Artificial Neural Network-Based Forecasting of Wind Speed and Wind Turbine Power: A Case Study in Balıkesir, Türkiye

In this study, artificial neural network (ANN)-based models were developed for wind speed forecasting and wind turbine power forecasting. The analysis was based on 10-minute measurement data collected from a wind power plant in Balıkesir, Türkiye, between January 1, 2023 and January 1, 2024. The models were implemented in MATLAB using the nntool interface, and 70% of the dataset was used for training, 15% for validation, and 15% for testing. First, wind speed was forecast using a nonlinear autoregressive (NAR)-based ANN with 10 delay values. Then, wind turbine power was forecast using a nonlinear autoregressive with exogenous input (NARX)-based ANN with 6 delay values under two input conditions, namely actual wind speed data and predicted wind speed data. This two-stage structure allowed a direct comparison of forecasting performance. For the representative months, monthly average deviations did not exceed 0.67% for the model based on actual wind speed data and 0.12% for the model based on predicted wind speed data. These results indicate that the proposed two-stage framework can provide wind turbine power predictions with low monthly average deviations by using predicted wind speed data under the investigated conditions.