CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM (http://www.cfd-online.com/Forums/openfoam/)
-   -   Spikes in gaps after snappyHexMesh (http://www.cfd-online.com/Forums/openfoam/119176-spikes-gaps-after-snappyhexmesh.html)

jackeyhust June 11, 2013 09:42

Spikes in gaps after snappyHexMesh
 
2 Attachment(s)
Hi, Foamers,

Recently I meet a problem for the a model. For geometry, there is a narrow gap, in which there will be the spikes in the gap. But actually it should be smooth, otherwise it causes problems. They should not be touched in the gap. You can see it in my attachment.

I have tried my methods:
1. I have tried to make the base size lower. The base size is lower than the gap distance, and it is somehow better, the spikes become less. But not all of them. But this also causes the increase of the cell numbers. I do not want to increase the cell number any more. So do you have any other suggestions on it ?

With kind regards,


Jackie

Attachment 22617

Attachment 22618

Irish09 June 11, 2013 14:00

More Information
 
Hello Jackie,

While I think I have an understanding of your problem, it would help if I could get some more information such as your snappyhexmeshdict file and possibly an image of the STL file you are trying to mesh. So to clarify, the gap is small and is suppose to be there?

jackeyhust June 11, 2013 15:16

yes, the gap is small, they should be there. That's difficult, I know. Do you have any suggestions?

Irish09 June 11, 2013 15:22

Potentially
 
Hello Jackie,

As I said before, it is difficult to say without knowing how you are setting up your snappyhexmesh case. One suggestion I could suggest is to use the surfaceFeatureExtract function, I talk about it for an internal geometry case here:

http://www.cfd-online.com/Forums/ope...-injector.html

niklas June 11, 2013 16:49

im pretty sure your cells are too large, just refine them abit more and it should work.
the first step of the meshing process is to refine the cells that contains the surfaces the number of times that you have specified in the dictionary.
If you have 2 surfaces that are very close and go through the same cells it will not refine those cells further because of this.

now, for the seconds step, the snap phase,
imagine that you have a cell containing 2 surfaces and ask yourself
which surface should the face of the cell snap to.
The result is clear from your pics, it varies, depending on which is closest (or which it finds first, not sure which)

You cant expect the mesh to have the details resolved on scale smaller than the cell size.


All times are GMT -4. The time now is 10:32.