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 The information provided may be out of date.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.
Microwave not heating, but cooling. Why?
Posted 17 sept. 2012, 15:01 UTC−4 RF & Microwave Engineering, Heat Transfer & Phase Change Version 4.2a, Version 4.3 7 Replies
Please login with a confirmed email address before reporting spam
I'm using the Microwave Heating (mh) physics, with a Frequency-stationary study. Temperature initial condition set to 293.15[K]. I have the model surrounded by thermal insulation, except for one part of the structure that is held at 293.15[K]. I inject 1Watt of microwave power and expect things to get warm, but they don't. They get colder!! Especially curious is the region I set to 293.15[K] appears to be at 253.88[K]. Why is this happening? Why don't I get the fixed temperature I set? Why is it almost 19 degrees colder? From what I can tell, this 19 degree offset is the same across the whole model.
When I do the transient study with exactly the same model, the initial temperature is a proper 293.15[K]. The fixed temperature region I set remains a proper 293.15[K], and as time progresses, the temperatures elsewhere rise as expected. The transient study works. The stationary seems to be all low by 19 degrees.
I've looked everywhere for an errant temperature setting. Is this a Comsol bug?
When I do the transient study with exactly the same model, the initial temperature is a proper 293.15[K]. The fixed temperature region I set remains a proper 293.15[K], and as time progresses, the temperatures elsewhere rise as expected. The transient study works. The stationary seems to be all low by 19 degrees.
I've looked everywhere for an errant temperature setting. Is this a Comsol bug?
7 Replies Last Post 27 sept. 2012, 18:39 UTC−4
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I'm not so sure you will find a "bug" that easily ;),
I assume rather a sign error or just a mesh density issue ,
what is your time stepping, and material thermal diffusivity alpha and how does it compare to your mesh density ?
--
Good luck
Ivar
I'm not so sure you will find a "bug" that easily ;),
I assume rather a sign error or just a mesh density issue ,
what is your time stepping, and material thermal diffusivity alpha and how does it compare to your mesh density ?
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Must've been a COMSOL bug. I deleted the study, and added it back in again. No other changes. Now the temperatures are sane.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
be aware that when you change a lot a model, including changing nodes in the solver, at some stage COMSOl cannot find its way and does not always continue to update the solver nodes, as it considers that the user has change so much that it's the user responsability to update the nodes.
Theerfore it's worth to always restart fresh, and if the solver is giving doubtful results, to delete the full solver sequence and restart a fres new "default " solver sequence, the only thing is that then mostly you loose all your postprocessing set-up too
--
Good luck
Ivar
be aware that when you change a lot a model, including changing nodes in the solver, at some stage COMSOl cannot find its way and does not always continue to update the solver nodes, as it considers that the user has change so much that it's the user responsability to update the nodes.
Theerfore it's worth to always restart fresh, and if the solver is giving doubtful results, to delete the full solver sequence and restart a fres new "default " solver sequence, the only thing is that then mostly you loose all your postprocessing set-up too
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Yes, this has been my experience too, and it's annoying to lose the postprocessing setup, but being sure the simulation is correct outweighs this convenience.
I can see getting errors. For example, if I change physics, I often get the "Unexpected error" dialog, which is easily resolved by clearing things out and reloading. It's a bug, but a benign bug.
On the other hand, it's decidedly non-benign when COMSOL gives totally incorrect answers with no error or warning, as happened to me when switching between the stationary and transient solvers in microwave heating. It's exactly that sort of non-benign problem that erodes trust in the software. After all, software like COMSOL is not useful to an engineer if the results are untrustworthy. I'm not talking about limitations of the finite element method itself. I'm talking about software bugs that radically alter numerical results. Scary.
I can see getting errors. For example, if I change physics, I often get the "Unexpected error" dialog, which is easily resolved by clearing things out and reloading. It's a bug, but a benign bug.
On the other hand, it's decidedly non-benign when COMSOL gives totally incorrect answers with no error or warning, as happened to me when switching between the stationary and transient solvers in microwave heating. It's exactly that sort of non-benign problem that erodes trust in the software. After all, software like COMSOL is not useful to an engineer if the results are untrustworthy. I'm not talking about limitations of the finite element method itself. I'm talking about software bugs that radically alter numerical results. Scary.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Updated:
5 years ago
Hi Chris,
Have you contacted technical support (support@comsol.com) with your question and model? We would very much like to investigate the problems you are having with this model. We do have plenty of customers simulating microwave heating successfully and we are not aware of any general issues.
Best regards,
Bjorn
Have you contacted technical support (support@comsol.com) with your question and model? We would very much like to investigate the problems you are having with this model. We do have plenty of customers simulating microwave heating successfully and we are not aware of any general issues.
Best regards,
Bjorn
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I'm also a "user" and live from delivering "correct" simulations, but I must say that for all of the numerous software I have used so far, the things you can do with COMSOL is far above average, if not unique. But, on the other hand, the complexity and number of buttons, tweaks etc also increases exponentially, therefore I only trust models build up rather directly.
Once I have tested out and run different combinations, if I have any doubts, I remake my model fresh to be sure I have all corret "default setings" and no mixed combinations from different tries.
But this was so also for other tools, for me this is no "bugs" when the software delivers probably "correct" result from some "incorrect" settings, the thing is to track and identify the "wrong" setting.
So with some perhaps strict methodology, and systematic validation of my models, I'm very happy, and get rapidly great results, I never had dreamed of possible, before ;)
--
Have fun Comsoling
Ivar
I'm also a "user" and live from delivering "correct" simulations, but I must say that for all of the numerous software I have used so far, the things you can do with COMSOL is far above average, if not unique. But, on the other hand, the complexity and number of buttons, tweaks etc also increases exponentially, therefore I only trust models build up rather directly.
Once I have tested out and run different combinations, if I have any doubts, I remake my model fresh to be sure I have all corret "default setings" and no mixed combinations from different tries.
But this was so also for other tools, for me this is no "bugs" when the software delivers probably "correct" result from some "incorrect" settings, the thing is to track and identify the "wrong" setting.
So with some perhaps strict methodology, and systematic validation of my models, I'm very happy, and get rapidly great results, I never had dreamed of possible, before ;)
--
Have fun Comsoling
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi,
I am working with light, so use refractive index for the materials. When the imaginary part of the refractive index was negative, I also got cooling. So it might be a sign problem.
The microwave heating node is behaving strangely in 4.3, with repeated crashes. I can't get Port to work, while it works fine with EM-node. I suspect Comsol mixes the variables between the two nodes, E and E2, as I have used both in the same model. Deleting EM-node and Reset history does not help. See attached example if you are interested. I will send it to support also.
Comsol is ok. The error-messages are not.
Olav
I am working with light, so use refractive index for the materials. When the imaginary part of the refractive index was negative, I also got cooling. So it might be a sign problem.
The microwave heating node is behaving strangely in 4.3, with repeated crashes. I can't get Port to work, while it works fine with EM-node. I suspect Comsol mixes the variables between the two nodes, E and E2, as I have used both in the same model. Deleting EM-node and Reset history does not help. See attached example if you are interested. I will send it to support also.
Comsol is ok. The error-messages are not.
Olav
Attachments: