Wind Speed Prediction Based on Statistical and Deep Learning Models

Wind is a dominant source of renewable energy with high sustainability potential. However, the intermittence and unstable nature wind affect efficiency reliability conversion systems. The prediction available potential also heavily flawed by its nature. Thus, evaluating trough speed prevision, crucial for adapting production to load shifting user demand rates. This work aims forecast using statistical Seasonal Auto-Regressive Integrated Moving Average (SARIMA) model Deep Neural Network Long Short-Term Memory (LSTM). In order shed light on these methods, comparative analysis conducted select most appropriate prediction. errors metrics, mean square error (MSE), root (RMSE), absolute (MAE), percentage (MAPE) are used evaluate effectiveness each best model. Overall, obtained results showed that LSTM model, compared SARIMA, has shown leading performance an average 14.05%.