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 The information provided may be out of date.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.
"Segregated solver steps do not involve all components."
Posted 8 apr 2011, 19:36 GMT-4 Plasma Physics, Fluid & Heat, Heat Transfer & Phase Change, Studies & Solvers Version 4.2, Version 4.2a 6 Replies
Please login with a confirmed email address before reporting spam
I hope if this has been answered before that someone can direct me to the thread.
I am trying to solve a very simple system (attached) using electrostatics and solid mechanics. The deformation I expect to occur is solved for correctly until I add contact pairs. I add a contact pair to prevent the solid components from intersecting but once this pair is added I get the error message "Segregated solver steps do not involve all components." from the stationary solver.
Can anyone shed some light on why this may be happening.
Thank you very much,
Greg
Attachments:
Please login with a confirmed email address before reporting spam
first of all update to the latest patch it would make you less frustrated :)
When you change your physics and have already solved a case you might be in the position of having added a dependent variable, and sometimes COMSOL doesn't notice and does not update the solver sequence.
Therefore it's often wisely to delete and regenerate a default solver, BUT this deletes all your results nodes too, if you do not connect them to another study, and then manually reconnect them back later. ANother less drastic way is to right click the Solver and ask for a default sequence (not always enough)
I suspect that you have added the contact, and I see that the lumped block with the contact pressure is not present in your list. It will appear if you clear and regenerate (sometimes also if you ask for a new default sequence, from ym 4.1.0.185 it seems to work)
Always check that the dependent solver variables are all present in a segregated sequence, somewhere in there.
By the way you are missing the ALE, or for such small displacement adding the air to the structural and setting the air "material" to E=1[atm], nu=0, rho=1[kg/m^3] and use the spatial deformation of the "air" spring is also possible
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Thanks for your first reply.
I updated my version of COMSOL but I am still experiencing difficulties.
First of all, my solid structures are still able to pass through one another (contact pairs are not adhered to). One thing I have noticed is that the boundaries are listed as "not applicable" in the contact physics. I read elsewhere that with internal contact pairs the assembly finalization method should be used instead of a union. While this did alleviate the problem of the boundaries being called "not applicable", the solution does not result in any deformation even though the static potentials and forces are still associated with the correct boundaries/domains.
Secondly, when I include air in the structural mechanics and have the contact physics enabled I am told I have a singular matrix so no solution is found. There is no trouble when the contact physics are not enabled.
I have attached the model finalized as a union, with air included in the structural mechanics, and with the contact physics enabled (but with the contact pair called "not applicable"). If you have any other thoughts on solutions I would greatly appreciate hearing them.
Thank you,
Greg Reynen
Attachments:
Please login with a confirmed email address before reporting spam
its the contact pair that does not accept the interiour boundaries. forgotten about that.. If you disable it you will see the displacement from the ES coupling.
That is my fault, I told you to try without ALE, this will then not work with the CONTACT, because the "air spring" cannot fully collapse without generating a topology change which COMSOL cannot handle. Sorry for that.
So you need to remove the air from the solid physics selection. Add your ALE, that could be restricted to only the air domain. Define it as free, and select all boundaries and enforce the displacements driven by u,v (to simplify for the solver you can also force the external boundaries to "0", those you know are not moving).
Now with ALE you cannot allow the interface to fully collapse, so you need to keep a small gap to allow the mesh to survive, this need some test and fail approach and parameter tweaking
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
I face the same problem with Greg, "Segregated solver steps do not involve all components."
Should I define a new domain for the air? Could this be the solution for my problem?
Another problem I am having is that I did what you have adviced. " to right click the Solver and ask for a default sequence (not always enough)" and now the solution takes a very long time and after that it does not converge. How can I get back to my normal solution where I get the error "Segregated solver steps do not involve all components." ?
Thanks and best regards,
Erol J.
Attachments:
Please login with a confirmed email address before reporting spam
(I do not have access to comsol just now, anyhow I do not have the plasma physics module) but probably this comes from the fact that you have changed your model a few times and you are using the lagrange multipliers (or something calling them). When you resolve, and change something, COMSOL cannot be sure how to order all variables, so you need to go through the segregated steps and manually define where and which variables to set into the different steps. Check the segregation nodes and click "add"
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
This problem is not solved by comsol from so long. There are several threads with contact +electromechanics problems in which user experiences difficulties. The physics users want to explore is the After pull-in response of the beams as attached in the papers.
Please release a tutorial in comsol for these problems.
Otherwise users have to use Intellisuite, coventorware and mems+2.0 to solve such MEMS problems
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
- KNOWLEDGE BASE Segregated solver steps do not involve all components
- BLOG Multiphysics Modeling and Standalone Simulation Apps Drive Innovation
- BLOG Veryst Combines Material Testing and Simulation for Reliable Results
- FORUM How to set up different components of the material
- FORUM Failed to evaluate AWE expression. Not all parameter steps returned.