STATISTICAL COMPARISONS OF CLEARANCE-INDEX-BASED DIFFUSE SOLAR RADIATION CORRELATIONS IN ALJOUF, SAUDI ARABIA

摘要

Both global and solar radiation data are necessary for many applications, including the design of solar energy systems. In comparison to global radiation data, diffuse radiation data are, however, rarely available. Several empirical models have been developed to calculate diffuse radiation using various parameters, such as the clearance index, sunshine duration and the atmospheric water content. In this study, we tested the performance of 22 empirical correlations suggested by various authors for estimating monthly mean daily diffuse radiation from the clearance index, using parameter values measured at Aljouf, Saudi Arabia. These correlations parameterize the diffuse radiation component either linearly (Page type) or by second, third (Liu and Jordan type) or fourth degree polynomials of the clearance index. We compared the models on the basis of statistical errors, such as the mean bias error (MBE), mean percentage error (MPE), root mean square error (RMSE) and t-statistics. Using the above statistical parameters, we considered the overall performances of the models in order to select the best one. The correlations suggested by Bahl, Singh and Bhati, and Benganam and Jorid showed the best performance. The Bahl model, using a fourth degree polynomial of the clearance index, showed the best results with MPE ＝ 0.96%, RMSE ＝ 0.062 and MBE ＝ 0.002. This was followed by the model of Singh and Bhati, with MPE ＝ -1.4%, RMSE ＝ 0.08 and MBE ＝ -0.003, and then by the model of Benganam and Jorid, which was developed for a neighboring site (Madina, Saudi Arabia), with MPE = 2.38%, RMSE ＝ 0.06 and MBE ＝ -0.004. On the other hand, the models suggested by Tuller and Rao showed the worst performance, with MPE ＝ 57.42%, RMSE ＝ 0.15 and MBE ＝ 0.15 for the former and MPE ＝ 41.98%, RMSE ＝ 0.141 and MBE = 0.123 for the latter.