CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Main CFD Forum (https://www.cfd-online.com/Forums/main/)
-   -   Xflow (https://www.cfd-online.com/Forums/main/242561-xflow.html)

Alnemr April 27, 2022 15:06

Xflow
 
I am a beginner in CFD world, however I know the basics of each blackbox of any commercial software. but while I take an overview of xflow software I did not see any mesh generation and while I was searching about that I found that, this software is based on the Lattice Boltzmann method, thus I want why there is no meshing and also what is the relation between the mesh and that method.

thank you

sbaffini April 28, 2022 04:18

I am no expert in Xflow, but I am pretty sure that, in fact, it uses a mesh, it just does it for you. I am also no expert in LBM, so I don't know why most of these tools use this approach and, seemingly, no tool allows the user to make his own mesh (probably because they all work with cartesian lattices only, so it really is best done automatically).

However, the most important thing to note is that a LATTICE is there, which is a mesh indeed. What the method does, very roughly, is moving particles on this lattice.

flotus1 April 28, 2022 05:27

Can confirm, LATTICE Boltzmann Methods need a mesh.
I haven't used any commercial LBM solvers myself, but as far as I am aware, they all try to keep the meshing process as far away from the user as possible. I guess that's where allegations of LBM being a "meshless" method come from.
If it's a single grid level, I can see why. More or less the only free parameter is the lattice spacing, since LBM uses Cartesian grids (with a few exceptions). Not much to do for a user, and grid generation is very straightforward.
If it's multiple grid levels with interfaces between them however, grid generation is more involved. And the same rules apply as with other CFD methods: mesh topology has an impact on results.


All times are GMT -4. The time now is 07:42.