CFD Online Discussion Forums

CFD Online Discussion Forums (
-   FLUENT (
-   -   Continuously growing cells with Gambit? (

Regert May 7, 2001 11:09

Continuously growing cells with Gambit?
Dear Gambit User Community,

I've got the following problem: I would like to create an unstructured mesh with tetrahedral cells, for external flow problem (3D flow about a body), the complexity of my geometry makes the application of structured mesh very difficult. With tetrahedral unstructured mesh, the cell sizes are very rapidly growing towards the interior space, so I could not set the fineness of the grid in the interior (far from the edges and surfaces)on the way, how it can be with structured mesh (I have tried to apply interior surfaces as well). Could you please help me, if any possibility for changing the parameters of the automatic meshing algorithm, or an interior 3D growth factor can be set somewhere. (I have already tried Volume Adaption in Fluent But it was not successful, and it does not that that I want). Summarized: one wall:given cell size, opposit wall:given (bigger) cell size; I want the interior cells continuously grow from one wall to the opposit one.

Thank you very much for your help. Regards, Regert

alain May 9, 2001 04:26

Re: Continuously growing cells with Gambit?
You can use several approach in order to control the size of the cells in the interior part of your geometry :

extrude prismatic boundary layer from the wall

decompose your interior part in subdomain and control mesh size with interior face/edges

I don't know if it is possible to do it in gambit but you can change the tetra volumetric mesh technic from the default delaunay method to advancing front method which generaly give a regulier and finner mesh



manaf hasan May 9, 2001 08:50

Re: Continuously growing cells with Gambit?
hello i don't know if you are familar with t grid or not if so you can creat a B.L prisims around your interior domain and all the other area will be cuirser meash .

you can do that in gambit by creating meach in edges or faces where you need fine mesh with small space than the defolt which is one then creat your volume meash it will be very fine near these faces and geting cuirser while you going away from these surface .

good luck

Regert May 9, 2001 12:16

Re: Continuously growing cells with Gambit?
I already know, how to make dense grid on given surfaces, but the problem is not that. If you make a very dense distribution of nodes on an edge, or on a face, the mesh on the face will really be dense, but if we make an interior volume mesh (using the pre-definded dense and coarse face meshes) the cells will change there volume towards the interior very rapidly. The main problem is, that the densening effect is only happens in the vicinity of the dense meshed surface. The volume of the cells can be maximized (somehow), but maximizing does not result a continuously growing cell distribution from the dense meshed surface until the coarser meshed surface. I hope, that it can easily be understood. I would like to thank you all very much for your help. Regards, Tamas Regert

Lanre May 9, 2001 13:18

Re: Continuously growing cells with Gambit?
The transition in tet element volumes from a finely meshed surface to an outer coarsely mesh surface will not be uniformly varying as you have noticed.

An earlier suggestion by Alain to break up your geometry to stage the progressive growth of your mesh is worth investigating.

On the other hand, if you cannot refine your outer surface mesh any more than you currently have it, you can consider refining the tet mesh further by changing the default 2 levels of refinement to 4 using, in the Edit-->Defaults panel, MESH.TETMESH.REFINE_LEVELS.

Regert May 9, 2001 13:37

Re: Continuously growing cells with Gambit?
I have tried to brake up my domain into several subdomains but in the middle of each subdomain, the cells were very big and only towards the sides of the domains was a rapid densening, that means the effect of all interior surfaces does not reach far from them. However, I haven't tried to change the refinement level yet, thank you very much for this suggession indeed. I will try it at once. Regards, Tamas Regert

Regert May 10, 2001 04:01

Thank you for the advices...
I think it is not worth to waste any more time on this question because this undesireable structure of grid is a special eigen property of the unstructured, automatic mesh algorithms. My question can not be solved, I think the only possibility is applying structured mesh at least in the wake region. I would like to thank you for all your advices. Best wishes and Regards from: Tamas Regert

All times are GMT -4. The time now is 03:47.