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.
Undefined value found error on coordinates not existed
Posted May 11, 2023, 8:08 p.m. EDT 6 Replies
Please login with a confirmed email address before reporting spam
Hi, I am trying to run a simulation of thermoelectrical cooling system. I got an error and I don't know how to fix it because it is strange.
the error message is as follows:
Undefined value found. - Detail: Undefined value found in the equation residual vector. There are 134 degrees of freedom giving NaN/Inf in the vector for the variable comp1.T. at coordinates: (0.0474,0.06187,-0.050753), (0.04195,0.06187,-0.050753), (0.0484038,0.06187,-0.0486638), (0.0504096,0.063,-0.050753), (0.0454387,0.06187,-0.0487679), ...
If you look at my model, those coordinates are not even in my model. How can I trace the problem?
Attachments:
Please login with a confirmed email address before reporting spam
Looks to me like those coordinates just might actually be in your model. Bear in mind that Comsol's default length units are meters. If you need more advice/suggestions than that, then I recommend you upload your .mph file to the forum.
-------------------Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
Please login with a confirmed email address before reporting spam
Is there any way to trace down those mesh nodes or elements by their coordinate? If so, how?
Please login with a confirmed email address before reporting spam
Some error messages occur in contexts where, with a little experimenting, you may be able to select/display the corresponding troublesome geometric locations/feature(s) to be highlighted and/or have the graphics display zoom in on the relevant details. I recommend you upload your .mph file to the forum.
-------------------Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
Please login with a confirmed email address before reporting spam
Attached is the file.
I tried to follow the example as shown below:
https://www.comsol.com/model/lumped-thermoelectric-module-with-pid-control-102421
I found that those nodes are on the fins of the heat sink. It seems that their temperatures quickly reach infinite and cause the problem. I have tried to use turbulant flow as well. It did not yield error but also did not converge. Since my model and the example look so similar, I could not figure out what made mine fail to work.
The major difference is that I only simulate half of my model by using symmetry.
Attachments:
Please login with a confirmed email address before reporting spam
That's not a simple model and I doubt I can address it all. But the first thing I notice is that your domain #4 "TEC_effective" doesn't seem to have any boundary specification on what must surely be a symmetry-plane face, and I don't see any physics assigned to that volume region either, although you have meshed that volume. (Usually, if you are deliberately not assigning physics to a volume in your model, then you wouldn't normally mesh that domain either.) Might these thoughts help/suggest part of your problem? Also, your mesh looks rather coarse to me, in a number of locations.
-------------------Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
Please login with a confirmed email address before reporting spam
Sorry, that was some legacy from different trials. I used that domain with turbulent flow before, which did not give me good solution. Then I switched to lump model. I must have forgot the disable that domain.
PS. I tried to modefy the example by changing the number of fins from 8 to 18. It did not coverge any more. I think you may be right, the coarse mesh may cause the problem.