Report Title: Mesoscopic Modeling Methods in Fluid Dynamics and their High-Performance Simulation

Report Time: August 13, 2024, 15:00 - 16:30

Report Location: Room 316, Civil Engineering Research Building, Second Campus

Organizing Unit: Smart Infrastructure Research Center

Report Abstract:

Unsteady time-dependent fluid flows are notoriously difficult to simulate efficiently especially at high Reynolds numbers and with dynamic fluid-structure interaction. In this talk, I will introduce the non-traditional mesoscopic methods for fluid simulation based on the lattice Boltzmann modeling and their high-performance GPU-based implementation to tackle this problem. More specifically, I will introduce advanced collision modeling and boundary treatment methods for high Reynolds number flows, as well as the multi-resolution simulation techniques. In addition, I will also introduce the mesoscopic and hybrid methods for simulating two-phase and multiphase fluid flows. Optimizations for GPU implementation will be briefly mentioned. Validations over different experimental setups will be shown, indicating the efficiency and accuracy of our method. Finally, as demonstrations, I will show a variety of practical applications of our method in multiple domains.

Introduction of the Speaker:

Associate Professor Xiaopei Liu is currently a tenured associate professor and doctoral advisor at the Visual and Data Intelligence Center, School of Information Science and Technology, ShanghaiTech University. He obtained his Ph.D. in Computer Science and Engineering from The Chinese University of Hong Kong and served as a postdoctoral researcher at Nanyang Technological University in Singapore, specializing in the simulation and visualization of classical and quantum fluids. He integrates GPU-based high-performance computing and machine learning into fluid simulations. His research findings have been published in numerous prestigious journals and conferences across various fields, including ACM SIGGRAPH, ACM TOG, JCP, PRD, POF, IEEE TVCG, and IEEE ICRA. He has led projects funded by the National Natural Science Foundation of China for young researchers and general projects, and has participated in key R&D projects from the Shanghai Municipal Science and Technology Commission and Hong Kong ITF. He has also been involved in virtual aircraft turbofan engine simulation R&D projects with Nanyang Technological University and Rolls-Royce in the UK. Dr. Liu is currently focusing on high-performance fluid simulation and modeling, large-scale scientific visualization and rendering, high-performance parallel computing, machine learning, and the dynamics simulation, design, and control of unmanned aerial vehicles/submersibles. He leads a team that independently develops fully autonomous domestically-produced fluid simulation industrial software, which is gradually being applied in practical engineering R&D and manufacturing.