Spectral Corrections Based on Air Mass, Aerosol Optical Depth, and Precipitable Water for PV Performance Modeling

Broadband incident global irradiance must be corrected when modeling the performance of a photovoltaic (PV) module to account for the impact of spectrum. However, most existing correction methods exhibit both climate and site dependencies. This paper presents a new set of analytical equations aimed at modeling the spectral effects exhibited by six PV single-junction materials under all operating conditions. Thus, the equations proposed here involve the three parameters that mostly influence the spectrum shape: air mass (AM), aerosol optical depth, and precipitable water. The extraction of the equations and their coefficients is then presented. Next, the method is empirically validated over the course of a 12-mo experimental campaign. Thus, coefficients of determination ranging from 0.87 (multicrystalline silicon) to 0.92 (amorphous silicon) prove that modeled values of the spectral mismatch factor (MM) are well aligned with experimental ones. In addition to the values of mean bias error that are virtually equal to zero, this new procedure also yields values of root mean square error that stay below 0.85% when modeling MM for all the technologies under investigation. Such figures are far better than those achieved by using the popular Sandia Labs method, based solely on AM.