Thin Layers in 2D Multiphysics Simulations

General
1

Hi all,

I am working on a multiphysics problem involving two domains in a 2D simulation. I am particularly interested in understanding how the thickness of Domain 2 influences the results. While I can reduce the thickness down to a 𝑑/β„Ž ratio of 0.01, the results become inconsistent when I go below this threshold. My goal is to reach a ratio as small as 0.0005. Although I have already implemented a very fine mesh around this thin layer, the results remain unreliable.
Does anyone have suggestions for accurately solving multiphysics problems involving such extreme aspect ratios?

Also, if there are any available examples in FEniCSx that involve similar setups (e.g., thin layers or high aspect ratio domains), I would greatly appreciate any pointers.

model
model450Γ—157 5.17 KB

Thanks,

2

Are you using a triangular mesh, or a mesh consisting of quadrilaterals? Quadrilateral cells deal better with high aspect ratios

3

I am using Quadrilateral cells. I would like to know the best way to simulate models with such extreme aspect ratios.

Thanks

4

without an example that illustrates the issue, it is hard to give guidance.

Without the PDEs you are solving in each domain, how they are coupled, the chosen function spaces and how many cells there are in the thin domain, it’s hard to be more concrete.

5

Thanks, I will start preparing a minimal example and will post it here.