CFD tetramesh Boundary layer bug circular pipe
Hi,
I am using hypermesh to generate CFD tetramesh and boundary layer to model a flow into pipes. I am first working on a simple pipe with an inlet and an outlet. I am using the cfd tetra mesh panel to generate the boundary layer and the tetramesh core.
The boundary layer obtained is good for the first couple layers but the elements start to skew at some point which result in a very bad boundary layer mesh. This is extremely disappointing since this is one of the simplest case I can think of. I am planning to use the cfd mesher for a much complex piping case and this kind of mesh is not acceptable. Please help me with this issue. I used hypermesh 12.0.117.
I attached a screenshot of the parameters used for the boundary layer. I also tried several other combination of the parameters. I also tried selecting float or fixed boundary layer without success. I also attached screenshots of the resulting elements to illustrate the problem. I also attach a .txt file containing the model of the pipe. (just change de extension to .fem and import it in hypermesh, I had to change extension otherwise the upload was not authorized). The pipe and inlet/outlet components are used as 2d input mesh for the boundary layer mesh generation.
Thanks in advance for your help, I really need to make this work!
<?xml version="1.0" encoding="UTF-8"?>
Answers
-
-
Hello tinh!
Thanks for your fast reply!
I did select the two tips as 'W/o BL' in the model I provided. Actually it is mandatory else the inlet and outlet would also have a boundary layer.
I checked the elem offset panel and it provides pretty nice tools to do the job. I never explored this panel, thank you for pointing it out! It does the job for a simple case but I doubt it will do the job for what I want to really do within hypermesh. I have a tcl code that automatically mesh a complex piping using the cfd tetra mesh. It seems like the elem offset panel does not provide simple tool to have a tetra meshabe volume representing a fluid inside a complex pipe.
I found by trial and error that the behavior I mentionned is reduced as the number of elements in the circumference of the pipe is increased. The skew in the elements completely disappears when the pipe has more than 16 elements in its circumference. I guess that there is something in the boundary layer algorithm that is sensitive. Can anyone try the same operation in hypermesh 13.0 to see if the behavior was corrected. I don't like it because you cannot see this behavior on the surface of the mesh and you need to look inside the pipe to see it. I somehow lost my trust in the cfd tetra mesher since I don't know the real source of this behavior.
Thanks again for your help! I also welcome any other potential explanations to this problem!?
0