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 The information provided may be out of date.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.

Model heat transfer in block but can't get the solution

Please login with a confirmed email address before reporting spam

I modelled the heat transfer within a block with some air gaps. In the model, heat is transferred by conduction in the solid domain, and in these air voids, convection and radiation are also involved. I have attached a picture of the block.

That problem is that the computational process is extremely slow. With very coarse mesh, it takes nearly 2 hours to get the simulation result.

I then refined the meshes for the air gaps to improve the accuracy of convection, and resolve the problem. After 4-hour computation, the solver sent this message;

Failed to find a solution.
The relative residual (0.28) is greater than the relative tolerance. Returned solution is not converged.

I don't think the heat transfer within a block will be a very complex model. It is astonishing to see COMSOL takes such a long time to solve it. My laptop is Macbook pro with Intel i7, 4GB RAM. Are there any problems in my model or is my laptop not powerful enough to run COMSOL?
Can anyone help me?


4 Replies Last Post 14.08.2012, 08:23 GMT-4
Ivar KJELBERG COMSOL Multiphysics(r) fan, retired, former "Senior Expert" at CSEM SA (CH)

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 09.08.2012, 08:46 GMT-4
Hi

the only point where I suspect you are wrong, is that heat transfer, including convection and radiation IS complex, and particularly non-linear, probably your coarse mesh is even making things worse, and you do probably not have enough RAM to include a much finer mesh.

You should rather ask yourself if you can simplify the problem. i.e. use 2D instead of 3D, remove radiation and even convection and use a simpler energy exchange to start, then add them one by one

--
Good luck
Ivar
Hi the only point where I suspect you are wrong, is that heat transfer, including convection and radiation IS complex, and particularly non-linear, probably your coarse mesh is even making things worse, and you do probably not have enough RAM to include a much finer mesh. You should rather ask yourself if you can simplify the problem. i.e. use 2D instead of 3D, remove radiation and even convection and use a simpler energy exchange to start, then add them one by one -- Good luck Ivar

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 09.08.2012, 13:37 GMT-4
If you are modeling radiation using the surface-to-surface radiation feature then that can really slow down your problem. It increases the problem size and its nonlinearity but specially its size. Start with removing that if you have it, and also follow Ivar’s suggestions.

Nagi Elabbasi
Veryst Engineering
If you are modeling radiation using the surface-to-surface radiation feature then that can really slow down your problem. It increases the problem size and its nonlinearity but specially its size. Start with removing that if you have it, and also follow Ivar’s suggestions. Nagi Elabbasi Veryst Engineering

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 12.08.2012, 15:48 GMT-4
Yes, you're right. After I add the surface-to-surface radiation part, the solution was obviously slowed down. Do you mean that I can simulate the model without radiation at first, and then go to the solver manager, change the initial value to current solution. After adding the part of radiation, simulate the model again? By the way, if I compute the model in this way, is the result the same as the model which simulates radiation, convection and conduction at one time?

Thanks for your time.
Yes, you're right. After I add the surface-to-surface radiation part, the solution was obviously slowed down. Do you mean that I can simulate the model without radiation at first, and then go to the solver manager, change the initial value to current solution. After adding the part of radiation, simulate the model again? By the way, if I compute the model in this way, is the result the same as the model which simulates radiation, convection and conduction at one time? Thanks for your time.

Nagi Elabbasi Facebook Reality Labs

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 14.08.2012, 08:23 GMT-4
Hi Xiqi,

If you solve the problem in the two steps you described the final solution should be the same. Convergence may be easier to reach, however, the second step will still be slow. I’m not sure what surface-to-surface radiation features where available in version 3.5a that you are using, but in 4.3 there are options within surface-to-surface radiation that simplify and speed up the radiation calculation.

Nagi Elabbasi
Veryst Engineering
Hi Xiqi, If you solve the problem in the two steps you described the final solution should be the same. Convergence may be easier to reach, however, the second step will still be slow. I’m not sure what surface-to-surface radiation features where available in version 3.5a that you are using, but in 4.3 there are options within surface-to-surface radiation that simplify and speed up the radiation calculation. Nagi Elabbasi Veryst Engineering

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.