A sample practice about estimation of wind speed with different methods

Since fossil fuels, which are used primarily in the world to meet the demand for electric energy, generate excessive amounts of greenhouse gases in the production of electricity, people face health problems. Procuring of fossil-based resources will be more expensive as their reserves will decrease and deplete in the next years Renewable energy sources are clean, harmless to the environment and do not require any raw material for production. Wind power generation systems, which come to the forefront in the production of renewable energy sources, gain importance when our country's existing wind energy potential is taken into consideration. on. Although electricity generation from wind energy has increased, it is still not seen as a safe energy source for the electrical grid. Since wind is a volatile (unstable and erratic) resource, it is difficult to predict. Wind generation estimates are needed to ensure that the generated energy does not have grid adaptation problems and to benefit from the generated energy effectively. In this study, the future wind speed estimation has been tried to be made by using eleven-year wind and meteorological values between 2000 and 2010 for the Mardin Region by using artificial neural networks, radial based neural networks, recurrent neural networks and regression analysis methodsThe methods we used were compared by finding the best Mean Absolute Percent Error values. As a result, when we compare the Regression and Recurrent Neural Networks according to the annual Mean Absolute Percent Errorvalues, it is seen that this method gives better results than other methods