Modelling photovoltaic soiling losses through optical characterization

Data and additional plots for the Scientific Reports paper titled, "Modelling photovoltaic soiling losses through optical characterization". This includes the spectral transmittance data and all of the optical microscopy. The manuscript presents the results of an international collaboration that investigated the spectral effects of soiling and dust naturally deposited on PV glass exposed outside at seven locations worldwide. These places were chosen to represent a wide variety of climates and environmental conditions. They are also relevant to the continued growth of installed PV systems. We report on one of the largest pools of naturally deposited soiling on glass. While most of the published soiling work focuses on a single site, there are few comparative studies of soiling from multiple locations. The glass coupons were analyzed after an 8-week outdoor exposure, but were not meant to reflect a complete picture of soiling at a given site. Our focus was on the fundamental aspects of soiling from the perspective of both the spectral shape of the transmittance and the corresponding particle size distribution. The optical transmittance and the particle size distribution of the soiling were compared in order to find correlations that could be universally valid, and that could open possibilities to modelling PV soiling losses. Together with this, we present the first effort where empirical models from other disciplines are applied in order to describe both the spectral characteristics of the soiled samples and their corresponding particle size distributions. The spectral hemispherical transmittance of the soiled coupons was taken using a Cary 500 spectrophotometer. The transmittance was fit to a modified form of the Ångström turbidity equation. Microscope images were captured at 100x and 500x magnification for the coupons soiled at each of the seven sites. Micrographs for each glass coupon were taken using a Keyence VHX-5000 microscope at a resolution of 1200 pixels x 800 pixels. The micrographs were analyzed with ImageJ to obtain the particle size distribution