Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations waross on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Meshing Issues using Tet elements

Status
Not open for further replies.

grnlan03

Mechanical
Feb 1, 2010
30
0
0
US
I have been having trouble getting two annular parts to mesh. I am trying to produce a mesh using tet- elements which will be used in a thermal-mechanical analysis. I can mesh the model when the seed size is large but when I lower it to around 0.005 the model will not mesh and I get an error message that says there was a problem with the boundary triangles in the model. I believe theat it has something to do with the partitions I have applied to several surfaces and cells in the parts.
My questions are:
1. Is there something I am missing in the model that will not allow it to mesh? I.e. a control or setting that I should be changed?
2. I need to apply a surface heat flux to only part of the model which is why I have done the partitioning. Is there a different way to do this without partitioning that will allow me to still only apply a heat flux to part of a surface?
I have attached the .cae file with the seeds and partitions that will not mesh.
Thanks
 
Replies continue below

Recommended for you

If you mesh with a smaller seed it can create the mesh. With 0.005 the mesh is very coarse and would probably yield poor results. I tried 0.004 which was still coarse but could mesh.

Since this is basically coaxial cable I would suggest partitioning with the X and Y planes and then you can mesh with a structured hex mesh.

I hope this helps.

Rob Stupplebeen
 
Ohh I was actually using 0.0005 not 0.005. That was my mistake in the post. The .cae file should be 0.0005 unless I messed that up as well. Could you maybe look at with the 0.0005 seed value?

I'll also look into the Hex mesh.

Thanks for the help.
 
I can't open the file as it's 6.9.EF-1 or something. Not knowing the boundary conditions though, isn't this a plane strain problem? At the very least, using a hex mesh will halve the number of elements.

Tata
 
Rob,
When you partitioned the model on the X and Y plane did you partition the cells in the models or the surfaces? I assume it’s the cells but just want to make sure.

I know the seed size is a bit of an over kill but I'm trying to get mesh independence so I'm just taking it down to explore the numbers. In the final model I should be using something more course.

Will the structured hex mesh have problems with the curved nature of the cylinders? I know there is a curvature control in the seed part dialogue box and I believe this is what controls that. Do you know if that is correct or not?

Corus,
Sorry you could not open the model but sounds like I need to try hex elements.

Thank you all for your help
 
I partitioned the cells.

I usually start coarse and then refine once everything is working.

I did not see any issues with the hex mesh. I believe the curvature control is to add more elements were curvature is changing rapidly.

I hope this helps.

Rob Stupplebeen
 
Partitioning it along the X and Y plane helped a lot, I was able to use a structured Hex-mesh.

I was wondering how you knew to use that approach? I have only been using Abaqus since the beginning of the year and my experience is a little limited. I would like to know for future models.

I am starting a new thread related to convergence issues I am having in a model. It would be great if either one of you could take a look at it.

Thanks for all of the help.
 
Meshing is one of the 'art' areas of FEA. Basically you are trying to help the auto mesher recognize prismatic shapes. By cutting the tube it created topology with 6 sides similar to a cube. I hope this helps.

Rob Stupplebeen
 
Status
Not open for further replies.
Back
Top