A PRELIMINARY ASSESSMENT OF A DETAILED 2 STREAM SHORT-WAVE NARROW-BAND MODEL USING SPECTRAL RADIATION MEASUREMENTS

Citation
Ve. Cachorro et al., A PRELIMINARY ASSESSMENT OF A DETAILED 2 STREAM SHORT-WAVE NARROW-BAND MODEL USING SPECTRAL RADIATION MEASUREMENTS, Solar energy, 61(4), 1997, pp. 265-273
Citations number
42
Journal title
ISSN journal
0038092X
Volume
61
Issue
4
Year of publication
1997
Pages
265 - 273
Database
ISI
SICI code
0038-092X(1997)61:4<265:APAOAD>2.0.ZU;2-1
Abstract
A data bank of measurements of global, direct and diffuse solar spectr al irradiances at ground level for clear skies has been compiled for V alencia (Spain) dating back to December 1992. The measurements were ma de with a commercial Li-cor 1800 spectroradiometer with a range of 300 -1100 nm and a spectral resolution of 6 nm. A preliminary comparative assessment has been carried out between the experimental data and mode l data. The chosen model was a detailed narrow-band model (208 spectra l intervals from 0.2 to 4 mu m) developed at the ''Laboratoire d'Optiq ue Atmospherique (LOA)'' of the University of Lille (France). This pla ne-parallel multilayer model uses a two-flux method to solve the radia tive transfer equation and an exponential sum-fitting procedure to sol ve the absorption-scattering problem for finite spectral intervals. Fo r this first comparative assessment we focused our attention on the ca pability of the LOA model to predict irradiance data (direct, global a nd diffuse) using four values of visibility (40, 23, 14 and 5 km) for two aerosol models (maritime and continental) in the boundary layer. T hese first results show the low sensitivity of global irradiance to di fferent turbidity conditions. Conversely, the spectral direct and diff use irradiances were highly influenced by the chosen aerosol model tak ing into account the visibility values. The spectral distribution of p redicted global and direct irradiances are in relatively good agreemen t with the observed values. The diffuse data show larger discrepancies , which are in part due to the nature of the measurement process itsel f. However, the observed differences can be partially explained by tak ing into account the associated errors of the measured data, the elaps ed time between the measured spectra and the prediction power of the L OA model. (C) 1997 Elsevier Science Ltd.