FR: Cubed Sphere Lattice about fluidx3d

[Meerkov]

That guide doesn't look like it would work. A Cubed Sphere requires special boundary conditions, and gravity to be in the center of the cube.

[ProjectPhysX]

Hi @Meerkov,

you can load multiple .stl meshes for composite geometries. Please follow this section in the documentation.

Kind regards,
Moritz

[ProjectPhysX]

Ah, now I get what you mean, stitching together and warping six 2D grids to simulate fluid flow on the surface of a sphere.
This is not really possible with LBM, as cell size has to be uniform Cartesian and cannot accomodate warping.

[Meerkov]

For what it's worth, there actually isn't anything that prevents stitching together warped 2D grids. It's possible, it just introduces error.

To enhance physical accuracy, you can alter the density calculation to take into account the non-uniform grid size, or other enhancements.

From what I understand, the non-uniform grid only leads to perhaps a 3% error on some simulations (e.g. a vortex moving across the corner of the cube).