报告一：知识和数据驱动的湍流研究

报告人：何国威  院士  （中国科学院力学研究所）

报告摘要：人工智能（AI）正在改变科学研究的范式。湍流两次被《Science》杂志列为未解决的125个重大科学问题之一。知识和数据驱动湍流研究的新范式，不仅推动了湍流理论和方法的发展，而且增强了计算机模拟工程湍流的能力。本次报告将介绍力学所湍流课题组在该项研究的进展: （1）针对湍流的时间与空间尺度耦合问题，我们将深度学习的思想引入预解分析，提出了深度预解方法，重构湍流的时空能谱；(2) 针对水下高速航行器的湍流噪声问题，在壁面解析大涡模拟的基础上，发展了基于集合卡尔曼方法的雷诺应力张量基神经网络方法，研究具有分离特点的工程湍流。最后探讨机器学习应用于湍流研究的三个典型案例。

报告二：Recent Progress on Aircraft/Aeroengine Icing Physics and Anti-/De-icing Studies

报告人：Prof. Hui Hu （Iowa State University）

报告摘要：Aircraft/Aero-engine icing is widely recognized as a significant hazard to aviation safety in cold climates. The speaker will introduce the recent research progress made in his group on aircraft/aero-engine icing physics and anti-/de-icing. By leveraging the unique Icing Research Tunnel available at Iowa State University (i.e., ISU-IRT), comprehensive investigations are conducted to characterize the important micro-physical processes pertinent to aircraft/aeroengine icing phenomena. A suite of advanced flow diagnostic techniques, including molecular tagging velocimetry and thermometry (MTV&T), particle image velocimetry (PIV), digital image projection (DIP), and high-speed infrared (IR) imaging thermometry techniques, are developed and applied to quantify water droplet impinging dynamics, transient behaviors of wind-driven water runback flows, unsteady heat transfer and dynamic solidification processes over ice accreting wing surfaces. Anti-/de-icing performances of various “state-of-the-art” hydro-/ice-phobic coatings are evaluated quantitatively under different icing conditions (i.e., ranged from dry rime icing to wet glaze icing conditions). The recent efforts on studying aero-engine icing due to inhalation of in-cloud ice crystals and developing a novel DBD-plasma-based anti-/de-icing strategy will also be introduced to ensure safer and more efficient operation of aircraft/aeroengine in cold climates.

报告三：非球形颗粒两相流研究进展：从微小颗粒到有限尺寸颗粒

报告人：赵立豪  长聘副教授 （清华大学）

报告摘要：非球形颗粒两相流常见于自然环境及工业过程中，相关研究过去十年里受到较多关注。本次报告将介绍我们在非球形颗粒两相流方向上的数值模拟研究工作。第一部分是针对微小非球形颗粒，我们采用拉格朗日点颗粒方法对非球形颗粒在壁湍流中的统计行为进行了研究，发现了不同形状颗粒具有不同的倾向性取向分布及转动行为。第二部分是针对有限尺寸非球形颗粒沉降问题，我们采用浸没边界方法完全解析模拟了颗粒群沉降行为，发现了颗粒沉降速度对颗粒体积分数的依赖性、颗粒自发聚团现象以及颗粒诱导的大尺度流动结构。