[snejball91]

Hello dear users,

I am currently trying to set up a simulation in CFX 15. The y+ spacing calculator told me that I need a spacing of 6E-6 meters to get the y+ value 1. So I edited the mesh in ICEM and reloaded it in CFX-Pre.

But there is a problem because an additional domain called "Primitive 2D" appeared which I did not create. It causes problems because the black arrows that indicate inlet and outlet are now in absolutely weird positions like at the outer walls. So I guess that my simulation will give me wrong solutions.


How do I get rid of these strange things just because my mesh is too fine?


Thank you

[ghorrocks]

You need to look at the mesh in ICEM and check the boundary patches are correctly defined and the mesh is joined together correctly.

[attiquejavaid08]

Originally Posted by ghorrocks

You need to look at the mesh in ICEM and check the boundary patches are correctly defined and the mesh is joined together correctly.

i need to ask here, sometimes a whole surface is defined in one part, say wall1. let's say there is another perpendicular surface which is circular in shape, it's also a wall but defined in some other part (say wall2) for post-processing point of view. but when we create mesh (converting hexa to unstruc) in ICEM, a meshed part of wall2 partially appear in wall1.

why does it happen in ICEM? what's remedy? please

AJ

[snejball91]

Originally Posted by ghorrocks

You need to look at the mesh in ICEM and check the boundary patches are correctly defined and the mesh is joined together correctly.

For me it seems like everything is correct. I just don't get it.
With bigger cells the problems disappear but I think I need the small sizes of cells to get correct results.

edit:
This is how the mesh looks like with the problematic "primitive 2d"-domains.
http://i.imgur.com/hdrhwyy.png

[singer1812]

Did you run Check Mesh in ICEM? I bet you might have a volume orientation issue or such.

[snejball91]

Originally Posted by singer1812

Did you run Check Mesh in ICEM? I bet you might have a volume orientation issue or such.

You are right. I have a problem with volume orientations but my mesh quality is too bad after refining the mesh, so ICEM can't fix it. ("elements could not be fixed -- probably the element quality is too bad.")

My fault: I checked the mesh quality at the beginning of creating the mesh with big cells and it was good. I thought it will stay in that quality no matter how fine I make the mesh.

But to be honest, I don't know how I could solve that whole problem. The geometry is quite simple and I have no idea how to get better mesh quality with such small mesh sizes.

[singer1812]

I take it that this is a hex mesh? You increased the divisions such that you have a very small first layer height?

I recall running into this a while back (years ago). I did solve the issue, but dont recall what I did.

If I run into it, I will post the solution.

[snejball91]

Originally Posted by singer1812

I take it that this is a hex mesh? You increased the divisions such that you have a very small first layer height?

I recall running into this a while back (years ago). I did solve the issue, but dont recall what I did.

If I run into it, I will post the solution.

Yes, it is a hexa mesh and I want to get very small size for the boundary layer.

Thank you so far.

[snejball91]

I found a reason for my bad quality in determinant 3x3x3 and angles.

On the first boundary layer cell ICEM creates triangular cells instead of foursquare cells.

http://i.imgur.com/h8JGn2V.jpg