Wind power prediction using Artificial Neural Network

This project reports the application of Artificial Neural Network (ANN) in wind power prediction based on historical meteorological data. ANN which is inspired by the functional aspects of biological neural networks is employed in this project due to its strong pattern recognition capabilities and its ability to model flexible linear or non-linear relationship among variables. A three layers feed-forward back-propagation neural network has been developed to predict the wind power for the next hour. In order to get an accurate wind power prediction, several network structures, training algorithms and transfer functions have been developed and tested with different sets of data. The performance of a network will be determined by its convergence capability, and only the network with the best performance will be selected. As a result, an ANN with the regression value of 0.81881 was developed, which has the ability to predict the wind power for the next hour with 81.881 % accuracy.