Wind speed forecasting model for northern-western region of India using decision tree and multilayer perceptron neural network approach

Abstract

Power production by wind energy with the increase in renewable energy sources, plays an important role in India due to its critical location. In this paper, using the input variables like latitude, longitude, cooling design temperature, relative humidity, air temperature, atmospheric pressure, daily solar radiation – horizontal, Earth temperature amplitude, Earth temperature, heating degree-days, cooling degree-days, elevation, heating design temperature, frost days at site, monthly wind power density and air density, wind speed is predicted by multilayer perceptron in 17 cities of India. The varying number of hidden neurons helps in calculation of accurate forecasting. It is found that prediction accuracy is highest for six hidden neurons in training and testing phase which is 99.14% and 96.116%, respectively.