How Important Is The Resolution Of Atmospheric Data In Calculations Of Spectral Irradiance And Energy Yield For (III–V) Triple-Junction Cells?

We use the SMARTS 2.9.5 model to simulate hourly spectral irradiance over a 1 year period at two different geographic locations. Hourly, daily, monthly and yearly averages of cloud‐screened atmospheric data from the AERONET network were used to generate 4 sets of spectra for each location. A detailed balance model was then used to investigate how differences in the temporal resolution of atmospheric inputs affect calculations of the energy harvest efficiency of a specific III–V multi‐junction cell design. The calculations were repeated using a single yearly averaged spectrum. The modelling results show that changing the time resolution of atmospheric data inputs to the SMARTS model does not cause large changes in calculated energy harvest efficiency for the specific cell design used. This suggests that in the absence of high‐resolution atmospheric data, efficiency predictions utilizing spectra derived from yearly or monthly averages will yield similar results to those using daily or hourly averaged data. Use of a single yearly averaged spectrum did however result in an increase of nearly 2% in modelled energy harvest efficiency meaning that temporally varying spectra should be used in calculations of energy harvest efficiency and III–V multi‐junction tuning rather than annual averages.