Menu

Discussion Forum

New Discussion
Filter

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.

Conductive heat transfer in a via/cylinder

Posted Oct 20, 2020, 4:36 a.m. EDT Heat Transfer Version 5.5 7 Replies

Guilherme Theis
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Hello,

I am trying to learn COMSOL modelling and simulation. As my final objective is to simulate PCBs heat transfer, I am starting with a simple via. The idea is then to set the convection and radiation from my cylinder to be non-existant. I would like to then set a temperature and a heat flux to further find a thermal resistance of this via by using:

It seems that I don't understand how to set the boundaries for this. My idea was to set actually a "room temperature" of 25[degC] as initial condition and add a heat flux of Q at one of the edges of the cylinder. However, it returns the following error:

Failed to find a solution. The relative residual (4.1) is greater than the relative tolerance. There was a warning message from the linear solver. Iterative refinement triggered. Returned solution is not converged. Not all parameter steps returned.

I've set both my via radius and my heat flux to be parametric (from 0.05 mm to 0.15mm and 1 to 10W, respectively). You have my COMSOL file attached to this.

Thank you in advance.

Attachments:

7 Replies Last Post Oct 22, 2020, 4:14 a.m. EDT
Edgar J. Kaiser Certified Consultant
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Oct 20, 2020, 7:24 a.m. EDT

Guilherme,

in your model all boundaries except the heat sources are insulated. Therefore the temperature rises to infinity and breaks the solver. You need a heat sink. This can e.g. be a temperature boundary. The heat you bring in must go out somewhere.

Cheers Edgar

-------------------
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Guilherme, in your model all boundaries except the heat sources are insulated. Therefore the temperature rises to infinity and breaks the solver. You need a heat sink. This can e.g. be a temperature boundary. The heat you bring in must go out somewhere. Cheers Edgar
Guilherme Theis
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Oct 20, 2020, 11:14 a.m. EDT
Updated: 5 years ago Oct 20, 2020, 11:23 a.m. EDT

Dear Edgard,

Thank you for the answer. Just so I see if I understand I should set one of my edges with a heat flux Q that I want to vary from 1 to 10 W and then set my remaining boundaries to a temperature? (e.g: 25 degC)

How I would see the temperature rise caused by this Q then?

EDIT: by doing as I exemplified above I can now properly simulate. My new question now is if I can somehow evaluate the points and values of my min and max temperatures (e.g: (1,1,1, 25) where 1,1,1 would be x,y,z and 25 would be the temperature).

Dear Edgard, Thank you for the answer. Just so I see if I understand I should set one of my edges with a heat flux Q that I want to vary from 1 to 10 W and then set my remaining boundaries to a temperature? (e.g: 25 degC) How I would see the temperature rise caused by this Q then? EDIT: by doing as I exemplified above I can now properly simulate. My new question now is if I can somehow evaluate the points and values of my min and max temperatures (e.g: (1,1,1, 25) where 1,1,1 would be x,y,z and 25 would be the temperature).
Edgar J. Kaiser Certified Consultant
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Oct 20, 2020, 3:31 p.m. EDT

You can check the max/min operations in 'Nonlocal couplings' and respective Max/Min datasets in the postprocessing. Also point probes and point cut datasets may be of interest.

Cheers Edgar

-------------------
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
You can check the max/min operations in 'Nonlocal couplings' and respective Max/Min datasets in the postprocessing. Also point probes and point cut datasets may be of interest. Cheers Edgar
Guilherme Theis
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Oct 21, 2020, 6:59 a.m. EDT
Updated: 5 years ago Oct 21, 2020, 7:13 a.m. EDT

Dear Edgar,

I actually ran the simulation and extracted the results using a Derived Value and created a Surface Maximum. However this doesn't agree with my analytical model and I am not sure if then I am correctly extracting the value. In theory my thermal resistance would be given by

which for d = 0.1mm, k = 401 and t = 1mm would yield a thermal resistance of 317 degC/W, while from simulation I get around 2degC/W.

What I am doing wrongly here?

Dear Edgar, I actually ran the simulation and extracted the results using a **Derived Value** and created a **Surface Maximum**. However this doesn't agree with my analytical model and I am not sure if then I am correctly extracting the value. In theory my thermal resistance would be given by R_{thermal} = \frac{t}{k_{cu}*pi*\frac{d}{4}^2} which for d = 0.1mm, k = 401 and t = 1mm would yield a thermal resistance of 317 degC/W, while from simulation I get around 2degC/W. What I am doing wrongly here?

Attachments:

Edgar J. Kaiser Certified Consultant
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Oct 21, 2020, 11:29 a.m. EDT

Check if your model realizes this kind of thermal resistor. I think the barrel of the cylinder should be thermally isolated in this case.

-------------------
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Check if your model realizes this kind of thermal resistor. I think the barrel of the cylinder should be thermally isolated in this case.
Guilherme Theis
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Oct 22, 2020, 3:10 a.m. EDT

Thank you Edgar, it seems now that this works. This is my first try with COMSOL and in my mind I was doing so (but I ovewrote the isolation by setting a temperature boundary).

Thank you Edgar, it seems now that this works. This is my first try with COMSOL and in my mind I was doing so (but I ovewrote the isolation by setting a temperature boundary).
Edgar J. Kaiser Certified Consultant
Send Private Message Flag post as spam

Please login with a confirmed email address before reporting spam

Posted: 5 years ago Oct 22, 2020, 4:14 a.m. EDT

Good to hear that it works for you now.

-------------------
Edgar J. Kaiser
emPhys Physical Technology
www.emphys.com
Good to hear that it works for you now.

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