Hello community!
Currently I'm working on a 2D model of human vocal folds in COMSOL by using the MEMS – toolbox. The problem is very similar to that presented within the FSI tutorial. The only differences are:
1) the beam is a layered / composite structure
2) the fluid is air (instead of water like fluid)
3) under certain conditions (large deformations) the beam may collide with the opposite wall dividing the fluid domain in two parts
4) the opposite wall fulfils the function of a symmetry line, namely, slip can occur along it
I use contact pairs for contact modelling, fully coupled solver, ALE method and automatic re-meshing. Nevertheless, I do often get convergence and re-meshing issues, especially when big deformation of the solid structure occur.
Is COMSOL an appropriate software to model the problem mentioned above by using only the standard toolboxes (w/o additional programming effort)?
Is COMSOL able to simulate FSI and contact within the same domain?
What must attention be paid on when modelling such kind of a problem?
Many thanks in advance and with kind regards
AW
https://www.comsol.de/model/fluid-structure-interaction-361
Conference Paper Numerical simulation of videokymographic images from the res...
Dear Alexander,
regarding your first question,yes COMSOL is an appropriate software to model your problem as the large deformations with geometirc nonlinearity is enabled in the structural analysis Module with multiphysics Mechanical Contact enabled. which means that your second question is also answered.
Moreover, the beam can be defined either by isotropic or composite material in COMSOL.
You have to pay attention to the coupling time-step as the convergence highly affected by it.
I hope that would help you
Best regards
Mohamed
Dear Alexander,
I have some experience with FSI.
Here, are my points:
Regarding to contact, I suggest you may need to use a constraint to hold the mesh or surface near the opposing wall as when the beam reaches to a surface the interface mesh goes smaller and smaller until it diverge.
Regarding to re-meshing you need to set re-meshing on general quality and use values such as 10 or larger to get good results (this works for me most times).
As your model is somehow complex more details or figures of your problem or errors could help the readers to provide better help.
Hello Mohamed, hello Mohammad,
thanks for quick response! I've attached a simulation with glottal half width of approximatelly 0.25mm. At this conditions no collision occur as well as a self excited vibration with a stable limit cycle. This problem converges fast, even if I use triangular elements and a coarse mesh for the whole domain.
I've also attached a link to a paper from czech researchers. They solve the problem by using ANSYS 15 but with some programming effort (they implemented the contact algorithm and fsi by hand). I'm trying to solve the symmetrical problem by using standard toolboxes in COMSOL only.
I should bring up that the structure is very flexible / compliant. The Young's moduli of the layers are between 3 [kPa] and 65 [kPa].
With kind regards, Alexander
I do not have too much experience with this, but regarding 3) you may be able to use Events Interface https://www.comsol.com/blogs/implementing-a-thermostat-with-the-events-interface/. I have seen it demonstrated to model a ball bouncing around a 2D box-domain.
hello,
how did you go about the glottal half with definition pls? i'm currently facing a similar fsi-contact issue as shown the attached image
anyone with any suggestions pls. the simulation would not just converge once those two membranes touch each other under a high inlet flow pressure
- Badges
- Science topic