Determination of Appropriate Distribution Functions for the Wind Speed Data Using the R Language

Year 2019, Volume: 3 Issue: 2, 131 - 138, 10.10.2019

İsmail Kırbaş

Abstract

Accurate determination of the proper distribution and parameters of this distribution according to the wind characteristics of the zone is vital for wind energy investment.  In determining a wind energy potential belonging to a region, meteorological wind speed measurements have a great proposition to take place within a certain statistical distribution. In our study, the wind speed data obtained from the metrology station within 1 year was evaluated and it was determined using the R language, which is an open source statistical programming language, which is better suited to distributions such as Weibull, gamma, lognormal and logistic. The Akaike Information Criterion and Schwarz-Bayesian Information Criterion (SBIC) scores were calculated as the performance parameters of the distributions and the distribution performances were compared graphically. While gamma and lognormal distributions have better results at low wind speeds, Weibull distribution achieves higher performance for higher wind speeds.

Keywords

gamma, logistic, lognormal, R language, statistical distribution, Weibull, wind speed

References

• [1] I. Kirbas and A. Kerem, “Short-Term Wind Speed Prediction Based on Artificial Neural Network Models,” Meas. Control, vol. 49, no. 6, pp. 183–190, 2016. [2] A. Kerem, I. Kirbas, and A. Saygın, “Performance Analysis of Time Series Forecasting Models for Short Term Wind Speed Prediction,” presented at the International Conference on Engineering and Natural Sciences (ICENS), 2016, pp. 2733–2739. [3] P. Bhattacharya and R. Bhattacharjee, “A Study On Weibull Dıstrıbutıon For Estimating The Parameters,” J. Appl. Quant. Methods, vol. 5, no. 2, pp. 234–241, 2010. [4] M. Kurban, Y. M. Kantar, and F. O. Hocaoğlu, “Weibull Dağılımı Kullanılarak Rüzgar Hız ve Güç Yoğunluklarının İstatistiksel Analizi,” Afyon Kocatepe Univ. J. Sci., vol. 7, no. 2, pp. 205–218. [5] W.-Y. Chang, “A Literature Review of Wind Forecasting Methods,” J. Power Energy Eng., vol. 2, no. 4, pp. 161–168, 2014. [6] T. P. Chang, “Estimation of wind energy potential using different probability density functions,” Appl. Energy, vol. 88, no. 5, pp. 1848–1856, 2011. [7] A. F. Özdemir, E. Yıldıztepe, and M. Binar, “İstatistiksel Yazılım Geliştirme Ortamı: R,” presented at the XII. Akademik Bilişim Konferansı, Muğla, 2010, vol. 1, pp. 375–379. [8] R Core Team, R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing, 2014. [9] P. Dalgaard, Introductory statistics with R. Springer Science & Business Media, 2008. [10] İ. Kırbaş, “Wind Speed Distribution Dataset and R Source Codes,” GitHub Page, 23-Apr-2017. [Online]. Available: https://github.com/ismkir/windSpeedDistribution/. [Accessed: 24-Apr-2017]. [11] O. Elitok, “Weibull Distributions and Its Applications,” M. Sc. Thesis, Kırıkkale University Institute of Science and Technology, Kırıkkale, 2006. [12] D. Indhumathy, C. V. Seshaiah, and K. Sukkiramathi, “Estimation of Weibull Parameters for Wind speed calculation at Kanyakumari in India,” Int. J. Innov. Res. Sci. Eng. Technol., vol. 3, no. 1, pp. 8340–8345, Jan. 2014. [13] A.-A. Bromideh, “Discriminating Between Weibull and Log-Normal Distributions Based on Kullback-Leibler Divergence,” Istanb. Ünversitesi İktisat Fakültesi Ekonom. Ve Istat. Derg., no. 16, pp. 44–54, 2012.