In this article, we propose a deep learning based semantic segmentation model that identifies and segments defects in electroluminescence (EL) images of silicon photovoltaic (PV) cells. The proposed model can differentiate between cracks, contact interruptions, cell interconnect failures, and contact corrosion for both multicrystalline and monocrystalline silicon cells. Our model utilizes a segmentation Deeplabv3 model with a ResNet-50 backbone. It was trained on 17,064 EL images including 256 physically realistic simulated images of PV cells generated to deal with class imbalance. While performing semantic segmentation for five defect classes, this model achieves a weighted F1-score of 0.95, an unweighted F1-score of 0.69, a pixel-level global accuracy of 95.4%, and a mean intersection over union score of 57.3%. In addition, we introduce the UCF EL Defect dataset, a large-scale dataset consisting of 17,064 EL images, which will be publicly available for use by the PV and computer vision research communities.

The cost of photovoltaic (PV) modules has declined by 85% since 2010. To achieve this reduction, manufacturers altered module designs and bill of materials; changes that could affect module durability and reliability. To determine if these changes have affected module durability, we measured the performance degradation of 834 fielded PV modules representing 13 module types from 7 manufacturers in 3 climates over 5 years. Degradation rates (Rd) are highly nonlinear over time, and seasonal variations are present in some module types. Mean and median degradation rate values of −0.62%/year and −0.58%/year, respectively, are consistent with rates measured for older modules. Of the 23 systems studied, 6 have degradation rates that will exceed the warranty limits in the future, whereas 13 systems demonstrate the potential of achieving lifetimes beyond 30 years, assuming Rd trends have stabilized.

The ultraviolet fluorescence (UVF) imaging method has been widely used as a rapid and economic field inspection tool for investigating encapsulant discoloration of field-aged photovoltaic (PV) modules. In field-aged PV modules, encapsulant discoloration can result in a pronounced decrease in the short-circuit current. The spatial distribution of environmental stressors and therefore discoloration pattern over each cell of a module depends on the climate. In this article, we investigated the discoloration patterns (and their correlation with electrical parameters of modules and the extracted-cells from the modules) of ten modules produced at the same manufacturing facility and retrieved respectively from two distinct climates of Arizona (hot-dry; seven modules; 18 years of exposure) and Florida (hot-humid; three modules; ten years of exposure). In the Arizona modules, three distinct spatial patterns were observed in the UVF images: cell center region (intense browning); cell edge region (no browning); region between cell center; and cell edge region (light browning). However, in the Florida modules, only two distinct spatial distributions were observed: nonedge cell region (medium browning) and cell edge region (no browning). Furthermore, we employed a high-contrast UVF imaging technique to correlate the interdependence between the encapsulant discoloration pattern and the metallization/contact degradation pattern. For this correlative study, various experimental characterizations were carried out on both field-aged (Arizona and Florida) and unexposed modules and the cells extracted from the modules of the same model and they include: processed UVF and EL images, and I-V, TLM, optical microscopy, and XPS measurements.