A Geometry-Adaptive Mesh Refinement Framework for Immersed Boundary Methods on Hybrid Unstructured Meshes

summary

The gist

This work describes a new preprocessing tool for mesh adaptation on hybrid unstructured meshes specifically designed for immersed boundary methods to refine meshes around immersed geometries in an…

In short

This work introduces a new preprocessing tool for adapting meshes on hybrid unstructured meshes used with immersed boundary methods. The tool refines the mesh around complex geometries efficiently using a multi-step algorithm, ensuring accurate and flexible simulations of fluid flow past bodies with moving boundaries.

Key concepts

Hybrid Unstructured Meshes
These are unstructured meshes that combine different element types, such as tetrahedra, hexahedra, prisms, and pyramids. They allow for a flexible mesh design where one part of the geometry is fixed (body-fitted) and another part moves (handled by the immersed boundary method).
Immersed Boundary Method
This is a technique used in computational fluid dynamics where the fluid domain is fixed, but moving or complex solid geometries are represented within that fixed grid using special methods. The mesh refinement tool specifically supports this approach.
Mesh Refinement Algorithm
The tool uses a six-step process to adapt the hybrid mesh around an immersed geometry. This sequence includes importing the mesh and geometry, overlapping them, flagging elements, balancing the mesh structure, and finally refining it to achieve a high-quality simulation grid.
Navier–Stokes Equations
These are the fundamental equations governing fluid motion. The simulations use these equations to model how fluids move through various geometries under different conditions like subsonic flow past cylinders or airfoils.

Terminology used across episodes

This episode discusses

The paper

A Geometry-Adaptive Mesh Refinement Framework for Immersed Boundary Methods on Hybrid Unstructured Meshes · Read on arXiv

School of Aeronautics, Universidad Politécnica de Madrid · Center for Computational Simulation, Universidad Politécnica de Madrid

In this work, we describe a new preprocessing tool for geometry-adaptive mesh refinement on hybrid unstructured meshes with a target application on immersed boundary methods. The tool has as input an unstructured, hybrid, and conforming mesh generated by an external mesh generation software, and the main goal is to refine this mesh around immersed geometries in such a way that the CFD solver using the immersed boundary method can simulate flow problems in an accurate and efficient manner. The input background mesh can be made of different types of elements, like tetrahedra, hexahedra, prisms, and pyramids, which, unlike Cartesian meshes, permit for a more flexible mesh. Hybrid unstructured meshes enable one to use the immersed boundary technology in a new class of flow problems where the full geometry is decomposed into a fixed geometry part and a changing geometry part. A body-fitted mesh is generated for the fixed geometry while for the changing one is used the immersed boundary method. We simulate several flow problems to test the new meshes, including subsonic flow past a cylinder and subsonic flow past an NACA0012 airfoil, both using finite volume and discontinuous Galerkin methods and solving the Navier-Stokes equations. As an industrial example of our mesh generation, we consider the simulation of a multi-element airfoil: in this case, a mesh generation software generates an unstructured conforming background mesh for the slat and main airfoil, while the flap is placed as immersed geometry in this body-fitted mesh. As accurate and efficient results are sought, this mesh is refined around the flap and then the subsonic flow at high-lift flow conditions is simulated with a finite volume method coupled with an immersed boundary method and using the Reynolds-averaged Navier-Stokes equations.

Transcript

Introduction to the show: ident: Physics. The day in general physics: instruments, methods and results across the field.

Kai: We're looking at the paper: "A Geometry-Adaptive Mesh Refinement Framework for Immersed Boundary Methods on Hybrid Unstructured Meshes".

Mira: Here's the gist of the paper. This work describes a new preprocessing tool for mesh adaptation on hybrid unstructured meshes specifically designed for immersed boundary methods to refine meshes around immersed geometries in an accurate and efficient manner <ref:2607.27580#pg14>.

Kai: Tool and Mesh Structure The tool has as input an unstructured, hybrid, and conforming mesh generated by an external mesh generation software <ref:2607.27580#pg14> The input background mesh can be made of different types of elements, like tetrahedra, hexahedra, prisms, and pyramids which permit for a more flexible mesh <ref:2607.27580#pg14> Hybrid unstructured meshes enable one to use the immersed boundary technology in a new class of flow problems where the full geometry is decomposed into a fixed geometry part and a changing geometry part <ref:2607.27580#pg14> A body-fitted mesh is generated for the fixed geometry while for the changing one is used the immersed boundary method <ref:2607.27580#pg14>.

Mira: Governing Equations and Numerical Schemes The simulations are based on the Navier–Stokes and Reynolds-Averaged Navier–Stokes equations <ref:2607.27580#pg14> The spatial discretization schemes used in CODA to solve the Navier–Stokes and Reynolds-Averaged Navier–Stokes equations are the finite volume method, the modal discontinuous Galerkin method, and the discontinuous Galerkin spectral element method <ref:2607.27580#pg14> For the time discretization, an implicit backward-Euler scheme based on preconditioned matrix-free GMRES is employed <ref:2607.27580#pg14>.

Kai: Mesh Refinement Algorithms The mesh refinement preprocessing tool follows a multi-step algorithm to refine the hybrid mesh around an immersed geometry <ref:2607.27580#pg14> The process involves several key steps, including: 1.

Paper summary: Mira: So to wrap up this paper, "A Geometry-Adaptive Mesh Refinement Framework for Immersed Boundary Methods on Hybrid Unstructured Meshes," what it boils down to is a preprocessing tool that takes any unstructured, hybrid mesh and uses a specific algorithm—importing geometry, overlapping it, flagging elements, balancing the mesh, and then refining—to prepare it for an immersed boundary method simulation.

Kai: The implication is that we can get better flow simulations using these methods without having to spend all our time generating incredibly complex body-fitted meshes for every single shape we want to study sixty-two <ref:2607.27580#pg2>. It makes the setup more flexible because the input mesh itself can be made of various element types like tetrahedra and prisms.

Lev: For us, on real hardware, it means that if you have a fixed geometry part and a moving geometry part, this tool gives you a structured way to prepare the mesh for those two parts interacting via an immersed boundary approach sixty-two <ref:2607.27580#pg2>. The goal is to get that accuracy we need when simulating things like high-lift flow conditions on multi-element airfoils.

Mira: It means the numerical scheme itself, even if it’s finite volume or discontinuous Galerkin, benefits from this better mesh quality near those boundaries, which helps in avoiding boundary condition headaches sixty-two <ref:2607.27580#pg2>. The authors show that they can get results that agree with full body-fitted simulations and even experimental data for things like subsonic flow past a cylinder or an NACA zero zero one two airfoil sixty-two <ref:2607.27580#pg2>.

Kai: It’s about bridging the gap between the complex geometry we want to study and the numerical methods we use. The tool provides a systematic way to refine that mesh around the immersed geometries so you get an accurate and efficient simulation setup sixty-two <ref:2607.27580#pg2>.

Lev: One thing they point out as a limitation is that this is just a preprocessing tool; it doesn't solve all the problems in the CFD solver itself, but it makes the input to the solver much better. It’s not a complete solution for every single problem.

Mira: So in short, this framework gives people a solid way to adapt hybrid unstructured meshes for immersed boundary methods, making complex geometry problems manageable numerically sixty-two <ref:2607.27580#pg2>. It’s about making sure the mesh quality is high enough so the CFD solver can actually deliver good physics.

Conclusion: Kai: So this paper is about a new way to make computer meshes better for immersed boundary methods using hybrid unstructured meshes sixty-two.

Mira: It’s taking this idea of refining the mesh around tricky shapes and applying it specifically to these hybrid setups, which are getting more common in flow problems sixty-two.

Lev: The framework has six steps: importing the mesh, importing the geometry, overlapping them, flagging elements, balancing the mesh, and then finally refining it sixty-two.

Kai: And what they’re showing us is that this systematic approach works well for things like subsonic flow past a cylinder and an NACA zero zero one two airfoil sixty-two.

Mira: The authors confirm these numerical results match both the full body-fitted simulations and even some real experimental data sixty-two.

Lev: That's good because it means the quality of the input mesh isn't just theoretical; it actually helps get accurate fluid flow solutions sixty-two.

Kai: So what does this mean for people who are doing quantum hardware or complex fluid dynamics simulations? It means less time spent on manual mesh generation and more time focused on the actual physics of what you're trying to study.

Mira: I think it shows that you can still get high accuracy even when you’re dealing with these hybrid geometries where one part is fixed and another part is moving sixty-two.

Lev: From an engineering standpoint, this preprocessing step reduces the overhead because a cleaner mesh means the solvers don't choke on bad element quality sixty-two.

Kai: It’s about making sure the geometry setup doesn't become a bottleneck in your larger computation pipeline. This kind of systematic refinement could really speed up how we test new boundary conditions.

More episodes

← Home