Indoor validation of a multiwavelength measurement approach to estimate soiling losses in photovoltaic modules

Soiling is a factor that impacts the performance of photovoltaic (PV) modules. Nowadays, the research related to PV soiling monitoring is focused on optical sensors, which estimate the soiling loss through a monochromatic transmittance or reflectance measurement. However, these typically neglect the spectral profile of soiling transmittance, which tends to absorb shorter wavelengths more than the longer ones. This leads to a spectral red shift of the light that is transmitted to the PV cells of a module. Therefore, if the spectral component of soiling is not considered, the estimated soiling losses are not fully representative of those occurring in the real PV modules. This investigation aims to address this issue by modeling the full soiling transmittance spectrum using several monochromatic light sources in a new version of a previously presented optical soiling sensor, called "DUSST ". Four different combinations of mono-wavelength light-emitting diodes have been used to model the full spectral transmittance profile of artificially soiled PV glass coupons and to estimate the electrical losses of distinct PV technologies. The results show that the errors in soiling estimation can be minimized by using an appropriate wavelength combination. The difference between the measured and the estimated soiling losses can be lower than 3% if the most convenient wavelength combination is utilized. In the case of m-Si, which is the prevalent PV technology nowadays, the application of the optimum wavelength combination is found to reduce the maximum measurement error to 2.6%, from the initial 7.7% returned when a single wavelength was employed.