CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Remeshing in FSI problem (

tctczxh February 18, 2011 17:17

Remeshing in FSI problem
If CFX is used alone,the remeshing options can be set in
Simulation Control -> Configurations

however in FSI project created in workbench, the configurations submodule is disabled.

Does it mean automatic remeshing is still not possible in this kind
of FSI problem?

tctczxh February 19, 2011 00:00

hopefully there is a way to use ICEM CFD to remesh in FSI problems:)

singer1812 February 21, 2011 16:26

FSI with remesh is not supported.

It can be done outside of WB using alot of Mechanical APDL and advanced modeling techniques that I care not to go into with someone not expert in both CFX and ANSYS APDL.

tctczxh February 22, 2011 14:28

Currently I deal with those FSI problems in workbench by enabling the mesh deformation option.

actually by adjusting the mesh stiffness and refining the mesh, the simulation works in some cases, but it's quite limited. the mesh quality degrades continuously then the solver is likely to be terminated by a folded mesh error.

so I'm confused that what should I do if I really need some structure displacement in FSI problems?

stumpy February 22, 2011 17:15

There are certainly some cases that can't be solved with 2-way FSI because re-meshing is not really available. However there's a lot that can be done using mesh deformation. You can add sliding interfaces, control the stiffness, add subdomains with defined motion etc. Do you have some images of your initial and deformed mesh?

vmlxb6 February 28, 2011 11:41

heres the link for re-meshing:

@ Singer1812: Why do you say that re-meshing is not possible in FSI Problem ?
@ Kesife: There are a couple of post where people have addressed the issue of folding mesh. It can be solved by reducing time step, better converging of displacement Eq. etc etc Even I had folding mesh issues but that was because I was using unrealistic values. For structural deformation, try 2-was FSI

@ ALL : Looks like you guys are pretty familiar with FSI. Am modeling a one way FSI for a flow over a cylinder and have a few questions. I will post it in a new thread. I hope to get them answered.

Thank you,

Prof. Chaos

singer1812 February 28, 2011 11:45

I didnt say remeshing is not possbile, remeshing in FSI witin ANSYS is not supported.

You will have to use alot of APDL and model manipulation in order to get a remeshed FSI problem to work.

singer1812 February 28, 2011 11:49

Ahh, I see where the confusion is.

I took this FSI question to be a case where structural deformation was the underlying cause of the fluid mesh movement, with a coupled structral FEA and CFD analysis used.

Just predicting rigid body motion in CFD can use remeshing with no issues.

vmlxb6 February 28, 2011 11:58

Ahh...........I get it now.
Correct me if am wrong.

Re-meshing in a One way FSI is possible, where the motion of the rigid body motion can be predicted.

But Re-meshing in a 2 way FSI (structural deformation) is difficult.

Am I right ??????????

BTW wats APDL ????

singer1812 February 28, 2011 12:07

What you said is correct.

APDL is the programming (scripting) language that the mechanical side (finite element) of ANSYS uses.

It stands for ANSYS Parametric Design Language

vmlxb6 February 28, 2011 12:14

Cool !!!!!!!!! Thanks Dude........... have a nice day.

Another favor: I am trying to model a 1-way FSI problem of a flow over a cylinder. Am getting very low values of displacement. I will be posting a thread. I hope you could answer a few of them.

gwen520 September 25, 2014 11:10

hello,vmlxb6,recently,I tried the example posted above

but after lots of trails, I still failed.Did you ever do that example? if you do,can you give your CFX file to me for reference. my email is
thank you very much!

gwen520 September 25, 2014 11:11

hello,singer1812,recently,I tried the example posted above

but after lots of trails, I still failed.Did you ever do that example? if you do,can you give your CFX file to me for reference. my email is
thank you very much!

All times are GMT -4. The time now is 00:12.