CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   TGRID & Meshing (http://www.cfd-online.com/Forums/fluent/29340-tgrid-meshing.html)

Eric February 8, 2002 11:57

TGRID & Meshing
 
Hello everyone,

I am trying to mesh helical coils inside of a tank. In laying down a boundary layer around the faces that make up the coil, I first have to specify a number of node points when meshing the edges of the faces. When I try to mesh the faces, whatever size I choose for TRI/PAVE does not work because it is limited to the node spacing on the edges; i.e. I can choose sizes of 1, 0.1, 0.001 - and they all give the same mesh. I also mesh the tank wall faces in order to TGRID the volume between the coils and the wall. However, this is not working since I either get low on virtual memory (which I shouldn't!) or skewness errors.

Can someone give me advice on how TGRID works in this type of geometry and possibly recommend a course of action? I think the TGRID needs existing surface meshes to get it going. In addition, I think the sizing of the cells on the coil and the cells on the wall have to be somewhat similar to avoid size-jump errors.

Any help would be appreciated (especially by former TGRID'ers!). Have a great day.

-Eric

Eric February 8, 2002 12:04

Read this too! NOT "TGRID" software
 
Hey,

When I mention TGRID, I actually mean TET/Hybrid meshing with type TGRID. Some people may be confused and think I mean the software "TGRID." I am using Gambit and Fluent.

Eric

An Modh Coinniolach February 9, 2002 11:34

Re: TGRID & Meshing
 
With regard to your skewness errors, I had a similar problem with a boundary layer at a turbine blade.

The node density on the trailing edge of the blade was high in order to capture the geometry (a semi-circle) and coarse along the blade radially, resulting in a face mesh cells that had one fairly short side and one long one. Trying to grid this with tet/hybrid in Gambit resulted in cells that had unacceptably high skewness.

What I did was to split the volume around the blade and mesh the new "near blade region" with hex cells only. The I meshed the rest of the volume with tet/hybrid cells. By displacing the interface of the hex and tet cells, I increased the length of the short side of those cells near the trailing edge that had to interface with the pyramid/tet cells. In this way, high skewness was eliminated.

Are you meshing the faces on the coil with quad cells? It you are using the boundary layer tool, I believe that the faces that you are attaching the b.l. to have to be meshed with quad cells.

Eric February 13, 2002 11:52

Re: TGRID & Meshing
 
Thank you. I tried meshing all faces with TRI/PAVE and then did the volume with TET/Hybrid (TGRID) and it finally worked. However, I got 3.2 million cells and need to trim that number down to about 400,000-1 million. I need to increase the sizing of the mesh on my coil and wall faces, however I cannot go much higher on the coil because it doesn't capture the geometry well and gives errors.

I tried a variety of things and QUAD/PAVE on the coils (with a boundary layer) doesn't work because the computer runs out of memory when meshing. I also heard that you have less errors when meshing a volume if you use TRI/PAVE on the surface meshes.

-Eric


All times are GMT -4. The time now is 08:59.