Forecasting of Wind Turbine Output Power with Artificial Neural Network in Izmir, Türkiye

This study focuses on one of the most common issues in wind energy, which is the forecasting of wind turbine output power. Using wind speed and turbine output power data collected from a wind turbine in Seferihisar, Izmir at 10-minute intervals over a year, an artificial neural network (ANN) model was trained in MATLAB. In the model, while the wind speed and turbine output power are taken as inputs, the output layer forecasts the wind turbine's output power. With the ANN model, short, medium, and long-term forecasts of the wind turbine's output power were forecasted. The forecasting results of the ANN model exhibit high correlation with the measured values, with regression values for long-term predictions on a monthly basis and medium-term predictions on a daily basis being, respectively, around R=0.97 and R=0.98. These results emphasize that artificial intelligence-based methods can be used as a reliable tool in wind energy forecasts.