Question about computation of wall heat flux in OpenFOAM
 

Hello,

I am computing a 2D-axisymmetric case of a rotating plate in ambient air. My domain is a square. The bottom boundary is the axisymmetric axis, the right boundary is the plate, with a rotation speed. The other 2 boundaries are ambient air. Furthermore I am using a k-omega SST turbulence model. My mesh at the wall has a yplus < 1.

I would like to compute the wall heat flux and validate that it is the right value. I am using the postprocessing utility 'wallHeatFlux' and want to validate it with the following formula. q = k_f * (dT/dn), where I use the postprocessing utility 'grad(T)' to compute dT/dn and where k_f is the thermal conductivity of air (which is constant). However when I do this, my results will not match. Does anybody maybe have an idea what I am doing wrong here? Or whether the postProcess utilities are trustworthy for this case?

Thank you in advance!

Kind regards,

Cars

Hi!


Your k_f is not constant. You have turbulence. You need the "effective thermal diffusivity", not only the thermal diffusivity of the fluid.


But if you want to validate your results, you should create a really simple case where you can calculate the analytical solution, and compare that to OpenFOAM. Or you will just do the same what OpenFOAM does for you but on a painful way, and of course your results will be the same. It won't give you much more information about the accuracy of the solver.

Hey!

Thank you for you reply! I have also checked the case with a laminar model (where the k_f is constant) and this didn't work too. However the idea of a simple case with an analytical solution is a good one! I will try that! Thank you.

Kind regards,

Cars

Hey,

A last update of the results.

I have done some simple cases and calculated the wall heat flux manually, by using the temperature at the wall and at the cell center next to it. It seems the wallHeatflux utility I was using was incorrect, however the gradient of the temperature utility I used seemed to predict it correct!

Hi!


I think you miss something. The wallHeatFlux utility should be correct. I had to create a small verification case for that in the past and the utility was really close to the analytical solution. It was 1-2 years ago but I don't think if the utility is changed since then...
And as I know it is widely used so I'm pretty sure it is correct... (or there should be a bug report already)


You have to create a case which has an analytical solution. (Heated flat plate for example or heated pipe. And with the simulation stay in the valid range for the analytical solution.) And not calculating in manually, but calculate the analytical solution for that case and calculate the wall heat flux using the utility and compare them.
As I understand you did the same thing again as previously. Not what I suggested you to check if the utility is correct or not. As you wrote you just calculated the heat flux by hand again...