Application of Artificial Neural Network to Predict Wind Speed: Case Study in Duhok City, Iraq

Abstract

Wind speed prediction is very critical for clean energy electricity generation, commitment decision-making, and wind farms planning strategy studies. It is also important for the wind energy industry to determine the characteristics of wind speed for site selection and to know the output of the wind turbine. A prediction of Daily Average Wind Speed (DAWS) for Duhok city, Iraq using Feed Forward (FF) Artificial Neural Network (ANN) is investigated using weather records for Duhok city, Iraq. To build and train the suggested network, MATLAB software is used. The variables that used as inputs are a daily average of the Humidity (H), Pressure (P), Minimum Temperature (Tmin), Solar Radiation (SR), Maximum Temperature (Tmax), day (D) and month (M) to estimate DAWS. The suggested networks are analyzed using Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) as statistical values. The proposed network forecasts accurate daily wind speed values based on the outcomes. This suggested method helps to predict the weather and to estimate the output strength of wind turbines.