풍력에너지 추정을 위한 와이블 분포의 적합도 평가

This paper presents a model selecting procedure for proper probability distribution of wind speed. In order to assess the goodness-of-fit of various Weibull distributions, this study introduced AIC(Akaike Information criterion) and BIC(Bayesian Information criterion) in addition to the conventional K-S test and χ2 test. The results show that 3-parameters Weibull is good for modeling ordinary wind data but mixture Weilbull reveals better performance at probability model with two peaks.