报告时间：2026年6月17日（周三）下午 15:30
报告地点：91国产 本部天赐庄校区精正楼103
报告人：Yongxing Wang 博士（讲师），英国利兹大学计算机学院
报告摘要：
In this talk, I will present recent work on the mathematical formulation and numerical simulation of fluid flow on curved manifolds. Starting from the classical Navier-Stokes equations in Euclidean space, I will introduce the differential geometric framework required to generalise fluid dynamics to non-flat geometries, including metric tensors, covariant derivatives, and intrinsic differential operators such as the Laplace operator.
I will then discuss incompressible surface Navier-Stokes equations on manifolds and their steady-state solutions. Numerical simulations will be presented to illustrate the behaviour of surface flows and the influence of geometry on fluid dynamics, including steady-state solutions associated with Killing vector fields arising from manifold symmetries. I will compare three numerical approaches: surface finite element methods, intrinsic finite element methods in parameter spaces, and numerical solutions from the corresponding eigenvalue problems.
Finally, I will briefly discuss extensions to relativistic fluid dynamics and the coupling of fluids with Einstein's field equations. In particular, I will outline the basic ideas of the 3+1 decomposition used in numerical relativity and discuss possible directions for combining geometric fluid mechanics with modern numerical methods for general relativity.
To explore potential collaborations, I will also briefly introduce several of my current research projects, including large-scale simulations of fluid-structure interaction problems with adaptive mesh refinement; optimal control of fluid-structure interaction systems with applications to the reconstruction of C. elegans locomotion; and physics-informed neural networks for fluid dynamics.
报告人简介：
Yongxing Wang 博士，英国利兹大学计算机学院讲师。研究方向包括流形上的流体力学、数值方法、流固耦合、最优控制、物理信息神经网络等。
邀请人：曹永罗教授、丁睿教授