Md.F. Islam and A.M.T. Oo (Australia)
Artificial Intelligence, Forecasting, Wind, Energy
This paper presents the experimental results and analysis of artificial neural network (ANN) models to forecast wind speed for wind turbine generation. A modified cascade correlated (MCC) training algorithm was developed for forecasting wind speeds and its performance is compared with those of the existing well established backpropagation with momentum (BPM) and backpropagaion with Bayesian regularization (BR) training algorithms. Results are analysed in the standardized methodology of prediction accuracy to have a clear idea about the model skills. It shows that MCC model performs better with respect to the BPM and BR for the wind speed forecasting in this event of three hourly prediction spheres.
Important Links:
Go Back
