Correlation of I-V Curve Parameters with Module-Level Electroluminescent Image Data Over 3000 Hours Damp-Heat Exposure

Four module brands of three samples each were exposed to 3000 hours of accelerated damp-heat testing. Current voltage (I-V) curve tracing and electroluminescent (EL) imaging were performed at 500 hour steps. The I-V curves were analyzed to get 8 common I-V parameters: maximum power $(P_{mp})$ , fill factor (FF), current at maximum power $(I_{mp})$ , voltage at maximum power $(V_{mp})$ , current at short circuit $(I_{sc})$ , voltage at open circuit $(V_{oc})$ , series resistance $(R_{s})$ , and shunt resistance $(R_{sh})$ . The EL images were processed with a data processing pipeline using the open source coding language Python, employing techniques such as filtering, thresholding, convex Hull, regression fitting, and perspective transformation. This pre-processing was necessary to re-orient the images in a uniform manner as there was variance involved in the alignment of the modules during the measurement process. With the preprocessed data in hand, module-level parameters of median, mean, and standard deviation were calculated from the image pixel intensity distributions. These image data parameters were plotted against all the I-V curve parameters to identify trends in these data types. Correlation heat maps were generated depicting the relationships between data type parameters from IV and EL measurements.