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 The information provided may be out of date.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.
How to control the material parameter with solution?
Posted Dec 7, 2010, 10:17 p.m. EST Heat Transfer & Phase Change Version 4.1 3 Replies
Please login with a confirmed email address before reporting spam
Dear Experts,
I used Joule heating model of 2D problem. The geometry is a rectangle with a circle in the center. The resistivity of the circle is very high. Therefore, the current density will be increased around the circle. The resistivity of the rectangle is related with the current density. Simply, I set the resistivity rho_rect = 1e5 ohm.m at the area that current density above a certain value while the rho_rect at other area is zero.
When I used a expression with the variable ec.normJ in the heating model to define the material's resistivity, an error jumped like this
"Circular variable dependency detected.
- Variable: mod1.ec.normJ
- Geometry: 1
- Domain: 1"
Then I added mathmatics, using a PDE equation to describe the rho_rect. When I first set resistivity of rectangle constantly in Joule heating model, I can get the result that around the circle rho_rect was increased. However, when I set resistivity of rectangle from PDE. The solver cannot be convergence.
How can I resolve this problem? Is it a hard problem? I spent a lot of time on it.
Many Thanks!
Zhao
I used Joule heating model of 2D problem. The geometry is a rectangle with a circle in the center. The resistivity of the circle is very high. Therefore, the current density will be increased around the circle. The resistivity of the rectangle is related with the current density. Simply, I set the resistivity rho_rect = 1e5 ohm.m at the area that current density above a certain value while the rho_rect at other area is zero.
When I used a expression with the variable ec.normJ in the heating model to define the material's resistivity, an error jumped like this
"Circular variable dependency detected.
- Variable: mod1.ec.normJ
- Geometry: 1
- Domain: 1"
Then I added mathmatics, using a PDE equation to describe the rho_rect. When I first set resistivity of rectangle constantly in Joule heating model, I can get the result that around the circle rho_rect was increased. However, when I set resistivity of rectangle from PDE. The solver cannot be convergence.
How can I resolve this problem? Is it a hard problem? I spent a lot of time on it.
Many Thanks!
Zhao
3 Replies Last Post Feb 2, 2015, 7:32 a.m. EST
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi
I hope I have understood you well, but when you couple many parameters and typically a few physics it is easy to get circular references.
You have to help COMSOL sort it out where to start in its solver loop, the best is to think over the physics and how they behave one with respect of the other.
In your case (if I have got it right) you have electric joule heating from a curent flow that heats up a material then the temperature rise changes the resisitivity of the material that loops back to change the current density, hence the dissipated power and so on.
To get out, try to define a "segregated solver case" meaning you solve first one physics (by keeping the other vlaues constant) and then the next one in sequence, and then loop around until the sequence converges. In some cases you do not even need to loop around, but that is when you know that one physics is more or less correct without any feedback from the other (or the time constants are very different, its all model dependent. In V4 COMSOL proposes
ofen by default a segregated sequence but sometimes it is not in the correct order, as you know mostly morethan what you tend to tell the software (luckily we are still the chief ;)
Try some of the examples in the model library (print out the pdf file and repeat the exercice, going with yur own examples is nice but often its worth to train well with step to step examples until one catches the use of most of the many buttons inside COMSOL
--
Good luck
Ivar
I hope I have understood you well, but when you couple many parameters and typically a few physics it is easy to get circular references.
You have to help COMSOL sort it out where to start in its solver loop, the best is to think over the physics and how they behave one with respect of the other.
In your case (if I have got it right) you have electric joule heating from a curent flow that heats up a material then the temperature rise changes the resisitivity of the material that loops back to change the current density, hence the dissipated power and so on.
To get out, try to define a "segregated solver case" meaning you solve first one physics (by keeping the other vlaues constant) and then the next one in sequence, and then loop around until the sequence converges. In some cases you do not even need to loop around, but that is when you know that one physics is more or less correct without any feedback from the other (or the time constants are very different, its all model dependent. In V4 COMSOL proposes
ofen by default a segregated sequence but sometimes it is not in the correct order, as you know mostly morethan what you tend to tell the software (luckily we are still the chief ;)
Try some of the examples in the model library (print out the pdf file and repeat the exercice, going with yur own examples is nice but often its worth to train well with step to step examples until one catches the use of most of the many buttons inside COMSOL
--
Good luck
Ivar
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
Hi,
Thanks for your help. I do as you suggested. The loop of my physics problem is that first calculate the current density and then the resistivity is changed with current density, then loop back to change the current density.
I tried to use segregated slover and make the sequence that solve current density firstly and resistivity secondly. But I still met error.
If I set the resistivity constanly, I can get the expected result. However, when I change the material of superconducting as 1/u where u is defined in PDE model, error occurs.
I put on my project. Could you help me to find the problem?
Many thanks!
Zhao
Thanks for your help. I do as you suggested. The loop of my physics problem is that first calculate the current density and then the resistivity is changed with current density, then loop back to change the current density.
I tried to use segregated slover and make the sequence that solve current density firstly and resistivity secondly. But I still met error.
If I set the resistivity constanly, I can get the expected result. However, when I change the material of superconducting as 1/u where u is defined in PDE model, error occurs.
I put on my project. Could you help me to find the problem?
Many thanks!
Zhao
Attachments:
Please login with a confirmed email address before reporting spam
Posted:
10 years ago
Hi,
Have you been solved the problem? I have the same question with you.
Could you help me?
Thank you.
Best regards,
Jing Wang
Have you been solved the problem? I have the same question with you.
Could you help me?
Thank you.
Best regards,
Jing Wang