SPECTRALLY RESOLVED SOLAR IRRADIANCE FROM METEOSAT CLOUD -METHODS AND VALIDATION
Jethro Betcke, Tanja Behrendt, Jan Kühnert, Annette Hammer, Elke Lorenz, Detlev Heinemann Carl von Ossietzky University Oldenburg, Energy and Semiconductor Laboratory, Energy Meteorology Group
The varying spectral distribution of the terrestrial irradiance can cause significant differences between the actual and the rated efficiency of solar cells. To account for this effect it is necessary to calculate the weighted irradiance which depends on the solar cell spectral response and on the spectral distribution of the irradiance. Because ground based spectral measurements are sparse, satellite based spectral data can provide a good alternative. In this work we have compared two approaches. The SOLIS method can calculate both broadband and spectrally resolved terrestrial solar irradiance from satellite data. The clear sky module of SOLIS is based on a partial parametriation of radiative transfer calculations, and uses atmospheric data from climatologies or satellite based methods. The cloud module is based on empirical relationships between the cloud albedo values determined from Meteosat visual channels and the cloud transmittance. In parallel we developed a simpler empirical method, that can be more easily implemented in photovoltaic modeling software. It is based on principal component analysis of measured spectrally resolved irradiance and empirical relationships between the principal components and satellite derived clear sky index and sun elevation. Both methods have been validated using the same ground measured data from Stuttgart Germany. The weighted irradiance has been used as metrics to compare the results of the models. It was found that the use of both models gave a significant improvement compared to the use of a constant AM1.5 spectrum. The use of the empirical model resulted in lower bias values, the use of SOLIS resulted in lower RMSE values. Apparatus used StellarNet BLACK-Comet, RED-Wave InGaS.
