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 The information provided may be out of date.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.
Time dependent Problem
Posted Mar 7, 2011, 3:36 p.m. EST Low-Frequency Electromagnetics, Structural Mechanics Version 4.0a 5 Replies
Please login with a confirmed email address before reporting spam
Hi,
I'm having some troubles with time dependent solving for a 2D cantilever. I define a step function and multiply it by a constant input Vin. The step transition is about 6e-9seconds, then it stabilizes at the constant value Vin. Ok, that's fine, so I would expect the cantilever bend during the transition of the step, and then stabilize at the position corresponding to the suitable voltage (Vin). However, what happens is that the more time I simulate, it means the more the step remains at Vin, the more the cantilever deflects. That's not coherent at all, I mean, If I reach a voltage and stay ther for to years I would expect the system to remain at the same position (deflection). This is really important for me, cause I'm studying the switching interactions on the cantilever, so I need the problem to be as close as posible to reality. This is very weird, cause the electrostatic force depends on the voltage I introduce, so if the voltage remains constant, the es Force should too.
Thanks in advance, any help will be greatly appreciated.
I'm having some troubles with time dependent solving for a 2D cantilever. I define a step function and multiply it by a constant input Vin. The step transition is about 6e-9seconds, then it stabilizes at the constant value Vin. Ok, that's fine, so I would expect the cantilever bend during the transition of the step, and then stabilize at the position corresponding to the suitable voltage (Vin). However, what happens is that the more time I simulate, it means the more the step remains at Vin, the more the cantilever deflects. That's not coherent at all, I mean, If I reach a voltage and stay ther for to years I would expect the system to remain at the same position (deflection). This is really important for me, cause I'm studying the switching interactions on the cantilever, so I need the problem to be as close as posible to reality. This is very weird, cause the electrostatic force depends on the voltage I introduce, so if the voltage remains constant, the es Force should too.
Thanks in advance, any help will be greatly appreciated.
5 Replies Last Post Apr 9, 2011, 10:13 a.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
if you cantilever is between one or two electrodes, and you apply a too large voltage you get the pull-in effect, at deflections above 2/3 (if I remember right) if the gap your system becomes unstable and the cantilever will pull-in and collapse on the electrode (if you have no contact the FEM model will overlap). So this might be (all depends on your model) a natural effect !
Take a look, try with a smaller step
--
Good luck
Ivar
if you cantilever is between one or two electrodes, and you apply a too large voltage you get the pull-in effect, at deflections above 2/3 (if I remember right) if the gap your system becomes unstable and the cantilever will pull-in and collapse on the electrode (if you have no contact the FEM model will overlap). So this might be (all depends on your model) a natural effect !
Take a look, try with a smaller step
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar,
Thanks for the response. Here's what's happening: I can smoothly run my time dependent simulations just before the pull-in voltage, however, as soon as I reach pull-in (or what I believe is pull-in), the simulation freezes at certain value, which I would expect, given that the lack of convergence at that point might drive comsol crazy. However, It just stays ther for ages, and in some cases when I leave comsol computing, it closes suddenly as if it had an error. My question is if there's anyway to tell comsol: "OK, that's enough, let's end this simulation and show some no convergence error message" or something like that. Which is the maximum simulation time for comsol to "give up" ? Can I define that time to avoid this ugly freezing?
Than you.
Thanks for the response. Here's what's happening: I can smoothly run my time dependent simulations just before the pull-in voltage, however, as soon as I reach pull-in (or what I believe is pull-in), the simulation freezes at certain value, which I would expect, given that the lack of convergence at that point might drive comsol crazy. However, It just stays ther for ages, and in some cases when I leave comsol computing, it closes suddenly as if it had an error. My question is if there's anyway to tell comsol: "OK, that's enough, let's end this simulation and show some no convergence error message" or something like that. Which is the maximum simulation time for comsol to "give up" ? Can I define that time to avoid this ugly freezing?
Than you.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
in the solver settings you have different stop conditions you can add or modify, and limit the number of iterations, it's worth to take a closer look.
But its absolutely worth to use the latest version, where a lot of early issues with the first 4.0 have been solved
--
Good luck
Ivar
in the solver settings you have different stop conditions you can add or modify, and limit the number of iterations, it's worth to take a closer look.
But its absolutely worth to use the latest version, where a lot of early issues with the first 4.0 have been solved
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi Ivar,
I managed to use the stop condition, however, my real problem is to make the contact problem work with time dependent simulation, do you know if it is feasible? I 've already seen some of the threads about it and followed the steps but I haven't got any results. As soon as I include contact and run the simulation, I get the error: segregated solver does not involve all components, even though I disabled the contact pressure from the list of variables to solve for.
I would greatly appreciate any help you can give me in this regard.
I managed to use the stop condition, however, my real problem is to make the contact problem work with time dependent simulation, do you know if it is feasible? I 've already seen some of the threads about it and followed the steps but I haven't got any results. As soon as I include contact and run the simulation, I get the error: segregated solver does not involve all components, even though I disabled the contact pressure from the list of variables to solve for.
I would greatly appreciate any help you can give me in this regard.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
normally it means that the solver segregated tree does not include all dependent variable items.
Do a manual check: open the Dependent variable node and check which variables are tagged "solved for" then open the segregated nodes and check that you have them all. If not, try to select the solver node and generate a default sequence, then check again, or update by hand.
It could well be an issue with the early v4.0. I'm using the latest version and I get very seldom such errors, it means that I have really messed up with the model and that COMSOL has not managed to follow. Then by regenerating the sequence everything is fine again ;)
--
Good luck
Ivar
normally it means that the solver segregated tree does not include all dependent variable items.
Do a manual check: open the Dependent variable node and check which variables are tagged "solved for" then open the segregated nodes and check that you have them all. If not, try to select the solver node and generate a default sequence, then check again, or update by hand.
It could well be an issue with the early v4.0. I'm using the latest version and I get very seldom such errors, it means that I have really messed up with the model and that COMSOL has not managed to follow. Then by regenerating the sequence everything is fine again ;)
--
Good luck
Ivar