Estimation of wind speeds of Fethiye district with artificial neural networks

Today, meeting the energy need has become one of the biggest problems of the countries. To resolve this,Turkey like the countries of the world is rapidly increasing day by day the use of renewable energy sources rather than fossil fuels. The fact that these resources are clean, environmentally friendly and inexhaustible has made their use more important. In this direction, wind energy potential is of great importance. However, geographical differences and temporal variability make wind power difficult for use. Therefore, wind power and speed prediction planning was needed. In this study, wind speed estimation of Fethiye district of Muğla province was made. For this purpose, the average hourly wind speed data, obtained from the metereology station of Muğla State General Directorate of Metereology Affairs located in Fethiye district, covering the years 2011-2013, was obtained from 10 m height and Artificial Neural Networks (ANN) model was established for estimation. The ANN model was created with a 3-layer feed-forward, back propagation Levenberg-Marquard learning algorithm. 80% of the data used in the calculation was used for education and 20% for testing. Within the scope of the thesis, wind speed estimation was made for 2013 year and for 4 months (March, June, September, December), representing 4 seasons. The performance of the results obtained was evaluated and it was determined that the model created was in harmony with the actual data.