Applying Ferrite.jl to solve Stokes equation in an irregular 2D domain

[gnaprs]

Hi!

I'm a bioengineer with a little bit of experience with fluid mechanics. I'm currently working on a problem where the end goal would be to solve a diffusion-convection problem. In the first instance, I would like to get some experience solving Stokes equation in my geometry. I've tried to summarize the problem in the following figure (apologies if my notation is incorrect):

In this problem, I'm planning to vary L_2 such that I can explore many contact angles theta.

Would you have any advice for solving this problem with Ferrite.jl?

Thanks!

[KnutAM]

Would you have any advice for solving this problem with Ferrite.jl?

Regarding the Ferrite-usage, there is the Stokes flow tutorial which is very similar to what you have. I haven't tried but I think using the rotation matrix with rotation $\pi$ should give the correct periodic boundary conditions for your problem. For different angles, the setup won't change as long as you have the same names for the different facetsets.

Regarding the parametric study, if your angle $\theta$ is not too close to 0 or $\pi$, I think you can get away with using a regular grid = generate_grid(Quadrilateral, (nelx, nely), x0, x1) followed by transform_coordinates!(grid, f) to map the grid onto the correct dimension.

For a more general solution for larger angle variation, you could use FerriteGmsh to mesh using gmsh and parse this into a Ferrite grid.