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
23 janv. 2012, 01:48 UTC−5
Hi
I believe you have one more phenomena:
curent flow Q[A/m2], => hence joule heating T[K],
heating T[K] => hence material expansion u[m] and sigma[Pa]
heating T[K] => hence material resisitivity/conductance change related to T (loops back to influence Joule heating, if you have a volage source, or changes the voltage if you have a current source)
material stress sigma[Pa] => hence PZT effets will/might (materil dependent) generate voltage and resistivity/conductivity changes, hence loops back to joule exchange
Nice mdel in perspective, many BC's to set correctly, normally with thermal stress and joule heating you have the physics, the rest you should be able to introduce in the material properties, I assume, but your model will be rather non-linear, tweaking the segregated solver order might be required.
Then when you talk about resonance freuencies: there you have a linear analysis, so to get the change in resonance freuqncies for different loading cases you need to run your time series to a given time = load case, then stop the time solver, and run a new study eigenfrequency WITH the time series results as initial conditions for the linearization point.
Then run further the time series, stop and redo an eigenfrequency analysis.
This is probably easier set up with matlab and livelink, but I suppose you should be able to tweak the GUI to loop around with a parametric sweep and a few study cases. At most you might need to run 1-2 cases manually to ge the loop running
--
Good luck
Ivar
Hi
I believe you have one more phenomena:
curent flow Q[A/m2], => hence joule heating T[K],
heating T[K] => hence material expansion u[m] and sigma[Pa]
heating T[K] => hence material resisitivity/conductance change related to T (loops back to influence Joule heating, if you have a volage source, or changes the voltage if you have a current source)
material stress sigma[Pa] => hence PZT effets will/might (materil dependent) generate voltage and resistivity/conductivity changes, hence loops back to joule exchange
Nice mdel in perspective, many BC's to set correctly, normally with thermal stress and joule heating you have the physics, the rest you should be able to introduce in the material properties, I assume, but your model will be rather non-linear, tweaking the segregated solver order might be required.
Then when you talk about resonance freuencies: there you have a linear analysis, so to get the change in resonance freuqncies for different loading cases you need to run your time series to a given time = load case, then stop the time solver, and run a new study eigenfrequency WITH the time series results as initial conditions for the linearization point.
Then run further the time series, stop and redo an eigenfrequency analysis.
This is probably easier set up with matlab and livelink, but I suppose you should be able to tweak the GUI to loop around with a parametric sweep and a few study cases. At most you might need to run 1-2 cases manually to ge the loop running
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
23 janv. 2012, 11:36 UTC−5
Thanks Ivar! You're right about the additional coupling, but in my case, I wish to try out the simplest case first with only one loop back.
I have a couple of questions.
1) You mentioned that many BCs have to be set. I've set the structural BCs (fixed etc.), current density, and the thermal ones (sink temperature, surface radiation etc.) Are there any other BCs you have in mind?
2) I've so far used only default solvers. Could you please tell me what would be a good place to start reading about how to tweak the solvers?
Thanks Ivar! You're right about the additional coupling, but in my case, I wish to try out the simplest case first with only one loop back.
I have a couple of questions.
1) You mentioned that many BCs have to be set. I've set the structural BCs (fixed etc.), current density, and the thermal ones (sink temperature, surface radiation etc.) Are there any other BCs you have in mind?
2) I've so far used only default solvers. Could you please tell me what would be a good place to start reading about how to tweak the solvers?
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
24 janv. 2012, 02:11 UTC−5
Hi
for me the doc is the best place, but I must admit I have read them some 5-6 times as you do not catch everything on the first reading, too much material ;)
Then practising on simple models, and taking some time to answer on the Forum, that's an excellent training ;)
--
Good luck
Ivar
Hi
for me the doc is the best place, but I must admit I have read them some 5-6 times as you do not catch everything on the first reading, too much material ;)
Then practising on simple models, and taking some time to answer on the Forum, that's an excellent training ;)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
6 févr. 2012, 22:07 UTC−5
Hello Ivar,
I tried performing the eigenfrequency analysis of my structure after I did a stationary analysis to get the linearization point. I also made the eignefrequency solver use the results of the stationary solver as inputs. However, when I run the eigenfrequency study, the values are all NaN.
I get a file extension error when I try to upload my modle here, do you have any suggestion on a common reason on why I might be getting the NaN results? My stationary solver works fine.
Thanks!
Hello Ivar,
I tried performing the eigenfrequency analysis of my structure after I did a stationary analysis to get the linearization point. I also made the eignefrequency solver use the results of the stationary solver as inputs. However, when I run the eigenfrequency study, the values are all NaN.
I get a file extension error when I try to upload my modle here, do you have any suggestion on a common reason on why I might be getting the NaN results? My stationary solver works fine.
Thanks!
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
6 févr. 2012, 22:15 UTC−5
I put up my file here.
http://www.filedropper.com/copymph
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
8 févr. 2012, 17:30 UTC−5
Hi
take a lok at the "buckling" example of the model library, that is also a stationary + eigenmode solver sequence, and this gives you info about solver sub-node settings (the default used by COMSOL)
--
Good luck
Ivar
Hi
take a lok at the "buckling" example of the model library, that is also a stationary + eigenmode solver sequence, and this gives you info about solver sub-node settings (the default used by COMSOL)
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
8 févr. 2012, 17:32 UTC−5
Thanks a lot, Ivar!
Thanks a lot, Ivar!