I am a UG student doing a final year project about the "Fluent" with my partner. We would like to form grids to study the wind flow around a single building. We know that we have to form fine grid around the building but coarse grid far from the building. Could anybody tell me how to form grid with increasing size as the grid is farer from the building?

http://ihome.ust.hk/~ma_ccc/fig4.gif

Could anyone teach us how to form this grid?

It's going to be difficult to give you a Gambit training over the forum...

but a good start: for good grading in a hexmesh using Gambit, the keyword is 'decomposition'.

You want to reconstruct the mesh in your url, but I wouldn't do that. If you have Fluent and Gambit available, you can construct meshes that are a lot better than that! I'd create a good 2D quad-paved mesh with a sizing function on the ground, and extrude that in a 'cooper mesh' to the ceiling. Of couse your geometry needs to be decomposed in a way that suits the cooper meshing scheme.

Be carefull with the turbulence modelling around buildings!

were are you based?

good luck!

Laika, still orbiting

With Gambit, there are some tutorials where you could find some explanations about grading meshes, ...

Thank you for your reply.

However, we can't form a good 2D quad-paved mesh with a sizing function on the ground.

http://ihome.ust.hk/~ma_ccc/Gambit_ImageFile1.tif

Source: Edges of the ground of inner block Attachment: Edges of the ground of outer block

When we change the attachment into ground face, similar result occurs.

Should we decompose the ground face first?

Thank you very much!!

CCC

Better is to play with your size limit. It is way to big. If you consider further decomposition, cut the model in the middle first. That will remove faces that have a 'hole' in it, like your ground face. Faces with a hole will often create problems when you want to have full control of the mesh cel size distribution.

L