Abstract
The paper includes two parts. The first part is the analysis of the operating data of wind turbines in Penghu from 2003 to 2005. The second part is the simulation of the aerodynamic characteristics of a wind turbine using STAR-CD computational fluid dynamic (CFD) code. In the first part of analysis, the operating data of Penghu wind turbines are analyzed using Excel. The relations of power, wind velocity, and rotational speed of wind turbines are displayed graphically. From these relations, the performance of these turbines in a particular year can be observed qualitatively. From the results of the analysis of first part, based on the statistical chi-square test, it can be concluded that neither the Weibull nor the Gamma probability mode can be used to represent the annular distribution of wind velocity. Nevertheless, the Weibull probability provides a better fit of the wind velocity of whole year, spring, autumn and winter. The Gamma distribution gives a better result for the wind velocity distribution during summer. We also find that the power generation of the wind turbines can be reasonably predicted when the wind velocity is represented by the Weibull probability model. Based on the Weibull probability distribution of wind velocity, and the characteristics of these wind turbines, it can be concluded that part of the wind energy of the site is wasted. The percentage of wind power that is wasted is calculated. Based on the Weibull probability model of wind distribution, we can define a factor to represent the variation of wind energy in different year. Based on the new factor, we can have a better assessment of the mechanical performance of the wind turbines. In the second part of the analysis, the aerodynamic behavior of NACA-4412 wind blade is simulated using SRAR-CD code. The results are displayed as the relation between the power vs wind velocity under different rotational speed of turbine, and power vs wind velocity with a specific rotational speed. Two different turbulent models are used in the study. These two models are K-Epsilon and Spalart-Allmaras turburlence model. The results of these two models are very similar. The wind power predicted by Spalart-Allmaras turburlence model is more than the power as predicted by K-Epsilon turburlence model. The study has demonstrated that the STAR-CD CFD code can be used to simulate the aerodynamic behavior of wind turbine.