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.

convergence issues in cojugate heat transfer problem

Please login with a confirmed email address before reporting spam

Hi,

I am trying to solve a conjugate fluid flow and heat transfer problem. The problem has a channel with fluid flowing through it and surrounding media which has some heat sources.
I am having issues when I increase the inlet velocity. Everything is laminar as it is a microchannel. As the inlet velocity is increased the solution fails to convergence.
I have tried including the results from lower velocity as initial condition for higher velocity but that is still not helping in convergence.

Any suggestion will be highly appreciated.


Regards,
Sumeet

9 Replies Last Post 15.06.2017, 12:16 GMT-4
Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15.07.2011, 11:44 GMT-4
Increasing the fluid velocity increases the Peclet number for the heat equation (ratio of advection to thermal diffusion). That necessitates the use of numerical stabilization, which COMSOL allows you to modify. You should check if the cell Peclet number is >= 1.0.

Nagi Elabbasi
Veryst Engineering
Increasing the fluid velocity increases the Peclet number for the heat equation (ratio of advection to thermal diffusion). That necessitates the use of numerical stabilization, which COMSOL allows you to modify. You should check if the cell Peclet number is >= 1.0. Nagi Elabbasi Veryst Engineering

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 15.07.2011, 18:18 GMT-4
Hi Nagi,

I agree that Peclet number is > 1. The objective of the study is to capture the effect as Pe increases. Do you have suggestion for numerical stabilization? I have been trying iterative solver and increasing the factor in error estimate but that is still not working.

Thank you,
Sumeet
Hi Nagi, I agree that Peclet number is > 1. The objective of the study is to capture the effect as Pe increases. Do you have suggestion for numerical stabilization? I have been trying iterative solver and increasing the factor in error estimate but that is still not working. Thank you, Sumeet

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 16.07.2011, 21:54 GMT-4
You can try two things: (i) use a finer mesh so that the element/cell Peclet number is smaller, or (ii) change the stabilization settings (streamline, crosswind and isotropic diffusion). The manual has some details on the different stabilization techniques.

Nagi Elabbasi
Veryst Engineering
You can try two things: (i) use a finer mesh so that the element/cell Peclet number is smaller, or (ii) change the stabilization settings (streamline, crosswind and isotropic diffusion). The manual has some details on the different stabilization techniques. Nagi Elabbasi Veryst Engineering

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 22.07.2011, 16:42 GMT-4
Hello Dr. Nagi Elabbasi,

I saw your reply to another problem, I hope you have some solution to my problem too. I'm modeling conjugate Heat Transfer problem too. It is periodic segment of cooling channel of gas turbine blade. with one rib on top and one at the bottom (staggered). When i give small i/p velocity it runs OK, but if I increase, it wont converge. let me know how I can solve this problem....

Thank you Very much for your Help in Advance

Raju
Hello Dr. Nagi Elabbasi, I saw your reply to another problem, I hope you have some solution to my problem too. I'm modeling conjugate Heat Transfer problem too. It is periodic segment of cooling channel of gas turbine blade. with one rib on top and one at the bottom (staggered). When i give small i/p velocity it runs OK, but if I increase, it wont converge. let me know how I can solve this problem.... Thank you Very much for your Help in Advance Raju

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 24.07.2011, 18:00 GMT-4
Dear Raju,

Try the same things I mentioned above and also check that Reynold’s number is not too high. Also try temporarily reducing the complexity of the problem by solving the fluid flow problem and then using that known velocity/pressure solution for the heat transfer problem.

Nagi Elabbasi
Veryst Engineering
Dear Raju, Try the same things I mentioned above and also check that Reynold’s number is not too high. Also try temporarily reducing the complexity of the problem by solving the fluid flow problem and then using that known velocity/pressure solution for the heat transfer problem. Nagi Elabbasi Veryst Engineering

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 25.07.2011, 14:11 GMT-4
Thank you Dr.Nagi,

I'm just wondering, you asked me to solve fluid flow first and introduce heat transfer later. To do this step I store the initial flow analysis solution by going to the menu Solve --> Solver manager and in this menu I click on store solution tab and select current solution at the intial value region, am I right?!.

Thank you

Raju
Thank you Dr.Nagi, I'm just wondering, you asked me to solve fluid flow first and introduce heat transfer later. To do this step I store the initial flow analysis solution by going to the menu Solve --> Solver manager and in this menu I click on store solution tab and select current solution at the intial value region, am I right?!. Thank you Raju

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26.07.2011, 08:49 GMT-4
Hi Raju,

I am sure there are several ways to do that. The way I do it is through the Study node, Solver Configuration, Dependent Variables. In the first line I change the “Defined by study step” to User defined. Then I go to the sub-node with the individual variables and unselect the “solve for this field” for those I don’t want solved. Then in a duplicate Study node, I do the opposite and set Method equal to the previous solution in “Values of Variables Not Solved For” and in “Initial values of Variables Solved For”.

If that is not clear you can check the COMSOL examples. The peristaltic pump example (Model ID 985) does something very similar.

Nagi Elabbasi
Veryst Engineering
Hi Raju, I am sure there are several ways to do that. The way I do it is through the Study node, Solver Configuration, Dependent Variables. In the first line I change the “Defined by study step” to User defined. Then I go to the sub-node with the individual variables and unselect the “solve for this field” for those I don’t want solved. Then in a duplicate Study node, I do the opposite and set Method equal to the previous solution in “Values of Variables Not Solved For” and in “Initial values of Variables Solved For”. If that is not clear you can check the COMSOL examples. The peristaltic pump example (Model ID 985) does something very similar. Nagi Elabbasi Veryst Engineering

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 26.07.2011, 11:14 GMT-4
Thank you so much Dr.Nagi,

I will try as per your advice...

Regards

Raju
Thank you so much Dr.Nagi, I will try as per your advice... Regards Raju

Please login with a confirmed email address before reporting spam

Posted: 7 years ago 15.06.2017, 12:16 GMT-4
Dear Nagi,

I'm modelling heat transfer in porous media couple with darcy flow module.
The goal of this simulation is investigating the influence of peclet number to the heat transfer result.

I read your comments in this thread but I don't understand how to obtain the numerical peclet number (Pe>1).
In my calculation, the Pe number always has a very large value (1e4 ~ 1e7).

Could you please advice me against this matter?

Thank you,

Best regards,
JN
Dear Nagi, I'm modelling heat transfer in porous media couple with darcy flow module. The goal of this simulation is investigating the influence of peclet number to the heat transfer result. I read your comments in this thread but I don't understand how to obtain the numerical peclet number (Pe>1). In my calculation, the Pe number always has a very large value (1e4 ~ 1e7). Could you please advice me against this matter? Thank you, Best regards, JN

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.