CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   OpenFOAM Post-Processing (https://www.cfd-online.com/Forums/openfoam-post-processing/)
-   -   foamCalc/heat flux problem (https://www.cfd-online.com/Forums/openfoam-post-processing/94148-foamcalc-heat-flux-problem.html)

 greel November 7, 2011 12:04

foamCalc/heat flux problem

Hi foamers!
I'm trying to validate and report the heat flux using foamCalc for a "flow of oil over plate" problem, using the buoyantBoussinesqSimpleFoam solver. But I'm getting different values after used foamCalc magGrad T utility.
http://img3.imageshack.us/img3/3787/caso.gif

http://img695.imageshack.us/img695/6104/placa.gif
L=5m for length and h= 1m width =1m but I'm testing with 2d case.

I have done several meshes, from a super coarse to ultra refined mesh with scaling, I also have calculated the first cell height.
The problem is that with some meshes I get values really similar to the analitic solution (about 3% error) but If I change the mesh dimension I get other values of magGrad T, this values are "similar" to the correct values but it seems that this values are multiplicated/divided by an natural number.
The analitic solution gives:
http://img404.imageshack.us/img404/4576/unledz.gif
My calc;
Fine mesh of 7500 volumes ( 150 for x and 50 for y with scalling first cell heigth =0.00507462)
Plate: magGrad T=15763.186 K/m --> multiplication by k=0.144 and the plate lenght gives q"=11349.4945152 W ( unitary lenght) this is 2.8% of error.

20000 Volumes mesh (200*100 with scalling FCL=0.000331119)
Plate: magGrad T=183675.96985 K/m absolutly out of merge, but if you divide this value by 12 (why 12!!! :D) get 15306.33 K/m --> multiplication by k=0.144 and the plate lenght gives q"=11020.558191 this is 0.178% of error

I will upload the case file soon!

Thanks for reading and sorry about my english!!!

http://www.mediafire.com/?qpddha0wd7br1a2
http://www.mediafire.com/?cnl5a4vrx1nch5n

 hugo17 November 17, 2011 03:45

Re: heat flux problem

Hi Andrés!

The different magGrad T values really seem to be strange and I don't know the reason. Do you get the same results with wallHeatFluxLaminar instead of magGrad?
I experienced that magGrad and wallHeatFluxLaminar give wrong results when a groovyBC is applied to the surface - but with a fixedValue boundary condition, it works fine also for different meshes in my studies.

Regards
Hugo

 All times are GMT -4. The time now is 12:17.