作者: S.A. Akdağ , H.S. Bagiorgas , G. Mihalakakou
DOI: 10.1016/J.APENERGY.2010.02.033
关键词: Wind speed 、 Mathematics 、 Coefficient of determination 、 Frequency distribution 、 Probability distribution 、 Probability density function 、 Wind power 、 Weibull distribution 、 Mean squared error 、 Statistics
摘要: Abstract The statistical characteristics of wind speed data recorded at nine buoys, located in Ionian and Aegean Sea (Eastern Mediterranean) are analyzed this paper, order to present a more accurate method for estimation characteristics, according the suitability probability distribution functions (pdf). This article has focussed on regimes that nearly zero percentages null speeds. selected distributions examination typical two-parameter Weibull (W-pdf) two-component mixture (WW-pdf), involving five parameters (two shape parameters, two scale one proportionality parameter). Suitable software, based maximum likelihood method, is used estimate aforementioned two-parameters W-pdf mixed WW-pdf. judged from coefficient determination (R2) fit standard error (RMSE) tests, which had been carried out between each theoretical corresponding experimental cumulative frequencies sites. From these tests it clear that, most cases (six stations - having either unimodal or bimodal frequency distributions), mixed-Weibull provides highest degree fit. In other three cases, mixing weight p density function equals (p = 0), so transformed Simple-Weibull distribution. Hence, general conclusion suitable description such conditions could offer less relative errors determining annual mean power density.