Indian Journal of Science and Technology
DOI: 10.17485/ijst/2015/v8i25/87891
Year: 2015, Volume: 8, Issue: 25, Pages: 1-10
Case Report
C. Balakrishna Moorthy1 *, Ankur Agrawal 2 and M. K. Deshmukh1
1 Department of EEE and E&I, BITS Pilani, K. K. Birla Goa Campus, Sancoale - 403726, Goa, India; [email protected], [email protected]
2 Department of Mathematics, BITS Pilani, K. K. Birla Goa Campus, Sancoale - 403726, Goa, India; [email protected]
In this paper, the wind power is predicted using artificial intelligence techniques. The wind speed data as input and wind power data as output, measured at ten minutes interval for a period of eleven months are used for the study. The artificial intelligence techniques such as artificial neural network and genetic algorithm are used for prediction of wind power. The Genetic Algorithm (GA) and Back Propagation Algorithm (BPA) are used as learning algorithm in the artificial neural network. Different parameters such as learning rate, momentum coefficient and epochs are varied in the back propagation algorithm to obtain the best architecture in the neural network. Similarly, crossover fraction and elite count are varied along with number of generations in the genetic algorithm learning technique to select the best model. Furthermore, two hybrid methods such as combination of the two algorithms, namely BPA-GA and GA-BPA are also proposed for prediction of wind power. The performance of the four models are compared in terms of Mean Square Error (MSE). From the results, it is observed that genetic algorithm outperforms the rest in terms of accuracy in prediction.
Keywords: Artificial Neural Network, Back Propagation Algorithm, Genetic Algorithm, Prediction, Wind Power
Subscribe now for latest articles and news.