Deterministic and probabilistic analysis of different empirical models to estimate monthly mean diffuse solar radiation for composite climatic region of India

This research compares the performance of the models to estimate monthly mean diffuse solar radiation (DSR) on a horizontal surface for composite climatic region of India. The goal is to identify the most accurate model using deterministic and probabilistic analysis through two levels of uncertainty in daily sunshine period to examine the accuracy of the model that will be used to estimate DSR in the region under consideration. The meteorological data were collected from Indian Meteorological Department for the city of New Delhi (28 degrees 34 ' N, 77 degrees 12 ' E) which comes under composite climatic region prescribed by the Energy Conservation Building Code for India. From the literature, 150 typical models were chosen and categorized into three categories correlating diffuse fraction to sky-clearness index and relative sunshine period. The models are statistically evaluated using well-known statistical indicators in a unique way. In addition, the global performance indicator (GPI) is calculated using scaled values of indicators. The GPI of the chosen models ranged from -2.2912 to 0.2584, with the greatest value indicating the best model. Following that, the models are then ranked in decreasing order of GPI. Finally, the performances of the models are also checked for different locations having similar climatic conditions. Thus, results of this work are useful for impoverished countries as well as remote areas having similar environment conditions.