[BarCFD]

Hello FOAMers,

I want to impose a boundary condition which is an exponential function of the Field. Lets assume that I want to solve Laplace equation for temperature (thermal diffusion in solid). How can I have following boundary conditions?

[MSF]

Hi,

I think the easiest way would be groovyBC.

https://openfoamwiki.net/index.php/Contrib/groovyBC

Best,
Moritz

[LeeRuns]

Couldn't we just add this to the openfoam build

Sent from my SM-G930V using CFD Online Forum mobile app

[babakflame]

This a robin type boundary condition in OpenFOAM.

Search a little bit before asking. Here is the link:

http://www.modlab.lv/docs/2011/OpenF...Vilums_pdf.pdf

How to write exponential in boundary condition plus laplace samples see these threads:

codedFixedValue boundary condition

source term modification in poisson equation does not affect results

[BarCFD]

Originally Posted by MSF

Hi,

I think the easiest way would be groovyBC.

https://openfoamwiki.net/index.php/Contrib/groovyBC

Best,
Moritz

Thanks Moritz,

I tired groovyBC. It works for easier robin type boundary condition such as :
k*grad(n,T)+alpha*(T-Tinf)=0;
However, for my case I have an exponential function and groovy cannot find the root.

[BarCFD]

Originally Posted by babakflame

This a robin type boundary condition in OpenFOAM.

Search a little bit before asking. Here is the link:

http://www.modlab.lv/docs/2011/OpenF...Vilums_pdf.pdf

How to write exponential in boundary condition plus laplace samples see these threads:

codedFixedValue boundary condition

source term modification in poisson equation does not affect results

Thanks bobi,

First example is an ordinary robin boundary condition which is possible by groovy. I have already tried this one.

Last link was about source term in Laplace equation. I could not find the application for my case as boundary condition.

However, I think I should try codedFixedValue to extract internalField and to find root of boundary equation -[ k*grad(n,T)+exp(alpha*(T-Tinf)) ]- iteratively. Thanks for your help.

[babakflame]

I have written another sample for Robin Type boundary condition without Groovy. Just using the basics of c++ and OpenFoam.

Find it here:

Snippet for redefining fixedGradient boundary condition for a patch inside solver

[Zhiheng Wang]

Originally Posted by babakflame

I have written another sample for Robin Type boundary condition without Groovy. Just using the basics of c++ and OpenFoam.

Find it here:

Snippet for redefining fixedGradient boundary condition for a patch inside solver

[QUOTE=babakflame;665758]I have written another sample for Robin Type boundary condition without Groovy. Just using the basics of c++ and OpenFoam.

Wiith All regards to bobi aka babakFlame for excellent and nice explanation

i can Just add few lines

scalarField GradT = T.boundaryFieldRef()[patchID].snGrad();
scalarField Tpatch = T.boundaryFieldRef()[patchID];
forAll(GradT,I)
{
scalar CONV = kappa*GradT[I] - Foam::exp(alpha*(Tpatch - Tinf));

}
if it is iterative bc you can add scalar CONVERGE = kappa*GradT[I] - Foam::exp(alpha*(Tpatch - Tinf)); with convergence 10-4 or 10-3 some thing under while loop which will give you Tnew for equation
But what is need of k*dT/dn if delta T = 0 at upper domain ???? what you want to find out ????
Here kappa and alpha are assumed to be const for variable add
volScalarField& Kapp = thermo.kappa(); use Kappa.boundaryfieldRef()[patchID][faceI] in code or reference it in scalarField as previously done for GradT and Tpatch
with thermo as basicThermo or psiReactingThermo object.
But Here kdT/dn doent have units same as exp(alpha(T-Tinf)) what you want to find Heat Flux or Temperature satisfies the condition ??

[BarCFD]

[QUOTE=Zhiheng Wang;665901][QUOTE=babakflame;665758]

Thanks a lot Zhiheng!

Actually I have made an example here. My goal was to find the way to impose boundary condition which is a function of Field (exponential func). My Field in fact is not Temperate. I as solving potential field and boundary condition is local current density which linked to potential by Bulter-Volmer equation.

Thanks again for your comments and your concern