CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Pointwise & Gridgen (https://www.cfd-online.com/Forums/pointwise/)
-   -   Gridgen Layer Problem (https://www.cfd-online.com/Forums/pointwise/78288-gridgen-layer-problem.html)

harishameed33 July 17, 2010 00:37

Gridgen Layer Problem
 
I have the following geometry

http://4.bp.blogspot.com/_gFgbUG_ehs...vM/s1600/1.jpg
I am using Gridgen V 15.10, below is the Unstructured mesh…
http://2.bp.blogspot.com/_gFgbUG_ehs...8/s1600/12.bmp

I know that it can be structured mesh but my requirement is unstructured (the above grid is not my case and I am only using the above to explain the problem)



After running the unstructured solver for the above I get

http://1.bp.blogspot.com/_gFgbUG_ehs...0/s1600/15.jpg



The layers at the top and bottom surface converge (as shown in the above figure)
Now I have tried different techniques
By changing Collusion buffer aniso-iso blend values
Redistributing connectors making them close to the edge
Forming layers in the top and bottom domain

But got no success

Kindly help me in this matter

rmatus July 17, 2010 12:03

Make layers in domains to
 
You just need to run the T-Rex (anisotropic) solver on the top and bottom domains with the same spacing and growth that you will use in the block.

In the existing grid, there is an inconsistency in the grid spacing on the boundary domains and the volume grid. Try running the DLR-F6 tutorial in Gridgen to see how this works. In that example, the symmetry domain is run with anisotropic layers to match the volume grid, just as you want to do here.

If you need help, let me know.

harishameed33 July 17, 2010 14:01

I have tried this option but it gives the same result ....
may be i am doing something wrong

http://3.bp.blogspot.com/_gFgbUG_ehs...A/s1600/12.bmp

http://4.bp.blogspot.com/_gFgbUG_ehs...w/s1600/13.bmp

but can you please elaborate about the "grid spacing on the boundary domain and volume grid"

i have run the solver on the top and bottom domains with same spacing n GR as for the block but still getting same results

kindly tell where i am wrong

rmatus July 17, 2010 16:57

Use symmetry on domains on side of growth layers
 
Sorry, I gave an incomplete answer before. There are two main considerations.

First, the domains (surface grids) at the top and bottom of your picture need to have anisotropic triangles on them, so they are compatible with the anisotropic tetrahedra growing in the volume. You can do this in the Domain, Solver, Unstructured menu by setting the anisotropic layers, initial step size, and growth rate to be the same as you are using in the volume.

Second, in the Block, Solver, Unstructured menu set the top and bottom domains to a Symmetry boundary condition. That will let Gridgen know to have the adjacent tetrahedral layers grow along them and include their points in the anisotropic layers.

Gridgen's DLR-F6 tutorial has a good explanation of this set-up, so you might want to take a look at it.

I hope I am understanding your problem correctly. If this does not help, let me know and we will try something else.

harishameed33 July 18, 2010 01:09

Thankx Alot the symmetry boundary cond works perfect :)


All times are GMT -4. The time now is 19:43.