Global Solar Radiation Estimation Modeling Using Artificial Neural Network: A Case Study on Metro Cities of India

摘要

The environmental constraints and limited availability of traditional energy resources have made the twenty-first century for the optimization of renewable energy resources. Solar energy is solely related to the quantity of solar radiation to be received by solar panel/devices. Optimization of solar energy is best possible when solar radiation is estimated well before. To overcome the availability of solar radiation measuring devices at the location of interest, solar radiation estimation models are developed. In the present study, four metro cities (Bombay-Colaba, Calcutta-Alipore, Madras-Meenambakkam, and New Delhi-Safdarjung) of India have been selected. The data are downloaded from CROPWAT 8.0. The solar radiation is taken as output whereas latitude, longitude, altitude, months of a year, maximum temperature, minimum temperature, humidity, wind velocity, and the sunshine hour are considered as input. Simulation is executed with MATLAB R2016a with MLP and LM algorithm. The proposed model shows an overall regression value of 0.99178, and RMSE for training is 0.0961, for validation 0.3102, and for testing 0.5727.