetros da curva de Weibull utilizando dados de velocidade do vento e de potência eólica

Abstract

To determine the wind potential of a region is of paramount importance that a study be conducted on the intrinsic characteristics of the wind site. This paper address some numerical methods to be used in calculating the parameters of the Weibull distribution the aids in the study of wind speed, so there is a correct definition of natural conditions existing. The previous knowledge of this information assist in the decision-making process on the technical feasibility of installing industrial wind farms. It perform a statistical analysis of seven mathematical systems known in the literature for estimating the parameters ( k ) form and ( c ) scale of the frequency distribution curve of Weibull. Data are used for wind speed and wind power from two coastal cities of Ceará State belonging to Northeast Brazil, Icapuí and Camocim. The methods to be employed in the development of this research are: Graphical Method, Maximum Likelihood Method, Maximum Likelihood Modified Method, Empirical Method, Moment Method, Energy Pattern Factor Method and the Equivalent Energy Method. The realization of the comparative analysis of efficiency and accuracy among these include the application of the following statistical tests: ANOVA(R2 ) , Square Root of Average Quadratic Errors (RMSE) and Chi-square (X 2 ).