Generalized Regression Neural Network Based Wind Speed Prediction Model for Western Region of India

With the growing demand of power generated by wind energy, prediction of wind speed has become an important region for research. In this paper, wind speed is predicted using Generalized Regression Neural Network (GRNN) and Multi-layer perceptron (MLP) in 67 cities of India. The input variables used are: Longitude, Latitude, daily solar radiation- horizontal, air temperature, relative humidity, earth temperature, elevation, cooling degree-days, heating degree-days, atmospheric pressure. The MSE of the two models are compared and found that GRNN gives better result than MLP. The accuracy of GRNN and MLP are 99.99% and 97.974% for training phase and 98.85% and 95.23% for testing phase respectively.