文献类型：期刊文献

英文题名：Thermal Stress Buffering of Flexible Thin-Film Solar Cells Based on Machine Learning

作者：Yuan, Jiahao[1];Shi, Chunhao[1];Huang, Zhichao[1];Gao, Yang[1,2];Wu, Min[3];Qian, Min[4,5];Yan, Yabin[1,2]

机构：[1]East China Univ Sci & Technol, Shanghai Key Lab Intelligent Sensing & Detect Tech, Shanghai 200237, Peoples R China;[2]East China Univ Sci & Technol, Minist Educ, Sch Mech & Power Engn, Key Lab Pressure Syst & Safety, Shanghai 200237, Peoples R China;[3]Shanghai Inst Space Power Sources, State Key Lab Space Power Sources, Shanghai 200245, Peoples R China;[4]East China Univ Sci & Technol, Sch Phys, Shanghai 200237, Peoples R China;[5]Shanghai Aerosp Syst Engn Inst, Natl Key Lab Aerosp Mech, Shanghai 201109, Peoples R China

年份：2026

卷号：9

期号：2

外文期刊名：ADVANCED THEORY AND SIMULATIONS

收录：;EI(收录号：20254119318492);WOS:【SCI-EXPANDED(收录号:WOS:001587081300001)】；

基金：This work was supported by the Space Application System of China Manned Program, the National Natural Science Foundation of China (Grant Nos. 52275146, 52275149, and 12411530109) and the Opening Project of State Key Laboratory of Space-Power (YF07050124F1266). Shanghai Municipal Education Commission AI-Driven Reform of Research Paradigms to Empower Advancement of Disciplines (G100-2-24107).

语种：英文

外文关键词：GaAs solar cell; machine learning; thermal stress simulation

摘要：This study primarily focuses on the design of a thermal buffer layer for flexible GaAs-based thin-film solar cells. By employing Abaqus thermodynamic cell array modeling, the investigation prioritizes the thermal Mises stress between the encapsulation layer and GaAs cell layer, where it is most critical. Through computational analysis of the thermal buffer layer, key parameters including the coefficients of thermal expansion (CTE) of the buffer material, the elastic modulus (E), and buffer layer thickness are identified as the principal factors effectively influencing thermal stress mitigation. Furthermore, a double gradient buffer layer is implemented to alleviate thermal mismatch, supported by extreme gradient boosting and an interpretive algorithm to determine CTE value of the second buffer layer as the most dominant factor. The thermal stress reduction predicted by the genetic algorithm (GA) is validated against the pre-established simulation model in Abaqus. Subsequently, the material database is employed to identify similar optimized material combinations for investigation at other temperature points. This research provides valuable insights for the design of thermal buffer layers for solar cells.