Estimating solar radiation on tilted surfaces with various orientations : a study case in Karaj (Iran)

• Published in: Theoretical and applied climatology 2006-03, Vol.84 (4), p.235-241
• Main Authors: KAMALI, Gh. A, MORADI, I, KHALILI, A
• Format: Article
• Language: English
• Subjects: Earth, ocean, space; Exact sciences and technology; External geophysics; Light; Meteorology; Radiation; Radiative transfer. Solar radiation; Solar radiation; Sun
• ISSN: 0177-798X; 1434-4483
• DOI: 10.1007/s00704-005-0171-y

The aim of this study is the evaluation of models that estimate daily global solar radiation on tilted surfaces from that measured on horizontal surfaces. Global solar radiation incident on a tilted plane consists of three components: beam radiation, diffuse radiation and reflected radiation from the ground. The Klein (1977) method, modified by Andersen (1980), was used for estimating direct solar radiation incident on tilted surfaces and an isotropic model was used for estimating reflected solar radiation incident on a tilted plane. In contrast models for the diffuse radiation component show major differences, which justifies a validation study which has been done. Eight models for derivation of daily slope diffuse irradiance from daily horizontal diffuse irradiance were tested against recorded slope irradiances at Karaj (35 degrees 55[variant prime] N; 50 degrees 56[variant prime] E), Iran. The following models were included: Badescu (2002), Tian et al. (2001), Reindl et al. (1990), Skartveit and Olseth (1986), Koronakis (1986), Steven and Unsworth (1980), Hay (1979) and Liu and Jordan (1962). All the models use the same method for calculating beam radiation as well as ground reflected radiation. However, only diffuse component of radiation was compared. Statistical indices showed that Reindl's model gives the most accurate prediction for the south-facing surface and Koronakis's model performs best for the west-facing surface. The Relative Root Mean Square Errors (%RMSE), except for Steven and Unsworth's model that has unacceptable results, for whole data range from 1.02 to 10.42%. In general, Reindl's model produces the best agreement with the measured tilted data. [PUBLICATION ABSTRACT]