应美国航空航天学会会刊（AIAA Journal）主编 Tom I-P. Shih 教授邀请，西北工业大学航空学院张伟伟教授、单湘淋博士将于北京时间2026年7月7日（21:00-22:00）在线作特邀主题报告（Keynote Seminars）。
📌 报告详情
报告题目： From Turbulence Databases to Data-driven Models: Knowledge Discovery of Scaling Laws and Turbulence Modeling
时间： 北京时间（GMT+8）： 2026年7月7日 晚上 9:00 美国东部时间（EDT）： 2026年7月7日 上午 9:00
会议链接： https://cassyni.com/events/Q9vaXurQcbeHRhkCk5qZuP https://aiaa.zoom.us/j/89382735075?pwd=xDFoJbxhXYb5QUbaGL0yQLnt9HJZdy.1
📝 报告摘要
Artificial intelligence is reshaping research paradigms in fluid mechanics. However, the prediction of engineering high-Reynolds-number wall-bounded turbulence remains a major bottleneck in aerodynamic design, particularly for flows involving complex geometries, surface curvature, adverse pressure gradients, and massive separation. This challenge is largely associated with the scarcity of high-fidelity data, incomplete understanding of the underlying physical mechanisms, and the limited predictive capability and generalizability of existing turbulence models. To address these challenges, we constructed AeroFlowData, an aerospace-oriented, high-resolution, high-Reynolds-number turbulence database comprising more than 500 cases and approximately 100 TB of data. The database covers a broad range of canonical and engineering configurations, including hypersonic vehicles, civil aircraft, turbomachinery flows, and supercritical-fluid cases. Then we discovered new mean-velocity and mixing-length scaling laws for wall-bounded turbulence, providing interpretable physical insights into high-Reynolds-number turbulent flows. Furthermore, we developed a data-driven turbulence modeling framework spanning black-box deep-learning models and white-box physics-informed model discovery. By incorporating physical constraints and a partitioned ensemble modeling strategy, the proposed models achieve improved universality and generalization across different flow regimes. The resulting models exhibit enhanced interpretability, robustness, and predictive accuracy, leading to substantially improved predictions for complex separated flows, including high-Reynolds-number airfoils, helicopter rotors, and civil aircraft configurations.
📖 AIAA Journal Seminars 介绍
AIAA Journal Seminars是由 AIAA Journal 主办的线上系列学术研讨活动，旨在围绕已发表的高水平航空航天研究成果及前沿科学问题开展国际学术交流与深度讨论。本次报告属于 AIAA Journal Seminars 中的Keynote Seminars系列邀请报告。
该系列报告于2024年由 AIAA Journal 正式发起，分为作者报告（Author Seminars）与特邀主题报告（Keynote Seminars）两类形式：
Author Seminars： 面向 AIAA Journal 近五年内已发表论文的作者与读者，主要围绕具体论文内容展开更深入的技术阐释与补充讨论，重点在于推动既有研究成果的理解与延伸。
Keynote Seminars： 面向更广泛的前沿研究议题与领域性科学问题，由期刊编辑及编委会从全球范围内遴选具有代表性与引领性的研究成果进行邀请报告，具有更强的学术影响力与领域引领性。
AIAA Journal Seminars 报告由期刊编辑及编委会提名，并由 AIAA Journal Seminar Series Committee 评选产生（该委员会由剑桥大学 Paul Tucker 教授与杜克大学 Earl Dowell 院士联合主持）， 每年仅遴选少数高水平论文作者进行专题报告。
📚 参考文献
[1] Shan Xianglin, Liu Yilang, Cao Wenbo, Sun Xuxiang, Zhang Weiwei. Turbulence modeling via data assimilation and machine learning for separated flows over airfoils [J].AIAA Journal, 2023, 61(9): 3883-3899.
AIAA J | 西北工业大学单湘淋等：绕翼型分离流动的数据同化与机器学习湍流建模方法
[2] Shan Xianglin, Zhang Weiwei. Data-driven adverse pressure gradient correction for turbulence model [J].AIAA Journal, 2025, 63(7): 2780-2796.
AIAAJ| 西工大单湘淋、张伟伟：数据驱动的湍流模型逆压梯度修正方法
[3] Zhang Weiwei, Shan Xianglin, Liu Yilang, et al. High-Reynolds-number turbulence database: AeroFlowData [J].Scientific Data, 2025, 12(1): 1500.
Nature子刊 Scientific Data | 九家单位联合打造！AeroFlowData
[4] Yang Zhongxin, Shan Xianglin, Yang Xiang I.A., Zhang Weiwei. Data-enabled discovery of specific and generalisable turbulence closures [J].Journal of Fluid Mechanics, 2025, 1016: R1.
JFM | 西北工业大学杨忠鑫、张伟伟等：数据驱动的特定且可泛化的湍流封闭知识发现
