An Automated Algorithm for Quantifying Cracks in Photovoltaic Backsheets Under Accelerated and Real-World Exposures

Abstract

Extending the lifetime of photovoltaic modules is critical for the industry's success. Backsheet cracks are concerning since they compromise module safety and performance. In this study, 23 backsheets with 9 unique material combinations were exposed to four weathering conditions both indoors and outdoors. A previously developed algorithm utilizing optical profilometry data quantified the depth, width, area, spacing, and number of cracks. Logistic regression predicted the likelihood of cracking with 88% accuracy (0.713 kappa) based on the 1715 cm ketone carbonyl absorption band and percent crystallinity of polyethylene. Based on thermo-oxidation and thin film fracture mechanics literature, this model indicates that increases in the modulus of polyethylene exacerbate the elastic mismatch between backsheet layers, thereby allowing cracks to initiate.