Evaluating the Effect of Air Mass on I–V Curves for Different Photovoltaic Technologies Based on the NREL Database
Abstract
The IEC 60891 standard, widely used for photovoltaic (PV) module performance modeling, considers only irradiance and temperature as input variables, neglecting spectral effects related to air mass (AM). These effects, however, significantly influence the incident solar spectrum and may cause non-negligible deviations in the electrical response of PV modules, particularly in technologies with selective spectral response such as CIGS and a-Si. In this study, I–V curves from the NREL database were compared under nearly identical irradiance (G) and temperature (T) conditions but acquired at different times of day and year, thereby exposing each technology to distinct spectral distributions. The results reveal measurable differences in electrical parameters (ΔPmp, ΔIsc, ΔImpp, ΔVmpp, ΔVoc, ΔFF) and in the spectral factor (ΔSF). Thin-film technologies (CIGS and a-Si) exhibited the highest sensitivity, with ΔSF exceeding 5% and ΔFF variations up to 3% under low-irradiance conditions, indicating strong spectral dependence. In contrast, crystalline silicon-based modules (x-Si and HIT) showed smaller and more consistent deviations, typically below 2% for ΔSF and 1.5% for ΔFF, confirming their broader and more stable spectral response. These findings demonstrate that conventional transposition procedures based solely on G and T tend to underestimate performance deviations under real spectral conditions. Incorporating an explicit spectral term, through AM, a spectral factor (SF), or spectral reconstruction, can substantially enhance predictive robustness in PV modeling, testing, and field diagnostics.
Keywords
photovoltaic modules, air mass, spectral effects, spectral factor, thin-film technologies, crystalline silicon, PV performance modeling