CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   CFX error message (http://www.cfd-online.com/Forums/cfx/115781-cfx-error-message.html)

M_Tidswell April 6, 2013 14:04

CFX error message
 
Dear All

I am a relatively new user to Ansys CFX/ICEM and am getting a repeated error message from CFX when I import the mesh from ICEM - Hexahedral element 71277 will degenerate into an invalid element.

The number changes depending on what combination of mesh geometry I have tried, it seems to occur when I try make the first boundary layer element sub 9e-6.
Is there a setting I have missed in ICEM or something I need to do in CFX to accept smaller mesh geometry?
CFX is happy to accept a mesh with the same geometry generated in Ansys Mesh, I'm sure I must be missing something.

Hopefully someone can point me in the right direct

Many thanks

Mark

ghorrocks April 7, 2013 06:28

Have a look at element 71277 in ICEM. You will find it has degenerated to a line or a flat thing. Then either fix it (you can move individual nodes in ICEM), remesh so it does not occur or delete the element.

M_Tidswell April 7, 2013 06:40

Thank you for such a swift response, may I intrude on your time once again and ask if there is a good practice that one should consider in order to avoid this error? I take it increasing the growth ratio perpendicular to the wall should address the issue as this would ensure that the element had greater depth?

Many thanks

Mark

ghorrocks April 7, 2013 06:44

It has been a while since I used ICEM but I seem to recall it have mesh validity checking stuff which would identify these types of problems before the mesh was written. You should run those before you write a mesh.

I also recall it was easy to smooth a mesh into oblivion with ICEM. By this I mean the smoothing increases the average quality of the elements, but a very small number of elements gets smoothed until they are either flat or inside out - which means the mesh is invalid. So be careful with mesh smoothing.

M_Tidswell April 7, 2013 07:09

ICEM did flag up that the aspect ratio within the boundary layer was particularly poor, I know that this must be expected to an extent but I assume it was rather worse than I thought. Are there 'golden' rules one should follow to minimise this - certain growth rates should be used for particularly small first element sizes or the such?

Sorry I promise this is the last question

Many thanks

Mark

ghorrocks April 7, 2013 07:39

Firstly - there is no need to apologise, and feel free to ask as many questions as you wish. As long as your questions are interesting and intelligent then you are likely to get people interested in what you are doing and get useful answers.

When it flags "poor" aspect ratio does it just mean you have a few bad quality elements (like aspect ratio 10 or something like that) or that you have aspect ratio 0 elements (ie elements which have been flattened to oblivion). Many simulations will still run with a few bad quality elements. But a mesh with a single flattened element might fail so this would have to be fixed.

M_Tidswell April 8, 2013 08:03

Dear Glenn

Thank you for all your assistance, I think a bit more playing with ICEM is in order. I think I still prefer it to Ansys mesh tho!

I have another interesting stumbling block regarding error quantification for SAS - SST models but I think that would better sit in a different thread.

Many Thanks

Mark

ghorrocks April 8, 2013 08:07

ANSYS meshing works nicely when little user intervention is required. It gets very annoying when you are trying to craft a mesh manually. ICEM has a long learning curve but is much more flexible in the hands of an experienced user. You need to play around with ICEM for a while and try lots of options before you find something which works for your setup.


All times are GMT -4. The time now is 09:06.