AE Seminar
High-Resolution Simulations of Compressible Turbulent Flows: Methods, Deployment, and Scientific Insights
featuring
Hang Song
Tuesday, October 7
11:00 a.m. - 12:00 p.m.
Guggenheim 442
About the Seminar:
High-performance computing (HPC) is transforming the study of complex flow phenomena in aerospace engineering. With today’s powerful supercomputers, researchers can perform high-fidelity simulations of compressible, turbulent flows that were beyond reach as recently as a decade ago. Yet, this progress also brings new challenges: How can simulation accuracy be further improved? How can massive computing resources be used most effectively? And what new physical insights can be extracted from such simulations?
In this seminar, I will address these questions based on my research experience that spans the full pipeline of computational fluid dynamics (CFD) and compressible turbulent flow physics. First, a high-order simulation framework built on compact (Padé) numerical schemes will be presented. This method delivers both high spectral resolution and robustness without relying on excessive dissipation, solution filtering, or other compromises common in many other high-order CFD solvers. Next, I will detail the underlying parallel linear solver algorithm that makes this advanced framework scalable on modern heterogeneous computing architectures, enabling simulations across more than 24,000 GPUs. Lastly, insights from a wall-resolved large-eddy simulation of transonic shock buffet on a supercritical airfoil will be highlighted. A comprehensive data analysis reveals the low-dimensional dynamics underlying this complex flow system.
Bio:
Dr. Hang Song is currently a postdoctoral fellow at the Center for Turbulence Research at Stanford University. He received his Ph.D. in 2024 from the Department of Mechanical Engineering at Stanford University and was advised by Professor Sanjiva K. Lele. His research involves a broad range of topics in parallel computing, high-order numerical methods, fundamental physics of compressible turbulent flows, transonic aerodynamics, and compressible multiphase flows. He is also interested in the design and development of high-performance software packages for computational fluid dynamics.