Menu

Discussion Forum

New Discussion
フィルター

Note: This discussion is about an older version of the COMSOL Multiphysics® software. The information provided may be out of date.

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.

fsi problem

Posted 2010/08/03 6:23 GMT-4 Version 4.1 19 Replies

Maral Hassani
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

hi
My model is a 3Dcylinder with a thickness(the walls are the solid domain and inside the cylinder is fluid domain where the fluid flows). The inlet boundary is a sin function(0.219+0.132sin(18.2t)),I wrote this in scalar experession and then used it in the inlet>laminar velocity boundary.The outlet is pressure outlet(4140pa).I chose the MEMS module>Fluid-structure interaction>solid,stress-strain..>transient to consider fsi..but my model didnot converge..I solved it before with anouther inlet velocity function(U^2/*(sqrt(t^4)+0.07t^2+0.0016) with no problem.
Does anyone know where is my problem?it may be at defining the sin inlet function!?how can I solve it?
please help me...
God bless you
maral
19 Replies Last Post 2011/06/21 15:30 GMT-4
Danial Tz
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/03 7:52 GMT-4

hi
My model is a 3Dcylinder with a thickness(the walls are the solid domain and inside the cylinder is fluid domain where the fluid flows). The inlet boundary is a sin function(0.219+0.132sin(18.2t)),I wrote this in scalar experession and then used it in the inlet>laminar velocity boundary.The outlet is pressure outlet(4140pa).I chose the MEMS module>Fluid-structure interaction>solid,stress-strain..>transient to consider fsi..but my model didnot converge..I solved it before with anouther inlet velocity function(U^2/*(sqrt(t^4)+0.07t^2+0.0016) with no problem.
Does anyone know where is my problem?it may be at defining the sin inlet function!?how can I solve it?
please help me...
God bless you
maral


Can you write the dimensions of your problem (to have an estimate of Re), and a clear second inlet you used?

U^2/*(sqrt(t^4)+0.07t^2+0.0016 is not correct, apparently. Also, how much is U?
[QUOTE] hi My model is a 3Dcylinder with a thickness(the walls are the solid domain and inside the cylinder is fluid domain where the fluid flows). The inlet boundary is a sin function(0.219+0.132sin(18.2t)),I wrote this in scalar experession and then used it in the inlet>laminar velocity boundary.The outlet is pressure outlet(4140pa).I chose the MEMS module>Fluid-structure interaction>solid,stress-strain..>transient to consider fsi..but my model didnot converge..I solved it before with anouther inlet velocity function(U^2/*(sqrt(t^4)+0.07t^2+0.0016) with no problem. Does anyone know where is my problem?it may be at defining the sin inlet function!?how can I solve it? please help me... God bless you maral [/QUOTE] Can you write the dimensions of your problem (to have an estimate of Re), and a clear second inlet you used? U^2/*(sqrt(t^4)+0.07t^2+0.0016 is not correct, apparently. Also, how much is U?
Maral Hassani
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/03 9:36 GMT-4
Danial my model, that is an article work(they worked with ansys) and I want to validate this model with comsol and then go to my own model, is attached.The boundary conditions are
inlet: velocity(.219+.132sin(18.2t))
outlet:pressure outlet(4140pa)
when I couldnt get result with this,I switched to a simple cylinder because I guessed that the problem may be because of inverting mech in the stenosis region.
but even in cylinder it didnt solved.
I changed the velocity function to [(U^2/sqrt(t^4-.07t^2)+.0016 and U=.033m/s]and pressure outlet to zero and it converges.but these changes are just to find the problem!!(indeed I didnt)
you know I think It is because Im not good familiar with comsol.
I need your help to continue please...
Danial my model, that is an article work(they worked with ansys) and I want to validate this model with comsol and then go to my own model, is attached.The boundary conditions are inlet: velocity(.219+.132sin(18.2t)) outlet:pressure outlet(4140pa) when I couldnt get result with this,I switched to a simple cylinder because I guessed that the problem may be because of inverting mech in the stenosis region. but even in cylinder it didnt solved. I changed the velocity function to [(U^2/sqrt(t^4-.07t^2)+.0016 and U=.033m/s]and pressure outlet to zero and it converges.but these changes are just to find the problem!!(indeed I didnt) you know I think It is because Im not good familiar with comsol. I need your help to continue please...

Attachments:

Danial Tz
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/03 12:05 GMT-4
again, you have written the inlet partially correct.

[(U^2/sqrt(t^4-.07t^2)+.0016

Do you mean:

Case A : U^2/(sqrt(t^4-.07t^2)+.0016)
Case B: (U^2/sqrt(t^4-.07t^2))+.0016

The important part is that you check that at t=0 you have a reasonable inlet velocity. It is possible that it is too high. With the sin function you have at t=0, u=.219, and Re of 340 roughly. In my simulations it will not converge. So, I have to ramp the velocity first. Even if you have a Sin inlet you can still ramp your flow:

U=(0.219+0.132sin(18.2t))*flc2hs(t-0.1,.1)

Just try it. If not we can move forward.
again, you have written the inlet partially correct. [(U^2/sqrt(t^4-.07t^2)+.0016 Do you mean: Case A : U^2/(sqrt(t^4-.07t^2)+.0016) Case B: (U^2/sqrt(t^4-.07t^2))+.0016 The important part is that you check that at t=0 you have a reasonable inlet velocity. It is possible that it is too high. With the sin function you have at t=0, u=.219, and Re of 340 roughly. In my simulations it will not converge. So, I have to ramp the velocity first. Even if you have a Sin inlet you can still ramp your flow: U=(0.219+0.132sin(18.2t))*flc2hs(t-0.1,.1) Just try it. If not we can move forward.
Maral Hassani
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/03 15:09 GMT-4
I tried it.in cylinder model the solution does 13% and after that it diverges and reciprocal of step sizes go to ...e11 and fluctuate in this limit.
but in main model(stenosis) it diverges from the first...
what should I do Danial??
I tried it.in cylinder model the solution does 13% and after that it diverges and reciprocal of step sizes go to ...e11 and fluctuate in this limit. but in main model(stenosis) it diverges from the first... what should I do Danial??
Danial Tz
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/04 4:06 GMT-4

I tried it.in cylinder model the solution does 13% and after that it diverges and reciprocal of step sizes go to ...e11 and fluctuate in this limit.
but in main model(stenosis) it diverges from the first...
what should I do Danial??


OK. One step ahead.

Now, there can be many possibilities. How about you upload your model and we take a look?
If you are already solving for FSI, did you try to solve it only for fluid, and no other physics? once the fluid is converged the rest of physics will follow.
[QUOTE] I tried it.in cylinder model the solution does 13% and after that it diverges and reciprocal of step sizes go to ...e11 and fluctuate in this limit. but in main model(stenosis) it diverges from the first... what should I do Danial?? [/QUOTE] OK. One step ahead. Now, there can be many possibilities. How about you upload your model and we take a look? If you are already solving for FSI, did you try to solve it only for fluid, and no other physics? once the fluid is converged the rest of physics will follow.
Maral Hassani
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/04 12:24 GMT-4
yes I solved for the fluid with rigid walls with no problem.
unfortunately I couldntt upload my model..it give me a message about license..I dont know how to get it.
any idea Danial please!!
yes I solved for the fluid with rigid walls with no problem. unfortunately I couldntt upload my model..it give me a message about license..I dont know how to get it. any idea Danial please!!
Danial Tz
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/04 15:53 GMT-4

yes I solved for the fluid with rigid walls with no problem.
unfortunately I couldntt upload my model..it give me a message about license..I dont know how to get it.
any idea Danial please!!


You don't have to upload it here. You can use file sharing services and paste the link here.

Hint: lifehacker.com/388284/best-online-file-sharing-services
[QUOTE] yes I solved for the fluid with rigid walls with no problem. unfortunately I couldntt upload my model..it give me a message about license..I dont know how to get it. any idea Danial please!! [/QUOTE] You don't have to upload it here. You can use file sharing services and paste the link here. Hint: http://lifehacker.com/388284/best-online-file-sharing-services
Maral Hassani
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/04 23:45 GMT-4
Thanks Danial
here is the file: www.mediafire.com/?4jhr7opo3a2k4kd
Im waiting...
Thanks Danial here is the file: http://www.mediafire.com/?4jhr7opo3a2k4kd Im waiting...
Danial Tz
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/05 3:31 GMT-4

Thanks Danial
here is the file: www.mediafire.com/?4jhr7opo3a2k4kd
Im waiting...


OK, here we go:

1. Abs tol set to 1e-5, relative to 1e-3
2. Redo your geometry with Comsol lines. There was something wrong with it, so I redid it.
3. If you want to use mapped mesh, then mesh as follow by dividing your domain into subdomains. You should do a finer meshing of course than what I did.
cl.ly/30d3aeacbf23b0517672

4. cl.ly/33823d72905351bdc895

have fun!
Danial
[QUOTE] Thanks Danial here is the file: http://www.mediafire.com/?4jhr7opo3a2k4kd Im waiting... [/QUOTE] OK, here we go: 1. Abs tol set to 1e-5, relative to 1e-3 2. Redo your geometry with Comsol lines. There was something wrong with it, so I redid it. 3. If you want to use mapped mesh, then mesh as follow by dividing your domain into subdomains. You should do a finer meshing of course than what I did. http://cl.ly/30d3aeacbf23b0517672 4. http://cl.ly/33823d72905351bdc895 have fun! Danial
Maral Hassani
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/05 7:23 GMT-4
Dear Danial
I appreciate u..
I will try it..but if no matter whould u please send me the model u made changes?becouse I have some problem to draw geometry with lines and also to submaped mesh...
best regards
Dear Danial I appreciate u.. I will try it..but if no matter whould u please send me the model u made changes?becouse I have some problem to draw geometry with lines and also to submaped mesh... best regards
Danial Tz
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/06 6:19 GMT-4

Dear Danial
I appreciate u..
I will try it..but if no matter whould u please send me the model u made changes?becouse I have some problem to draw geometry with lines and also to submaped mesh...
best regards


Dear Maral,

Sorry, but I did the model on fly without saving. It was pretty easy to draw your model with lines and also mapped mesh. There are very nice tutorials in Comsol docs to learn from.
[QUOTE] Dear Danial I appreciate u.. I will try it..but if no matter whould u please send me the model u made changes?becouse I have some problem to draw geometry with lines and also to submaped mesh... best regards [/QUOTE] Dear Maral, Sorry, but I did the model on fly without saving. It was pretty easy to draw your model with lines and also mapped mesh. There are very nice tutorials in Comsol docs to learn from.
Maral Hassani
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/06 7:54 GMT-4
ok..
thanks again
ok.. thanks again
Maral Hassani
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/07 7:40 GMT-4
Hi Danial
sorry but I did all the changes as u said but again Im in trouble...no convergence
would u please check my file please: www.mediafire.com/download.php?7zaav4p70k0fa45
maybe its the problem with comsol3.5!!!
I got confused..
Hi Danial sorry but I did all the changes as u said but again Im in trouble...no convergence would u please check my file please: http://www.mediafire.com/download.php?7zaav4p70k0fa45 maybe its the problem with comsol3.5!!! I got confused..
Maral Hassani
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/09 10:59 GMT-4
Hi Danial
sorry but I did all the changes as u said but again Im in trouble...no convergence
would u please check my file please: www.mediafire.com/download.php?7zaav4p70k0fa45
maybe its the problem with comsol3.5!!!
I got confused..

Hi Danial sorry but I did all the changes as u said but again Im in trouble...no convergence would u please check my file please: http://www.mediafire.com/download.php?7zaav4p70k0fa45 maybe its the problem with comsol3.5!!! I got confused..
Danial Tz
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/09 12:27 GMT-4
Salam Maral,

Your model looks fine. To my guess, the problem lies in the model definition of yours. Normally, as I said before, setting pressure does not mean much for Navier-stokes, as dP is important. BUT, not in the case when you deal with FSI, as fluid pressure exerts forces on the structure. Just to see what I mean, run a simulation at steady-state with your max velocity. You will see that The vein nearly explodes, due to very high internal forces. Either you have assigned a very low E, or you should rethink what you're doing with your boundary conditions. If you have a paper reference, go through it and see if you've missed a detail (and maybe post the ref here).

movaffagh bashi,
Danial
Salam Maral, Your model looks fine. To my guess, the problem lies in the model definition of yours. Normally, as I said before, setting pressure does not mean much for Navier-stokes, as dP is important. BUT, not in the case when you deal with FSI, as fluid pressure exerts forces on the structure. Just to see what I mean, run a simulation at steady-state with your max velocity. You will see that The vein nearly explodes, due to very high internal forces. Either you have assigned a very low E, or you should rethink what you're doing with your boundary conditions. If you have a paper reference, go through it and see if you've missed a detail (and maybe post the ref here). movaffagh bashi, Danial
Maral Hassani
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/10 1:05 GMT-4
salaam
but you could solve it as u sent me the animation...???
salaam but you could solve it as u sent me the animation...???
Danial Tz
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/10 2:09 GMT-4

salaam
but you could solve it as u sent me the animation...???


I am not sure that coupling in that model was right. The displacement was 1e-8, practically nothing, and especially I looked at the flow only.

Danial

ps. which paper you're trying to model?
[QUOTE] salaam but you could solve it as u sent me the animation...??? [/QUOTE] I am not sure that coupling in that model was right. The displacement was 1e-8, practically nothing, and especially I looked at the flow only. Danial ps. which paper you're trying to model?
Maral Hassani
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2010/08/10 9:57 GMT-4
this is the paper Danial..
thank you very much for your attention..
this is the paper Danial.. thank you very much for your attention..

Attachments:

Hamid Ghasemi
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011/06/21 15:30 GMT-4
salam danial aziz
dar iran aziz man kasi ro peyda nakardam dar comsol b man komak kone.

I am trying to simulate natural heart valve in 2D fsi mode. when leaflets contact an error occurred.

Repeated error test failures. May have reached a singularity.
Time : 0.002592892726514386
Last time step is not converged.

When I define “contact pair” in definition and add “contact 1” in fsi node , I get error:

Failed to find consistent initial values.
Last time step is not converged.

How can I use contact pair? Or is there another way to modeling this example?
Please help me
Thank you so much!


salam danial aziz dar iran aziz man kasi ro peyda nakardam dar comsol b man komak kone. I am trying to simulate natural heart valve in 2D fsi mode. when leaflets contact an error occurred. Repeated error test failures. May have reached a singularity. Time : 0.002592892726514386 Last time step is not converged. When I define “contact pair” in definition and add “contact 1” in fsi node , I get error: Failed to find consistent initial values. Last time step is not converged. How can I use contact pair? Or is there another way to modeling this example? Please help me Thank you so much!

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.

Suggested Content