Jiqi Liu to defend Ph.D. thesis

Jiqi Liu

On May 26, Jiqi Liu will defend her Ph.D. thesis, "Degradation of Bifacial & Monofacial, Double Glass & Glass-backsheet, Photovoltaic Modules with Multiple Packaging Combinations."

While growing up in Heihe, China, which Liu described as having “a climate that is even colder than Cleveland,” her father bought her a set of encyclopedias and she immediately found herself interested in natural science. She went on to earn a B.S. in Materials Science and Engineering at Southeast University in her native country before coming to Case Western Reserve University to pursue her M.S. and Ph.D.

As her time at CWRU comes to an end, Liu is proud of her contributions to DOE funded projects, both of which she believes "bring good analytical methods and study protocols to the PV industry." She praised her advisor, Professor Roger French, for having “immense knowledge and plentiful experience that have encouraged me all the time in my academic research and daily life.”

To prepare for her thesis defense, Liu made many small PV modules through soldering and lamination, putting some of them in chambers for accelerated exposures to multiple characterizations, including current-voltage curves, Suns-Voc curves, electroluminescence, photoluminescent to characterize changes in the exposed samples. She mounted the remaining samples at the SDLE solar farm for an outdoor exposure of around 1.6 years, tracking and analyzing time-series electrical features, module temperature and weather variables through modeling, with a goal of quantifying degradation and identifying activated degradation mechanisms.

"Commercial PV modules contain several layers to protect the internal solar cells from multiple environmental stressors,” said Liu. “The packaging strategies in the current PV market exhibit more diversity due to technology innovation, which bring challenges on reliability studies. While the material of a specific layer is changed or modified for some benefits, it is the whole packaging system that determines the long-term reliability of PV modules.” She pointed out that it is unclear if different packaging strategies bring significant differences in PV module long-term reliability and that studies examining samples under indoor accelerated exposures have often failed to report the statistical significance of results. The aforementioned studies also frequently failed to use identical materials and fabrication processes for different studied samples. Having seen even less outdoor reliability data, Liu decided to focus her thesis on “examining whether statistically significant differences exist for different module variants under different exposure conditions, and extensive characterization data collection in this study brings us the opportunity to develop neural network models to predict the degradation of electrical features.”

After Liu graduates from CWRU in August with her Ph.D. in Materials Science and Engineering, she will move to the Bay Area, where she will become a packaging reliability engineer at Apple. As her postgraduate career begins, Liu hopes to “develop more reliable products and become a very professional engineer trusted by my colleagues.” She also hopes to use the knowledge she learned from books to solve real-world problems and to help achieve a greener economy.

"I want to thank current and former members in my research group,” said Liu. “Their kind help and support made my study and life here a wonderful time.”

Liu’s abstract:

Annual installed capacity of solar energy has grown rapidly in recent years and reached 773.3 GW by the end of 2020, providing 3.1% of global electricity demand. The levelized cost of electricity (LCOE) of solar energy has been continuously decreasing since 2009 and reached $0.037/kWh in 2020. Improving the reliability of photovoltaic (PV) modules and reducing their degradation rates are critical for further decreasing the LCOE and maintaining the competitiveness. The degradation of PV modules depends on their interaction with exposure conditions and is strongly influenced by their packaging materials and combinations. In recent years, modules using polyolefin elastomer (POE), double glass (DG) module architecture, or transparent backsheet have been gaining market share and have become strong competitors to conventional monofacial ethylene-vinyl acetate (EVA) glass-backsheet (GB) modules. However, the reliability performance data of these emerging packaging strategies were lacking. This work used statistical analysis to compare the degradation behaviors of sixteen module variants under two indoor accelerated exposures and 1.6 years of outdoor exposure. The two indoor accelerated exposures included modified damp heat (80 °C, 85% relative humidity) and modified damp heat with full-spectrum light, up to 2,520 hours. The EVA+GB modules with opaque rear encapsulant exhibited a significantly greater power loss, and the dominant degradation mechanism was identified as interconnection corrosion. The outdoor exposure location was in the Dfa climate zone (continental, no dry season, hot summer). Significant differences in the average power loss were identified between three module variants and the other two. The dominant power loss factor for most module variants was uniform current power loss, followed by power loss due to increased series resistance. This work developed a cross-correlation algorithm to quantify the similarity of degradation behaviors under different exposures, considering the power loss rates and the similarity in trends for various electrical features over time. Enabled by extensive characterization data collected, various neural network models were explored to predict the change in electrical features based on images. Recurrent neural network (RNN) models outperformed convolution neural network (CNN) models, emphasizing the importance of utilizing measurements for the same sample taken at different exposure times to improve prediction accuracy.