Short-term wind speed prediction based on artificial neural network models
Citation
Kirbas, I., Kerem, A., (2016). Short-term wind speed prediction based on artificial neural network models. Measurement and Control, 49(6), 183-190. DOI: 10.1177/0020294016656891Abstract
Wind energy has an important place in renewable energy sources. Biggest challenges in wind energy production are the variability of the wind and difficulty of estimation of true wind speed. In this study, 25,777 records have been taken from wind measurements carried out at Mehmet Akif Ersoy University campus. Records include meteorological data such as wind speed at different heights/altitudes, wind direction, temperature, pressure and humidity. In order to estimate the wind speed that may occur at 61m altitude, multilayer perceptron and radial basis function methods have been used. During the application phase, 100 artificial neural networks were trained and performance evaluations of these networks were done. The obtained results show that the wind speed at 61m can be estimated with 99% accuracy using artificial neural network when other meteorological data are taken as input.