Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.

Fluid Structure Interaction - Important

Please login with a confirmed email address before reporting spam

I am trying to simulate the harmonica, where the fluid load causes the beam to vibrate. I am using the Fluid-Structure-Interaction module from MEMS for simulation. I am doing a 2D simulation with plane-strain for sturcture, incompressible navier-stokes for fluid and ale for moving mesh. I am doing it the same way as the example in MEMS module does, however I changed the constants and dimensions to suit my model. The problem I face is that the solution doesnot converge or the solver exits saying inconsistent initial conditions. If I reduce the dimensions to suit the dimensions given in example it solves but stops in middle after 0.04 time step again giving same errors. I see the result and beam only bends but never vibrates, and it is same the example also even if you increase inlet velocity. The beam however vibrates if your inlet velocity is more than the threshold value. I am unable to find an answer for this question. Plz help me, I am struggling a lot with this and unable to find the answer. Is this fsi mems model capable of producing vibrations in beam o should i use another model. Thanks in advance and hope someone helps me out.

4 Replies Last Post 2010年2月4日 GMT-5 11:42

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010年1月31日 GMT-5 08:15
What is your Re?
some screenshots of the model or the mesh may help as well.
What is your Re? some screenshots of the model or the mesh may help as well.

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010年1月31日 GMT-5 11:45
I am attaching an image of my problem. This is the solution when I reduce the dimensions of my model, it exits after solving for 0.04 seconds. It exits for first time step, if I use orginal dimensions.
I am attaching an image of my problem. This is the solution when I reduce the dimensions of my model, it exits after solving for 0.04 seconds. It exits for first time step, if I use orginal dimensions.


Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010年2月4日 GMT-5 04:13
Hi,
did you check relative and, above all, absolute tolerances, in your time dependent solver?
It is often better to set absolute tolerance for each variable you solve for. Within Comsol manuals (time dependent solver part), you see how to do this.
Please always set an absolute tolerance with is, at least, one order of magnitude lower then the magnitude of your variable, so that the time dependent scheme can go on safetly.
Clearly you have to know in advance, roughly, what is the magnitude of your displacements, velocities, and so on...
Moreover,you have also to check time step. If you want to correctly simulate vibration, you have to use a time step which at least about 1/10 of the period of your vibration. If you use therefore a too large time step, the time dependent scheme will probably fail, since the solver is not able to satisfay all tolerances within the time step you request.
I saw from your pic, that you are using the ALE application mode.
This is correct, but you are just interested in the *linear* vibration of your structure, you can avoid such a module and save computational cost.
Finally, please, be always aware that Comsol tutorials have always settings which are customized according to the problem under analysis. For instance, if you scale down the dimensions of the solid, it will be probabily vibrate at higher frequencies and therefore the default time step suggested by Comsol will be no more valid for your case.
I hope this helps.

Alessandro
Hi, did you check relative and, above all, absolute tolerances, in your time dependent solver? It is often better to set absolute tolerance for each variable you solve for. Within Comsol manuals (time dependent solver part), you see how to do this. Please always set an absolute tolerance with is, at least, one order of magnitude lower then the magnitude of your variable, so that the time dependent scheme can go on safetly. Clearly you have to know in advance, roughly, what is the magnitude of your displacements, velocities, and so on... Moreover,you have also to check time step. If you want to correctly simulate vibration, you have to use a time step which at least about 1/10 of the period of your vibration. If you use therefore a too large time step, the time dependent scheme will probably fail, since the solver is not able to satisfay all tolerances within the time step you request. I saw from your pic, that you are using the ALE application mode. This is correct, but you are just interested in the *linear* vibration of your structure, you can avoid such a module and save computational cost. Finally, please, be always aware that Comsol tutorials have always settings which are customized according to the problem under analysis. For instance, if you scale down the dimensions of the solid, it will be probabily vibrate at higher frequencies and therefore the default time step suggested by Comsol will be no more valid for your case. I hope this helps. Alessandro

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010年2月4日 GMT-5 11:42
Thank you very much, I will try the things you told.
Thank you very much, I will try the things you told.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.