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 The information provided may be out of date.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.
Initial values from solution
Posted 15.03.2011, 04:40 GMT-4 Interfacing, Heat Transfer & Phase Change, Modeling Tools & Definitions, Parameters, Variables, & Functions Version 4.1 5 Replies
Please login with a confirmed email address before reporting spam
I'm struggeling with the same problem for quite a while. The thing I want to do is:
1) Solve a time dependet heat transfer problem with two simple domains (initiated with a different temperature) for e.g. t_start = 0 to t_end = 1 second.
2) Use the solution of step 1 and solve again for t_start = 1 to t_end = 2 second but now with a third domain added.
The new (third) domain has T_initial = 20[degC] and the two other domains should have the same temperature as at the end of step 1, so T = 'T' .
The problem is that on 1 vertex a really low temperature appears without any reason (way lower than any initial temperature in the simulation (see attached figure)). Does someone has any clou how this happens?
Regards,
Peter
Attachments:
Please login with a confirmed email address before reporting spam
indeed, not really what is expected.
One thing to know: points, (as edges) take their properties, such as temperature, from an average of the surrounding domains. Therefore it's always delicate to impose values on points (in 2D or 3D) or even edges in 3D.
Could it be that you have a point BC set up ?
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Thanks for your reply. No, I don't have point boundary condition.
I actually think I know the source of the problem, however I do not know how to solve it.
Since I add a new domain to the geometry, COMSOL automatically re-meshes the geometry. During the re-mesh (I guess) COMSOL has to interpolate in order to fit the solution as new initial value on the new mesh. During the interpolation something strange happens... because when I try the same without modifying the geometry (so no re-mesh is needed) the effect does not appear.
Could this be the source of the problem and do you perhaps know a way to solve it?
Regards,
Peter.
Please login with a confirmed email address before reporting spam
be careful when adding domains or boundaries (= topology changes) as this renumbers all internal id numbers, and your BC and material settings pointing to given Entity ID's might get all wrong, always carefully check all BC's and domain definitions after having changed your geometry
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
That's a good point, however, I have a very simple geometry and I checked all the boundaries, domain numbers etc. by hand afterwards, and there was no error in that.
I have send a copy to support.
Thanks for your help!
Ragards,
Peter
Please login with a confirmed email address before reporting spam
I am having the same problem you had months ago...
Actually, it is also an heat transfer problem which I have coupled with a PDE module.
I have 2 dependent variables: temperature (T) and hydration degree (xi), which I have to save at the end of each calculation and then input them in the following one.
I did that using interpolation functions (by browsing the .txt file). However, I don't know if this is the right way to do it.
The problem is that when I run the study, there is an error: it finds a negative xi I don't know where and it can't proceed with the calculation...
Have you got any idea of what this could be due?
Thank you,
Ana