[manuc]

Hello All
I am calculating Nusselt #at the wall using multiregionfoam during runtime.
swak4foam is used for calculating the wall averaged Nusselt #.

Issue
When I use vanleer scheme for divSchemes my results are close to experimental results.

When I use limitedlinear scheme for divSchemes my results are off by 20% w.r.t experimental results. [ tried with factors 0.33, 0.5 and 1]

I would like to have a pseudo-DNS kind simulation, because of which I prefer to use limitedlinear.
[Though both the schemes are 2nd order accurate I believe limitedlinear is more close to CDS]
Does the current results indicate that I need fine mesh for limitedlinear.
Nu=L*snGrad(T)/dT , where L=characterisitc length, T=Temperatute, dT =delta Temeprature (between hot and cold walls)

Any suggestions would be of great help.

Thanks

[svramana]

Originally Posted by manuc

Hello All
I am calculating Nusselt #at the wall using multiregionfoam during runtime.
swak4foam is used for calculating the wall averaged Nusselt #.

Issue
When I use vanleer scheme for divSchemes my results are close to experimental results.

When I use limitedlinear scheme for divSchemes my results are off by 20% w.r.t experimental results. [ tried with factors 0.33, 0.5 and 1]

I would like to have a pseudo-DNS kind simulation, because of which I prefer to use limitedlinear.
[Though both the schemes are 2nd order accurate I believe limitedlinear is more close to CDS]
Does the current results indicate that I need fine mesh for limitedlinear.


Any suggestions would be of great help.

Thanks

Hi,

i am dealing with a problem where i need to calculate local Nusselt number variation along the surface of the cube.
.can i have access to your code

Regards,
Ramana

[piu58]

I always use Gauss limited linear with factor 1 (which is not limited, in effect). With these schemes I calculated a few cases for which theoretical solutions exist. I was always satisfied with the results. My cases included heat transfer and Nusselt numbers.

May it be, that your material constants are slightly off and you "compensated" this by the scheme?

[manuc]

Dear Pilz

Thanks for your reply. Could you please let me know what you mean by
material constants are slightly off

Thanks

[piu58]

In your simulation yu have to use the thermal conductivity, heat capacity, Prandl number and perhaps some more. Most of these are temperature dependent, and in many cases not known to the very last decimale. It may be that the real values differs from the ones you set in the simulation.

[manuc]

Oh..I am using the values from experiments that we did and they are giving good results with other schemes. So I doubt that could be an issue. Thank you for the msg. I would certainly look if I messed up something.