The power generated by a PV cell depends on both the intensity and spectrum of the light incident on the cell. Silicon solar cells, for example, respond to both visible and near-infrared light, but differently: red light has more impact on PV power output than blue light. Accurately predicting power output requires accounting for the effect of spectrum on the PV cell’s power. However, measuring the entire incident solar spectrum is an expensive and data intensive task.

We explored whether PV performance modeling can be improved by using spectral measurements at only a few wavelengths rather than measuring the full spectrum. The analysis used data collected at Los Alamos and Albuquerque, New Mexico.  When irradiance measurements at two wavelengths were included in the performance models, errors were reduced by up to a factor of two compared to modeling that used air mass as a proxy for the effects of spectrum, as is current practice in the PV modeling community. The important wavelengths were relatively consistent across different cell technologies suggesting that measurements at a few specific spectral wavelengths may improve the accuracy of performance modeling for all types of PV cells.  

The plot shows curves of spectral intensity of incident sunlight at each wavelength, with colors indicating the short circuit current (Isc, an input to PV power models) at the instant each spectral intensity curve was captured. Strong vertical color gradients (dark to light, or light to dark) at a specific wavelength (e.g., 600nm) indicate that wavelength is important for PV performance modeling. Measurements of spectral intensity at only a few of these important wavelengths can help reduce PV performance modeling errors. This analysis will be presented at the Photovoltaic Specialists Conference (PVSC) in June 2016.