CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   STAR-CCM+ (http://www.cfd-online.com/Forums/star-ccm/)
-   -   CCM+ Abaqus FSI coupling (http://www.cfd-online.com/Forums/star-ccm/90048-ccm-abaqus-fsi-coupling.html)

stumpy June 29, 2011 09:25

CCM+ Abaqus FSI coupling
 
Hi all,
I'm a long time FSI user of CFX with ANSYS, but am looking at CCM+ due to a need to couple with Abaqus. After reading the info on the STAR website I'm left with a few questions on this capability.
- How does this capability scale, can both codes run in parallel?
- Is the force transfer to Abaqus conservative?
- Has anybody used this? Good/bad experiences?
Thanks

abdul099 July 10, 2011 06:23

I can just answer the first question. As both codes run independendly and just exchance information after every (or a few) time steps / iterations, both should keep all capabilities. ccm+ is able to run in parallel and I assume, abaqus as well.

daveatstyacht July 25, 2011 06:36

My experience with the two together is that they can both run in parallel, however some operations in abaqus appear to be performed on a single thread. Getting the two programs to start talking for a cosimulation is pretty straight forward with a decent example provided in the starccm+ help files of setting up and running them together. They seem fairly robust so long as the geometries are reasonably close matches and you adhere to the rules listed in star's help file (namely not using hybrid tet or mixed elements in abaqus). Also it is unclear to me if double sided shells can be included (ie sails). If you are doing structural deformation you will probably find the morphing of the domain in star is the slowest part of the process.

My biggest issues with the cosimulation process are that all data transfer happens in the global lab coordinate system (problematic when one wants rigid body motion in addition to deformation) and that as far as I know it is impossible as far as I know to get star to perform the exchanges less frequently than every time step (which means I am performing far more FEA steps and attendant morphing steps than I would like). Overall though I think it is a pretty good setup since the nodes of each domain need not be conformal and the process is reasonably robust considering.

smak.quadri March 1, 2012 14:28

Need help in FSI,
 
dear stumpy,

i always see your reply on most of FSI related problems, but still i am stuck in my prob from 1 year.

May i Know your maild id so that i can mail u in details my problems. Please help me out , I have to submit my thesis by april.

FSI coupling b/w ANSYS CFX and workbench input file is happening wit no prob for a simple straight tube with time varying velocity at inlet and time varying pressure at outlet.

But the structural deformation is not in radial directio, but it looks like there is a bending in x direction. i tired with cylindrical coordinate system, but stil its same prob.

I feel there is not proper mapping of nodes at interface. The mesh is fine. I tired contacting many people, but my prob still remains the same.

Please help me out and mail me on smak.quadri@gmail.com


All times are GMT -4. The time now is 11:43.