Temperature based radiation models for the estimation of global solar radiation at horizontal surface in India

Abstract

Background/Objectives: The objective of this study is to compare five different temperature based empirical models and to select the most accurate model to estimate the monthly average Global Solar Radiation (GSR) in India. Methods/Analysis: Five empirical equations namely Hargreaves and Samani, Bristow and Campbell, Pandey and Katiyar First, second and third order models have been employed to estimate monthly average GSR on horizontal surface using minimum and maximum temperature. Using these equations GSR is estimated at New Delhi (Latitude 28.61° N, Longitude 77.20° E) and Chennai (Latitude 13.08° N, Longitude 80.27° E), India and the meteorological data for this work have been obtained from India Meteorological Department (IMD), Pune from 2002-2012. Findings: Solar radiation data are not easily available in all locations due to higher cost and difficulty in measurement. It can be estimated by empirical equations using meteorological parameters like sunshine hour, temperature and relative humidity, out of which temperature is the most commonly available meteorological data. Empirical coefficients appeared in correlation equations based on temperature have been found using the latest computing MATLAB software. Estimated GSR values were compared with measured values based on statistical measures such as Root Mean Square Error (RMSE) and correlation coefficient (R). Comparing five equations, it is found that third order correlation provides good results with R = 0.9882 and RMSE = 0.840. From the results, it is concluded that the temperature based models has good potential for estimating GSR for any locations where measurement of sunshine duration data are not possible. Improvements: Further, in future, these empirical equations and temperature based ANN models would be applied for different places having good solar potential such as Hyderabad, Bhubaneswar and other states of India would be reported.