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 The information provided may be out of date.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.

Model coupling

Please login with a confirmed email address before reporting spam

Hi everyone

I just ran a model of a stirring tank with the Rotating Machinery physics. My next goal is to study the structural behavior of the impellers. Since this objective was not important during the development of the rotating model, I did not include the geometry of the impeller as a solid. The rotating model took 4 days to solve, so I don't want to run it again. Also, I don't have the fluid-structure interaction module, nor the structural mechanics module.
I thought I could add a second model in which the geometry would consist of the impeller only. With some kind of model coupling, I could "import" the total stress from the rotating model as boundary loads on the structural model. I tried the identity mapping model coupling, but when I try to solve it, I get a message saying that it didn't find a certain point on the source. Both geometries are on the same coordinates. Is this the correct type of coupling? Is this proposed methodology ok?
A second issue is that when adding the physics of structural mechanics, I don't have the option of having a stationary model, but a transient one. For this model's study, I just defined the same time range than in Study 1. Is this ok?
Any observations, suggestions and comments are very welcome!

Thank you!
Sylvana

1 Reply Last Post 2012年3月10日 GMT-5 04:12
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2012年3月10日 GMT-5 04:12
Hi

What I understand is that if you have run atransient model of the fuid, assuming the solid as fully rigid, then you should have the forces interacting on the surfaces. these are probably quite constant after some time in the rotation frame, so you need to extract those w.r.t. the geoemtry, as you state, to load your solid model in a stationary case. But to link the time series results to the stationary, you need to give as load values a given time to extract only one section of your data set array.

Even if a stationary solver is not proposed, you can define one manually, deselect the CFD and define the stationary to get loads from the previously calculated CFD case, UT defining only 1 given time step.

To make the tests easier, I would propose that you extract only the force lods in a table and import tht into a new model. Try it out on a simple canteliever beam model with only a few mesh elements to have it operating quickly and smoothly, then apply it to your mode complex model.

THis corresponds to y couple CFD + solid model using by ignoring the "inertial" terms, or in Quasi static, as the dynamic coupling between the fluid and the solid are not taken into account. Mostly valid for many studies.

You can anycase add a eignefrequency analsis of your solid, and then see how the fluid load (added surface mass) wil laffectthe eigenfrequency, just to know where these are, to compare with the oscillations you might see in your fluid model, if you have here a close match, it might indicate some true problems in real "fully coupled" life ;)

--
Good luck
Ivar
Hi What I understand is that if you have run atransient model of the fuid, assuming the solid as fully rigid, then you should have the forces interacting on the surfaces. these are probably quite constant after some time in the rotation frame, so you need to extract those w.r.t. the geoemtry, as you state, to load your solid model in a stationary case. But to link the time series results to the stationary, you need to give as load values a given time to extract only one section of your data set array. Even if a stationary solver is not proposed, you can define one manually, deselect the CFD and define the stationary to get loads from the previously calculated CFD case, UT defining only 1 given time step. To make the tests easier, I would propose that you extract only the force lods in a table and import tht into a new model. Try it out on a simple canteliever beam model with only a few mesh elements to have it operating quickly and smoothly, then apply it to your mode complex model. THis corresponds to y couple CFD + solid model using by ignoring the "inertial" terms, or in Quasi static, as the dynamic coupling between the fluid and the solid are not taken into account. Mostly valid for many studies. You can anycase add a eignefrequency analsis of your solid, and then see how the fluid load (added surface mass) wil laffectthe eigenfrequency, just to know where these are, to compare with the oscillations you might see in your fluid model, if you have here a close match, it might indicate some true problems in real "fully coupled" life ;) -- Good luck Ivar

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.