Statistical Approach to select the Best Suitable Solar Model for Global Radiation: Case Study of Karachi, Pakistan

Utilization of solar energy for the fulfillment of daily routine tasks has gained a lot of interest in the recent past. The complete utilization can be accomplished only if proper solar data is available for a particular site. The measurement of solar data is expensive and time consuming. Different empirical correlations are available for the estimation of solar radiations on a particular site. In this research, monthly average daily solar radiation on horizontal surface is compared with measured data available as typical metrological year, TMY2, and estimated from the different solar models available for Karachi. The models studied are Ahmad models, Raja Models, Hargreaves Model and Black model. The estimation of different empirical correlations and NASA are compared with the measured values. The comparison was done by using statistical techniques like MAPE, MBE, RMSE and t-stats. The results revealed that Hargreaves model is the best model to estimate global radiation for Karachi with minimum value of t-stats while Raja 3 model is the most inappropriate model with highest value of MAPE, RMSE, MBE and t-stat. Overall, the already established empirical correlations can be used for the estimation of solar energy for Karachi for complete utilization of solar energy with minimum errors.