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 The information provided may be out of date.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.

Computing flux in conduction convection module

Please login with a confirmed email address before reporting spam

Hi everybody,

I'm facing a quite anoying problem.

In the model i've attached to this post i cannot find the flux through outer boundaries equal to the integration over a subdomain of the heat source.

First attempt : subdomain 4
On one side i calculate the integration over a subdomain 4 of Q_cc
On the other side i calculate every ntflux_T_cc on every outer boundaries of the subdomain 4.
The result is not good and i cannot understand why ?

Second attempt : I want to extend this to a bigger domain : the domain made of domains 2 and 4
On one side i calculate the integration over a subdomains (subdomain 2 and 4) of Q_cc
On the other side i calculate every ntflux_T_cc on every outer boundaries of the subdomain formed by subdomain 2 and 4.
The result is not good and i cannot understand why ?

It probably comes from the computation of the flux on the boundary 4 and 21.

I've paid attention to the signs of the normal flux.
I've check the mesh dependancy and it's independant.
I've tried the lagrange multiplers by i'm not so familiar with this.
I've tried the reactf method but not succes with this too.

Do someone has already solved that sort of problem ?

Thanks a lot for your help.

Thomas


2 Replies Last Post 2011年3月2日 GMT-5 11:05
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 2011年3月2日 GMT-5 10:41
Hi

your model is not that trivial to understand. You have high density "fluids" and not solids at around normal temperature, all flowing at 2m/s along X, the temperature field solved is only fractions of a degree, with between 6.075 to 5.9057 W respectively solid heat generation versus total integrated heat flux (steady state analysis)

if you turn off the 2m/s flow of all 4 domains the heat profile enlarges and the total normal heat Flux increases to 6.0392 W

which for me tends to indicate that you are missing the energy lost through the flow. A last check at u=1[m/s] the total normal heat flux is 5.9522 W, at 10[m/s] 5.8358 W

Obviously we need to tale into account the energy transported out by your central thin layer, as the top bottom ones hardly have any temperature difference, hence the Tin/Tou ratio is roughly 1

By the way I'm running your model in v4.1

--
Good luck
Ivar
Hi your model is not that trivial to understand. You have high density "fluids" and not solids at around normal temperature, all flowing at 2m/s along X, the temperature field solved is only fractions of a degree, with between 6.075 to 5.9057 W respectively solid heat generation versus total integrated heat flux (steady state analysis) if you turn off the 2m/s flow of all 4 domains the heat profile enlarges and the total normal heat Flux increases to 6.0392 W which for me tends to indicate that you are missing the energy lost through the flow. A last check at u=1[m/s] the total normal heat flux is 5.9522 W, at 10[m/s] 5.8358 W Obviously we need to tale into account the energy transported out by your central thin layer, as the top bottom ones hardly have any temperature difference, hence the Tin/Tou ratio is roughly 1 By the way I'm running your model in v4.1 -- Good luck Ivar

Please login with a confirmed email address before reporting spam

Posted: 1 decade ago 2011年3月2日 GMT-5 11:05
Thank you Ivar,

The model I've uploaded is a simplification of my real model. My model actually use many functions from txt files and I was not able to upload them.

Is there another way of computing the flux better than ntflux_T_cc ? Using reacf(T) or lm1 for example ?

Best regards.

PS : i will try to open it in 4.1 to see if the flux are better calculated.
Thank you Ivar, The model I've uploaded is a simplification of my real model. My model actually use many functions from txt files and I was not able to upload them. Is there another way of computing the flux better than ntflux_T_cc ? Using reacf(T) or lm1 for example ? Best regards. PS : i will try to open it in 4.1 to see if the flux are better calculated.

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.